female_voxcpm_1: 90 Hz high-pass post-filter (removes p pops learned from its VoxCPM2 reference clip): highpass_hz in voices_index/voices.json, note in the card, filtered samples

README.md

@@ -27,6 +27,13 @@ this repo:
 questions (4 wh-, 4 yes/no) vs the teacher reading the same questions. The voices are synthetic — designed from a text
 description with VoxCPM2 (OpenBMB, Apache-2.0) and frozen as prompt caches; not recordings or clones of real people.
 
+> **Note on `female_voxcpm_1`.** The VoxCPM2 reference clip this voice was frozen from has strong *p* pops (low-frequency
+> thumps on a *p* before a vowel), and the model learned them. Its output is therefore high-passed at **90 Hz** (zero-phase,
+> 4th-order Butterworth): `onnx/voices_index.json` and `onnx/female_voxcpm_1/voices.json` carry `"highpass_hz": 90`, and the
+> Space applies it automatically. Apply the same filter when you use this voice's ONNX graph or checkpoint directly
+> (see the quick start). The other voices need no post-filter.
+
+
 - **Modern Polish of a humanoid assistant** — about 15 h of training speech per voice, the same texts for every voice:
   assistant speech in 18 everyday domains, questions of every kind, modern corpora (NKJP, Wikinews, KPWr, ParlaMint,
   OpenAssistant, Wikibooks, Wikivoyage) and reviewed YouTube transcripts; two thirds of it in multi-sentence passages
@@ -74,7 +81,11 @@ vj = json.load(open("onnx/voices.json"))           # other voices: onnx/<voice>/
 spk = np.array(vj["speakers"]["0"]["emb"], np.float32) + np.array(vj["styles"]["8"]["emb"], np.float32)   # 8 = neutral style (optional)
 sess = ort.InferenceSession("onnx/matcha_pl_t4.onnx")
 wav, n = sess.run(None, {"x": ids[None], "x_lengths": np.array([ids.size]), "scales": np.array([1.0, 0.8], np.float32), "spk_emb": spk[None]})
-sf.write("out.wav", wav[0, :n[0]], 22050)   # scales = [temperature, length_scale (tempo)]
+out = wav[0, :n[0]]
+if vj.get("defaults", {}).get("highpass_hz"):   # female_voxcpm_1: remove the learned "p" pops
+    from scipy.signal import butter, sosfiltfilt
+    out = sosfiltfilt(butter(4, vj["defaults"]["highpass_hz"], "highpass", fs=22050, output="sos"), out)
+sf.write("out.wav", out, 22050)   # scales = [temperature, length_scale (tempo)]
 ```
 
 Text front end: `mark_english()` wraps English words in braces (`Po {meeting}u wyślę ci {feedback}.`), the cleaner

onnx/female_voxcpm_1/voices.json

@@ -1 +1 @@
-{"default": "female_voxcpm_1", "targets": [], "presets": {"female_voxcpm_1": {"label": "female_voxcpm_1 (synthetic voice)", "weights": {"0": 1.0}}}, "speakers": {"0": {"name": "female_voxcpm_1", "emb": [0.690957, 0.698754, 0.22082, -0.115173, 0.250336, -0.852718, 0.362078, 0.451939, 0.178783, -0.170042, -1.624426, 1.016987, 0.740448, -0.041979, -0.901563, 0.79981, 0.201421, -0.808403, -0.637875, -0.434832, 0.189175, 0.144373, -0.371677, 0.075151, 0.281995, -1.017027, -0.27039, -0.41271, 0.24927, 1.745257, -1.275067, 0.048505, 0.357803, 1.115419, 0.212917, -0.169497, 1.017062, 1.481566, -0.530176, -0.057421, -0.620267, -0.397663, -1.363262, 0.921226, 0.506389, -0.025009, -0.195288, 0.787793, 0.925348, -0.392237, 1.943252, 2.008754, -0.395257, -0.690557, 1.274446, 0.468927, 0.222545, -0.709589, -0.279138, 1.389906, 0.937078, -0.416544, 0.893186, 0.433271]}}, "styles": {"0": {"label": "flat-range · slow", "emb": [-0.14398, 0.030409, 0.023366, 0.212853, 0.130899, -0.0853, 0.020606, -0.056369, 0.124819, -0.032304, -0.040329, -0.121085, -0.075622, 0.041307, 0.028547, 0.002017, 0.018731, -0.006378, 0.016171, -0.01124, 0.089603, 0.047565, -0.010464, -0.020278, 0.00249, -0.125436, -0.044735, -0.023403, -0.046575, 0.03308, 0.031525, -0.005011, -0.062249, 0.105864, -0.001503, -0.013248, -0.097321, -0.034723, -0.050534, 0.06587, 0.027612, 0.100292, 0.113216, 0.097034, -0.034022, 0.017676, 0.079936, 0.014128, 0.095708, -0.12326, 0.010466, -0.020589, -0.017111, 0.050987, 0.079795, 0.143887, -0.021827, -0.025365, 0.131949, -0.05286, -0.098738, -0.007769, -0.028433, 0.108708]}, "1": {"label": "flat-range · slow · question", "emb": [-0.061033, 0.178381, -0.051662, 0.106432, 