Dataset Viewer
Auto-converted to Parquet Duplicate
audio
stringlengths
43
57
duration_s
float64
1.66k
1.8k
end_samples
listlengths
0
223
ends
listlengths
0
223
level_dbfs
listlengths
0
223
onset_samples
listlengths
0
223
onsets
listlengths
0
223
snr_db
listlengths
0
223
sources
listlengths
0
223
spoken_s
listlengths
0
223
category
stringclasses
6 values
community-amelia/audio/ww_stream_00000.flac
1,660.2165
[ 142169, 336218, 445895, 532218, 625548, 816000, 925505, 1008507, 1094931, 1253830, 1365914, 1445302, 1560936, 1759461, 1926223, 2167586, 2258500, 2338880, 2435787, 2554212, 2679620, 2763485, 2907928, 3012604, 3189770, 3273532, 3346125, 3410175, 3492826, 3676...
[ 8.8855625, 21.013625, 27.8684375, 33.263625, 39.09675, 51, 57.8440625, 63.0316875, 68.4331875, 78.364375, 85.369625, 90.331375, 97.5585, 109.9663125, 120.3889375, 135.474125, 141.15625, 146.18, 152.2366875, 159.63825, 167.47625, 172.7178125, 181.7455, 188.28775, 199.3...
[ -18.03, -22.72, -13.97, -17.54, -19.41, -12.37, -18.88, -19.18, -22.4, -18.65, -16.89, -14.46, -20.54, -15.26, -13.76, -13.23, -19.24, -14.82, -15.83, -15.35, -15.04, -19.44, -17.06, -14.44, -13.98, -22.01, -18.71, -16.42, -17.5, -15.87, -13.25, -14.07, ...
[ 115553, 296319, 433741, 502812, 612711, 780168, 894550, 976032, 1061356, 1218874, 1338874, 1432749, 1527652, 1731823, 1897559, 2156998, 2230856, 2318077, 2424038, 2543991, 2647367, 2750160, 2872751, 2983940, 3165202, 3248417, 3332260, 3401995, 3474614, 36579...
[ 7.2220625, 18.5199375, 27.1088125, 31.42575, 38.2944375, 48.7605, 55.909375, 61.002, 66.33475, 76.179625, 83.679625, 89.5468125, 95.47825, 108.2389375, 118.5974375, 134.812375, 139.4285, 144.8798125, 151.502375, 158.9994375, 165.4604375, 171.885, 179.5469375, 186.49625,...
[ 6.5, 1.8, 10.56, 6.99, 5.12, 12.16, 5.65, 5.34, 2.13, 5.88, 7.64, 10.07, 3.99, 9.27, 10.77, 11.3, 5.29, 9.71, 8.7, 9.18, 9.49, 5.09, 7.47, 10.09, 10.55, 2.52, 5.82, 8.11, 7.03, 8.66, 11.28, 10.46, 6.02, 8.31, 5.18, 4.99, 12.19, 8.6, 10.05, ...
[ "00075.wav", "00198.wav", "00176.wav", "00032.wav", "00056.wav", "00018.wav", "00027.wav", "00061.wav", "00016.wav", "00110.wav", "00030.wav", "00131.wav", "00017.wav", "00123.wav", "00201.wav", "00001.wav", "00040.wav", "00177.wav", "00156.wav", "00159.wav", "00175.wav", "...
[ 1.6635, 2.4936875, 0.759625, 1.837875, 0.8023125, 2.2395, 1.9346875, 2.0296875, 2.0984375, 2.18475, 1.69, 0.7845625, 2.08025, 1.727375, 1.7915, 0.66175, 1.72775, 1.3001875, 0.7343125, 0.6388125, 2.0158125, 0.8328125, 2.1985625, 1.7915, 1.5355, 1.5696875, 0.8665625...
null
negatives/en-US/read_speech/audio/ww_stream_00000.flac
1,800
[]
[]
[]
[]
[]
[]
[]
[]
read_speech
negatives/en-US/read_speech/audio/ww_stream_00001.flac
1,800
[]
[]
[]
[]
[]
[]
[]
[]
read_speech
negatives/en-US/read_speech/audio/ww_stream_00002.flac
1,800
[]
[]
[]
[]
[]
[]
[]
[]
read_speech
negatives/en-US/read_speech/audio/ww_stream_00003.flac
1,800
[]
[]
[]
[]
[]
[]
[]
[]
read_speech
negatives/en-US/read_speech/audio/ww_stream_00004.flac
1,800
[]
[]
[]
[]
[]
[]
[]
[]
read_speech
negatives/en-US/read_speech/audio/ww_stream_00005.flac
1,800
[]
[]
[]
[]
[]
[]
[]
[]
read_speech
negatives/en-US/read_speech/audio/ww_stream_00006.flac
