Add files using upload-large-folder tool

.gitattributes

@@ -53697,3 +53697,17 @@ EgoPressure/shared-v1.1/data/p_018_press_palm_high_x5_left/cam-d-rgb/observation
 EgoPressure/shared-v1.1/data/p_018_press_palm_high_x5_left/cam-d-rgb/observation/rgb.mkv filter=lfs diff=lfs merge=lfs -text
 EgoPressure/shared-v1.1/data/p_018_pull-towards_5x_left/cam-d-rgb/observation/depth_valid.mkv filter=lfs diff=lfs merge=lfs -text
 EgoPressure/shared-v1.1/data/p_018_pull-towards_5x_left/cam-d-rgb/observation/rgb.mkv filter=lfs diff=lfs merge=lfs -text
+EgoPressure/shared-v1.1/data/p_018_pull-towards_5x_right/cam-d-rgb/observation/depth_valid.mkv filter=lfs diff=lfs merge=lfs -text
+EgoPressure/shared-v1.1/data/p_018_pull-towards_5x_right/cam-d-rgb/observation/rgb.mkv filter=lfs diff=lfs merge=lfs -text
+EgoPressure/shared-v1.1/data/p_018_pull-towards_5x_right/cam-d-rgb/observation/depth.mkv filter=lfs diff=lfs merge=lfs -text
+EgoPressure/shared-v1.1/data/p_018_push-away_5x_right/cam-d-rgb/observation/depth_valid.mkv filter=lfs diff=lfs merge=lfs -text
+EgoPressure/shared-v1.1/data/p_018_push-away_5x_right/cam-d-rgb/observation/rgb.mkv filter=lfs diff=lfs merge=lfs -text
+EgoPressure/shared-v1.1/data/p_018_push-away_5x_right/cam-d-rgb/observation/depth.mkv filter=lfs diff=lfs merge=lfs -text
+EgoPressure/shared-v1.1/data/p_018_push-away_5x_left/cam-d-rgb/observation/depth.mkv filter=lfs diff=lfs merge=lfs -text
+EgoPressure/shared-v1.1/data/p_018_push-away_5x_left/cam-d-rgb/observation/depth_valid.mkv filter=lfs diff=lfs merge=lfs -text
+EgoPressure/shared-v1.1/data/p_018_push-away_5x_left/cam-d-rgb/observation/rgb.mkv filter=lfs diff=lfs merge=lfs -text
+EgoPressure/shared-v1.1/data/p_018_type_ipad_5x_left/cam-d-rgb/observation/depth_valid.mkv filter=lfs diff=lfs merge=lfs -text
+EgoPressure/shared-v1.1/data/p_018_type_ipad_5x_left/cam-d-rgb/observation/rgb.mkv filter=lfs diff=lfs merge=lfs -text
+EgoPressure/shared-v1.1/data/p_018_type_ipad_5x_right/cam-d-rgb/observation/depth.mkv filter=lfs diff=lfs merge=lfs -text
+EgoPressure/shared-v1.1/data/p_018_type_ipad_5x_left/cam-d-rgb/observation/depth.mkv filter=lfs diff=lfs merge=lfs -text
+EgoPressure/shared-v1.1/data/p_019_calibration_routine_left/cam-d-rgb/observation/depth.mkv filter=lfs diff=lfs merge=lfs -text

EgoPressure/shared-v1.1/data/p_018_pull-towards_5x_right/cam-d-rgb/camera/camera.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:42d3b56bcf3cb1ad367d02adb3b026374c3d78a4d9a3a989cf59f142ece752f3
+size 29657

EgoPressure/shared-v1.1/data/p_018_pull-towards_5x_right/cam-d-rgb/observation/depth.mkv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e42bf54cd3b78d1c1b53647edf93c1be18076ec7e236e8364c049e6c9391025a
+size 197164248

EgoPressure/shared-v1.1/data/p_018_pull-towards_5x_right/cam-d-rgb/observation/depth_valid.mkv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2a528f73497659099a598718262b2661197966098dec99bdb88bcdc3c4844985
+size 6511863

EgoPressure/shared-v1.1/data/p_018_pull-towards_5x_right/cam-d-rgb/observation/rgb.mkv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:382577d4c778bd87f366f19417810682ae0b8a8a85d8c841f78884ae4b2df2ba
+size 74923077

EgoPressure/shared-v1.1/data/p_018_push-away_5x_left/cam-d-rgb/camera/camera.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9639b4fca669713bcf4c96e721ab129dd61a7e87cce85d8f3c03759685fead96
+size 31769

EgoPressure/shared-v1.1/data/p_018_push-away_5x_left/cam-d-rgb/observation/depth.mkv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e5f4e65ce70006ec6e551a5b08d8df62c22e4ea747816e95bd5ed10020f845a1
+size 203075962

EgoPressure/shared-v1.1/data/p_018_push-away_5x_left/cam-d-rgb/observation/depth_valid.mkv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c059e5bde670832c6f6a9b37d3ec0d1096953ea08557fb588c6355c7f4965255
+size 7026888

EgoPressure/shared-v1.1/data/p_018_push-away_5x_left/cam-d-rgb/observation/rgb.mkv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e1a6aab17abd04fdfd1b7409d5d1d0a1ea3abd74eb54a974c1c1143d898c18cb
+size 81930016

EgoPressure/shared-v1.1/data/p_018_push-away_5x_right/cam-d-rgb/camera/camera.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:418dc08aa6f90dd5c391ab588540294b33519e93d1ddf005feee87d32d8714f0
+size 28377

EgoPressure/shared-v1.1/data/p_018_push-away_5x_right/cam-d-rgb/observation/depth.mkv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f7cbb0ea91da01b62f525e89e0c4ecfa9ea2ee1bb939c67c4d3d0d00650c174f
+size 186633888

EgoPressure/shared-v1.1/data/p_018_push-away_5x_right/cam-d-rgb/observation/depth_valid.mkv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c32b12e5b0a1e27880c6d70ebdbeaad3dd0118c373e71521cb6d2716f0049487
+size 6210011

EgoPressure/shared-v1.1/data/p_018_push-away_5x_right/cam-d-rgb/observation/rgb.mkv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:edfa14fe809cffe07ee70aaf1a97392639772cbc8329734ad9be0d1e7bd835fe
+size 71275455

EgoPressure/shared-v1.1/data/p_018_push-away_5x_right/cam-d-rgb/observation/rgb_valid.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:35e65a443c20c599008c2d29796db372074db70a98434438c2f02630338407fc
+size 1009

EgoPressure/shared-v1.1/data/p_018_type_ipad_5x_left/cam-d-rgb/annotation/mano/mano.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f21e4f16f6e3d1fe1f9aee08e639b3cf950f4012c8c4face9290ca5394fa14f1
+size 281243

EgoPressure/shared-v1.1/data/p_018_type_ipad_5x_left/cam-d-rgb/camera/camera.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8e98f749f9271af2424df2fc875fa9bd0946d97ed93494d2e73c4e81e244b00d
+size 59673

EgoPressure/shared-v1.1/data/p_018_type_ipad_5x_left/cam-d-rgb/observation/depth.mkv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4324075d5cc8e78e557da3bff6ef5275d0bf26915e566a7e7644c7ff70d91046
+size 387535600

EgoPressure/shared-v1.1/data/p_018_type_ipad_5x_left/cam-d-rgb/observation/depth_valid.mkv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:53ffab8c650b80004aa2754b9a71eaa2ad6cfa68be46e7993f3d58e19620f336
+size 13458209

EgoPressure/shared-v1.1/data/p_018_type_ipad_5x_left/cam-d-rgb/observation/rgb.mkv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fcc8b6e834c48484d8ba4a774766e7cb277b673e9d7bc3ef62ede0a0f344986b
+size 158370967

EgoPressure/shared-v1.1/data/p_018_type_ipad_5x_left/cam-d-rgb/observation/rgb_valid.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dbd717ed4403d3faed4b5b58f9a29c82d61f8bae29a72e3e098b4dea39c46095
+size 1265

EgoPressure/shared-v1.1/data/p_018_type_ipad_5x_right/cam-d-rgb/annotation/mano/mano.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:27553db11c755574357251dc10fb8f315a5735b5ca254e826033b8ab8df8d27c
+size 270491

EgoPressure/shared-v1.1/data/p_018_type_ipad_5x_right/cam-d-rgb/camera/camera.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:333fcdc6eb51e5692e13effc964f13f1eb8b6b23e8a91ac5a761aa5e676fac67
+size 57497

EgoPressure/shared-v1.1/data/p_018_type_ipad_5x_right/cam-d-rgb/observation/depth.mkv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3f060b0ee9226a690c43fcfaf4276e92351a5ba55952b5727001a683dba7f77d
+size 379968017

EgoPressure/shared-v1.1/data/p_018_type_ipad_5x_right/cam-d-rgb/observation/rgb_valid.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6c3b7bc80f77fcca04e90f684bbf7779c4574609a6c2e04f93ccc993ea93ccee
+size 1265

EgoPressure/shared-v1.1/data/p_019_calibration_routine_left/cam-d-rgb/annotation/mano/mano.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:68ebdfda6edad6deb59aa533ba9ef9379b86c70b4e47ba73ddb1844fa3d04bac
+size 215579

EgoPressure/shared-v1.1/data/p_019_calibration_routine_left/cam-d-rgb/camera/camera.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:05d3ec4f2132570439b945295d4a76d7dc83283a50dca0437481b6dc25898353
+size 46041

EgoPressure/shared-v1.1/data/p_019_calibration_routine_left/cam-d-rgb/observation/depth.mkv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d276bd9e382e1ea22c076fb62b54f88036bd1a1c9335aaf14973ffe84b565a8a
+size 242447742

EgoPressure/shared-v1.1/data/p_019_calibration_routine_left/cam-d-rgb/observation/rgb_valid.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e2808576d4f761ad71e12c3fe5a692670a96ee1efb7f839e49f5bd9edcebbfef
+size 1137

EgoPressure/shared-v1.1/data/p_019_calibration_routine_right/cam-d-rgb/annotation/mano/mano.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:da431d469b89e0cb842546b1448ace439aca69d5c2850e5ded3d3e831e93f189
+size 260187

EgoPressure/shared-v1.1/data/p_019_calibration_routine_right/cam-d-rgb/observation/rgb_valid.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:97d312b165d66ec24daa134f2f04920b539ac2da015f0c3521fd87d3969d2d72
+size 1265

EgoPressure/shared-v1.1/data/p_019_draw_word_5x_left/cam-d-rgb/annotation/mano/mano.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3a5ebdba8a102b4af7e0355af7d57e621c5f1cd2b0e3c79c92a8febd5dcc7aed
+size 516955

EgoPressure/shared-v1.1/data/p_019_grasp-edge_curled_thumb-up_5x_right/cam-d-rgb/metainfo.json