0.165465, -0.037532, -0.059462, -0.024972, -0.06046, -0.074313, -0.098681, -0.183874, -0.077843, 0.062629, -0.036618, -0.148147, 0.080639, -0.043882, 0.059626, 0.074587, 0.186477, 0.049013, -0.119929, -0.063622, 0.057287, -0.005503, 0.038537, -0.053421, 0.011411, -0.064609, -0.186102, 0.071441, 0.033673, 0.00155, 0.007699, -0.082893, 0.015258, -0.200058, -0.042355, -0.00729, -0.093148, 0.143498, -0.037434, 0.019407, 0.120094, 0.087321, -0.146154, 0.058989, -0.106186, 0.00464, 0.02066, -0.063478, 0.027542, 0.055835, 0.018486, 0.086953, 0.11426, 0.020635, 0.172064, -0.0852, -0.082508, 0.031616, -0.154232, 0.095821]}, "2": {"label": "flat-range · normal", "emb": [0.017568, 0.00187, -0.075945, 0.088486, 0.043719, -0.070199, -0.045886, -0.027174, 0.078851, 0.03833, -0.003377, -0.028247, -0.049341, -0.06619, 0.081511, 0.018393, 0.111666, 0.018756, -0.012152, 0.034477, -0.018171, 0.014237, -0.044843, -0.021744, 0.023323, -0.054454, -0.071864, 0.019111, -0.027989, 0.004519, 0.015141, 0.012132, -0.034578, -0.033658, -0.109154, 0.057147, -0.070521, -0.012242, 0.015799, 0.062432, -0.002357, 0.117376, 0.089492, 0.057641, 0.064168, 0.007646, 0.083124, -0.036106, 0.024839, -0.114304, -0.067804, 0.057919, 0.021429, 0.010382, -0.028461, 0.036117, 0.075226, 0.014568, 0.068641, 0.033758, 0.00923, -0.072297, 0.062861, -0.016484]}, "3": {"label": "flat-range · normal · question", "emb": [-0.046016, 0.142228, -0.119134, 0.100926, 0.169199, 0.052183, -0.140385, -0.135899, 0.030201, 0.028536, 0.09377, -0.117906, -0.021374, 0.056857, 0.198338, -0.056057, 0.139625, 0.065561, 0.136326, -0.007006, -0.030562, 0.10232, 0.119253, -0.128263, 0.072766, -0.094819, 0.129702, -0.015778, 0.134753, -0.090945, -0.105393, 0.127585, -0.028045, -0.109548, -0.017027, 0.092973, -0.074519, -0.166098, -0.078535, -0.059869, -0.137866, 0.07222, -0.007684, -0.045438, -0.040227, 0.007188, -0.225349, 0.051961, -0.082411, -0.076947, 0.011075, 0.006199, 0.007904, -0.062004, 0.028387, -0.006425, 0.113841, 0.023016, 0.152413, 0.057874, -0.02501, 0.098548, -0.036147, 0.022801]}, "4": {"label": "flat-range · fast", "emb": [0.084065, -0.047289, -0.151934, -0.012176, -0.08274, -0.015073, -0.131593, -0.073581, 0.020055, 0.10816, 0.053061, -0.023593, 0.050535, -0.101353, 0.104773, 0.11488, 0.041148, 0.107364, -0.085312, -0.006108, -0.089819, 0.085837, -0.085279, 0.050884, 0.084257, 0.108905, -0.111029, 0.06566, 0.128958, 0.022899, 0.069944, 0.028806, 0.058162, -0.090036, -0.131395, 0.081167, 0.022148, 0.029939, 0.067136, -0.029729, -0.033066, 0.123732, 0.034436, -0.013833, 0.059808, -0.110927, 0.013954, -0.170464, -0.059715, -0.183083, -0.110209, 0.108558, 0.041299, 0.035497, -0.067958, -0.012705, 0.142512, 0.058855, -0.024762, 0.004824, 0.128863, 0.026725, 0.128141, -0.102997]}, "5": {"label": "flat-range · fast · question", "emb": [0.07575, 0.035663, -0.196381, -0.169104, 0.035841, -0.042252, -0.185077, 0.168694, -0.072301, 0.053597, 0.069834, 0.109655, 0.017557, -0.100262, 0.1293, -0.058751, -0.045794, -0.007651, -0.008669, 0.005847, -0.008637, 0.071527, -0.042025, -0.082946, 0.084372, 0.184644, 0.19884, 0.09944, 0.159555, 0.051217, -0.150325, 0.119318, 0.129863, -0.133552, -0.003474, 0.033602, 0.00258, -0.180632, -0.061197, -0.193996, -0.043192, 0.05113, -0.098845, -0.105947, -0.0022, 0.003738, -0.248226, -0.134244, -0.095087, -0.055485, -0.019917, 0.157606, 0.103913, -0.058439, -0.036395, -0.021091, 0.197977, 0.191157, 0.066361, 0.147967, -0.022513, 0.036815, 0.006301, -0.044516]}, "6": {"label": "mid-range · slow", "emb": [-0.104293, -0.011849, 0.044505, 0.096407, 0.096089, 0.02188, 0.085944, 0.019154, 0.083896, -0.04856, -0.062448, -0.027363, -0.051385, 0.002331, 0.024914, -0.044651, -0.048886, -0.03537, -0.020656, -0.059121, 0.064784, -0.102973, 0.035275, 0.025956, -0.042609, -0.061044, -0.091292, 0.015171, -0.105764, 0.070032, 0.024454, -0.062987, -0.088504, 0.111437, 0.037099, -0.022023, -0.046536, -0.125111, -0.011192, 0.081757, 0.041246, -0.035027, 0.065546, 0.056607, 0.065176, 0.10712, 0.075505, 0.086796, 0.115501, 0.072506, 0.044448, -0.079467, 0.027319, 0.011517, 0.116007, 0.058862, -0.115388, -0.011914, 0.057195, 0.01596, -0.097552, -0.037067, -0.095568, 