dghadiya commited on
Commit
64ca30c
·
verified ·
1 Parent(s): 0b83c2a

Add annotation f30f4fbfd4324c2ebd8c31c200d3a48b

Browse files
annotations/20260718T064000956196Z_help-with-thumbs_up__sync_count__MOVA_Adi_f30f4fbfd4324c2ebd8c31c200d3a48b.json ADDED
@@ -0,0 +1,151 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "id": "f30f4fbfd4324c2ebd8c31c200d3a48b",
3
+ "video_id": "help-with-thumbs_up__sync_count__MOVA",
4
+ "user": "Adi",
5
+ "annotations": {
6
+ "responses": {
7
+ "events": {
8
+ "speech": {
9
+ "a": {
10
+ "label": "FAIL",
11
+ "rationale": "extra speech"
12
+ },
13
+ "v": {
14
+ "label": "FAIL",
15
+ "rationale": "extra"
16
+ },
17
+ "avb": {
18
+ "label": "FAIL",
19
+ "rationale": ""
20
+ },
21
+ "tcCount": {
22
+ "value": "1",
23
+ "rationale": ""
24
+ }
25
+ },
26
+ "silent_action": {
27
+ "v": {
28
+ "label": "FAIL",
29
+ "rationale": "wrong hand"
30
+ },
31
+ "tcCount": {
32
+ "value": "0",
33
+ "rationale": ""
34
+ }
35
+ }
36
+ },
37
+ "video": {
38
+ "tcRel": {
39
+ "label": "FAIL",
40
+ "rationale": ""
41
+ },
42
+ "cc": {
43
+ "label": "PASS",
44
+ "rationale": ""
45
+ }
46
+ }
47
+ },
48
+ "axis_definitions": [
49
+ {
50
+ "id": "a",
51
+ "key": "A",
52
+ "title": "Audio Event Correctness",
53
+ "granularity": "per-event",
54
+ "summary": "Whether the required speech or sound event is present and correct in the audio.",
55
+ "detail": "PASS when every required speech/sound component is present, clear, and correct. FAIL when a required audio event is missing, the content or sound type is wrong, or the audio is too unclear to identify. Judge audibility leniently (a soft or quiet sound still counts); judge identity leniently within an acoustic family (clap/slap/tap/thud/pat) and strictly across families."
56
+ },
57
+ {
58
+ "id": "v",
59
+ "key": "V",
60
+ "title": "Visual Event Correctness",
61
+ "granularity": "per-event",
62
+ "summary": "Whether the required visual event is present and correct. For speech events this evaluates visible speaking behavior (mouth/lip motion), not the spoken content.",
63
+ "detail": "PASS when every required visual component is present, visible, and correct in actor, action, side, body part, object, and contact target. FAIL when it's missing; the wrong actor/side/body part/object/action is used; occlusion, cropping, or blur prevents verification; or (for speech) the mouth movement is absent or inconsistent with the required utterance."
64
+ },
65
+ {
66
+ "id": "avb",
67
+ "key": "AVB",
68
+ "title": "Audio-Visual Binding",
69
+ "granularity": "per-event",
70
+ "summary": "Whether the audio and visual components within the same event are correctly synchronized and causally aligned.",
71
+ "detail": "Applies only to events with both audio and visual components (speech, sounding action \u2014 not silent action). PASS when the sound is synchronized with the corresponding visible action, impact, or mouth movement. FAILS if A or V fails, or the two are clearly misaligned in time. Temporal relations between different events are handled under TC_REL, not AVB."
72
+ },
73
+ {
74
+ "id": "tcRel",
75
+ "key": "TC_REL",
76
+ "title": "Temporal Relation",
77
+ "granularity": "per-video",
78
+ "summary": "Whether the specified relationship between events holds - synchronization, order, and absolute timing.",
79
+ "detail": "Compare the defined start times of the required events. If any required event fails A, V, or AVB, TC_REL FAILS. Synchronization: PASS when events begin at the same moment (simple overlap is not enough). Order: PASS when events clearly begin in the required order (overlap allowed). Absolute timing: PASS when the event starts within about +/-0.3s of the required timestamp."