1,800
[]
[]
[]
[]
[]
[]
[]
[]
read_speech
negatives/en-US/read_speech/audio/ww_stream_00007.flac
1,800
[]
[]
[]
[]
[]
[]
[]
[]
read_speech
negatives/en-US/read_speech/audio/ww_stream_00008.flac
1,800
[]
[]
[]
[]
[]
[]
[]
[]
read_speech
negatives/en-US/read_speech/audio/ww_stream_00009.flac
1,800
[]
[]
[]
[]
[]
[]
[]
[]
read_speech
negatives/en-US/read_speech/audio/ww_stream_00010.flac
1,800
[]
[]
[]
[]
[]
[]
[]
[]
read_speech
negatives/en-US/read_speech/audio/ww_stream_00011.flac
1,800
[]
[]
[]
[]
[]
[]
[]
[]
read_speech
negatives/en-US/conversation/audio/ww_stream_00000.flac
1,800
[]
[]
[]
[]
[]
[]
[]
[]
conversation
negatives/en-US/conversation/audio/ww_stream_00001.flac
1,800
[]
[]
[]
[]
[]
[]
[]
[]
conversation
negatives/en-US/conversation/audio/ww_stream_00002.flac
1,800
[]
[]
[]
[]
[]
[]
[]
[]
conversation
negatives/en-US/conversation/audio/ww_stream_00003.flac
1,800
[]
[]
[]
[]
[]
[]
[]
[]
conversation
negatives/en-US/conversation/audio/ww_stream_00004.flac
1,800
[]
[]
[]
[]
[]
[]
[]
[]
conversation
negatives/en-US/conversation/audio/ww_stream_00005.flac
1,800
[]
[]
[]
[]
[]
[]
[]
[]
conversation
negatives/en-US/conversation/audio/ww_stream_00006.flac
1,800
[]
[]
[]
[]
[]
[]
[]
[]
conversation
negatives/en-US/conversation/audio/ww_stream_00007.flac
1,800
[]
[]
[]
[]
[]
[]
[]
[]
conversation
negatives/en-US/broadcast/audio/ww_stream_00000.flac
1,800
[]
[]
[]
[]
[]
[]
[]
[]
broadcast
negatives/en-US/broadcast/audio/ww_stream_00001.flac
1,800
[]
[]
[]
[]
[]
[]
[]
[]
broadcast
negatives/en-US/broadcast/audio/ww_stream_00002.flac
1,800
[]
[]
[]
[]
[]
[]
[]
[]
broadcast
negatives/en-US/broadcast/audio/ww_stream_00003.flac
1,800
[]
[]
[]
[]
[]
[]
[]
[]
broadcast
negatives/en-US/broadcast/audio/ww_stream_00004.flac
1,800
[]
[]
[]
[]
[]
[]
[]
[]
broadcast
negatives/en-US/broadcast/audio/ww_stream_00005.flac
1,800
[]
[]
[]
[]
[]
[]
[]
[]
broadcast
negatives/en-US/broadcast/audio/ww_stream_00006.flac
1,800
[]
[]
[]
[]
[]
[]
[]
[]
broadcast
negatives/en-US/broadcast/audio/ww_stream_00007.flac
1,800
[]
[]
[]
[]
[]
[]
[]
[]
broadcast
negatives/en-US/foreign_speech/audio/ww_stream_00000.flac
1,800
[]
[]
[]
[]
[]
[]
[]
[]
foreign_speech
negatives/en-US/foreign_speech/audio/ww_stream_00001.flac
1,800
[]
[]
[]
[]
[]
[]
[]
[]
foreign_speech
negatives/en-US/foreign_speech/audio/ww_stream_00002.flac
1,800
[]
[]
[]
[]
[]
[]
[]
[]
foreign_speech
negatives/en-US/foreign_speech/audio/ww_stream_00003.flac
1,800
[]
[]
[]
[]
[]
[]
[]
[]
foreign_speech
negatives/shared/music/audio/ww_stream_00000.flac
1,800
[]
[]
[]
[]
[]
[]
[]
[]
music
negatives/shared/music/audio/ww_stream_00001.flac
1,800
[]
[]
[]
[]
[]
[]
[]
[]
music
negatives/shared/music/audio/ww_stream_00002.flac
1,800
[]
[]
[]
[]
[]
[]
[]
[]
music
negatives/shared/music/audio/ww_stream_00003.flac
1,800
[]
[]
[]
[]
[]
[]
[]
[]
music
negatives/shared/noise/audio/ww_stream_00000.flac
1,800
[]
[]
[]
[]
[]
[]
[]
[]
noise
negatives/shared/noise/audio/ww_stream_00001.flac
1,800
[]
[]
[]
[]
[]
[]
[]
[]
noise
negatives/shared/noise/audio/ww_stream_00002.flac
1,800
[]
[]
[]
[]
[]
[]
[]
[]
noise
negatives/shared/noise/audio/ww_stream_00003.flac
1,800
[]
[]
[]
[]
[]
[]
[]
[]
noise