0.028525]}, "7": {"label": "mid-range · slow · question", "emb": [-0.028034, 0.15156, 0.096824, 0.117999, 0.112664, -0.004051, -0.033125, -0.129233, -0.100515, -0.118795, -0.146304, -0.117443, -0.082994, 0.147749, -0.062633, -0.021608, 0.00172, -0.02625, 0.088702, -0.000753, 0.087401, 0.026152, 0.044008, -0.037227, -0.072107, -0.027871, 0.051427, -0.128747, 0.023097, 0.026014, -0.082626, 0.094185, -0.039407, 0.064518, 0.103513, -0.03244, -0.040414, -0.236961, -0.127738, -0.02446, 0.009733, 0.003609, 0.035045, 0.110125, -0.031441, 0.030818, -0.240785, 0.110666, -0.030312, -0.042909, 0.015134, -0.071783, -0.016083, 0.12851, -0.017322, -0.057818, 0.072893, -0.142502, 0.14394, -0.066953, -0.175165, 0.084044, -0.164857, -0.046082]}, "8": {"label": "mid-range · normal", "emb": [0.040325, -0.034942, 0.015546, -0.01229, -0.030611, -0.002711, 0.018887, 0.057815, 0.094502, 0.013878, -0.010343, -0.051775, 0.030958, 0.00577, 0.021913, 0.04246, -0.012915, 0.001159, -0.001509, 0.006532, -0.029781, -0.053317, 0.015617, -0.005636, -0.027244, 0.033852, -0.073096, 0.060299, -0.014383, 0.026542, 0.005064, -0.029692, 0.012371, 0.002884, -0.037527, 0.027464, -0.051884, -0.001538, 0.034305, 0.007618, -0.001371, -0.067319, 0.015439, 0.010267, 0.040606, 0.029444, 0.062596, 0.024093, 0.002269, -0.035851, -0.006044, 0.013311, -0.005803, -0.024032, 0.024276, 0.033736, 0.003688, 0.028817, -0.007366, 0.003345, 0.03725, -0.025171, 0.029667, -0.041007]}, "9": {"label": "mid-range · normal · question", "emb": [0.091698, -0.002524, -0.127841, -0.050386, -0.016272, 0.038396, -0.067007, -0.070621, -0.054103, -0.05023, -0.081954, -0.05582, -0.077297, -0.04669, -0.078975, -0.078728, 0.180836, -0.007532, -0.033479, 0.047668, 0.140138, 0.043127, -0.095214, -0.101, 0.055808, 0.001689, 0.222507, -0.051993, -0.071796, 0.022182, -0.187381, -0.044044, 0.030333, 0.06024, 0.046027, 0.073969, 0.039415, -0.042523, -0.091135, 0.023762, -0.039292, -0.02791, -0.019416, -0.045677, 0.035068, 0.015388, -0.271587, 0.064824, -0.16825, 0.087909, -0.150038, -0.043371, 0.203787, 0.047497, -0.002261, -0.098378, 0.081042, -0.070636, -0.003774, 0.000536, -0.018059, 0.013428, -0.042822, -0.005598]}, "10": {"label": "mid-range · fast", "emb": [0.147626, -0.116868, -0.112384, -0.07741, -0.134946, -0.014118, -0.061164, 0.011408, 0.002045, 0.121829, 0.027229, 0.079871, 0.109609, -0.067337, -0.002079, 0.078531, -0.020449, 0.078481, -0.021227, -0.041845, -0.087226, 0.01001, -0.037358, 0.006407, 0.00434, 0.125369, -0.068892, 0.026876, 0.0818, -0.040373, 0.046662, 0.074545, 0.071587, -0.088522, -0.071004, 0.130188, 0.008398, 0.055909, 0.039382, 0.019951, -0.03084, -0.003464, 0.006201, -0.096781, -0.031668, -0.052106, 0.032113, -0.067205, -0.074113, -0.110714, -0.073812, 0.043254, 0.041009, -0.014633, -0.088342, -0.020701, 0.005664, 0.069283, -0.07389, 0.041757, 0.080699, -0.005939, 0.147918, -0.106558]}, "11": {"label": "mid-range · fast · question", "emb": [0.082875, 0.032858, -0.040144, -0.06111, -0.281192, 0.075835, -0.172437, -0.000165, -0.017487, 0.106903, -0.056778, 0.076691, 0.194586, -0.020857, 0.001188, 0.006087, 0.067932, 0.063724, 0.135053, 0.146208, -0.028184, 0.126007, -0.238533, 0.004558, 0.146236, 0.076227, 0.051399, -0.088933, 0.052541, -0.177903, -0.110675, 0.066582, 0.117934, -0.011973, -0.10133, 0.150195, 0.182504, 0.00459, -0.036472, -0.217204, -0.151273, -0.006601, -0.270032, -0.21586, -0.166893, -0.025444, -0.103912, -0.03936, -0.081979, 0.012386, 0.011253, 0.143898, 0.010106, -0.070927, -0.063447, -0.075639, 0.084309, 0.073778, 0.032439, 0.104562, 0.101209, 0.207349, -0.054258, -0.121938]}, "12": {"label": "wide-range · slow", "emb": [-0.121648, 0.028001, 0.143882, 0.017797, 0.034657, 0.0679, 0.179828, -0.036035, 0.044968, -0.079819, -0.018535, -0.00426, -0.040887, -0.032543, -0.082234, 0.024187, -0.125506, -0.075957, 0.02109, -0.027279, 0.012867, -0.113219, 0.081668, -0.03067, -0.107726, -0.023187, -0.032821, -0.059522, -0.110489, -0.003947, 0.057858, -0.09206, -0.102589, 0.137226, 0.091372, -0.106958, -0.011663, 0.084092, -0.001009, 0.008224, 0.058258, -0.073818, 0.00679, 0.076668, -0.013991, 0.067037, 