@@ -0,0 +1,481 @@
+{
+  "schema_version": "hand-warehouse/1.1",
+  "dataset_id": "EgoPressure",
+  "release_id": "shared-v1.1",
+  "sequence_id": "p_019_grasp-edge_curled_thumb-up_5x_right",
+  "camera_id": "cam-d-rgb",
+  "physical_camera_id": "d",
+  "view_type": "ego",
+  "sample_type": "video",
+  "num_frames": 277,
+  "hand_slots": [
+    "left",
+    "right"
+  ],
+  "sampling": {
+    "type": "dense",
+    "source_fps": 30,
+    "frames_ref": "camera/frames.jsonl",
+    "policy": "Every source frame, exact frame-ID joins, no true timestamps available; presentation 30Hz only"
+  },
+  "camera": {
+    "model": "pinhole",
+    "schema_id": "camera/pinhole/v1",
+    "image_size": [
+      1920,
+      1080
+    ]
+  },
+  "components": {
+    "rgb": {
+      "available": true,
+      "schema_id": "rgb/uint8/v1",
+      "data_ref": "observation/rgb.mkv",
+      "valid_ref": "observation/rgb_valid.pt",
+      "axes": [
+        "T",
+        "H",
+        "W",
+        "C"
+      ],
+      "valid_axes": [
+        "T"
+      ],
+      "storage": {
+        "profile": "mkv/mjpeg_copy/v1",
+        "codec": "mjpeg",
+        "pixel_format": "yuvj420p",
+        "encoded_size": [
+          1920,
+          1080
+        ],
+        "source_lossy": true,
+        "decoder": {
+          "name": "PyAV/libavcodec",
+          "pyav_version": "12.3.0",
+          "library_versions": {
+            "libavutil": [
+              58,
+              29,
+              100
+            ],
+            "libavcodec": [
+              60,
+              31,
+              102
+            ],
+            "libavformat": [
+              60,
+              16,
+              100
+            ],
+            "libavdevice": [
+              60,
+              3,
+              100
+            ],
+            "libavfilter": [
+              9,
+              12,
+              100
+            ],
+            "libswscale": [
+              7,
+              5,
+              100
+            ],
+            "libswresample": [
+              4,
+              12,
+              100
+            ]
+          },
+          "pixel_interpretation": "rgb24 for RGB; native integer planes for depth and bool; no autorotation"
+        },
+        "encoder_parameters": {
+          "stream_copy": true
+        }
+      },
+      "provenance": {
+        "kind": "format_conversion",
+        "method": "Original JPEG packets copied, no resize/crop/re-encoding"
+      },
+      "source_locator": {
+        "file": "raw/data/data/p_019/p_019_grasp-edge_curled_thumb-up_5x_right/cam-d.color.parquet",
+        "record_key": "frame"
+      }
+    },
+    "depth": {
+      "available": true,
+      "schema_id": "depth/z_float32/v1",
+      "data_ref": "observation/depth.mkv",
+      "valid_ref": "observation/depth_valid.mkv",
+      "axes": [
+        "T",
+        "H",
+        "W"
+      ],
+      "valid_axes": [
+        "T",
+        "H",
+        "W"
+      ],
+      "depth_type": "z",
+      "depth_origin": "measured",
+      "unit": "m",
+      "storage": {
+        "profile": "mkv/depth_float32_bits/v1",
+        "codec": "ffv1",
+        "pixel_format": "gray16le",
+        "encoded_size": [
+          3840,
+          1080
+        ],
+        "source_lossy": false,
+        "decoder": {
+          "name": "PyAV/libavcodec",
+          "pyav_version": "12.3.0",
+          "library_versions": {
+            "libavutil": [
+              58,
+              29,
+              100
+            ],
+            "libavcodec": [
+              60,
+              31,
+              102
+            ],
+            "libavformat": [
+              60,
+              16,
+              100
+            ],
+            "libavdevice": [
+              60,
+              3,
+              100
+            ],
+            "libavfilter": [
+              9,
+              12,
+              100
+            ],
+            "libswscale": [
+              7,
+              5,
+              100
+            ],
+            "libswresample": [
+              4,
+              12,
+              100
+            ]
+          },
+          "pixel_interpretation": "rgb24 for RGB; native integer planes for depth and bool; no autorotation"
+        },
+        "encoder_parameters": {
+          "name": "ffv1",
+          "level": 3,
+          "coder": 1,
+          "context": 1,
+          "slicecrc": 1,
+          "slices": 4
+        }
+      },
+      "valid_storage": {
+        "profile": "mkv/bool_ffv1/v1",
+        "codec": "ffv1",
+        "pixel_format": "gray",
+        "encoded_size": [
+          1920,
+          1080
+        ],
+        "source_lossy": false,
+        "decoder": {
+          "name": "PyAV/libavcodec",
+          "pyav_version": "12.3.0",
+          "library_versions": {
+            "libavutil": [
+              58,
+              29,
+              100
+            ],
+            "libavcodec": [
+              60,
+              31,
+              102
+            ],
+            "libavformat": [
+              60,
+              16,
+              100
+            ],
+            "libavdevice": [
+              60,
+              3,
+              100
+            ],
+            "libavfilter": [
+              9,
+              12,
+              100
+            ],
+            "libswscale": [
+              7,
+              5,
+              100
+            ],
+            "libswresample": [
+              4,
+              12,
+              100
+            ]
+          },
+          "pixel_interpretation": "rgb24 for RGB; native integer planes for depth and bool; no autorotation"
+        },
+        "encoder_parameters": {
+          "name": "ffv1",
+          "level": 3,
+          "coder": 1,
+          "context": 1,
+          "slicecrc": 1,
+          "slices": 4
+        }
+      },
+      "provenance": {
+        "kind": "derived",
+        "method": "Official k4a_transformation_depth_image_to_color_camera surface rasterization; SDK uint16 millimetres normalized to float32 metres; zero invalid. No external hole filling."
+      },
+      "source_locator": {
+        "file": "raw/data/data/p_019/p_019_grasp-edge_curled_thumb-up_5x_right/cam-d.depth.parquet",
+        "record_key": "frame"
+      }
+    },
+    "intrinsic": {
+      "available": true,
+      "schema_id": "camera/pinhole/v1",
+      "data_ref": "camera/camera.pt",
+      "data_key": [
+        "intrinsic",
+        "data"
+      ],
+      "valid_ref": "camera/camera.pt",
+      "valid_key": [
+        "intrinsic",
+        "valid"
+      ],
+      "axes": [
+        "T",
+        "row",
+        "column"
+      ],
+      "valid_axes": [
+        "T"
+      ],
+      "storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "valid_storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "provenance": {
+        "kind": "format_conversion",
+        "method": "Original undistorted color K"
+      },
+      "source_locator": {
+        "file": "raw/data/configs/p_019/p_019_grasp-edge_curled_thumb-up_5x_right.json"
+      }
+    },
+    "extrinsic": {
+      "available": true,
+      "schema_id": "transform/se3/v1",
+      "data_ref": "camera/camera.pt",
+      "data_key": [
+        "extrinsic",
+        "data"
+      ],
+      "valid_ref": "camera/camera.pt",
+      "valid_key": [
+        "extrinsic",
+        "valid"
+      ],
+      "axes": [
+        "T",
+        "row",
+        "column"
+      ],
+      "valid_axes": [
+        "T"
+      ],
+      "storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "valid_storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "provenance": {
+        "kind": "format_conversion",
+        "method": "Source ego_R/ego_T unchanged"
+      },
+      "source_locator": {
+        "file": "raw/data/data/p_019/p_019_grasp-edge_curled_thumb-up_5x_right/annotation.parquet",
+        "record_key": "frame"
+      },
+      "direction": "T_camera_from_world",
+      "convention": "T_camera_from_world"
+    },
+    "mano": {
+      "available": true,
+      "schema_id": "mano/axis_angle61/v1",
+      "data_ref": "annotation/mano/mano.pt",
+      "data_key": [
+        "motion",
+        "data"
+      ],
+      "valid_ref": "annotation/mano/mano.pt",
+      "valid_key": [
+        "motion",
+        "valid"
+      ],
+      "axes": [
+        "T",
+        "hand",
+        "parameter"
+      ],
+      "valid_axes": [
+        "T",
+        "hand"
+      ],
+      "storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "valid_storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "provenance": {
+        "kind": "format_conversion",
+        "method": "Standard MANO, no per-vertex refinement; camera root rotation composed; additive translation compensated around shaped root joint"
+      },
+      "source_locator": {
+        "file": "raw/data/data/p_019/p_019_grasp-edge_curled_thumb-up_5x_right/annotation.parquet",
+        "record_key": "frame"
+      },
+      "model_profile": {
+        "model": "MANO",
+        "vertices": 778,
+        "flat_hand_mean": true,
+        "use_pca": false,
+        "left_shapedirs_fix_applied": false,
+        "per_vertex_refinement": false,
+        "assets_sha256": {
+          "left": "c4022f7083f2ca7c78b2b3d595abbab52debd32b09d372b16923a801f0ea6a30",
+          "right": "45d60aa3b27ef9107a7afd4e00808f307fd91111e1cfa35afd5c4a62de264767"
+        }
+      },
+      "valid_policy": {
+        "source_annotation_present": "Source has_annotation=true and exact frame ID join",
+        "numerical_checks": "Finite shapes and geometry checked; failure aborts",
+        "official_qa": "not available; visibility not inferred"
+      }
+    }
+  },
+  "split": "unknown",
+  "split_reason": "Official split not provided in selected source",
+  "source_revision": "cea37de99be44078f0e4b9796fd6f614d9110dc2",
+  "calibration": {
+    "baseline": "config DepthToColor",
+    "offset_rgb_optical_z_mm": -6.84644549381777,
+    "offset_estimation": "100-sequence participant-balanced pad fit",
+    "T_rgb_from_depth_rotation": [
+      [
+        0.999999,
+        0.000952568,
+        0.0010665
+      ],
+      [
+        -0.00105482,
+        0.994948,
+        0.10039
+      ],
+      [
+        -0.000965488,
+        -0.100391,
+        0.994948
+      ]
+    ],
+    "T_rgb_from_depth_translation_m": [
+      -0.0335589,
+      -0.00214962,
+      -0.00945180549381777
+    ],
+    "status": "User-approved empirical offset; not official recalibration"
+  },
+  "registration_sdk": {
+    "sdk": "Microsoft.Azure.Kinect.Sensor",
+    "version": "1.4.1",
+    "dll_sha256": "1c6bf5aabb62d4907aa66da92448aebe5f1b8b59f81f2099bc1d9f989cf44cb6",
+    "interfaces": [
+      "k4a_calibration_get_from_raw",
+      "k4a_transformation_create",
+      "k4a_transformation_depth_image_to_color_camera"
+    ],
+    "depth_mode": "WFOV_2X2BINNED",
+    "color_resolution": "1080P",
+    "calibration_adaptation": "Source RGB is already undistorted: config RGB K, zero color distortion, color metric radius 10; depth factory calibration parsed by SDK; config DepthToColor plus approved optical-Z offset; SDK translations in mm.",
+    "T_rgb_from_depth_mm": [
+      [
+        0.999999,
+        0.000952568,
+        0.0010665,
+        -33.5589
+      ],
+      [
+        -0.00105482,
+        0.994948,
+        0.10039,
+        -2.14962
+      ],
+      [
+        -0.000965488,
+        -0.100391,
+        0.994948,
+        -9.45180549381777
+      ],
+      [
+        0.0,
+        0.0,
+        0.0,
+        1.0
+      ]
+    ],
+    "native_depth_K": [
+      [
+        252.52279663085938,
+        0.0,
+        257.9237060546875
+      ],
+      [
+        0.0,
+        252.60423278808594,
+        251.84109497070312
+      ],
+      [
+        0.0,
+        0.0,
+        1.0
+      ]
+    ],
+    "calls": {
+      "depth_to_color": 277,
+      "color_to_depth": 0
+    },
+    "depth_output": "Official uint16 mm, zero invalid; losslessly normalized to float32 metres."
+  },
+  "processing": {
+    "adapter": "egopressure-official-k4a-rgb-only-full/1.0",
+    "contract_repo_revision": "9e89722fc3935d1e16aeb9eeeffa4cd5c9d93383",
+    "scope": "All 1344 selected ego d sequences; original RGB plus official SDK depth-to-RGB",
+    "validation_policy": "All frame IDs, tensor validity and packet counts; deterministic representative decode checks, not full decode or SDK replay"
+  }
+}

EgoPressure/shared-v1.1/data/p_019_grasp-edge_uncurled_thumb-down_5x_left/cam-d-rgb/metainfo.json

@@ -0,0 +1,481 @@
+{
+  "schema_version": "hand-warehouse/1.1",
+  "dataset_id": "EgoPressure",
+  "release_id": "shared-v1.1",
+  "sequence_id": "p_019_grasp-edge_uncurled_thumb-down_5x_left",
+  "camera_id": "cam-d-rgb",
+  "physical_camera_id": "d",
+  "view_type": "ego",
+  "sample_type": "video",
+  "num_frames": 377,
+  "hand_slots": [
+    "left",
+    "right"
+  ],
+  "sampling": {
+    "type": "dense",
+    "source_fps": 30,
+    "frames_ref": "camera/frames.jsonl",
+    "policy": "Every source frame, exact frame-ID joins, no true timestamps available; presentation 30Hz only"
+  },
+  "camera": {
+    "model": "pinhole",
+    "schema_id": "camera/pinhole/v1",
+    "image_size": [
+      1920,
+      1080
+    ]
+  },
+  "components": {
+    "rgb": {
+      "available": true,
+      "schema_id": "rgb/uint8/v1",
+      "data_ref": "observation/rgb.mkv",
+      "valid_ref": "observation/rgb_valid.pt",
+      "axes": [
+        "T",
+        "H",
+        "W",
+        "C"
+      ],
+      "valid_axes": [
+        "T"
+      ],
+      "storage": {
+        "profile": "mkv/mjpeg_copy/v1",
+        "codec": "mjpeg",
+        "pixel_format": "yuvj420p",
+        "encoded_size": [
+          1920,
+          1080
+        ],
+        "source_lossy": true,
+        "decoder": {
+          "name": "PyAV/libavcodec",
+          "pyav_version": "12.3.0",
+          "library_versions": {
+            "libavutil": [
+              58,
+              29,
+              100
+            ],
+            "libavcodec": [
+              60,
+              31,
+              102
+            ],
+            "libavformat": [
+              60,
+              16,
+              100
+            ],
+            "libavdevice": [
+              60,
+              3,
+              100
+            ],
+            "libavfilter": [
+              9,
+              12,
+              100
+            ],
+            "libswscale": [
+              7,
+              5,
+              100
+            ],
+            "libswresample": [
+              4,
+              12,
+              100
+            ]
+          },
+          "pixel_interpretation": "rgb24 for RGB; native integer planes for depth and bool; no autorotation"
+        },
+        "encoder_parameters": {
+          "stream_copy": true
+        }
+      },
+      "provenance": {
+        "kind": "format_conversion",
+        "method": "Original JPEG packets copied, no resize/crop/re-encoding"
+      },
+      "source_locator": {
+        "file": "raw/data/data/p_019/p_019_grasp-edge_uncurled_thumb-down_5x_left/cam-d.color.parquet",
+        "record_key": "frame"
+      }
+    },
+    "depth": {
+      "available": true,
+      "schema_id": "depth/z_float32/v1",
+      "data_ref": "observation/depth.mkv",
+      "valid_ref": "observation/depth_valid.mkv",
+      "axes": [
+        "T",
+        "H",
+        "W"
+      ],
+      "valid_axes": [
+        "T",
+        "H",
+        "W"
+      ],
+      "depth_type": "z",
+      "depth_origin": "measured",
+      "unit": "m",
+      "storage": {
+        "profile": "mkv/depth_float32_bits/v1",
+        "codec": "ffv1",
+        "pixel_format": "gray16le",
+        "encoded_size": [
+          3840,
+          1080
+        ],
+        "source_lossy": false,
+        "decoder": {
+          "name": "PyAV/libavcodec",
+          "pyav_version": "12.3.0",
+          "library_versions": {
+            "libavutil": [
+              58,
+              29,
+              100
+            ],
+            "libavcodec": [
+              60,
+              31,
+              102
+            ],
+            "libavformat": [
+              60,
+              16,
+              100
+            ],
+            "libavdevice": [
+              60,
+              3,
+              100
+            ],
+            "libavfilter": [
+              9,
+              12,
+              100
+            ],
+            "libswscale": [
+              7,
+              5,
+              100
+            ],
+            "libswresample": [
+              4,
+              12,
+              100
+            ]
+          },
+          "pixel_interpretation": "rgb24 for RGB; native integer planes for depth and bool; no autorotation"
+        },
+        "encoder_parameters": {
+          "name": "ffv1",
+          "level": 3,
+          "coder": 1,
+          "context": 1,
+          "slicecrc": 1,
+          "slices": 4
+        }
+      },
+      "valid_storage": {
+        "profile": "mkv/bool_ffv1/v1",
+        "codec": "ffv1",
+        "pixel_format": "gray",
+        "encoded_size": [
+          1920,
+          1080
+        ],
+        "source_lossy": false,
+        "decoder": {
+          "name": "PyAV/libavcodec",
+          "pyav_version": "12.3.0",
+          "library_versions": {
+            "libavutil": [
+              58,
+              29,
+              100
+            ],
+            "libavcodec": [
+              60,
+              31,
+              102
+            ],
+            "libavformat": [
+              60,
+              16,
+              100
+            ],
+            "libavdevice": [
+              60,
+              3,
+              100
+            ],
+            "libavfilter": [
+              9,
+              12,
+              100
+            ],
+            "libswscale": [
+              7,
+              5,
+              100
+            ],
+            "libswresample": [
+              4,
+              12,
+              100
+            ]
+          },
+          "pixel_interpretation": "rgb24 for RGB; native integer planes for depth and bool; no autorotation"
+        },
+        "encoder_parameters": {
+          "name": "ffv1",
+          "level": 3,
+          "coder": 1,
+          "context": 1,
+          "slicecrc": 1,
+          "slices": 4
+        }
+      },
+      "provenance": {
+        "kind": "derived",
+        "method": "Official k4a_transformation_depth_image_to_color_camera surface rasterization; SDK uint16 millimetres normalized to float32 metres; zero invalid. No external hole filling."
+      },
+      "source_locator": {
+        "file": "raw/data/data/p_019/p_019_grasp-edge_uncurled_thumb-down_5x_left/cam-d.depth.parquet",
+        "record_key": "frame"
+      }
+    },
+    "intrinsic": {
+      "available": true,
+      "schema_id": "camera/pinhole/v1",
+      "data_ref": "camera/camera.pt",
+      "data_key": [
+        "intrinsic",
+        "data"
+      ],
+      "valid_ref": "camera/camera.pt",
+      "valid_key": [
+        "intrinsic",
+        "valid"
+      ],
+      "axes": [
+        "T",
+        "row",
+        "column"
+      ],
+      "valid_axes": [
+        "T"
+      ],
+      "storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "valid_storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "provenance": {
+        "kind": "format_conversion",
+        "method": "Original undistorted color K"
+      },
+      "source_locator": {
+        "file": "raw/data/configs/p_019/p_019_grasp-edge_uncurled_thumb-down_5x_left.json"
+      }
+    },
+    "extrinsic": {
+      "available": true,
+      "schema_id": "transform/se3/v1",
+      "data_ref": "camera/camera.pt",
+      "data_key": [
+        "extrinsic",
+        "data"
+      ],
+      "valid_ref": "camera/camera.pt",
+      "valid_key": [
+        "extrinsic",
+        "valid"
+      ],
+      "axes": [
+        "T",
+        "row",
+        "column"
+      ],
+      "valid_axes": [
+        "T"
+      ],
+      "storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "valid_storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "provenance": {
+        "kind": "format_conversion",
+        "method": "Source ego_R/ego_T unchanged"
+      },
+      "source_locator": {
+        "file": "raw/data/data/p_019/p_019_grasp-edge_uncurled_thumb-down_5x_left/annotation.parquet",
+        "record_key": "frame"
+      },
+      "direction": "T_camera_from_world",
+      "convention": "T_camera_from_world"
+    },
+    "mano": {
+      "available": true,
+      "schema_id": "mano/axis_angle61/v1",
+      "data_ref": "annotation/mano/mano.pt",
+      "data_key": [
+        "motion",
+        "data"
+      ],
+      "valid_ref": "annotation/mano/mano.pt",
+      "valid_key": [
+        "motion",
+        "valid"
+      ],
+      "axes": [
+        "T",
+        "hand",
+        "parameter"
+      ],
+      "valid_axes": [
+        "T",
+        "hand"
+      ],
+      "storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "valid_storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "provenance": {
+        "kind": "format_conversion",
+        "method": "Standard MANO, no per-vertex refinement; camera root rotation composed; additive translation compensated around shaped root joint"
+      },
+      "source_locator": {
+        "file": "raw/data/data/p_019/p_019_grasp-edge_uncurled_thumb-down_5x_left/annotation.parquet",
+        "record_key": "frame"
+      },
+      "model_profile": {
+        "model": "MANO",
+        "vertices": 778,
+        "flat_hand_mean": true,
+        "use_pca": false,
+        "left_shapedirs_fix_applied": false,
+        "per_vertex_refinement": false,
+        "assets_sha256": {
+          "left": "c4022f7083f2ca7c78b2b3d595abbab52debd32b09d372b16923a801f0ea6a30",
+          "right": "45d60aa3b27ef9107a7afd4e00808f307fd91111e1cfa35afd5c4a62de264767"
+        }
+      },
+      "valid_policy": {
+        "source_annotation_present": "Source has_annotation=true and exact frame ID join",
+        "numerical_checks": "Finite shapes and geometry checked; failure aborts",
+        "official_qa": "not available; visibility not inferred"
+      }
+    }
+  },
+  "split": "unknown",
+  "split_reason": "Official split not provided in selected source",
+  "source_revision": "cea37de99be44078f0e4b9796fd6f614d9110dc2",
+  "calibration": {
+    "baseline": "config DepthToColor",
+    "offset_rgb_optical_z_mm": -6.84644549381777,
+    "offset_estimation": "100-sequence participant-balanced pad fit",
+    "T_rgb_from_depth_rotation": [
+      [
+        0.999999,
+        0.000952568,
+        0.0010665
+      ],
+      [
+        -0.00105482,
+        0.994948,
+        0.10039
+      ],
+      [
+        -0.000965488,
+        -0.100391,
+        0.994948
+      ]
+    ],
+    "T_rgb_from_depth_translation_m": [
+      -0.0335589,
+      -0.00214962,
+      -0.00945180549381777
+    ],
+    "status": "User-approved empirical offset; not official recalibration"
+  },
+  "registration_sdk": {
+    "sdk": "Microsoft.Azure.Kinect.Sensor",
+    "version": "1.4.1",
+    "dll_sha256": "1c6bf5aabb62d4907aa66da92448aebe5f1b8b59f81f2099bc1d9f989cf44cb6",
+    "interfaces": [
+      "k4a_calibration_get_from_raw",
+      "k4a_transformation_create",
+      "k4a_transformation_depth_image_to_color_camera"
+    ],
+    "depth_mode": "WFOV_2X2BINNED",
+    "color_resolution": "1080P",
+    "calibration_adaptation": "Source RGB is already undistorted: config RGB K, zero color distortion, color metric radius 10; depth factory calibration parsed by SDK; config DepthToColor plus approved optical-Z offset; SDK translations in mm.",
+    "T_rgb_from_depth_mm": [
+      [
+        0.999999,
+        0.000952568,
+        0.0010665,
+        -33.5589
+      ],
+      [
+        -0.00105482,
+        0.994948,
+        0.10039,
+        -2.14962
+      ],
+      [
+        -0.000965488,
+        -0.100391,
+        0.994948,
+        -9.45180549381777
+      ],
+      [
+        0.0,
+        0.0,
+        0.0,
+        1.0
+      ]
+    ],
+    "native_depth_K": [
+      [
+        252.52279663085938,
+        0.0,
+        257.9237060546875
+      ],
+      [
+        0.0,
+        252.60423278808594,
+        251.84109497070312
+      ],
+      [
+        0.0,
+        0.0,
+        1.0
+      ]
+    ],
+    "calls": {
+      "depth_to_color": 377,
+      "color_to_depth": 0
+    },
+    "depth_output": "Official uint16 mm, zero invalid; losslessly normalized to float32 metres."
+  },
+  "processing": {
+    "adapter": "egopressure-official-k4a-rgb-only-full/1.0",
+    "contract_repo_revision": "9e89722fc3935d1e16aeb9eeeffa4cd5c9d93383",
+    "scope": "All 1344 selected ego d sequences; original RGB plus official SDK depth-to-RGB",
+    "validation_policy": "All frame IDs, tensor validity and packet counts; deterministic representative decode checks, not full decode or SDK replay"
+  }
+}