OVOS streaming wake-word benchmark

Continuous recordings for testing a wake-word detector the way it runs on a live microphone. A detector streams each recording in 80 ms blocks. The manifest gives the time of every real wake word, so a test counts the wake words it missed and the times it woke with no wake word. The OVOS plugin arena scores its ww_stream boards from this repo. It also ranks every detector of a language on how often per hour it wakes on the negatives of that language.

Layout

Path Contents
negatives/<language>/<kind>/audio/ Recordings with no wake word, 30 minutes each
negatives/shared/<kind>/audio/ Music and sounds, part of the negatives of every language
<set>/audio/ One recording per test set: its clips with the phrase, with speech between them
manifest-negatives-<language>.jsonl Every recording with no wake word of a language
manifest-<set>.jsonl The test set's recording, then the negatives of its language
dropped-utterances.txt The source rows left out because their transcript holds a phrase

A manifest row has audio, duration_s, onsets and ends (seconds), sources (the clip behind each onset), snr_db and level_dbfs. A negative recording has empty onsets and a category: the kind of audio it holds.

The audio with no wake word

Hours of each kind in the negatives of each language:

Language read speech meeting conversation radio, parliament and talk recordings music household and outdoor sounds speech in other languages Total
Aragonese (an-ES) 0.0 0.0 0.0 1.0 1.0 5.0 7.0
Catalan (ca-ES) 4.0 0.0 0.0 1.0 1.0 1.0 7.0
Danish (da-DK) 4.0 0.0 0.0 1.0 1.0 1.0 7.0
German (de-DE) 4.0 0.0 0.0 1.0 1.0 1.0 7.0
English (en-US) 6.0 4.0 4.0 2.0 2.0 2.0 20.0
Spanish (es-ES) 4.0 0.0 0.0 1.0 1.0 1.0 7.0
Basque (eu-ES) 3.0 0.0 1.5 1.0 1.0 1.0 7.5
French (fr-FR) 4.0 0.0 0.0 1.0 1.0 1.0 7.0
Italian (it-IT) 4.0 0.0 0.0 1.0 1.0 1.0 7.0
Dutch (nl-NL) 4.0 0.0 0.0 1.0 1.0 1.0 7.0
Portuguese (pt-PT) 4.0 0.0 0.0 1.0 1.0 1.0 7.0

English has its own conversation, broadcast and other-language parts. Every other language gets the first hour of the shared music and sounds, and the first hour of the English read speech as speech in another language. No open Aragonese speech set was found, so the Aragonese negatives hold the Spanish speech, counted as speech in another language.