0.095726, 0.131537, 0.078093, 0.15102, 0.114423, -0.045809, -0.066403, -0.095368, 0.079705, 0.027117, -0.100365, -0.040727, -0.095362, 0.00104, -0.11955, 0.022527, -0.040875, 0.054085]}, "13": {"label": "wide-range · slow · question", "emb": [-0.039876, 0.101239, 0.136608, -0.035049, 0.020774, 0.056521, 0.072213, -0.042389, -0.080864, -0.110971, -0.067826, 0.063813, -0.063138, 0.100039, -0.113617, 0.029626, -0.002378, -0.12996, 0.186369, -0.085064, 0.018133, -0.042033, -0.024354, -0.094368, -0.010489, -0.004548, 0.149845, -0.042959, 0.052134, -0.099539, -0.084581, 0.060076, -0.026368, 0.113323, 0.213498, -0.098277, 0.033725, -0.177773, -0.069319, -0.097462, 0.005022, -0.056218, -0.074701, 0.00245, -0.076689, 0.059776, -0.125242, 0.128964, -0.058445, 0.134477, 0.086712, -0.183893, 0.126434, 0.092436, 0.069446, 0.024698, 0.102874, -0.092387, -0.095791, 0.078669, -0.006563, 0.138836, -0.104855, 0.025124]}, "14": {"label": "wide-range · normal", "emb": [0.00269, -0.032637, 0.018865, -0.135973, -0.020501, 0.077561, 0.093044, -0.018687, 0.069897, -0.039896, 0.002043, 0.055893, 0.040581, -0.020882, -0.061842, 0.067586, -0.083576, -0.047541, -0.024197, -0.023202, -0.103058, -0.081448, 0.05513, 0.066875, -0.10657, 0.091467, 0.034306, 0.062071, 0.036955, 0.031461, -0.012932, -0.014451, 0.012463, -0.043957, 0.029662, 0.013052, -0.008069, 0.169887, 0.001168, 0.003871, 0.037584, -0.126336, -0.075458, -0.066546, -0.046716, -0.029741, 0.072363, 0.031657, 0.01183, 0.092434, 0.044211, -0.033157, 0.017229, -0.056768, -0.021254, -0.003255, -0.092903, 0.010752, -0.138317, 0.007128, -0.032302, 0.054666, 0.02941, -0.046007]}, "15": {"label": "wide-range · normal · question", "emb": [0.024907, 0.107957, 0.094111, -0.114399, 0.016736, 0.084958, -0.011846, -0.073102, -0.099056, -0.12738, -0.006912, 0.194047, -0.023969, -0.06025, -0.209493, 0.058988, -0.139523, -0.087111, 0.267993, 0.100402, 0.036756, -0.022336, -0.105278, -0.082592, -0.043519, 0.159112, 0.189347, -0.031447, 0.114585, -0.184224, -0.106783, 0.095705, 0.129988, -0.158636, 0.106278, -0.167184, -0.043115, -0.049744, -0.078937, -0.15529, -0.086491, -0.187866, -0.073565, -0.106086, -0.001153, 0.05245, -0.177366, 0.03337, -0.052094, 0.008313, 0.00722, -0.050813, 0.045557, -0.097276, 0.021969, 0.10629, 0.178822, -0.032506, 0.01667, 0.136303, -0.016842, 0.03437, 0.036771, 0.036015]}, "16": {"label": "wide-range · fast", "emb": [0.137901, -0.152386, -0.03001, -0.197125, -0.195452, 0.125978, -0.032157, -0.066356, -0.012404, 0.053708, 0.071885, 0.108499, 0.171788, -0.039429, -0.040344, 0.106061, -0.060987, -0.025344, -0.05523, -0.046877, -0.127055, -0.03118, -0.071329, 0.045965, -0.022887, 0.167557, 0.053117, 0.040285, 0.119252, -0.021179, 0.052676, 0.068854, 0.093127, -0.041842, 0.013186, 0.131871, 0.09236, 0.192715, 0.030429, -0.078171, 0.031806, -0.074955, -0.047324, -0.128185, -0.081195, -0.119995, 0.075459, -0.069516, -0.073624, 0.003047, -0.001286, 0.034415, -0.010288, -0.040276, -0.13021, -0.033989, 0.023113, 0.069603, -0.23938, 0.051563, 0.128871, 0.050944, 0.142987, -0.065172]}, "17": {"label": "wide-range · fast · question", "emb": [0.121581, -0.051034, 0.024167, -0.075757, -0.194037, 0.163611, -0.03362, -0.091433, -0.086895, -0.004009, 0.039961, 0.19085, 0.01348, -0.105522, 0.01225, -0.055131, 0.023322, 0.008695, 0.164979, 0.012105, -0.090392, 0.013387, 0.007591, 0.040956, 0.160792, 0.140925, 0.073347, -0.024174, 0.128721, -0.085713, -0.138871, -0.046943, 0.077839, -0.052017, -0.000503, -0.025841, 0.068672, -0.247879, -0.069911, -0.230609, -0.033162, -0.116996, -0.070535, -0.140787, -0.03132, -0.027647, -0.102118, 0.024192, -0.165969, 0.154944, -0.056996, 0.0511, 0.033867, -0.177098, -0.059519, -0.0789, 0.059972, 0.088305, -0.051595, 0.060956, 0.00328, 0.051113, 0.01897, -0.010034]}}, "formula": "spk_emb = speakers['0'].emb (+ styles[k].emb for a style token; 8 = neutral: mid range, normal tempo)", "defaults": {"temperature": 1.0, "length_scale": 0.85, "style": 8}}