EgoPressure/shared-v1.1/data/p_019_grasp-edge_uncurled_thumb-down_5x_right/cam-d-rgb/metainfo.json

@@ -0,0 +1,481 @@
+{
+  "schema_version": "hand-warehouse/1.1",
+  "dataset_id": "EgoPressure",
+  "release_id": "shared-v1.1",
+  "sequence_id": "p_019_grasp-edge_uncurled_thumb-down_5x_right",
+  "camera_id": "cam-d-rgb",
+  "physical_camera_id": "d",
+  "view_type": "ego",
+  "sample_type": "video",
+  "num_frames": 332,
+  "hand_slots": [
+    "left",
+    "right"
+  ],
+  "sampling": {
+    "type": "dense",
+    "source_fps": 30,
+    "frames_ref": "camera/frames.jsonl",
+    "policy": "Every source frame, exact frame-ID joins, no true timestamps available; presentation 30Hz only"
+  },
+  "camera": {
+    "model": "pinhole",
+    "schema_id": "camera/pinhole/v1",
+    "image_size": [
+      1920,
+      1080
+    ]
+  },
+  "components": {
+    "rgb": {
+      "available": true,
+      "schema_id": "rgb/uint8/v1",
+      "data_ref": "observation/rgb.mkv",
+      "valid_ref": "observation/rgb_valid.pt",
+      "axes": [
+        "T",
+        "H",
+        "W",
+        "C"
+      ],
+      "valid_axes": [
+        "T"
+      ],
+      "storage": {
+        "profile": "mkv/mjpeg_copy/v1",
+        "codec": "mjpeg",
+        "pixel_format": "yuvj420p",
+        "encoded_size": [
+          1920,
+          1080
+        ],
+        "source_lossy": true,
+        "decoder": {
+          "name": "PyAV/libavcodec",
+          "pyav_version": "12.3.0",
+          "library_versions": {
+            "libavutil": [
+              58,
+              29,
+              100
+            ],
+            "libavcodec": [
+              60,
+              31,
+              102
+            ],
+            "libavformat": [
+              60,
+              16,
+              100
+            ],
+            "libavdevice": [
+              60,
+              3,
+              100
+            ],
+            "libavfilter": [
+              9,
+              12,
+              100
+            ],
+            "libswscale": [
+              7,
+              5,
+              100
+            ],
+            "libswresample": [
+              4,
+              12,
+              100
+            ]
+          },
+          "pixel_interpretation": "rgb24 for RGB; native integer planes for depth and bool; no autorotation"
+        },
+        "encoder_parameters": {
+          "stream_copy": true
+        }
+      },
+      "provenance": {
+        "kind": "format_conversion",
+        "method": "Original JPEG packets copied, no resize/crop/re-encoding"
+      },
+      "source_locator": {
+        "file": "raw/data/data/p_019/p_019_grasp-edge_uncurled_thumb-down_5x_right/cam-d.color.parquet",
+        "record_key": "frame"
+      }
+    },
+    "depth": {
+      "available": true,
+      "schema_id": "depth/z_float32/v1",
+      "data_ref": "observation/depth.mkv",
+      "valid_ref": "observation/depth_valid.mkv",
+      "axes": [
+        "T",
+        "H",
+        "W"
+      ],
+      "valid_axes": [
+        "T",
+        "H",
+        "W"
+      ],
+      "depth_type": "z",
+      "depth_origin": "measured",
+      "unit": "m",
+      "storage": {
+        "profile": "mkv/depth_float32_bits/v1",
+        "codec": "ffv1",
+        "pixel_format": "gray16le",
+        "encoded_size": [
+          3840,
+          1080
+        ],
+        "source_lossy": false,
+        "decoder": {
+          "name": "PyAV/libavcodec",
+          "pyav_version": "12.3.0",
+          "library_versions": {
+            "libavutil": [
+              58,
+              29,
+              100
+            ],
+            "libavcodec": [
+              60,
+              31,
+              102
+            ],
+            "libavformat": [
+              60,
+              16,
+              100
+            ],
+            "libavdevice": [
+              60,
+              3,
+              100
+            ],
+            "libavfilter": [
+              9,
+              12,
+              100
+            ],
+            "libswscale": [
+              7,
+              5,
+              100
+            ],
+            "libswresample": [
+              4,
+              12,
+              100
+            ]
+          },
+          "pixel_interpretation": "rgb24 for RGB; native integer planes for depth and bool; no autorotation"
+        },
+        "encoder_parameters": {
+          "name": "ffv1",
+          "level": 3,
+          "coder": 1,
+          "context": 1,
+          "slicecrc": 1,
+          "slices": 4
+        }
+      },
+      "valid_storage": {
+        "profile": "mkv/bool_ffv1/v1",
+        "codec": "ffv1",
+        "pixel_format": "gray",
+        "encoded_size": [
+          1920,
+          1080
+        ],
+        "source_lossy": false,
+        "decoder": {
+          "name": "PyAV/libavcodec",
+          "pyav_version": "12.3.0",
+          "library_versions": {
+            "libavutil": [
+              58,
+              29,
+              100
+            ],
+            "libavcodec": [
+              60,
+              31,
+              102
+            ],
+            "libavformat": [
+              60,
+              16,
+              100
+            ],
+            "libavdevice": [
+              60,
+              3,
+              100
+            ],
+            "libavfilter": [
+              9,
+              12,
+              100
+            ],
+            "libswscale": [
+              7,
+              5,
+              100
+            ],
+            "libswresample": [
+              4,
+              12,
+              100
+            ]
+          },
+          "pixel_interpretation": "rgb24 for RGB; native integer planes for depth and bool; no autorotation"
+        },
+        "encoder_parameters": {
+          "name": "ffv1",
+          "level": 3,
+          "coder": 1,
+          "context": 1,
+          "slicecrc": 1,
+          "slices": 4
+        }
+      },
+      "provenance": {
+        "kind": "derived",
+        "method": "Official k4a_transformation_depth_image_to_color_camera surface rasterization; SDK uint16 millimetres normalized to float32 metres; zero invalid. No external hole filling."
+      },
+      "source_locator": {
+        "file": "raw/data/data/p_019/p_019_grasp-edge_uncurled_thumb-down_5x_right/cam-d.depth.parquet",
+        "record_key": "frame"
+      }
+    },
+    "intrinsic": {
+      "available": true,
+      "schema_id": "camera/pinhole/v1",
+      "data_ref": "camera/camera.pt",
+      "data_key": [
+        "intrinsic",
+        "data"
+      ],
+      "valid_ref": "camera/camera.pt",
+      "valid_key": [
+        "intrinsic",
+        "valid"
+      ],
+      "axes": [
+        "T",
+        "row",
+        "column"
+      ],
+      "valid_axes": [
+        "T"
+      ],
+      "storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "valid_storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "provenance": {
+        "kind": "format_conversion",
+        "method": "Original undistorted color K"
+      },
+      "source_locator": {
+        "file": "raw/data/configs/p_019/p_019_grasp-edge_uncurled_thumb-down_5x_right.json"
+      }
+    },
+    "extrinsic": {
+      "available": true,
+      "schema_id": "transform/se3/v1",
+      "data_ref": "camera/camera.pt",
+      "data_key": [
+        "extrinsic",
+        "data"
+      ],
+      "valid_ref": "camera/camera.pt",
+      "valid_key": [
+        "extrinsic",
+        "valid"
+      ],
+      "axes": [
+        "T",
+        "row",
+        "column"
+      ],
+      "valid_axes": [
+        "T"
+      ],
+      "storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "valid_storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "provenance": {
+        "kind": "format_conversion",
+        "method": "Source ego_R/ego_T unchanged"
+      },
+      "source_locator": {
+        "file": "raw/data/data/p_019/p_019_grasp-edge_uncurled_thumb-down_5x_right/annotation.parquet",
+        "record_key": "frame"
+      },
+      "direction": "T_camera_from_world",
+      "convention": "T_camera_from_world"
+    },
+    "mano": {
+      "available": true,
+      "schema_id": "mano/axis_angle61/v1",
+      "data_ref": "annotation/mano/mano.pt",
+      "data_key": [
+        "motion",
+        "data"
+      ],
+      "valid_ref": "annotation/mano/mano.pt",
+      "valid_key": [
+        "motion",
+        "valid"
+      ],
+      "axes": [
+        "T",
+        "hand",
+        "parameter"
+      ],
+      "valid_axes": [
+        "T",
+        "hand"
+      ],
+      "storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "valid_storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "provenance": {
+        "kind": "format_conversion",
+        "method": "Standard MANO, no per-vertex refinement; camera root rotation composed; additive translation compensated around shaped root joint"
+      },
+      "source_locator": {
+        "file": "raw/data/data/p_019/p_019_grasp-edge_uncurled_thumb-down_5x_right/annotation.parquet",
+        "record_key": "frame"
+      },
+      "model_profile": {
+        "model": "MANO",
+        "vertices": 778,
+        "flat_hand_mean": true,
+        "use_pca": false,
+        "left_shapedirs_fix_applied": false,
+        "per_vertex_refinement": false,
+        "assets_sha256": {
+          "left": "c4022f7083f2ca7c78b2b3d595abbab52debd32b09d372b16923a801f0ea6a30",
+          "right": "45d60aa3b27ef9107a7afd4e00808f307fd91111e1cfa35afd5c4a62de264767"
+        }
+      },
+      "valid_policy": {
+        "source_annotation_present": "Source has_annotation=true and exact frame ID join",
+        "numerical_checks": "Finite shapes and geometry checked; failure aborts",
+        "official_qa": "not available; visibility not inferred"
+      }
+    }
+  },
+  "split": "unknown",
+  "split_reason": "Official split not provided in selected source",
+  "source_revision": "cea37de99be44078f0e4b9796fd6f614d9110dc2",
+  "calibration": {
+    "baseline": "config DepthToColor",
+    "offset_rgb_optical_z_mm": -6.84644549381777,
+    "offset_estimation": "100-sequence participant-balanced pad fit",
+    "T_rgb_from_depth_rotation": [
+      [
+        0.999999,
+        0.000952568,
+        0.0010665
+      ],
+      [
+        -0.00105482,
+        0.994948,
+        0.10039
+      ],
+      [
+        -0.000965488,
+        -0.100391,
+        0.994948
+      ]
+    ],
+    "T_rgb_from_depth_translation_m": [
+      -0.0335589,
+      -0.00214962,
+      -0.00945180549381777
+    ],
+    "status": "User-approved empirical offset; not official recalibration"
+  },
+  "registration_sdk": {
+    "sdk": "Microsoft.Azure.Kinect.Sensor",
+    "version": "1.4.1",
+    "dll_sha256": "1c6bf5aabb62d4907aa66da92448aebe5f1b8b59f81f2099bc1d9f989cf44cb6",
+    "interfaces": [
+      "k4a_calibration_get_from_raw",
+      "k4a_transformation_create",
+      "k4a_transformation_depth_image_to_color_camera"
+    ],
+    "depth_mode": "WFOV_2X2BINNED",
+    "color_resolution": "1080P",
+    "calibration_adaptation": "Source RGB is already undistorted: config RGB K, zero color distortion, color metric radius 10; depth factory calibration parsed by SDK; config DepthToColor plus approved optical-Z offset; SDK translations in mm.",
+    "T_rgb_from_depth_mm": [
+      [
+        0.999999,
+        0.000952568,
+        0.0010665,
+        -33.5589
+      ],
+      [
+        -0.00105482,
+        0.994948,
+        0.10039,
+        -2.14962
+      ],
+      [
+        -0.000965488,
+        -0.100391,
+        0.994948,
+        -9.45180549381777
+      ],
+      [
+        0.0,
+        0.0,
+        0.0,
+        1.0
+      ]
+    ],
+    "native_depth_K": [
+      [
+        252.52279663085938,
+        0.0,
+        257.9237060546875
+      ],
+      [
+        0.0,
+        252.60423278808594,
+        251.84109497070312
+      ],
+      [
+        0.0,
+        0.0,
+        1.0
+      ]
+    ],
+    "calls": {
+      "depth_to_color": 332,
+      "color_to_depth": 0
+    },
+    "depth_output": "Official uint16 mm, zero invalid; losslessly normalized to float32 metres."
+  },
+  "processing": {
+    "adapter": "egopressure-official-k4a-rgb-only-full/1.0",
+    "contract_repo_revision": "9e89722fc3935d1e16aeb9eeeffa4cd5c9d93383",
+    "scope": "All 1344 selected ego d sequences; original RGB plus official SDK depth-to-RGB",
+    "validation_policy": "All frame IDs, tensor validity and packet counts; deterministic representative decode checks, not full decode or SDK replay"
+  }
+}