Source Repo and revision Split Licence Used as Hours Rows with a phrase left out
LibriSpeech openslr/librispeech_asr 71cacbfb test-clean, test-other CC BY 4.0 read speech 6.00 424
LibriSpeech openslr/librispeech_asr 71cacbfb dev-clean, dev-other CC BY 4.0 speech between the wake words 8.00 414
AMI meetings, distant microphone edinburghcstr/ami 46f28f25 sdm test CC BY 4.0 meeting conversation 4.00 660
People's Speech MLCommons/peoples_speech f10597c5 test CC BY and CC BY-SA radio, parliament and talk recordings 2.00 519
VoxPopuli en facebook/voxpopuli 42f01879 test CC0 radio, parliament and talk recordings 2.00 81
Multilingual LibriSpeech facebook/multilingual_librispeech 2e83e618 test, de/fr/es/it/nl/pt CC BY 4.0 speech in other languages 2.00 1407
MUSAN music corypaik/musan 76f9882c music CC BY 4.0, per-track terms music 2.05 0
MUSAN noise corypaik/musan 76f9882c noise CC BY 4.0, per-clip terms household and outdoor sounds 0.83 0
DEMAND philgzl/demand 29f46e61 16k CC BY 4.0 household and outdoor sounds 0.40 0
Public-domain sounds of pdsounds.org TigreGotico/public_domain_sounds_3secs f47136f9 all public domain household and outdoor sounds 0.40 0
ESC-50 TigreGotico/ESC-50 f4368b7d all CC BY-NC 3.0 (non-commercial) household and outdoor sounds 0.20 0
FSD50K, clips with no human voice Fhrozen/FSD50k ccf1acaf eval CC0, CC BY or CC BY-NC per clip (non-commercial) household and outdoor sounds 0.20 0
Speech-MASSIVE pt-PT TigreGotico/speech_MASSIVE_pt-PT 288e7b5f test CC BY-NC-SA 4.0 (non-commercial) read speech 2.50 267
Basque Parliament, Basque segments gttsehu/basque_parliament_1 46208b30 test CC0 radio, parliament and talk recordings 1.50 2
FLEURS de_de google/fleurs 70bb2e84 test CC BY 4.0 read speech 2.50 2
FLEURS es_419 google/fleurs 70bb2e84 test CC BY 4.0 read speech 2.30 164
FLEURS it_it google/fleurs 70bb2e84 test CC BY 4.0 read speech 2.51 12
FLEURS ca_es google/fleurs 70bb2e84 test CC BY 4.0 read speech 2.30 175
FLEURS da_dk google/fleurs 70bb2e84 test CC BY 4.0 read speech 2.71 0
FLEURS fr_fr google/fleurs 70bb2e84 test CC BY 4.0 read speech 1.70 2
Common Voice 17 pt fsicoli/common_voice_17_0 8262c16b test CC0 read speech 2.50 625
Common Voice 17 de fsicoli/common_voice_17_0 8262c16b test CC0 read speech 2.50 50
Common Voice 17 es fsicoli/common_voice_17_0 8262c16b test CC0 read speech 2.70 928
Common Voice 17 it fsicoli/common_voice_17_0 8262c16b test CC0 read speech 2.50 148
Common Voice 17 nl fsicoli/common_voice_17_0 8262c16b test CC0 read speech 5.00 18
Common Voice 17 ca fsicoli/common_voice_17_0 8262c16b test CC0 read speech 2.70 1186
Common Voice 17 da fsicoli/common_voice_17_0 8262c16b test CC0 read speech 2.30 4
Common Voice 17 eu fsicoli/common_voice_17_0 8262c16b test CC0 read speech 4.00 18
Common Voice 17 fr fsicoli/common_voice_17_0 8262c16b test CC0 read speech 3.30 34

Each recording keeps the licence of its sources. A source marked non-commercial holds clips that must not be used commercially. The other sources are open for any use under their licence.