+{"default": "female_voxcpm_1", "targets": [], "presets": {"female_voxcpm_1": {"label": "female_voxcpm_1 (synthetic voice)", "weights": {"0": 1.0}}}, "speakers": {"0": {"name": "female_voxcpm_1", "emb": [0.690957, 0.698754, 0.22082, -0.115173, 0.250336, -0.852718, 0.362078, 0.451939, 0.178783, -0.170042, -1.624426, 1.016987, 0.740448, -0.041979, -0.901563, 0.79981, 0.201421, -0.808403, -0.637875, -0.434832, 0.189175, 0.144373, -0.371677, 0.075151, 0.281995, -1.017027, -0.27039, -0.41271, 0.24927, 1.745257, -1.275067, 0.048505, 0.357803, 1.115419, 0.212917, -0.169497, 1.017062, 1.481566, -0.530176, -0.057421, -0.620267, -0.397663, -1.363262, 0.921226, 0.506389, -0.025009, -0.195288, 0.787793, 0.925348, -0.392237, 1.943252, 2.008754, -0.395257, -0.690557, 1.274446, 0.468927, 0.222545, -0.709589, -0.279138, 1.389906, 0.937078, -0.416544, 0.893186, 0.433271]}}, "styles": {"0": {"label": "flat-range · slow", "emb": [-0.14398, 0.030409, 0.023366, 0.212853, 0.130899, -0.0853, 0.020606, -0.056369, 0.124819, -0.032304, -0.040329, -0.121085, -0.075622, 0.041307, 0.028547, 0.002017, 0.018731, -0.006378, 0.016171, -0.01124, 0.089603, 0.047565, -0.010464, -0.020278, 0.00249, -0.125436, -0.044735, -0.023403, -0.046575, 0.03308, 0.031525, -0.005011, -0.062249, 0.105864, -0.001503, -0.013248, -0.097321, -0.034723, -0.050534, 0.06587, 0.027612, 0.100292, 0.113216, 0.097034, -0.034022, 0.017676, 0.079936, 0.014128, 0.095708, -0.12326, 0.010466, -0.020589, -0.017111, 0.050987, 0.079795, 0.143887, -0.021827, -0.025365, 0.131949, -0.05286, -0.098738, -0.007769, -0.028433, 0.108708]}, "1": {"label": "flat-range · slow · question", "emb": [-0.061033, 0.178381, -0.051662, 0.106432, 0.165465, -0.037532, -0.059462, -0.024972, -0.06046, -0.074313, -0.098681, -0.183874, -0.077843, 0.062629, -0.036618, -0.148147, 0.080639, -0.043882, 0.059626, 0.074587, 0.186477, 0.049013, -0.119929, -0.063622, 0.057287, -0.005503, 0.038537, -0.053421, 0.011411, -0.064609, -0.186102, 0.071441, 0.033673, 0.00155, 0.007699, -0.082893, 0.015258, -0.200058, -0.042355, -0.00729, -0.093148, 0.143498, -0.037434, 0.019407, 0.120094, 0.087321, -0.146154, 0.058989, -0.106186, 0.00464, 0.02066, -0.063478, 0.027542, 0.055835, 0.018486, 0.086953, 0.11426, 0.020635, 0.172064, -0.0852, -0.082508, 0.031616, -0.154232, 0.095821]}, "2": {"label": "flat-range · normal", "emb": [0.017568, 0.00187, -0.075945, 0.088486, 0.043719, -0.070199, -0.045886, -0.027174, 0.078851, 0.03833, -0.003377, -0.028247, -0.049341, -0.06619, 0.081511, 0.018393, 0.111666, 0.018756, -0.012152, 0.034477, -0.018171, 0.014237, -0.044843, -0.021744, 0.023323, -0.054454, -0.071864, 0.019111, -0.027989, 0.004519, 0.015141, 0.012132, -0.034578, -0.033658, -0.109154, 0.057147, -0.070521, -0.012242, 0.015799, 0.062432, -0.002357, 0.117376, 0.089492, 0.057641, 0.064168, 0.007646, 0.083124, -0.036106, 0.024839, -0.114304, -0.067804, 0.057919, 0.021429, 0.010382, -0.028461, 0.036117, 0.075226, 0.014568, 0.068641, 0.033758, 0.00923, -0.072297, 0.062861, -0.016484]}, "3": {"label": "flat-range · normal · question", "emb": [-0.046016, 0.142228, -0.119134, 0.100926, 0.169199, 0.052183, -0.140385, -0.135899, 0.030201, 0.028536, 0.09377, -0.117906, -0.021374, 0.056857, 0.198338, -0.056057, 0.139625, 0.065561, 0.136326, -0.007006, -0.030562, 0.10232, 0.119253, -0.128263, 0.072766, -0.094819, 0.129702, -0.015778, 0.134753, -0.090945, -0.105393, 0.127585, -0.028045, -0.109548, -0.017027, 0.092973, -0.074519, -0.166098, -0.078535, -0.059869, -0.137866, 0.07222, -0.007684, -0.045438, -0.040227, 0.007188, -0.225349, 0.051961, -0.082411, -0.076947, 0.011075, 0.006199, 0.007904, -0.062004, 0.028387, -0.006425, 0.113841, 0.023016, 0.152413, 0.057874, -0.02501, 0.098548, -0.036147, 0.022801]}, "4": {"label": "flat-range · fast", "emb": [0.084065, -0.047289, -0.151934, -0.012176, -0.08274, -0.015073, -0.131593, -0.073581, 0.020055, 0.10816, 0.053061, -0.023593, 0.050535, -0.101353, 0.104773, 0.11488, 0.041148, 0.107364, -0.085312, -0.006108, -0.089819, 0.085837, -0.085279, 0.050884, 0.084257, 0.108905, -0.111029, 0.06566, 0.128958, 0.022899, 0.069944, 0.028806, 0.058162, -0.090036, -0.131395, 0.081167, 0.022148, 0.029939, 0.067136, -0.029729, -0.033066, 0.123732, 0.034436, -0.013833, 0.059808, -0.110927, 0.013954, -0.170464, -0.059715, -0.183083, -0.110209, 