EgoPressure/shared-v1.1/data/p_019_index_press_high_x5_left/cam-d-rgb/metainfo.json

@@ -0,0 +1,481 @@
+{
+  "schema_version": "hand-warehouse/1.1",
+  "dataset_id": "EgoPressure",
+  "release_id": "shared-v1.1",
+  "sequence_id": "p_019_index_press_high_x5_left",
+  "camera_id": "cam-d-rgb",
+  "physical_camera_id": "d",
+  "view_type": "ego",
+  "sample_type": "video",
+  "num_frames": 325,
+  "hand_slots": [
+    "left",
+    "right"
+  ],
+  "sampling": {
+    "type": "dense",
+    "source_fps": 30,
+    "frames_ref": "camera/frames.jsonl",
+    "policy": "Every source frame, exact frame-ID joins, no true timestamps available; presentation 30Hz only"
+  },
+  "camera": {
+    "model": "pinhole",
+    "schema_id": "camera/pinhole/v1",
+    "image_size": [
+      1920,
+      1080
+    ]
+  },
+  "components": {
+    "rgb": {
+      "available": true,
+      "schema_id": "rgb/uint8/v1",
+      "data_ref": "observation/rgb.mkv",
+      "valid_ref": "observation/rgb_valid.pt",
+      "axes": [
+        "T",
+        "H",
+        "W",
+        "C"
+      ],
+      "valid_axes": [
+        "T"
+      ],
+      "storage": {
+        "profile": "mkv/mjpeg_copy/v1",
+        "codec": "mjpeg",
+        "pixel_format": "yuvj420p",
+        "encoded_size": [
+          1920,
+          1080
+        ],
+        "source_lossy": true,
+        "decoder": {
+          "name": "PyAV/libavcodec",
+          "pyav_version": "12.3.0",
+          "library_versions": {
+            "libavutil": [
+              58,
+              29,
+              100
+            ],
+            "libavcodec": [
+              60,
+              31,
+              102
+            ],
+            "libavformat": [
+              60,
+              16,
+              100
+            ],
+            "libavdevice": [
+              60,
+              3,
+              100
+            ],
+            "libavfilter": [
+              9,
+              12,
+              100
+            ],
+            "libswscale": [
+              7,
+              5,
+              100
+            ],
+            "libswresample": [
+              4,
+              12,
+              100
+            ]
+          },
+          "pixel_interpretation": "rgb24 for RGB; native integer planes for depth and bool; no autorotation"
+        },
+        "encoder_parameters": {
+          "stream_copy": true
+        }
+      },
+      "provenance": {
+        "kind": "format_conversion",
+        "method": "Original JPEG packets copied, no resize/crop/re-encoding"
+      },
+      "source_locator": {
+        "file": "raw/data/data/p_019/p_019_index_press_high_x5_left/cam-d.color.parquet",
+        "record_key": "frame"
+      }
+    },
+    "depth": {
+      "available": true,
+      "schema_id": "depth/z_float32/v1",
+      "data_ref": "observation/depth.mkv",
+      "valid_ref": "observation/depth_valid.mkv",
+      "axes": [
+        "T",
+        "H",
+        "W"
+      ],
+      "valid_axes": [
+        "T",
+        "H",
+        "W"
+      ],
+      "depth_type": "z",
+      "depth_origin": "measured",
+      "unit": "m",
+      "storage": {
+        "profile": "mkv/depth_float32_bits/v1",
+        "codec": "ffv1",
+        "pixel_format": "gray16le",
+        "encoded_size": [
+          3840,
+          1080
+        ],
+        "source_lossy": false,
+        "decoder": {
+          "name": "PyAV/libavcodec",
+          "pyav_version": "12.3.0",
+          "library_versions": {
+            "libavutil": [
+              58,
+              29,
+              100
+            ],
+            "libavcodec": [
+              60,
+              31,
+              102
+            ],
+            "libavformat": [
+              60,
+              16,
+              100
+            ],
+            "libavdevice": [
+              60,
+              3,
+              100
+            ],
+            "libavfilter": [
+              9,
+              12,
+              100
+            ],
+            "libswscale": [
+              7,
+              5,
+              100
+            ],
+            "libswresample": [
+              4,
+              12,
+              100
+            ]
+          },
+          "pixel_interpretation": "rgb24 for RGB; native integer planes for depth and bool; no autorotation"
+        },
+        "encoder_parameters": {
+          "name": "ffv1",
+          "level": 3,
+          "coder": 1,
+          "context": 1,
+          "slicecrc": 1,
+          "slices": 4
+        }
+      },
+      "valid_storage": {
+        "profile": "mkv/bool_ffv1/v1",
+        "codec": "ffv1",
+        "pixel_format": "gray",
+        "encoded_size": [
+          1920,
+          1080
+        ],
+        "source_lossy": false,
+        "decoder": {
+          "name": "PyAV/libavcodec",
+          "pyav_version": "12.3.0",
+          "library_versions": {
+            "libavutil": [
+              58,
+              29,
+              100
+            ],
+            "libavcodec": [
+              60,
+              31,
+              102
+            ],
+            "libavformat": [
+              60,
+              16,
+              100
+            ],
+            "libavdevice": [
+              60,
+              3,
+              100
+            ],
+            "libavfilter": [
+              9,
+              12,
+              100
+            ],
+            "libswscale": [
+              7,
+              5,
+              100
+            ],
+            "libswresample": [
+              4,
+              12,
+              100
+            ]
+          },
+          "pixel_interpretation": "rgb24 for RGB; native integer planes for depth and bool; no autorotation"
+        },
+        "encoder_parameters": {
+          "name": "ffv1",
+          "level": 3,
+          "coder": 1,
+          "context": 1,
+          "slicecrc": 1,
+          "slices": 4
+        }
+      },
+      "provenance": {
+        "kind": "derived",
+        "method": "Official k4a_transformation_depth_image_to_color_camera surface rasterization; SDK uint16 millimetres normalized to float32 metres; zero invalid. No external hole filling."
+      },
+      "source_locator": {
+        "file": "raw/data/data/p_019/p_019_index_press_high_x5_left/cam-d.depth.parquet",
+        "record_key": "frame"
+      }
+    },
+    "intrinsic": {
+      "available": true,
+      "schema_id": "camera/pinhole/v1",
+      "data_ref": "camera/camera.pt",
+      "data_key": [
+        "intrinsic",
+        "data"
+      ],
+      "valid_ref": "camera/camera.pt",
+      "valid_key": [
+        "intrinsic",
+        "valid"
+      ],
+      "axes": [
+        "T",
+        "row",
+        "column"
+      ],
+      "valid_axes": [
+        "T"
+      ],
+      "storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "valid_storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "provenance": {
+        "kind": "format_conversion",
+        "method": "Original undistorted color K"
+      },
+      "source_locator": {
+        "file": "raw/data/configs/p_019/p_019_index_press_high_x5_left.json"
+      }
+    },
+    "extrinsic": {
+      "available": true,
+      "schema_id": "transform/se3/v1",
+      "data_ref": "camera/camera.pt",
+      "data_key": [
+        "extrinsic",
+        "data"
+      ],
+      "valid_ref": "camera/camera.pt",
+      "valid_key": [
+        "extrinsic",
+        "valid"
+      ],
+      "axes": [
+        "T",
+        "row",
+        "column"
+      ],
+      "valid_axes": [
+        "T"
+      ],
+      "storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "valid_storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "provenance": {
+        "kind": "format_conversion",
+        "method": "Source ego_R/ego_T unchanged"
+      },
+      "source_locator": {
+        "file": "raw/data/data/p_019/p_019_index_press_high_x5_left/annotation.parquet",
+        "record_key": "frame"
+      },
+      "direction": "T_camera_from_world",
+      "convention": "T_camera_from_world"
+    },
+    "mano": {
+      "available": true,
+      "schema_id": "mano/axis_angle61/v1",
+      "data_ref": "annotation/mano/mano.pt",
+      "data_key": [
+        "motion",
+        "data"
+      ],
+      "valid_ref": "annotation/mano/mano.pt",
+      "valid_key": [
+        "motion",
+        "valid"
+      ],
+      "axes": [
+        "T",
+        "hand",
+        "parameter"
+      ],
+      "valid_axes": [
+        "T",
+        "hand"
+      ],
+      "storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "valid_storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "provenance": {
+        "kind": "format_conversion",
+        "method": "Standard MANO, no per-vertex refinement; camera root rotation composed; additive translation compensated around shaped root joint"
+      },
+      "source_locator": {
+        "file": "raw/data/data/p_019/p_019_index_press_high_x5_left/annotation.parquet",
+        "record_key": "frame"
+      },
+      "model_profile": {
+        "model": "MANO",
+        "vertices": 778,
+        "flat_hand_mean": true,
+        "use_pca": false,
+        "left_shapedirs_fix_applied": false,
+        "per_vertex_refinement": false,
+        "assets_sha256": {
+          "left": "c4022f7083f2ca7c78b2b3d595abbab52debd32b09d372b16923a801f0ea6a30",
+          "right": "45d60aa3b27ef9107a7afd4e00808f307fd91111e1cfa35afd5c4a62de264767"
+        }
+      },
+      "valid_policy": {
+        "source_annotation_present": "Source has_annotation=true and exact frame ID join",
+        "numerical_checks": "Finite shapes and geometry checked; failure aborts",
+        "official_qa": "not available; visibility not inferred"
+      }
+    }
+  },
+  "split": "unknown",
+  "split_reason": "Official split not provided in selected source",
+  "source_revision": "cea37de99be44078f0e4b9796fd6f614d9110dc2",
+  "calibration": {
+    "baseline": "config DepthToColor",
+    "offset_rgb_optical_z_mm": -6.84644549381777,
+    "offset_estimation": "100-sequence participant-balanced pad fit",
+    "T_rgb_from_depth_rotation": [
+      [
+        0.999999,
+        0.000952568,
+        0.0010665
+      ],
+      [
+        -0.00105482,
+        0.994948,
+        0.10039
+      ],
+      [
+        -0.000965488,
+        -0.100391,
+        0.994948
+      ]
+    ],
+    "T_rgb_from_depth_translation_m": [
+      -0.0335589,
+      -0.00214962,
+      -0.00945180549381777
+    ],
+    "status": "User-approved empirical offset; not official recalibration"
+  },
+  "registration_sdk": {
+    "sdk": "Microsoft.Azure.Kinect.Sensor",
+    "version": "1.4.1",
+    "dll_sha256": "1c6bf5aabb62d4907aa66da92448aebe5f1b8b59f81f2099bc1d9f989cf44cb6",
+    "interfaces": [
+      "k4a_calibration_get_from_raw",
+      "k4a_transformation_create",
+      "k4a_transformation_depth_image_to_color_camera"
+    ],
+    "depth_mode": "WFOV_2X2BINNED",
+    "color_resolution": "1080P",
+    "calibration_adaptation": "Source RGB is already undistorted: config RGB K, zero color distortion, color metric radius 10; depth factory calibration parsed by SDK; config DepthToColor plus approved optical-Z offset; SDK translations in mm.",
+    "T_rgb_from_depth_mm": [
+      [
+        0.999999,
+        0.000952568,
+        0.0010665,
+        -33.5589
+      ],
+      [
+        -0.00105482,
+        0.994948,
+        0.10039,
+        -2.14962
+      ],
+      [
+        -0.000965488,
+        -0.100391,
+        0.994948,
+        -9.45180549381777
+      ],
+      [
+        0.0,
+        0.0,
+        0.0,
+        1.0
+      ]
+    ],
+    "native_depth_K": [
+      [
+        252.52279663085938,
+        0.0,
+        257.9237060546875
+      ],
+      [
+        0.0,
+        252.60423278808594,
+        251.84109497070312
+      ],
+      [
+        0.0,
+        0.0,
+        1.0
+      ]
+    ],
+    "calls": {
+      "depth_to_color": 325,
+      "color_to_depth": 0
+    },
+    "depth_output": "Official uint16 mm, zero invalid; losslessly normalized to float32 metres."
+  },
+  "processing": {
+    "adapter": "egopressure-official-k4a-rgb-only-full/1.0",
+    "contract_repo_revision": "9e89722fc3935d1e16aeb9eeeffa4cd5c9d93383",
+    "scope": "All 1344 selected ego d sequences; original RGB plus official SDK depth-to-RGB",
+    "validation_policy": "All frame IDs, tensor validity and packet counts; deterministic representative decode checks, not full decode or SDK replay"
+  }
+}

EgoPressure/shared-v1.1/data/p_019_index_press_high_x5_right/cam-d-rgb/metainfo.json