These sources stay out:

  • Sound-alike and near-miss sets (not-wake-words-soundalikes-*, near-miss-wakewords, OpenVoiceOS/wake_word_noise). They are made to be hard, so a rate on them does not say how often a detector wakes in a home.
  • Sources a fighter trained on: TigreGotico/not-wake-words-speech-en, AudioSet, LibriSpeech train, and the train splits of Common Voice and FLEURS. Only the test splits of Common Voice and FLEURS are used.
  • Sources with no known licence: TigreGotico/ambient_noises and TigreGotico/building_106_kitchen_3secs.
  • TigreGotico/NAR: it is for research use only, and it holds the word "stop".

How it is made

  • No phrase in the speech: a row whose transcript holds any of the 109 phrases of the arena's wake-word test sets and fighters, as whole words, with case and accents ignored, is left out. That removes 7140 rows, named in dropped-utterances.txt. Music and sounds have no transcript and are not filtered. FSD50K clips with a human-voice label are left out.
  • Speech between the wake words: LibriSpeech dev-clean and dev-other for English. For another language, 1 hour of its own speech that its negatives do not use.
  • Wake words: up to 400 clips with the phrase from each test set, picked evenly across the set. Silence before and after each clip is cut, so its onset is where the phrase starts. Each clip is set 5 to 15 dB above the level of the speech in its recording. Consecutive clips are at least 3 seconds of speech apart, and 6 seconds on average.
  • The build is a pure function of the sources, the phrases above and seed 0: python3 scripts/publish_ww_stream_bench.py --work <folder> --seed 0 in the arena.
Test set Phrase Language Source Licence Clips Hours
community-amelia amelia en-US OpenVoiceOS/ovos-community-wakewords-dataset 1a3ad631 CC-BY-4.0 223 of 223 0.46
community-athena athena en-US OpenVoiceOS/ovos-community-wakewords-dataset 1a3ad631 CC-BY-4.0 400 of 482 0.85
community-computer computer en-US OpenVoiceOS/ovos-community-wakewords-dataset 1a3ad631 CC-BY-4.0 400 of 470 1.03
community-ey-ordenador ey_ordenador es-ES OpenVoiceOS/ovos-community-wakewords-dataset 1a3ad631 CC-BY-4.0 48 of 48 0.12
community-hey-chatterbox hey_chatterbox en-US OpenVoiceOS/ovos-community-wakewords-dataset 1a3ad631 CC-BY-4.0 116 of 116 0.24
community-hey-floyd hey_floyd en-US OpenVoiceOS/ovos-community-wakewords-dataset 1a3ad631 CC-BY-4.0 96 of 96 0.21
community-hey-k9 hey_k9 en-US OpenVoiceOS/ovos-community-wakewords-dataset 1a3ad631 CC-BY-4.0 46 of 46 0.11
community-hey-savant hey_savant en-US OpenVoiceOS/ovos-community-wakewords-dataset 1a3ad631 CC-BY-4.0 115 of 115 0.26
community-hey-scout hey_scout en-US OpenVoiceOS/ovos-community-wakewords-dataset 1a3ad631 CC-BY-4.0 34 of 34 0.07
community-hey-ziggy hey_ziggy en-US OpenVoiceOS/ovos-community-wakewords-dataset 1a3ad631 CC-BY-4.0 26 of 26 0.06
picovoice-alexa alexa en-US Picovoice/wake-word-benchmark f047dd73 apache-2.0 329 of 329 0.70