0.108558, 0.041299, 0.035497, -0.067958, -0.012705, 0.142512, 0.058855, -0.024762, 0.004824, 0.128863, 0.026725, 0.128141, -0.102997]}, "5": {"label": "flat-range · fast · question", "emb": [0.07575, 0.035663, -0.196381, -0.169104, 0.035841, -0.042252, -0.185077, 0.168694, -0.072301, 0.053597, 0.069834, 0.109655, 0.017557, -0.100262, 0.1293, -0.058751, -0.045794, -0.007651, -0.008669, 0.005847, -0.008637, 0.071527, -0.042025, -0.082946, 0.084372, 0.184644, 0.19884, 0.09944, 0.159555, 0.051217, -0.150325, 0.119318, 0.129863, -0.133552, -0.003474, 0.033602, 0.00258, -0.180632, -0.061197, -0.193996, -0.043192, 0.05113, -0.098845, -0.105947, -0.0022, 0.003738, -0.248226, -0.134244, -0.095087, -0.055485, -0.019917, 0.157606, 0.103913, -0.058439, -0.036395, -0.021091, 0.197977, 0.191157, 0.066361, 0.147967, -0.022513, 0.036815, 0.006301, -0.044516]}, "6": {"label": "mid-range · slow", "emb": [-0.104293, -0.011849, 0.044505, 0.096407, 0.096089, 0.02188, 0.085944, 0.019154, 0.083896, -0.04856, -0.062448, -0.027363, -0.051385, 0.002331, 0.024914, -0.044651, -0.048886, -0.03537, -0.020656, -0.059121, 0.064784, -0.102973, 0.035275, 0.025956, -0.042609, -0.061044, -0.091292, 0.015171, -0.105764, 0.070032, 0.024454, -0.062987, -0.088504, 0.111437, 0.037099, -0.022023, -0.046536, -0.125111, -0.011192, 0.081757, 0.041246, -0.035027, 0.065546, 0.056607, 0.065176, 0.10712, 0.075505, 0.086796, 0.115501, 0.072506, 0.044448, -0.079467, 0.027319, 0.011517, 0.116007, 0.058862, -0.115388, -0.011914, 0.057195, 0.01596, -0.097552, -0.037067, -0.095568, 0.028525]}, "7": {"label": "mid-range · slow · question", "emb": [-0.028034, 0.15156, 0.096824, 0.117999, 0.112664, -0.004051, -0.033125, -0.129233, -0.100515, -0.118795, -0.146304, -0.117443, -0.082994, 0.147749, -0.062633, -0.021608, 0.00172, -0.02625, 0.088702, -0.000753, 0.087401, 0.026152, 0.044008, -0.037227, -0.072107, -0.027871, 0.051427, -0.128747, 0.023097, 0.026014, -0.082626, 0.094185, -0.039407, 0.064518, 0.103513, -0.03244, -0.040414, -0.236961, -0.127738, -0.02446, 0.009733, 0.003609, 0.035045, 0.110125, -0.031441, 0.030818, -0.240785, 0.110666, -0.030312, -0.042909, 0.015134, -0.071783, -0.016083, 0.12851, -0.017322, -0.057818, 0.072893, -0.142502, 0.14394, -0.066953, -0.175165, 0.084044, -0.164857, -0.046082]}, "8": {"label": "mid-range · normal", "emb": [0.040325, -0.034942, 0.015546, -0.01229, -0.030611, -0.002711, 0.018887, 0.057815, 0.094502, 0.013878, -0.010343, -0.051775, 0.030958, 0.00577, 0.021913, 0.04246, -0.012915, 0.001159, -0.001509, 0.006532, -0.029781, -0.053317, 0.015617, -0.005636, -0.027244, 0.033852, -0.073096, 0.060299, -0.014383, 0.026542, 0.005064, -0.029692, 0.012371, 0.002884, -0.037527, 0.027464, -0.051884, -0.001538, 0.034305, 0.007618, -0.001371, -0.067319, 0.015439, 0.010267, 0.040606, 0.029444, 0.062596, 0.024093, 0.002269, -0.035851, -0.006044, 0.013311, -0.005803, -0.024032, 0.024276, 0.033736, 0.003688, 0.028817, -0.007366, 0.003345, 0.03725, -0.025171, 0.029667, -0.041007]}, "9": {"label": "mid-range · normal · question", "emb": [0.091698, -0.002524, -0.127841, -0.050386, -0.016272, 0.038396, -0.067007, -0.070621, -0.054103, -0.05023, -0.081954, -0.05582, -0.077297, -0.04669, -0.078975, -0.078728, 0.180836, -0.007532, -0.033479, 0.047668, 0.140138, 0.043127, -0.095214, -0.101, 0.055808, 0.001689, 0.222507, -0.051993, -0.071796, 0.022182, -0.187381, -0.044044, 0.030333, 0.06024, 0.046027, 0.073969, 0.039415, -0.042523, -0.091135, 0.023762, -0.039292, -0.02791, -0.019416, -0.045677, 0.035068, 0.015388, -0.271587, 0.064824, -0.16825, 0.087909, -0.150038, -0.043371, 0.203787, 0.047497, -0.002261, -0.098378, 0.081042, -0.070636, -0.003774, 0.000536, -0.018059, 0.013428, -0.042822, -0.005598]}, "10": {"label": "mid-range · fast", "emb": [0.147626, -0.116868, -0.112384, -0.07741, -0.134946, -0.014118, -0.061164, 0.011408, 0.002045, 0.121829, 0.027229, 0.079871, 0.109609, -0.067337, -0.002079, 0.078531, -0.020449, 0.078481, -0.021227, -0.041845, -0.087226, 0.01001, -0.037358, 0.006407, 0.00434, 0.125369, -0.068892, 0.026876, 0.0818, -0.040373, 0.046662, 0.074545, 0.071587, -0.088522, -0.071004, 0.130188, 