@@ -0,0 +1,481 @@
+{
+  "schema_version": "hand-warehouse/1.1",
+  "dataset_id": "EgoPressure",
+  "release_id": "shared-v1.1",
+  "sequence_id": "p_019_index_press_high_x5_right",
+  "camera_id": "cam-d-rgb",
+  "physical_camera_id": "d",
+  "view_type": "ego",
+  "sample_type": "video",
+  "num_frames": 281,
+  "hand_slots": [
+    "left",
+    "right"
+  ],
+  "sampling": {
+    "type": "dense",
+    "source_fps": 30,
+    "frames_ref": "camera/frames.jsonl",
+    "policy": "Every source frame, exact frame-ID joins, no true timestamps available; presentation 30Hz only"
+  },
+  "camera": {
+    "model": "pinhole",
+    "schema_id": "camera/pinhole/v1",
+    "image_size": [
+      1920,
+      1080
+    ]
+  },
+  "components": {
+    "rgb": {
+      "available": true,
+      "schema_id": "rgb/uint8/v1",
+      "data_ref": "observation/rgb.mkv",
+      "valid_ref": "observation/rgb_valid.pt",
+      "axes": [
+        "T",
+        "H",
+        "W",
+        "C"
+      ],
+      "valid_axes": [
+        "T"
+      ],
+      "storage": {
+        "profile": "mkv/mjpeg_copy/v1",
+        "codec": "mjpeg",
+        "pixel_format": "yuvj420p",
+        "encoded_size": [
+          1920,
+          1080
+        ],
+        "source_lossy": true,
+        "decoder": {
+          "name": "PyAV/libavcodec",
+          "pyav_version": "12.3.0",
+          "library_versions": {
+            "libavutil": [
+              58,
+              29,
+              100
+            ],
+            "libavcodec": [
+              60,
+              31,
+              102
+            ],
+            "libavformat": [
+              60,
+              16,
+              100
+            ],
+            "libavdevice": [
+              60,
+              3,
+              100
+            ],
+            "libavfilter": [
+              9,
+              12,
+              100
+            ],
+            "libswscale": [
+              7,
+              5,
+              100
+            ],
+            "libswresample": [
+              4,
+              12,
+              100
+            ]
+          },
+          "pixel_interpretation": "rgb24 for RGB; native integer planes for depth and bool; no autorotation"
+        },
+        "encoder_parameters": {
+          "stream_copy": true
+        }
+      },
+      "provenance": {
+        "kind": "format_conversion",
+        "method": "Original JPEG packets copied, no resize/crop/re-encoding"
+      },
+      "source_locator": {
+        "file": "raw/data/data/p_019/p_019_index_press_high_x5_right/cam-d.color.parquet",
+        "record_key": "frame"
+      }
+    },
+    "depth": {
+      "available": true,
+      "schema_id": "depth/z_float32/v1",
+      "data_ref": "observation/depth.mkv",
+      "valid_ref": "observation/depth_valid.mkv",
+      "axes": [
+        "T",
+        "H",
+        "W"
+      ],
+      "valid_axes": [
+        "T",
+        "H",
+        "W"
+      ],
+      "depth_type": "z",
+      "depth_origin": "measured",
+      "unit": "m",
+      "storage": {
+        "profile": "mkv/depth_float32_bits/v1",
+        "codec": "ffv1",
+        "pixel_format": "gray16le",
+        "encoded_size": [
+          3840,
+          1080
+        ],
+        "source_lossy": false,
+        "decoder": {
+          "name": "PyAV/libavcodec",
+          "pyav_version": "12.3.0",
+          "library_versions": {
+            "libavutil": [
+              58,
+              29,
+              100
+            ],
+            "libavcodec": [
+              60,
+              31,
+              102
+            ],
+            "libavformat": [
+              60,
+              16,
+              100
+            ],
+            "libavdevice": [
+              60,
+              3,
+              100
+            ],
+            "libavfilter": [
+              9,
+              12,
+              100
+            ],
+            "libswscale": [
+              7,
+              5,
+              100
+            ],
+            "libswresample": [
+              4,
+              12,
+              100
+            ]
+          },
+          "pixel_interpretation": "rgb24 for RGB; native integer planes for depth and bool; no autorotation"
+        },
+        "encoder_parameters": {
+          "name": "ffv1",
+          "level": 3,
+          "coder": 1,
+          "context": 1,
+          "slicecrc": 1,
+          "slices": 4
+        }
+      },
+      "valid_storage": {
+        "profile": "mkv/bool_ffv1/v1",
+        "codec": "ffv1",
+        "pixel_format": "gray",
+        "encoded_size": [
+          1920,
+          1080
+        ],
+        "source_lossy": false,
+        "decoder": {
+          "name": "PyAV/libavcodec",
+          "pyav_version": "12.3.0",
+          "library_versions": {
+            "libavutil": [
+              58,
+              29,
+              100
+            ],
+            "libavcodec": [
+              60,
+              31,
+              102
+            ],
+            "libavformat": [
+              60,
+              16,
+              100
+            ],
+            "libavdevice": [
+              60,
+              3,
+              100
+            ],
+            "libavfilter": [
+              9,
+              12,
+              100
+            ],
+            "libswscale": [
+              7,
+              5,
+              100
+            ],
+            "libswresample": [
+              4,
+              12,
+              100
+            ]
+          },
+          "pixel_interpretation": "rgb24 for RGB; native integer planes for depth and bool; no autorotation"
+        },
+        "encoder_parameters": {
+          "name": "ffv1",
+          "level": 3,
+          "coder": 1,
+          "context": 1,
+          "slicecrc": 1,
+          "slices": 4
+        }
+      },
+      "provenance": {
+        "kind": "derived",
+        "method": "Official k4a_transformation_depth_image_to_color_camera surface rasterization; SDK uint16 millimetres normalized to float32 metres; zero invalid. No external hole filling."
+      },
+      "source_locator": {
+        "file": "raw/data/data/p_019/p_019_index_press_high_x5_right/cam-d.depth.parquet",
+        "record_key": "frame"
+      }
+    },
+    "intrinsic": {
+      "available": true,
+      "schema_id": "camera/pinhole/v1",
+      "data_ref": "camera/camera.pt",
+      "data_key": [
+        "intrinsic",
+        "data"
+      ],
+      "valid_ref": "camera/camera.pt",
+      "valid_key": [
+        "intrinsic",
+        "valid"
+      ],
+      "axes": [
+        "T",
+        "row",
+        "column"
+      ],
+      "valid_axes": [
+        "T"
+      ],
+      "storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "valid_storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "provenance": {
+        "kind": "format_conversion",
+        "method": "Original undistorted color K"
+      },
+      "source_locator": {
+        "file": "raw/data/configs/p_019/p_019_index_press_high_x5_right.json"
+      }
+    },
+    "extrinsic": {
+      "available": true,
+      "schema_id": "transform/se3/v1",
+      "data_ref": "camera/camera.pt",
+      "data_key": [
+        "extrinsic",
+        "data"
+      ],
+      "valid_ref": "camera/camera.pt",
+      "valid_key": [
+        "extrinsic",
+        "valid"
+      ],
+      "axes": [
+        "T",
+        "row",
+        "column"
+      ],
+      "valid_axes": [
+        "T"
+      ],
+      "storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "valid_storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "provenance": {
+        "kind": "format_conversion",
+        "method": "Source ego_R/ego_T unchanged"
+      },
+      "source_locator": {
+        "file": "raw/data/data/p_019/p_019_index_press_high_x5_right/annotation.parquet",
+        "record_key": "frame"
+      },
+      "direction": "T_camera_from_world",
+      "convention": "T_camera_from_world"
+    },
+    "mano": {
+      "available": true,
+      "schema_id": "mano/axis_angle61/v1",
+      "data_ref": "annotation/mano/mano.pt",
+      "data_key": [
+        "motion",
+        "data"
+      ],
+      "valid_ref": "annotation/mano/mano.pt",
+      "valid_key": [
+        "motion",
+        "valid"
+      ],
+      "axes": [
+        "T",
+        "hand",
+        "parameter"
+      ],
+      "valid_axes": [
+        "T",
+        "hand"
+      ],
+      "storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "valid_storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "provenance": {
+        "kind": "format_conversion",
+        "method": "Standard MANO, no per-vertex refinement; camera root rotation composed; additive translation compensated around shaped root joint"
+      },
+      "source_locator": {
+        "file": "raw/data/data/p_019/p_019_index_press_high_x5_right/annotation.parquet",
+        "record_key": "frame"
+      },
+      "model_profile": {
+        "model": "MANO",
+        "vertices": 778,
+        "flat_hand_mean": true,
+        "use_pca": false,
+        "left_shapedirs_fix_applied": false,
+        "per_vertex_refinement": false,
+        "assets_sha256": {
+          "left": "c4022f7083f2ca7c78b2b3d595abbab52debd32b09d372b16923a801f0ea6a30",
+          "right": "45d60aa3b27ef9107a7afd4e00808f307fd91111e1cfa35afd5c4a62de264767"
+        }
+      },
+      "valid_policy": {
+        "source_annotation_present": "Source has_annotation=true and exact frame ID join",
+        "numerical_checks": "Finite shapes and geometry checked; failure aborts",
+        "official_qa": "not available; visibility not inferred"
+      }
+    }
+  },
+  "split": "unknown",
+  "split_reason": "Official split not provided in selected source",
+  "source_revision": "cea37de99be44078f0e4b9796fd6f614d9110dc2",
+  "calibration": {
+    "baseline": "config DepthToColor",
+    "offset_rgb_optical_z_mm": -6.84644549381777,
+    "offset_estimation": "100-sequence participant-balanced pad fit",
+    "T_rgb_from_depth_rotation": [
+      [
+        0.999999,
+        0.000952568,
+        0.0010665
+      ],
+      [
+        -0.00105482,
+        0.994948,
+        0.10039
+      ],
+      [
+        -0.000965488,
+        -0.100391,
+        0.994948
+      ]
+    ],
+    "T_rgb_from_depth_translation_m": [
+      -0.0335589,
+      -0.00214962,
+      -0.00945180549381777
+    ],
+    "status": "User-approved empirical offset; not official recalibration"
+  },
+  "registration_sdk": {
+    "sdk": "Microsoft.Azure.Kinect.Sensor",
+    "version": "1.4.1",
+    "dll_sha256": "1c6bf5aabb62d4907aa66da92448aebe5f1b8b59f81f2099bc1d9f989cf44cb6",
+    "interfaces": [
+      "k4a_calibration_get_from_raw",
+      "k4a_transformation_create",
+      "k4a_transformation_depth_image_to_color_camera"
+    ],
+    "depth_mode": "WFOV_2X2BINNED",
+    "color_resolution": "1080P",
+    "calibration_adaptation": "Source RGB is already undistorted: config RGB K, zero color distortion, color metric radius 10; depth factory calibration parsed by SDK; config DepthToColor plus approved optical-Z offset; SDK translations in mm.",
+    "T_rgb_from_depth_mm": [
+      [
+        0.999999,
+        0.000952568,
+        0.0010665,
+        -33.5589
+      ],
+      [
+        -0.00105482,
+        0.994948,
+        0.10039,
+        -2.14962
+      ],
+      [
+        -0.000965488,
+        -0.100391,
+        0.994948,
+        -9.45180549381777
+      ],
+      [
+        0.0,
+        0.0,
+        0.0,
+        1.0
+      ]
+    ],
+    "native_depth_K": [
+      [
+        252.52279663085938,
+        0.0,
+        257.9237060546875
+      ],
+      [
+        0.0,
+        252.60423278808594,
+        251.84109497070312
+      ],
+      [
+        0.0,
+        0.0,
+        1.0
+      ]
+    ],
+    "calls": {
+      "depth_to_color": 281,
+      "color_to_depth": 0
+    },
+    "depth_output": "Official uint16 mm, zero invalid; losslessly normalized to float32 metres."
+  },
+  "processing": {
+    "adapter": "egopressure-official-k4a-rgb-only-full/1.0",
+    "contract_repo_revision": "9e89722fc3935d1e16aeb9eeeffa4cd5c9d93383",
+    "scope": "All 1344 selected ego d sequences; original RGB plus official SDK depth-to-RGB",
+    "validation_policy": "All frame IDs, tensor validity and packet counts; deterministic representative decode checks, not full decode or SDK replay"
+  }
+}

EgoPressure/shared-v1.1/data/p_019_index_press_low_x5_left/cam-d-rgb/metainfo.json