picovoice-computer computer en-US Picovoice/wake-word-benchmark f047dd73 apache-2.0 400 of 411 0.78
picovoice-jarvis jarvis en-US Picovoice/wake-word-benchmark f047dd73 apache-2.0 384 of 384 0.74
picovoice-smart-mirror smart mirror en-US Picovoice/wake-word-benchmark f047dd73 apache-2.0 369 of 369 0.74
picovoice-snowboy snowboy en-US Picovoice/wake-word-benchmark f047dd73 apache-2.0 400 of 401 0.78
picovoice-view-glass view glass en-US Picovoice/wake-word-benchmark f047dd73 apache-2.0 399 of 399 0.81
sam-wake-word sam en-US quo-labs/sam-wake-word d5398f06 mit 399 of 1021 0.74
synthetic-wakewords-alexa alexa en-US OpenVoiceOS/synthetic-wakewords 73211e16 apache-2.0 400 of 2224 0.74
synthetic-wakewords-amelia amelia en-US OpenVoiceOS/synthetic-wakewords 73211e16 apache-2.0 400 of 2224 0.74
synthetic-wakewords-android android en-US OpenVoiceOS/synthetic-wakewords 73211e16 apache-2.0 400 of 2224 0.74
synthetic-wakewords-athena athena en-US OpenVoiceOS/synthetic-wakewords 73211e16 apache-2.0 400 of 2224 0.72
synthetic-wakewords-hey_andromeda hey_andromeda en-US OpenVoiceOS/synthetic-wakewords 73211e16 apache-2.0 400 of 2226 0.78
synthetic-wakewords-hey_chatterbox hey_chatterbox en-US OpenVoiceOS/synthetic-wakewords 73211e16 apache-2.0 400 of 2226 0.78
synthetic-wakewords-hey_computer hey_computer en-US OpenVoiceOS/synthetic-wakewords 73211e16 apache-2.0 400 of 2223 0.76
synthetic-wakewords-hey_eggbert hey_eggbert en-US OpenVoiceOS/synthetic-wakewords 73211e16 apache-2.0 400 of 2224 0.75
synthetic-wakewords-hey_floyd hey_floyd en-US OpenVoiceOS/synthetic-wakewords 73211e16 apache-2.0 400 of 2221 0.75
synthetic-wakewords-hey_jarvis hey_jarvis en-US OpenVoiceOS/synthetic-wakewords 73211e16 apache-2.0 400 of 2225 0.77
synthetic-wakewords-hey_marvin hey_marvin en-US OpenVoiceOS/synthetic-wakewords 73211e16 apache-2.0 400 of 2098 0.75
synthetic-wakewords-hey_mycroft hey_mycroft en-US OpenVoiceOS/synthetic-wakewords 73211e16 apache-2.0 400 of 2226 0.76
synthetic-wakewords-hey_neon hey_neon en-US OpenVoiceOS/synthetic-wakewords 73211e16 apache-2.0 400 of 2225 0.75
synthetic-wakewords-hey_robin hey_robin en-US OpenVoiceOS/synthetic-wakewords 73211e16 apache-2.0 400 of 2226 0.75
synthetic-wakewords-hey_ziggy hey_ziggy en-US OpenVoiceOS/synthetic-wakewords 73211e16 apache-2.0 400 of 2221 0.75
synthetic-wakewords-stop stop en-US OpenVoiceOS/synthetic-wakewords 73211e16 apache-2.0 400 of 2221 0.72
synthetic-wakewords-wake_up wake_up en-US OpenVoiceOS/synthetic-wakewords 73211e16 apache-2.0 400 of 2226 0.73
tigregotico-acorda-pt acorda pt-PT TigreGotico/synthetic-wakeword-acorda 185c2288 cc-by-4.0 360 of 360 0.68
tigregotico-afonsinho-pt afonsinho pt-PT TigreGotico/synthetic-wakeword-afonsinho 23be7866 cc-by-4.0 400 of 1830 0.77
tigregotico-alexa alexa en-US TigreGotico/synthetic-wakeword-alexa 3426bd3f cc-by-4.0 400 of 9076 0.74