0.008398, 0.055909, 0.039382, 0.019951, -0.03084, -0.003464, 0.006201, -0.096781, -0.031668, -0.052106, 0.032113, -0.067205, -0.074113, -0.110714, -0.073812, 0.043254, 0.041009, -0.014633, -0.088342, -0.020701, 0.005664, 0.069283, -0.07389, 0.041757, 0.080699, -0.005939, 0.147918, -0.106558]}, "11": {"label": "mid-range · fast · question", "emb": [0.082875, 0.032858, -0.040144, -0.06111, -0.281192, 0.075835, -0.172437, -0.000165, -0.017487, 0.106903, -0.056778, 0.076691, 0.194586, -0.020857, 0.001188, 0.006087, 0.067932, 0.063724, 0.135053, 0.146208, -0.028184, 0.126007, -0.238533, 0.004558, 0.146236, 0.076227, 0.051399, -0.088933, 0.052541, -0.177903, -0.110675, 0.066582, 0.117934, -0.011973, -0.10133, 0.150195, 0.182504, 0.00459, -0.036472, -0.217204, -0.151273, -0.006601, -0.270032, -0.21586, -0.166893, -0.025444, -0.103912, -0.03936, -0.081979, 0.012386, 0.011253, 0.143898, 0.010106, -0.070927, -0.063447, -0.075639, 0.084309, 0.073778, 0.032439, 0.104562, 0.101209, 0.207349, -0.054258, -0.121938]}, "12": {"label": "wide-range · slow", "emb": [-0.121648, 0.028001, 0.143882, 0.017797, 0.034657, 0.0679, 0.179828, -0.036035, 0.044968, -0.079819, -0.018535, -0.00426, -0.040887, -0.032543, -0.082234, 0.024187, -0.125506, -0.075957, 0.02109, -0.027279, 0.012867, -0.113219, 0.081668, -0.03067, -0.107726, -0.023187, -0.032821, -0.059522, -0.110489, -0.003947, 0.057858, -0.09206, -0.102589, 0.137226, 0.091372, -0.106958, -0.011663, 0.084092, -0.001009, 0.008224, 0.058258, -0.073818, 0.00679, 0.076668, -0.013991, 0.067037, 0.095726, 0.131537, 0.078093, 0.15102, 0.114423, -0.045809, -0.066403, -0.095368, 0.079705, 0.027117, -0.100365, -0.040727, -0.095362, 0.00104, -0.11955, 0.022527, -0.040875, 0.054085]}, "13": {"label": "wide-range · slow · question", "emb": [-0.039876, 0.101239, 0.136608, -0.035049, 0.020774, 0.056521, 0.072213, -0.042389, -0.080864, -0.110971, -0.067826, 0.063813, -0.063138, 0.100039, -0.113617, 0.029626, -0.002378, -0.12996, 0.186369, -0.085064, 0.018133, -0.042033, -0.024354, -0.094368, -0.010489, -0.004548, 0.149845, -0.042959, 0.052134, -0.099539, -0.084581, 0.060076, -0.026368, 0.113323, 0.213498, -0.098277, 0.033725, -0.177773, -0.069319, -0.097462, 0.005022, -0.056218, -0.074701, 0.00245, -0.076689, 0.059776, -0.125242, 0.128964, -0.058445, 0.134477, 0.086712, -0.183893, 0.126434, 0.092436, 0.069446, 0.024698, 0.102874, -0.092387, -0.095791, 0.078669, -0.006563, 0.138836, -0.104855, 0.025124]}, "14": {"label": "wide-range · normal", "emb": [0.00269, -0.032637, 0.018865, -0.135973, -0.020501, 0.077561, 0.093044, -0.018687, 0.069897, -0.039896, 0.002043, 0.055893, 0.040581, -0.020882, -0.061842, 0.067586, -0.083576, -0.047541, -0.024197, -0.023202, -0.103058, -0.081448, 0.05513, 0.066875, -0.10657, 0.091467, 0.034306, 0.062071, 0.036955, 0.031461, -0.012932, -0.014451, 0.012463, -0.043957, 0.029662, 0.013052, -0.008069, 0.169887, 0.001168, 0.003871, 0.037584, -0.126336, -0.075458, -0.066546, -0.046716, -0.029741, 0.072363, 0.031657, 0.01183, 0.092434, 0.044211, -0.033157, 0.017229, -0.056768, -0.021254, -0.003255, -0.092903, 0.010752, -0.138317, 0.007128, -0.032302, 0.054666, 0.02941, -0.046007]}, "15": {"label": "wide-range · normal · question", "emb": [0.024907, 0.107957, 0.094111, -0.114399, 0.016736, 0.084958, -0.011846, -0.073102, -0.099056, -0.12738, -0.006912, 0.194047, -0.023969, -0.06025, -0.209493, 0.058988, -0.139523, -0.087111, 0.267993, 0.100402, 0.036756, -0.022336, -0.105278, -0.082592, -0.043519, 0.159112, 0.189347, -0.031447, 0.114585, -0.184224, -0.106783, 0.095705, 0.129988, -0.158636, 0.106278, -0.167184, -0.043115, -0.049744, -0.078937, -0.15529, -0.086491, -0.187866, -0.073565, -0.106086, -0.001153, 0.05245, -0.177366, 0.03337, -0.052094, 0.008313, 0.00722, -0.050813, 0.045557, -0.097276, 0.021969, 0.10629, 0.178822, -0.032506, 0.01667, 0.136303, -0.016842, 0.03437, 0.036771, 0.036015]}, "16": {"label": "wide-range · fast", "emb": [0.137901, -0.152386, -0.03001, -0.197125, -0.195452, 0.125978, -0.032157, -0.066356, -0.012404, 0.053708, 0.071885, 0.108499, 0.171788, -0.039429, -0.040344, 