@@ -0,0 +1,481 @@
+{
+  "schema_version": "hand-warehouse/1.1",
+  "dataset_id": "EgoPressure",
+  "release_id": "shared-v1.1",
+  "sequence_id": "p_019_index_press_low_x5_left",
+  "camera_id": "cam-d-rgb",
+  "physical_camera_id": "d",
+  "view_type": "ego",
+  "sample_type": "video",
+  "num_frames": 315,
+  "hand_slots": [
+    "left",
+    "right"
+  ],
+  "sampling": {
+    "type": "dense",
+    "source_fps": 30,
+    "frames_ref": "camera/frames.jsonl",
+    "policy": "Every source frame, exact frame-ID joins, no true timestamps available; presentation 30Hz only"
+  },
+  "camera": {
+    "model": "pinhole",
+    "schema_id": "camera/pinhole/v1",
+    "image_size": [
+      1920,
+      1080
+    ]
+  },
+  "components": {
+    "rgb": {
+      "available": true,
+      "schema_id": "rgb/uint8/v1",
+      "data_ref": "observation/rgb.mkv",
+      "valid_ref": "observation/rgb_valid.pt",
+      "axes": [
+        "T",
+        "H",
+        "W",
+        "C"
+      ],
+      "valid_axes": [
+        "T"
+      ],
+      "storage": {
+        "profile": "mkv/mjpeg_copy/v1",
+        "codec": "mjpeg",
+        "pixel_format": "yuvj420p",
+        "encoded_size": [
+          1920,
+          1080
+        ],
+        "source_lossy": true,
+        "decoder": {
+          "name": "PyAV/libavcodec",
+          "pyav_version": "12.3.0",
+          "library_versions": {
+            "libavutil": [
+              58,
+              29,
+              100
+            ],
+            "libavcodec": [
+              60,
+              31,
+              102
+            ],
+            "libavformat": [
+              60,
+              16,
+              100
+            ],
+            "libavdevice": [
+              60,
+              3,
+              100
+            ],
+            "libavfilter": [
+              9,
+              12,
+              100
+            ],
+            "libswscale": [
+              7,
+              5,
+              100
+            ],
+            "libswresample": [
+              4,
+              12,
+              100
+            ]
+          },
+          "pixel_interpretation": "rgb24 for RGB; native integer planes for depth and bool; no autorotation"
+        },
+        "encoder_parameters": {
+          "stream_copy": true
+        }
+      },
+      "provenance": {
+        "kind": "format_conversion",
+        "method": "Original JPEG packets copied, no resize/crop/re-encoding"
+      },
+      "source_locator": {
+        "file": "raw/data/data/p_019/p_019_index_press_low_x5_left/cam-d.color.parquet",
+        "record_key": "frame"
+      }
+    },
+    "depth": {
+      "available": true,
+      "schema_id": "depth/z_float32/v1",
+      "data_ref": "observation/depth.mkv",
+      "valid_ref": "observation/depth_valid.mkv",
+      "axes": [
+        "T",
+        "H",
+        "W"
+      ],
+      "valid_axes": [
+        "T",
+        "H",
+        "W"
+      ],
+      "depth_type": "z",
+      "depth_origin": "measured",
+      "unit": "m",
+      "storage": {
+        "profile": "mkv/depth_float32_bits/v1",
+        "codec": "ffv1",
+        "pixel_format": "gray16le",
+        "encoded_size": [
+          3840,
+          1080
+        ],
+        "source_lossy": false,
+        "decoder": {
+          "name": "PyAV/libavcodec",
+          "pyav_version": "12.3.0",
+          "library_versions": {
+            "libavutil": [
+              58,
+              29,
+              100
+            ],
+            "libavcodec": [
+              60,
+              31,
+              102
+            ],
+            "libavformat": [
+              60,
+              16,
+              100
+            ],
+            "libavdevice": [
+              60,
+              3,
+              100
+            ],
+            "libavfilter": [
+              9,
+              12,
+              100
+            ],
+            "libswscale": [
+              7,
+              5,
+              100
+            ],
+            "libswresample": [
+              4,
+              12,
+              100
+            ]
+          },
+          "pixel_interpretation": "rgb24 for RGB; native integer planes for depth and bool; no autorotation"
+        },
+        "encoder_parameters": {
+          "name": "ffv1",
+          "level": 3,
+          "coder": 1,
+          "context": 1,
+          "slicecrc": 1,
+          "slices": 4
+        }
+      },
+      "valid_storage": {
+        "profile": "mkv/bool_ffv1/v1",
+        "codec": "ffv1",
+        "pixel_format": "gray",
+        "encoded_size": [
+          1920,
+          1080
+        ],
+        "source_lossy": false,
+        "decoder": {
+          "name": "PyAV/libavcodec",
+          "pyav_version": "12.3.0",
+          "library_versions": {
+            "libavutil": [
+              58,
+              29,
+              100
+            ],
+            "libavcodec": [
+              60,
+              31,
+              102
+            ],
+            "libavformat": [
+              60,
+              16,
+              100
+            ],
+            "libavdevice": [
+              60,
+              3,
+              100
+            ],
+            "libavfilter": [
+              9,
+              12,
+              100
+            ],
+            "libswscale": [
+              7,
+              5,
+              100
+            ],
+            "libswresample": [
+              4,
+              12,
+              100
+            ]
+          },
+          "pixel_interpretation": "rgb24 for RGB; native integer planes for depth and bool; no autorotation"
+        },
+        "encoder_parameters": {
+          "name": "ffv1",
+          "level": 3,
+          "coder": 1,
+          "context": 1,
+          "slicecrc": 1,
+          "slices": 4
+        }
+      },
+      "provenance": {
+        "kind": "derived",
+        "method": "Official k4a_transformation_depth_image_to_color_camera surface rasterization; SDK uint16 millimetres normalized to float32 metres; zero invalid. No external hole filling."
+      },
+      "source_locator": {
+        "file": "raw/data/data/p_019/p_019_index_press_low_x5_left/cam-d.depth.parquet",
+        "record_key": "frame"
+      }
+    },
+    "intrinsic": {
+      "available": true,
+      "schema_id": "camera/pinhole/v1",
+      "data_ref": "camera/camera.pt",
+      "data_key": [
+        "intrinsic",
+        "data"
+      ],
+      "valid_ref": "camera/camera.pt",
+      "valid_key": [
+        "intrinsic",
+        "valid"
+      ],
+      "axes": [
+        "T",
+        "row",
+        "column"
+      ],
+      "valid_axes": [
+        "T"
+      ],
+      "storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "valid_storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "provenance": {
+        "kind": "format_conversion",
+        "method": "Original undistorted color K"
+      },
+      "source_locator": {
+        "file": "raw/data/configs/p_019/p_019_index_press_low_x5_left.json"
+      }
+    },
+    "extrinsic": {
+      "available": true,
+      "schema_id": "transform/se3/v1",
+      "data_ref": "camera/camera.pt",
+      "data_key": [
+        "extrinsic",
+        "data"
+      ],
+      "valid_ref": "camera/camera.pt",
+      "valid_key": [
+        "extrinsic",
+        "valid"
+      ],
+      "axes": [
+        "T",
+        "row",
+        "column"
+      ],
+      "valid_axes": [
+        "T"
+      ],
+      "storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "valid_storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "provenance": {
+        "kind": "format_conversion",
+        "method": "Source ego_R/ego_T unchanged"
+      },
+      "source_locator": {
+        "file": "raw/data/data/p_019/p_019_index_press_low_x5_left/annotation.parquet",
+        "record_key": "frame"
+      },
+      "direction": "T_camera_from_world",
+      "convention": "T_camera_from_world"
+    },
+    "mano": {
+      "available": true,
+      "schema_id": "mano/axis_angle61/v1",
+      "data_ref": "annotation/mano/mano.pt",
+      "data_key": [
+        "motion",
+        "data"
+      ],
+      "valid_ref": "annotation/mano/mano.pt",
+      "valid_key": [
+        "motion",
+        "valid"
+      ],
+      "axes": [
+        "T",
+        "hand",
+        "parameter"
+      ],
+      "valid_axes": [
+        "T",
+        "hand"
+      ],
+      "storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "valid_storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "provenance": {
+        "kind": "format_conversion",
+        "method": "Standard MANO, no per-vertex refinement; camera root rotation composed; additive translation compensated around shaped root joint"
+      },
+      "source_locator": {
+        "file": "raw/data/data/p_019/p_019_index_press_low_x5_left/annotation.parquet",
+        "record_key": "frame"
+      },
+      "model_profile": {
+        "model": "MANO",
+        "vertices": 778,
+        "flat_hand_mean": true,
+        "use_pca": false,
+        "left_shapedirs_fix_applied": false,
+        "per_vertex_refinement": false,
+        "assets_sha256": {
+          "left": "c4022f7083f2ca7c78b2b3d595abbab52debd32b09d372b16923a801f0ea6a30",
+          "right": "45d60aa3b27ef9107a7afd4e00808f307fd91111e1cfa35afd5c4a62de264767"
+        }
+      },
+      "valid_policy": {
+        "source_annotation_present": "Source has_annotation=true and exact frame ID join",
+        "numerical_checks": "Finite shapes and geometry checked; failure aborts",
+        "official_qa": "not available; visibility not inferred"
+      }
+    }
+  },
+  "split": "unknown",
+  "split_reason": "Official split not provided in selected source",
+  "source_revision": "cea37de99be44078f0e4b9796fd6f614d9110dc2",
+  "calibration": {
+    "baseline": "config DepthToColor",
+    "offset_rgb_optical_z_mm": -6.84644549381777,
+    "offset_estimation": "100-sequence participant-balanced pad fit",
+    "T_rgb_from_depth_rotation": [
+      [
+        0.999999,
+        0.000952568,
+        0.0010665
+      ],
+      [
+        -0.00105482,
+        0.994948,
+        0.10039
+      ],
+      [
+        -0.000965488,
+        -0.100391,
+        0.994948
+      ]
+    ],
+    "T_rgb_from_depth_translation_m": [
+      -0.0335589,
+      -0.00214962,
+      -0.00945180549381777
+    ],
+    "status": "User-approved empirical offset; not official recalibration"
+  },
+  "registration_sdk": {
+    "sdk": "Microsoft.Azure.Kinect.Sensor",
+    "version": "1.4.1",
+    "dll_sha256": "1c6bf5aabb62d4907aa66da92448aebe5f1b8b59f81f2099bc1d9f989cf44cb6",
+    "interfaces": [
+      "k4a_calibration_get_from_raw",
+      "k4a_transformation_create",
+      "k4a_transformation_depth_image_to_color_camera"
+    ],
+    "depth_mode": "WFOV_2X2BINNED",
+    "color_resolution": "1080P",
+    "calibration_adaptation": "Source RGB is already undistorted: config RGB K, zero color distortion, color metric radius 10; depth factory calibration parsed by SDK; config DepthToColor plus approved optical-Z offset; SDK translations in mm.",
+    "T_rgb_from_depth_mm": [
+      [
+        0.999999,
+        0.000952568,
+        0.0010665,
+        -33.5589
+      ],
+      [
+        -0.00105482,
+        0.994948,
+        0.10039,
+        -2.14962
+      ],
+      [
+        -0.000965488,
+        -0.100391,
+        0.994948,
+        -9.45180549381777
+      ],
+      [
+        0.0,
+        0.0,
+        0.0,
+        1.0
+      ]
+    ],
+    "native_depth_K": [
+      [
+        252.52279663085938,
+        0.0,
+        257.9237060546875
+      ],
+      [
+        0.0,
+        252.60423278808594,
+        251.84109497070312
+      ],
+      [
+        0.0,
+        0.0,
+        1.0
+      ]
+    ],
+    "calls": {
+      "depth_to_color": 315,
+      "color_to_depth": 0
+    },
+    "depth_output": "Official uint16 mm, zero invalid; losslessly normalized to float32 metres."
+  },
+  "processing": {
+    "adapter": "egopressure-official-k4a-rgb-only-full/1.0",
+    "contract_repo_revision": "9e89722fc3935d1e16aeb9eeeffa4cd5c9d93383",
+    "scope": "All 1344 selected ego d sequences; original RGB plus official SDK depth-to-RGB",
+    "validation_policy": "All frame IDs, tensor validity and packet counts; deterministic representative decode checks, not full decode or SDK replay"
+  }
+}

EgoPressure/shared-v1.1/data/p_019_index_press_low_x5_right/cam-d-rgb/metainfo.json

@@ -0,0 +1,481 @@
+{
+  "schema_version": "hand-warehouse/1.1",
+  "dataset_id": "EgoPressure",
+  "release_id": "shared-v1.1",
+  "sequence_id": "p_019_index_press_low_x5_right",
+  "camera_id": "cam-d-rgb",
+  "physical_camera_id": "d",
+  "view_type": "ego",
+  "sample_type": "video",
+  "num_frames": 287,
+  "hand_slots": [
+    "left",
+    "right"
+  ],
+  "sampling": {
+    "type": "dense",
+    "source_fps": 30,
+    "frames_ref": "camera/frames.jsonl",
+    "policy": "Every source frame, exact frame-ID joins, no true timestamps available; presentation 30Hz only"
+  },
+  "camera": {
+    "model": "pinhole",
+    "schema_id": "camera/pinhole/v1",
+    "image_size": [
+      1920,
+      1080
+    ]
+  },
+  "components": {
+    "rgb": {
+      "available": true,
+      "schema_id": "rgb/uint8/v1",
+      "data_ref": "observation/rgb.mkv",
+      "valid_ref": "observation/rgb_valid.pt",
+      "axes": [
+        "T",
+        "H",
+        "W",
+        "C"
+      ],
+      "valid_axes": [
+        "T"
+      ],
+      "storage": {
+        "profile": "mkv/mjpeg_copy/v1",
+        "codec": "mjpeg",
+        "pixel_format": "yuvj420p",
+        "encoded_size": [
+          1920,
+          1080
+        ],
+        "source_lossy": true,
+        "decoder": {
+          "name": "PyAV/libavcodec",
+          "pyav_version": "12.3.0",
+          "library_versions": {
+            "libavutil": [
+              58,
+              29,
+              100
+            ],
+            "libavcodec": [
+              60,
+              31,
+              102
+            ],
+            "libavformat": [
+              60,
+              16,
+              100
+            ],
+            "libavdevice": [
+              60,
+              3,
+              100
+            ],
+            "libavfilter": [
+              9,
+              12,
+              100
+            ],
+            "libswscale": [
+              7,
+              5,
+              100
+            ],
+            "libswresample": [
+              4,
+              12,
+              100
+            ]
+          },
+          "pixel_interpretation": "rgb24 for RGB; native integer planes for depth and bool; no autorotation"
+        },
+        "encoder_parameters": {
+          "stream_copy": true
+        }
+      },
+      "provenance": {
+        "kind": "format_conversion",
+        "method": "Original JPEG packets copied, no resize/crop/re-encoding"
+      },
+      "source_locator": {
+        "file": "raw/data/data/p_019/p_019_index_press_low_x5_right/cam-d.color.parquet",
+        "record_key": "frame"
+      }
+    },
+    "depth": {
+      "available": true,
+      "schema_id": "depth/z_float32/v1",
+      "data_ref": "observation/depth.mkv",
+      "valid_ref": "observation/depth_valid.mkv",
+      "axes": [
+        "T",
+        "H",
+        "W"
+      ],
+      "valid_axes": [
+        "T",
+        "H",
+        "W"
+      ],
+      "depth_type": "z",
+      "depth_origin": "measured",
+      "unit": "m",
+      "storage": {
+        "profile": "mkv/depth_float32_bits/v1",
+        "codec": "ffv1",
+        "pixel_format": "gray16le",
+        "encoded_size": [
+          3840,
+          1080
+        ],
+        "source_lossy": false,
+        "decoder": {
+          "name": "PyAV/libavcodec",
+          "pyav_version": "12.3.0",
+          "library_versions": {
+            "libavutil": [
+              58,
+              29,
+              100
+            ],
+            "libavcodec": [
+              60,
+              31,
+              102
+            ],
+            "libavformat": [
+              60,
+              16,
+              100
+            ],
+            "libavdevice": [
+              60,
+              3,
+              100
+            ],
+            "libavfilter": [
+              9,
+              12,
+              100
+            ],
+            "libswscale": [
+              7,
+              5,
+              100
+            ],
+            "libswresample": [
+              4,
+              12,
+              100
+            ]
+          },
+          "pixel_interpretation": "rgb24 for RGB; native integer planes for depth and bool; no autorotation"
+        },
+        "encoder_parameters": {
+          "name": "ffv1",
+          "level": 3,
+          "coder": 1,
+          "context": 1,
+          "slicecrc": 1,
+          "slices": 4
+        }
+      },
+      "valid_storage": {
+        "profile": "mkv/bool_ffv1/v1",
+        "codec": "ffv1",
+        "pixel_format": "gray",
+        "encoded_size": [
+          1920,
+          1080
+        ],
+        "source_lossy": false,
+        "decoder": {
+          "name": "PyAV/libavcodec",
+          "pyav_version": "12.3.0",
+          "library_versions": {
+            "libavutil": [
+              58,
+              29,
+              100
+            ],
+            "libavcodec": [
+              60,
+              31,
+              102
+            ],
+            "libavformat": [
+              60,
+              16,
+              100
+            ],
+            "libavdevice": [
+              60,
+              3,
+              100
+            ],
+            "libavfilter": [
+              9,
+              12,
+              100
+            ],
+            "libswscale": [
+              7,
+              5,
+              100
+            ],
+            "libswresample": [
+              4,
+              12,
+              100
+            ]
+          },
+          "pixel_interpretation": "rgb24 for RGB; native integer planes for depth and bool; no autorotation"
+        },
+        "encoder_parameters": {
+          "name": "ffv1",
+          "level": 3,
+          "coder": 1,
+          "context": 1,
+          "slicecrc": 1,
+          "slices": 4
+        }
+      },
+      "provenance": {
+        "kind": "derived",
+        "method": "Official k4a_transformation_depth_image_to_color_camera surface rasterization; SDK uint16 millimetres normalized to float32 metres; zero invalid. No external hole filling."
+      },
+      "source_locator": {
+        "file": "raw/data/data/p_019/p_019_index_press_low_x5_right/cam-d.depth.parquet",
+        "record_key": "frame"
+      }
+    },
+    "intrinsic": {
+      "available": true,
+      "schema_id": "camera/pinhole/v1",
+      "data_ref": "camera/camera.pt",
+      "data_key": [
+        "intrinsic",
+        "data"
+      ],
+      "valid_ref": "camera/camera.pt",
+      "valid_key": [
+        "intrinsic",
+        "valid"
+      ],
+      "axes": [
+        "T",
+        "row",
+        "column"
+      ],
+      "valid_axes": [
+        "T"
+      ],
+      "storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "valid_storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "provenance": {
+        "kind": "format_conversion",
+        "method": "Original undistorted color K"
+      },
+      "source_locator": {
+        "file": "raw/data/configs/p_019/p_019_index_press_low_x5_right.json"
+      }
+    },
+    "extrinsic": {
+      "available": true,
+      "schema_id": "transform/se3/v1",
+      "data_ref": "camera/camera.pt",
+      "data_key": [
+        "extrinsic",
+        "data"
+      ],
+      "valid_ref": "camera/camera.pt",
+      "valid_key": [
+        "extrinsic",
+        "valid"
+      ],
+      "axes": [
+        "T",
+        "row",
+        "column"
+      ],
+      "valid_axes": [
+        "T"
+      ],
+      "storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "valid_storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "provenance": {
+        "kind": "format_conversion",
+        "method": "Source ego_R/ego_T unchanged"
+      },
+      "source_locator": {
+        "file": "raw/data/data/p_019/p_019_index_press_low_x5_right/annotation.parquet",
+        "record_key": "frame"
+      },
+      "direction": "T_camera_from_world",
+      "convention": "T_camera_from_world"
+    },
+    "mano": {
+      "available": true,
+      "schema_id": "mano/axis_angle61/v1",
+      "data_ref": "annotation/mano/mano.pt",
+      "data_key": [
+        "motion",
+        "data"
+      ],
+      "valid_ref": "annotation/mano/mano.pt",
+      "valid_key": [
+        "motion",
+        "valid"
+      ],
+      "axes": [
+        "T",
+        "hand",
+        "parameter"
+      ],
+      "valid_axes": [
+        "T",
+        "hand"
+      ],
+      "storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "valid_storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "provenance": {
+        "kind": "format_conversion",
+        "method": "Standard MANO, no per-vertex refinement; camera root rotation composed; additive translation compensated around shaped root joint"
+      },
+      "source_locator": {
+        "file": "raw/data/data/p_019/p_019_index_press_low_x5_right/annotation.parquet",
+        "record_key": "frame"
+      },
+      "model_profile": {
+        "model": "MANO",
+        "vertices": 778,
+        "flat_hand_mean": true,
+        "use_pca": false,
+        "left_shapedirs_fix_applied": false,
+        "per_vertex_refinement": false,
+        "assets_sha256": {
+          "left": "c4022f7083f2ca7c78b2b3d595abbab52debd32b09d372b16923a801f0ea6a30",
+          "right": "45d60aa3b27ef9107a7afd4e00808f307fd91111e1cfa35afd5c4a62de264767"
+        }
+      },
+      "valid_policy": {
+        "source_annotation_present": "Source has_annotation=true and exact frame ID join",
+        "numerical_checks": "Finite shapes and geometry checked; failure aborts",
+        "official_qa": "not available; visibility not inferred"
+      }
+    }
+  },
+  "split": "unknown",
+  "split_reason": "Official split not provided in selected source",
+  "source_revision": "cea37de99be44078f0e4b9796fd6f614d9110dc2",
+  "calibration": {
+    "baseline": "config DepthToColor",
+    "offset_rgb_optical_z_mm": -6.84644549381777,
+    "offset_estimation": "100-sequence participant-balanced pad fit",
+    "T_rgb_from_depth_rotation": [
+      [
+        0.999999,
+        0.000952568,
+        0.0010665
+      ],
+      [
+        -0.00105482,
+        0.994948,
+        0.10039
+      ],
+      [
+        -0.000965488,
+        -0.100391,
+        0.994948
+      ]
+    ],
+    "T_rgb_from_depth_translation_m": [
+      -0.0335589,
+      -0.00214962,
+      -0.00945180549381777
+    ],
+    "status": "User-approved empirical offset; not official recalibration"
+  },
+  "registration_sdk": {
+    "sdk": "Microsoft.Azure.Kinect.Sensor",
+    "version": "1.4.1",
+    "dll_sha256": "1c6bf5aabb62d4907aa66da92448aebe5f1b8b59f81f2099bc1d9f989cf44cb6",
+    "interfaces": [
+      "k4a_calibration_get_from_raw",
+      "k4a_transformation_create",
+      "k4a_transformation_depth_image_to_color_camera"
+    ],
+    "depth_mode": "WFOV_2X2BINNED",
+    "color_resolution": "1080P",
+    "calibration_adaptation": "Source RGB is already undistorted: config RGB K, zero color distortion, color metric radius 10; depth factory calibration parsed by SDK; config DepthToColor plus approved optical-Z offset; SDK translations in mm.",
+    "T_rgb_from_depth_mm": [
+      [
+        0.999999,
+        0.000952568,
+        0.0010665,
+        -33.5589
+      ],
+      [
+        -0.00105482,
+        0.994948,
+        0.10039,
+        -2.14962
+      ],
+      [
+        -0.000965488,
+        -0.100391,
+        0.994948,
+        -9.45180549381777
+      ],
+      [
+        0.0,
+        0.0,
+        0.0,
+        1.0
+      ]
+    ],
+    "native_depth_K": [
+      [
+        252.52279663085938,
+        0.0,
+        257.9237060546875
+      ],
+      [
+        0.0,
+        252.60423278808594,
+        251.84109497070312
+      ],
+      [
+        0.0,
+        0.0,
+        1.0
+      ]
+    ],
+    "calls": {
+      "depth_to_color": 287,
+      "color_to_depth": 0
+    },
+    "depth_output": "Official uint16 mm, zero invalid; losslessly normalized to float32 metres."
+  },
+  "processing": {
+    "adapter": "egopressure-official-k4a-rgb-only-full/1.0",
+    "contract_repo_revision": "9e89722fc3935d1e16aeb9eeeffa4cd5c9d93383",
+    "scope": "All 1344 selected ego d sequences; original RGB plus official SDK depth-to-RGB",
+    "validation_policy": "All frame IDs, tensor validity and packet counts; deterministic representative decode checks, not full decode or SDK replay"
+  }
+}