80
+ },
81
+ {
82
+ "id": "tcCount",
83
+ "key": "TC_COUNT",
84
+ "title": "Temporal Count",
85
+ "granularity": "per-event",
86
+ "summary": "Whether each required event occurs the specified number of times, with no missing or extra occurrences.",
87
+ "detail": "Enter how many times the event actually occurs in the video. If no count is given, expect one. If the event fails A, V, or its applicable AVB, record its count as 0. A continuously held pose counts as one occurrence; releasing and re-forming it (or a clear distinct repetition without full release) counts as another."
88
+ },
89
+ {
90
+ "id": "cc",
91
+ "key": "CC",
92
+ "title": "Constraint Compliance",
93
+ "granularity": "per-video",
94
+ "summary": "Non-core constraints: scene, camera, subject count, no subtitles/BGM, no extra actions or sounds, post-event stillness when required.",
95
+ "detail": "Everything not already covered above: scene and camera conditions; number and identity of subjects; initial pose and required visibility; absence of subtitles, on-screen text, or background music; absence of additional deliberate actions, speech, or salient sounds; and post-event stillness only when explicitly required. Allowed natural motion: blinking, breathing, minor posture adjustment, necessary preparation, and natural reset motion."
96
+ }
97
+ ],
98
+ "events": [
99
+ {
100
+ "id": "speech",
101
+ "modality": "speech",
102
+ "modalityLabel": "Speech",
103
+ "actor": "person_A",
104
+ "label": "Speech: say \"help\"",
105
+ "definition": "The person clearly says the single word \"help\". The speech is audible and easy to distinguish. No extra words are spoken unless explicitly required.",
106
+ "axes": {
107
+ "a": true,
108
+ "v": true,
109
+ "avb": true
110
+ }
111
+ },
112
+ {
113
+ "id": "silent_action",
114
+ "modality": "silent_action",
115
+ "modalityLabel": "Silent action",
116
+ "actor": "person_A",
117
+ "label": "Silent action: thumbs_up (right hand)",
118
+ "definition": "The person's right hand forms or holds a thumbs-up pose: a fist closed with the thumb extended upward, clearly recognizable as a thumbs-up. The target is the pose, not a separate upward hand-raising motion.",
119
+ "axes": {
120
+ "a": false,
121
+ "v": true,
122
+ "avb": false
123
+ }
124
+ }
125
+ ],
126
+ "temporal_constraint": {
127
+ "relationType": "sync_count",
128
+ "timingType": "relative",
129
+ "value": {
130
+ "events": [
131
+ "speech",
132
+ "silent_action"
133
+ ],
134
+ "count": 2
135
+ },
136
+ "expectedCounts": {
137
+ "speech": 2,
138
+ "silent_action": 2
139
+ },
140
+ "description": "Synchronization x2: Speech: say \"help\" + Silent action: thumbs_up (right hand) must occur together, repeated exactly 2 times."
141
+ },
142
+ "model": "MOVA",
143
+ "item_id": "help-with-thumbs_up__sync_count",
144
+ "canonical_id": "speech_action_silent_rel_sync_count",
145
+ "variant": "v1",
146
+ "scene_hint": "plain indoor room",
147
+ "camera_hint": "static medium shot",
148
+ "prompt": "In a plain indoor room, a seated person faces the camera in a static medium shot, with the full face, mouth, upper body, both hands, and thighs visible, and both hands initially resting on the knees or thighs. The person performs the synchronized bundle twice: each time, the person clearly says the single word \u201chelp\u201d while the right hand forms and holds a clearly recognizable thumbs-up pose with the thumb extended upward. The two synchronized repetitions are distinct, with a brief reset between them if needed, and no other words are spoken. After the final event, no new target action occurs."
149
+ },
150
+ "created_at": "2026-07-18T06:40:00.956196Z"
151
+ }