0.106061, -0.060987, -0.025344, -0.05523, -0.046877, -0.127055, -0.03118, -0.071329, 0.045965, -0.022887, 0.167557, 0.053117, 0.040285, 0.119252, -0.021179, 0.052676, 0.068854, 0.093127, -0.041842, 0.013186, 0.131871, 0.09236, 0.192715, 0.030429, -0.078171, 0.031806, -0.074955, -0.047324, -0.128185, -0.081195, -0.119995, 0.075459, -0.069516, -0.073624, 0.003047, -0.001286, 0.034415, -0.010288, -0.040276, -0.13021, -0.033989, 0.023113, 0.069603, -0.23938, 0.051563, 0.128871, 0.050944, 0.142987, -0.065172]}, "17": {"label": "wide-range · fast · question", "emb": [0.121581, -0.051034, 0.024167, -0.075757, -0.194037, 0.163611, -0.03362, -0.091433, -0.086895, -0.004009, 0.039961, 0.19085, 0.01348, -0.105522, 0.01225, -0.055131, 0.023322, 0.008695, 0.164979, 0.012105, -0.090392, 0.013387, 0.007591, 0.040956, 0.160792, 0.140925, 0.073347, -0.024174, 0.128721, -0.085713, -0.138871, -0.046943, 0.077839, -0.052017, -0.000503, -0.025841, 0.068672, -0.247879, -0.069911, -0.230609, -0.033162, -0.116996, -0.070535, -0.140787, -0.03132, -0.027647, -0.102118, 0.024192, -0.165969, 0.154944, -0.056996, 0.0511, 0.033867, -0.177098, -0.059519, -0.0789, 0.059972, 0.088305, -0.051595, 0.060956, 0.00328, 0.051113, 0.01897, -0.010034]}}, "formula": "spk_emb = speakers['0'].emb (+ styles[k].emb for a style token; 8 = neutral: mid range, normal tempo)", "defaults": {"temperature": 1.0, "length_scale": 0.85, "style": 8, "highpass_hz": 90}, "postfilter": "highpass_hz: apply a zero-phase 4th-order Butterworth high-pass at this frequency to the output (removes 'p' pops the voice inherited from its VoxCPM2 reference clip)"}

onnx/voices_index.json

@@ -22,8 +22,10 @@
    "defaults": {
     "temperature": 1.0,
     "length_scale": 0.85,
-    "style": 8
-   }
+    "style": 8,
+    "highpass_hz": 90
+   },
+   "highpass_hz": 90
   },
   {
    "id": "male_voxcpm_2",

samples/female_voxcpm_1/README.md

@@ -1,4 +1,4 @@
-Held-out test sentences (not in the training set), synthesized with `onnx/female_voxcpm_1/matcha_pl_t4.onnx` at the voice's defaults (temperature 1.0, tempo 0.85, style 8); teacher = VoxCPM2 with the frozen female_voxcpm_1 voice.
+Held-out test sentences (not in the training set), synthesized with `onnx/female_voxcpm_1/matcha_pl_t4.onnx` at the voice's defaults, high-passed at 90 Hz as the voice requires (temperature 1.0, tempo 0.85, style 8); teacher = VoxCPM2 with the frozen female_voxcpm_1 voice.
 
 | # | text | female_voxcpm_1 | VoxCPM2 teacher |
 |---|---|---|---|

samples/female_voxcpm_1/e05.wav

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4995ea8e670f38fd213a6069ed4ccd4d248bc6e3c218a06f092026cbcf48dffa
+oid sha256:cd9eb4758e71bd9e75f622e01b933f6205c35baf0f86b8c116b7eaea7503dde1
 size 187436

samples/female_voxcpm_1/e06.wav

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fca7227cc807091f2ce568389aa5fe3928dce601e2905e894c01486106ede37e
+oid sha256:89e0be2904377b008f0caaa75fa8e9b62f6f517767bb1bc3715fd34a3a413528
 size 187436

samples/female_voxcpm_1/e07.wav

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e10b354d6bed982e3ff13f5e01a2388bf159dc13dad98b4165619817650bb442
+oid sha256:f2e63333d71171ecf2f05a6d214ce8f81db9f7ca15a1da39e45eda3ce9759bae
 size 240172

samples/female_voxcpm_1/e09.wav

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a5e20629eda62915a706290351379882cf99163eae65d4c5aeb7c4cfc2d7a466
+oid sha256:089ccce3633718f579d50a8172724510e6b135b6557c0dc1f7c2877baaa7d0df
 size 133676

samples/female_voxcpm_1/e10.wav

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1dc34998c1da78cc1144aa480e87c3952eedc95043172d042e26a49aa8c1df8e
+oid sha256:3c3aafb117b0ceac48e192b0f3396e78a1dbb51b765679d209c633ddff95df1e
 size 445484

samples/female_voxcpm_1/e11.wav

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:21b89c9767d1adbe7528c4f2c9782f373b191ab8552a23e863b4e9bc8ac2192f
+oid sha256:271983e7430a0b905e0ddffcb7617d9d7bc4c457c6301446502fe1e720ac8be2
 size 570924

samples/female_voxcpm_1/e12.wav

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:043c1b171436459ac67a7c5cbabe8d1af4a14c0831d94f4d91ba3b563930be83
+oid sha256:5c084c26d638efbde1b0f162ef9ba69eab22e89b044668acb7f0dc648b23d7f8
 size 268332