EgoPressure/shared-v1.1/data/p_019_index_press_no-contact_x5_left/cam-d-rgb/metainfo.json

@@ -0,0 +1,481 @@
+{
+  "schema_version": "hand-warehouse/1.1",
+  "dataset_id": "EgoPressure",
+  "release_id": "shared-v1.1",
+  "sequence_id": "p_019_index_press_no-contact_x5_left",
+  "camera_id": "cam-d-rgb",
+  "physical_camera_id": "d",
+  "view_type": "ego",
+  "sample_type": "video",
+  "num_frames": 292,
+  "hand_slots": [
+    "left",
+    "right"
+  ],
+  "sampling": {
+    "type": "dense",
+    "source_fps": 30,
+    "frames_ref": "camera/frames.jsonl",
+    "policy": "Every source frame, exact frame-ID joins, no true timestamps available; presentation 30Hz only"
+  },
+  "camera": {
+    "model": "pinhole",
+    "schema_id": "camera/pinhole/v1",
+    "image_size": [
+      1920,
+      1080
+    ]
+  },
+  "components": {
+    "rgb": {
+      "available": true,
+      "schema_id": "rgb/uint8/v1",
+      "data_ref": "observation/rgb.mkv",
+      "valid_ref": "observation/rgb_valid.pt",
+      "axes": [
+        "T",
+        "H",
+        "W",
+        "C"
+      ],
+      "valid_axes": [
+        "T"
+      ],
+      "storage": {
+        "profile": "mkv/mjpeg_copy/v1",
+        "codec": "mjpeg",
+        "pixel_format": "yuvj420p",
+        "encoded_size": [
+          1920,
+          1080
+        ],
+        "source_lossy": true,
+        "decoder": {
+          "name": "PyAV/libavcodec",
+          "pyav_version": "12.3.0",
+          "library_versions": {
+            "libavutil": [
+              58,
+              29,
+              100
+            ],
+            "libavcodec": [
+              60,
+              31,
+              102
+            ],
+            "libavformat": [
+              60,
+              16,
+              100
+            ],
+            "libavdevice": [
+              60,
+              3,
+              100
+            ],
+            "libavfilter": [
+              9,
+              12,
+              100
+            ],
+            "libswscale": [
+              7,
+              5,
+              100
+            ],
+            "libswresample": [
+              4,
+              12,
+              100
+            ]
+          },
+          "pixel_interpretation": "rgb24 for RGB; native integer planes for depth and bool; no autorotation"
+        },
+        "encoder_parameters": {
+          "stream_copy": true
+        }
+      },
+      "provenance": {
+        "kind": "format_conversion",
+        "method": "Original JPEG packets copied, no resize/crop/re-encoding"
+      },
+      "source_locator": {
+        "file": "raw/data/data/p_019/p_019_index_press_no-contact_x5_left/cam-d.color.parquet",
+        "record_key": "frame"
+      }
+    },
+    "depth": {
+      "available": true,
+      "schema_id": "depth/z_float32/v1",
+      "data_ref": "observation/depth.mkv",
+      "valid_ref": "observation/depth_valid.mkv",
+      "axes": [
+        "T",
+        "H",
+        "W"
+      ],
+      "valid_axes": [
+        "T",
+        "H",
+        "W"
+      ],
+      "depth_type": "z",
+      "depth_origin": "measured",
+      "unit": "m",
+      "storage": {
+        "profile": "mkv/depth_float32_bits/v1",
+        "codec": "ffv1",
+        "pixel_format": "gray16le",
+        "encoded_size": [
+          3840,
+          1080
+        ],
+        "source_lossy": false,
+        "decoder": {
+          "name": "PyAV/libavcodec",
+          "pyav_version": "12.3.0",
+          "library_versions": {
+            "libavutil": [
+              58,
+              29,
+              100
+            ],
+            "libavcodec": [
+              60,
+              31,
+              102
+            ],
+            "libavformat": [
+              60,
+              16,
+              100
+            ],
+            "libavdevice": [
+              60,
+              3,
+              100
+            ],
+            "libavfilter": [
+              9,
+              12,
+              100
+            ],
+            "libswscale": [
+              7,
+              5,
+              100
+            ],
+            "libswresample": [
+              4,
+              12,
+              100
+            ]
+          },
+          "pixel_interpretation": "rgb24 for RGB; native integer planes for depth and bool; no autorotation"
+        },
+        "encoder_parameters": {
+          "name": "ffv1",
+          "level": 3,
+          "coder": 1,
+          "context": 1,
+          "slicecrc": 1,
+          "slices": 4
+        }
+      },
+      "valid_storage": {
+        "profile": "mkv/bool_ffv1/v1",
+        "codec": "ffv1",
+        "pixel_format": "gray",
+        "encoded_size": [
+          1920,
+          1080
+        ],
+        "source_lossy": false,
+        "decoder": {
+          "name": "PyAV/libavcodec",
+          "pyav_version": "12.3.0",
+          "library_versions": {
+            "libavutil": [
+              58,
+              29,
+              100
+            ],
+            "libavcodec": [
+              60,
+              31,
+              102
+            ],
+            "libavformat": [
+              60,
+              16,
+              100
+            ],
+            "libavdevice": [
+              60,
+              3,
+              100
+            ],
+            "libavfilter": [
+              9,
+              12,
+              100
+            ],
+            "libswscale": [
+              7,
+              5,
+              100
+            ],
+            "libswresample": [
+              4,
+              12,
+              100
+            ]
+          },
+          "pixel_interpretation": "rgb24 for RGB; native integer planes for depth and bool; no autorotation"
+        },
+        "encoder_parameters": {
+          "name": "ffv1",
+          "level": 3,
+          "coder": 1,
+          "context": 1,
+          "slicecrc": 1,
+          "slices": 4
+        }
+      },
+      "provenance": {
+        "kind": "derived",
+        "method": "Official k4a_transformation_depth_image_to_color_camera surface rasterization; SDK uint16 millimetres normalized to float32 metres; zero invalid. No external hole filling."
+      },
+      "source_locator": {
+        "file": "raw/data/data/p_019/p_019_index_press_no-contact_x5_left/cam-d.depth.parquet",
+        "record_key": "frame"
+      }
+    },
+    "intrinsic": {
+      "available": true,
+      "schema_id": "camera/pinhole/v1",
+      "data_ref": "camera/camera.pt",
+      "data_key": [
+        "intrinsic",
+        "data"
+      ],
+      "valid_ref": "camera/camera.pt",
+      "valid_key": [
+        "intrinsic",
+        "valid"
+      ],
+      "axes": [
+        "T",
+        "row",
+        "column"
+      ],
+      "valid_axes": [
+        "T"
+      ],
+      "storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "valid_storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "provenance": {
+        "kind": "format_conversion",
+        "method": "Original undistorted color K"
+      },
+      "source_locator": {
+        "file": "raw/data/configs/p_019/p_019_index_press_no-contact_x5_left.json"
+      }
+    },
+    "extrinsic": {
+      "available": true,
+      "schema_id": "transform/se3/v1",
+      "data_ref": "camera/camera.pt",
+      "data_key": [
+        "extrinsic",
+        "data"
+      ],
+      "valid_ref": "camera/camera.pt",
+      "valid_key": [
+        "extrinsic",
+        "valid"
+      ],
+      "axes": [
+        "T",
+        "row",
+        "column"
+      ],
+      "valid_axes": [
+        "T"
+      ],
+      "storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "valid_storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "provenance": {
+        "kind": "format_conversion",
+        "method": "Source ego_R/ego_T unchanged"
+      },
+      "source_locator": {
+        "file": "raw/data/data/p_019/p_019_index_press_no-contact_x5_left/annotation.parquet",
+        "record_key": "frame"
+      },
+      "direction": "T_camera_from_world",
+      "convention": "T_camera_from_world"
+    },
+    "mano": {
+      "available": true,
+      "schema_id": "mano/axis_angle61/v1",
+      "data_ref": "annotation/mano/mano.pt",
+      "data_key": [
+        "motion",
+        "data"
+      ],
+      "valid_ref": "annotation/mano/mano.pt",
+      "valid_key": [
+        "motion",
+        "valid"
+      ],
+      "axes": [
+        "T",
+        "hand",
+        "parameter"
+      ],
+      "valid_axes": [
+        "T",
+        "hand"
+      ],
+      "storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "valid_storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "provenance": {
+        "kind": "format_conversion",
+        "method": "Standard MANO, no per-vertex refinement; camera root rotation composed; additive translation compensated around shaped root joint"
+      },
+      "source_locator": {
+        "file": "raw/data/data/p_019/p_019_index_press_no-contact_x5_left/annotation.parquet",
+        "record_key": "frame"
+      },
+      "model_profile": {
+        "model": "MANO",
+        "vertices": 778,
+        "flat_hand_mean": true,
+        "use_pca": false,
+        "left_shapedirs_fix_applied": false,
+        "per_vertex_refinement": false,
+        "assets_sha256": {
+          "left": "c4022f7083f2ca7c78b2b3d595abbab52debd32b09d372b16923a801f0ea6a30",
+          "right": "45d60aa3b27ef9107a7afd4e00808f307fd91111e1cfa35afd5c4a62de264767"
+        }
+      },
+      "valid_policy": {
+        "source_annotation_present": "Source has_annotation=true and exact frame ID join",
+        "numerical_checks": "Finite shapes and geometry checked; failure aborts",
+        "official_qa": "not available; visibility not inferred"
+      }
+    }
+  },
+  "split": "unknown",
+  "split_reason": "Official split not provided in selected source",
+  "source_revision": "cea37de99be44078f0e4b9796fd6f614d9110dc2",
+  "calibration": {
+    "baseline": "config DepthToColor",
+    "offset_rgb_optical_z_mm": -6.84644549381777,
+    "offset_estimation": "100-sequence participant-balanced pad fit",
+    "T_rgb_from_depth_rotation": [
+      [
+        0.999999,
+        0.000952568,
+        0.0010665
+      ],
+      [
+        -0.00105482,
+        0.994948,
+        0.10039
+      ],
+      [
+        -0.000965488,
+        -0.100391,
+        0.994948
+      ]
+    ],
+    "T_rgb_from_depth_translation_m": [
+      -0.0335589,
+      -0.00214962,
+      -0.00945180549381777
+    ],
+    "status": "User-approved empirical offset; not official recalibration"
+  },
+  "registration_sdk": {
+    "sdk": "Microsoft.Azure.Kinect.Sensor",
+    "version": "1.4.1",
+    "dll_sha256": "1c6bf5aabb62d4907aa66da92448aebe5f1b8b59f81f2099bc1d9f989cf44cb6",
+    "interfaces": [
+      "k4a_calibration_get_from_raw",
+      "k4a_transformation_create",
+      "k4a_transformation_depth_image_to_color_camera"
+    ],
+    "depth_mode": "WFOV_2X2BINNED",
+    "color_resolution": "1080P",
+    "calibration_adaptation": "Source RGB is already undistorted: config RGB K, zero color distortion, color metric radius 10; depth factory calibration parsed by SDK; config DepthToColor plus approved optical-Z offset; SDK translations in mm.",
+    "T_rgb_from_depth_mm": [
+      [
+        0.999999,
+        0.000952568,
+        0.0010665,
+        -33.5589
+      ],
+      [
+        -0.00105482,
+        0.994948,
+        0.10039,
+        -2.14962
+      ],
+      [
+        -0.000965488,
+        -0.100391,
+        0.994948,
+        -9.45180549381777
+      ],
+      [
+        0.0,
+        0.0,
+        0.0,
+        1.0
+      ]
+    ],
+    "native_depth_K": [
+      [
+        252.52279663085938,
+        0.0,
+        257.9237060546875
+      ],
+      [
+        0.0,
+        252.60423278808594,
+        251.84109497070312
+      ],
+      [
+        0.0,
+        0.0,
+        1.0
+      ]
+    ],
+    "calls": {
+      "depth_to_color": 292,
+      "color_to_depth": 0
+    },
+    "depth_output": "Official uint16 mm, zero invalid; losslessly normalized to float32 metres."
+  },
+  "processing": {
+    "adapter": "egopressure-official-k4a-rgb-only-full/1.0",
+    "contract_repo_revision": "9e89722fc3935d1e16aeb9eeeffa4cd5c9d93383",
+    "scope": "All 1344 selected ego d sequences; original RGB plus official SDK depth-to-RGB",
+    "validation_policy": "All frame IDs, tensor validity and packet counts; deterministic representative decode checks, not full decode or SDK replay"
+  }
+}

EgoPressure/shared-v1.1/data/p_019_index_press_no-contact_x5_right/cam-d-rgb/metainfo.json

@@ -0,0 +1,481 @@
+{
+  "schema_version": "hand-warehouse/1.1",
+  "dataset_id": "EgoPressure",
+  "release_id": "shared-v1.1",
+  "sequence_id": "p_019_index_press_no-contact_x5_right",
+  "camera_id": "cam-d-rgb",
+  "physical_camera_id": "d",
+  "view_type": "ego",
+  "sample_type": "video",
+  "num_frames": 332,
+  "hand_slots": [
+    "left",
+    "right"
+  ],
+  "sampling": {
+    "type": "dense",
+    "source_fps": 30,
+    "frames_ref": "camera/frames.jsonl",
+    "policy": "Every source frame, exact frame-ID joins, no true timestamps available; presentation 30Hz only"
+  },
+  "camera": {
+    "model": "pinhole",
+    "schema_id": "camera/pinhole/v1",
+    "image_size": [
+      1920,
+      1080
+    ]
+  },
+  "components": {
+    "rgb": {
+      "available": true,
+      "schema_id": "rgb/uint8/v1",
+      "data_ref": "observation/rgb.mkv",
+      "valid_ref": "observation/rgb_valid.pt",
+      "axes": [
+        "T",
+        "H",
+        "W",
+        "C"
+      ],
+      "valid_axes": [
+        "T"
+      ],
+      "storage": {
+        "profile": "mkv/mjpeg_copy/v1",
+        "codec": "mjpeg",
+        "pixel_format": "yuvj420p",
+        "encoded_size": [
+          1920,
+          1080
+        ],
+        "source_lossy": true,
+        "decoder": {
+          "name": "PyAV/libavcodec",
+          "pyav_version": "12.3.0",
+          "library_versions": {
+            "libavutil": [
+              58,
+              29,
+              100
+            ],
+            "libavcodec": [
+              60,
+              31,
+              102
+            ],
+            "libavformat": [
+              60,
+              16,
+              100
+            ],
+            "libavdevice": [
+              60,
+              3,
+              100
+            ],
+            "libavfilter": [
+              9,
+              12,
+              100
+            ],
+            "libswscale": [
+              7,
+              5,
+              100
+            ],
+            "libswresample": [
+              4,
+              12,
+              100
+            ]
+          },
+          "pixel_interpretation": "rgb24 for RGB; native integer planes for depth and bool; no autorotation"
+        },
+        "encoder_parameters": {
+          "stream_copy": true
+        }
+      },
+      "provenance": {
+        "kind": "format_conversion",
+        "method": "Original JPEG packets copied, no resize/crop/re-encoding"
+      },
+      "source_locator": {
+        "file": "raw/data/data/p_019/p_019_index_press_no-contact_x5_right/cam-d.color.parquet",
+        "record_key": "frame"
+      }
+    },
+    "depth": {
+      "available": true,
+      "schema_id": "depth/z_float32/v1",
+      "data_ref": "observation/depth.mkv",
+      "valid_ref": "observation/depth_valid.mkv",
+      "axes": [
+        "T",
+        "H",
+        "W"
+      ],
+      "valid_axes": [
+        "T",
+        "H",
+        "W"
+      ],
+      "depth_type": "z",
+      "depth_origin": "measured",
+      "unit": "m",
+      "storage": {
+        "profile": "mkv/depth_float32_bits/v1",
+        "codec": "ffv1",
+        "pixel_format": "gray16le",
+        "encoded_size": [
+          3840,
+          1080
+        ],
+        "source_lossy": false,
+        "decoder": {
+          "name": "PyAV/libavcodec",
+          "pyav_version": "12.3.0",
+          "library_versions": {
+            "libavutil": [
+              58,
+              29,
+              100
+            ],
+            "libavcodec": [
+              60,
+              31,
+              102
+            ],
+            "libavformat": [
+              60,
+              16,
+              100
+            ],
+            "libavdevice": [
+              60,
+              3,
+              100
+            ],
+            "libavfilter": [
+              9,
+              12,
+              100
+            ],
+            "libswscale": [
+              7,
+              5,
+              100
+            ],
+            "libswresample": [
+              4,
+              12,
+              100
+            ]
+          },
+          "pixel_interpretation": "rgb24 for RGB; native integer planes for depth and bool; no autorotation"
+        },
+        "encoder_parameters": {
+          "name": "ffv1",
+          "level": 3,
+          "coder": 1,
+          "context": 1,
+          "slicecrc": 1,
+          "slices": 4
+        }
+      },
+      "valid_storage": {
+        "profile": "mkv/bool_ffv1/v1",
+        "codec": "ffv1",
+        "pixel_format": "gray",
+        "encoded_size": [
+          1920,
+          1080
+        ],
+        "source_lossy": false,
+        "decoder": {
+          "name": "PyAV/libavcodec",
+          "pyav_version": "12.3.0",
+          "library_versions": {
+            "libavutil": [
+              58,
+              29,
+              100
+            ],
+            "libavcodec": [
+              60,
+              31,
+              102
+            ],
+            "libavformat": [
+              60,
+              16,
+              100
+            ],
+            "libavdevice": [
+              60,
+              3,
+              100
+            ],
+            "libavfilter": [
+              9,
+              12,
+              100
+            ],
+            "libswscale": [
+              7,
+              5,
+              100
+            ],
+            "libswresample": [
+              4,
+              12,
+              100
+            ]
+          },
+          "pixel_interpretation": "rgb24 for RGB; native integer planes for depth and bool; no autorotation"
+        },
+        "encoder_parameters": {
+          "name": "ffv1",
+          "level": 3,
+          "coder": 1,
+          "context": 1,
+          "slicecrc": 1,
+          "slices": 4
+        }
+      },
+      "provenance": {
+        "kind": "derived",
+        "method": "Official k4a_transformation_depth_image_to_color_camera surface rasterization; SDK uint16 millimetres normalized to float32 metres; zero invalid. No external hole filling."
+      },
+      "source_locator": {
+        "file": "raw/data/data/p_019/p_019_index_press_no-contact_x5_right/cam-d.depth.parquet",
+        "record_key": "frame"
+      }
+    },
+    "intrinsic": {
+      "available": true,
+      "schema_id": "camera/pinhole/v1",
+      "data_ref": "camera/camera.pt",
+      "data_key": [
+        "intrinsic",
+        "data"
+      ],
+      "valid_ref": "camera/camera.pt",
+      "valid_key": [
+        "intrinsic",
+        "valid"
+      ],
+      "axes": [
+        "T",
+        "row",
+        "column"
+      ],
+      "valid_axes": [
+        "T"
+      ],
+      "storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "valid_storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "provenance": {
+        "kind": "format_conversion",
+        "method": "Original undistorted color K"
+      },
+      "source_locator": {
+        "file": "raw/data/configs/p_019/p_019_index_press_no-contact_x5_right.json"
+      }
+    },
+    "extrinsic": {
+      "available": true,
+      "schema_id": "transform/se3/v1",
+      "data_ref": "camera/camera.pt",
+      "data_key": [
+        "extrinsic",
+        "data"
+      ],
+      "valid_ref": "camera/camera.pt",
+      "valid_key": [
+        "extrinsic",
+        "valid"
+      ],
+      "axes": [
+        "T",
+        "row",
+        "column"
+      ],
+      "valid_axes": [
+        "T"
+      ],
+      "storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "valid_storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "provenance": {
+        "kind": "format_conversion",
+        "method": "Source ego_R/ego_T unchanged"
+      },
+      "source_locator": {
+        "file": "raw/data/data/p_019/p_019_index_press_no-contact_x5_right/annotation.parquet",
+        "record_key": "frame"
+      },
+      "direction": "T_camera_from_world",
+      "convention": "T_camera_from_world"
+    },
+    "mano": {
+      "available": true,
+      "schema_id": "mano/axis_angle61/v1",
+      "data_ref": "annotation/mano/mano.pt",
+      "data_key": [
+        "motion",
+        "data"
+      ],
+      "valid_ref": "annotation/mano/mano.pt",
+      "valid_key": [
+        "motion",
+        "valid"
+      ],
+      "axes": [
+        "T",
+        "hand",
+        "parameter"
+      ],
+      "valid_axes": [
+        "T",
+        "hand"
+      ],
+      "storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "valid_storage": {
+        "profile": "pt/tensor_bundle/v1"
+      },
+      "provenance": {
+        "kind": "format_conversion",
+        "method": "Standard MANO, no per-vertex refinement; camera root rotation composed; additive translation compensated around shaped root joint"
+      },
+      "source_locator": {
+        "file": "raw/data/data/p_019/p_019_index_press_no-contact_x5_right/annotation.parquet",
+        "record_key": "frame"
+      },
+      "model_profile": {
+        "model": "MANO",
+        "vertices": 778,
+        "flat_hand_mean": true,
+        "use_pca": false,
+        "left_shapedirs_fix_applied": false,
+        "per_vertex_refinement": false,
+        "assets_sha256": {
+          "left": "c4022f7083f2ca7c78b2b3d595abbab52debd32b09d372b16923a801f0ea6a30",
+          "right": "45d60aa3b27ef9107a7afd4e00808f307fd91111e1cfa35afd5c4a62de264767"
+        }
+      },
+      "valid_policy": {
+        "source_annotation_present": "Source has_annotation=true and exact frame ID join",
+        "numerical_checks": "Finite shapes and geometry checked; failure aborts",
+        "official_qa": "not available; visibility not inferred"
+      }
+    }
+  },
+  "split": "unknown",
+  "split_reason": "Official split not provided in selected source",
+  "source_revision": "cea37de99be44078f0e4b9796fd6f614d9110dc2",
+  "calibration": {
+    "baseline": "config DepthToColor",
+    "offset_rgb_optical_z_mm": -6.84644549381777,
+    "offset_estimation": "100-sequence participant-balanced pad fit",
+    "T_rgb_from_depth_rotation": [
+      [
+        0.999999,
+        0.000952568,
+        0.0010665
+      ],
+      [
+        -0.00105482,
+        0.994948,
+        0.10039
+      ],
+      [
+        -0.000965488,
+        -0.100391,
+        0.994948
+      ]
+    ],
+    "T_rgb_from_depth_translation_m": [
+      -0.0335589,
+      -0.00214962,
+      -0.00945180549381777
+    ],
+    "status": "User-approved empirical offset; not official recalibration"
+  },
+  "registration_sdk": {
+    "sdk": "Microsoft.Azure.Kinect.Sensor",
+    "version": "1.4.1",
+    "dll_sha256": "1c6bf5aabb62d4907aa66da92448aebe5f1b8b59f81f2099bc1d9f989cf44cb6",
+    "interfaces": [
+      "k4a_calibration_get_from_raw",
+      "k4a_transformation_create",
+      "k4a_transformation_depth_image_to_color_camera"
+    ],
+    "depth_mode": "WFOV_2X2BINNED",
+    "color_resolution": "1080P",
+    "calibration_adaptation": "Source RGB is already undistorted: config RGB K, zero color distortion, color metric radius 10; depth factory calibration parsed by SDK; config DepthToColor plus approved optical-Z offset; SDK translations in mm.",
+    "T_rgb_from_depth_mm": [
+      [
+        0.999999,
+        0.000952568,
+        0.0010665,
+        -33.5589
+      ],
+      [
+        -0.00105482,
+        0.994948,
+        0.10039,
+        -2.14962
+      ],
+      [
+        -0.000965488,
+        -0.100391,
+        0.994948,
+        -9.45180549381777
+      ],
+      [
+        0.0,
+        0.0,
+        0.0,
+        1.0
+      ]
+    ],
+    "native_depth_K": [
+      [
+        252.52279663085938,
+        0.0,
+        257.9237060546875
+      ],
+      [
+        0.0,
+        252.60423278808594,
+        251.84109497070312
+      ],
+      [
+        0.0,
+        0.0,
+        1.0
+      ]
+    ],
+    "calls": {
+      "depth_to_color": 332,
+      "color_to_depth": 0
+    },
+    "depth_output": "Official uint16 mm, zero invalid; losslessly normalized to float32 metres."
+  },
+  "processing": {
+    "adapter": "egopressure-official-k4a-rgb-only-full/1.0",
+    "contract_repo_revision": "9e89722fc3935d1e16aeb9eeeffa4cd5c9d93383",
+    "scope": "All 1344 selected ego d sequences; original RGB plus official SDK depth-to-RGB",
+    "validation_policy": "All frame IDs, tensor validity and packet counts; deterministic representative decode checks, not full decode or SDK replay"
+  }
+}