tigregotico-amalia-pt amalia pt-PT TigreGotico/synthetic-wakeword-amalia eb290457 cc-by-4.0 117 of 117 0.22
tigregotico-ambrosio-pt ambrosio pt-PT TigreGotico/synthetic-wakeword-ambrosio 05f52fef cc-by-4.0 400 of 1116 0.77
tigregotico-amelia amelia en-US TigreGotico/synthetic-wakeword-amelia e1d9ca19 cc-by-4.0 400 of 1000 0.75
tigregotico-android android en-US TigreGotico/synthetic-wakeword-android 9d293a50 cc-by-4.0 400 of 695 0.75
tigregotico-antonio-pt antonio pt-PT TigreGotico/synthetic-wakeword-antonio abc0e62a cc-by-4.0 133 of 133 0.25
tigregotico-apaga_a_luz-pt apaga_a_luz pt-PT TigreGotico/synthetic-wakeword-apaga_a_luz 40b7cc3c cc-by-4.0 80 of 80 0.16
tigregotico-athena athena en-US TigreGotico/synthetic-wakeword-athena 5768d6c5 cc-by-4.0 400 of 3235 0.73
tigregotico-aufwachen-de aufwachen de-DE TigreGotico/synthetic-wakeword-aufwachen e42cbe75 cc-by-4.0 390 of 390 0.75
tigregotico-christopher christopher en-US TigreGotico/synthetic-wakeword-christopher db12eed1 cc-by-4.0 147 of 147 0.27
tigregotico-computer computer en-US TigreGotico/synthetic-wakeword-computer eef6ff0c cc-by-4.0 400 of 16487 0.74
tigregotico-desperta-ca desperta ca-ES TigreGotico/synthetic-wakeword-desperta a03e094f cc-by-4.0 390 of 390 0.74
tigregotico-desperta_ca desperta ca-ES TigreGotico/synthetic-wakeword-desperta_ca 985495d5 cc-by-4.0 179 of 179 0.33
tigregotico-despierta-an despierta an-ES TigreGotico/synthetic-wakeword-despierta ca54d284 cc-by-4.0 267 of 267 0.51
tigregotico-despierta-es despierta es-ES TigreGotico/synthetic-wakeword-despierta ca54d284 cc-by-4.0 370 of 370 0.70
tigregotico-ei_computador-pt ei_computador pt-PT TigreGotico/synthetic-wakeword-computador d825d1ae cc-by-4.0 80 of 80 0.15
tigregotico-ei_jarbas-pt ei_jarbas pt-PT TigreGotico/synthetic-wakeword-jarbas 1d75ca8e cc-by-4.0 80 of 80 0.15
tigregotico-elsa-pt elsa pt-PT TigreGotico/synthetic-wakeword-elsa 15af7c2d cc-by-4.0 40 of 40 0.07
tigregotico-esnatu-eu esnatu eu-ES TigreGotico/synthetic-wakeword-esnatu 84ae1fba cc-by-4.0 291 of 291 0.58
tigregotico-espierta-an espierta an-ES TigreGotico/synthetic-wakeword-espierta 3e99b56a cc-by-4.0 123 of 123 0.23
tigregotico-evaristo-pt evaristo pt-PT TigreGotico/synthetic-wakeword-evaristo e9900f72 cc-by-4.0 218 of 218 0.42
tigregotico-ey_ordenador-es ey_ordenador es-ES TigreGotico/synthetic-wakeword-ey_ordenador 9a9a9e01 cc-by-4.0 400 of 587 0.77
tigregotico-firefox firefox en-US TigreGotico/synthetic-wakeword-firefox a7d20f6d cc-by-4.0 400 of 865 0.78
tigregotico-gervasio-pt gervasio pt-PT TigreGotico/synthetic-wakeword-gervasio 74e5ff73 cc-by-4.0 300 of 300 0.57
tigregotico-gloria-pt gloria pt-PT TigreGotico/synthetic-wakeword-gloria e616275d cc-by-4.0 400 of 633 0.74
tigregotico-hello_nabu hello_nabu en-US TigreGotico/synthetic-wakeword-hello_nabu 4aa7e353 cc-by-4.0 400 of 11100 0.76
tigregotico-hey_android hey_android en-US TigreGotico/synthetic-wakeword-hey_android 1560e157 cc-by-4.0 400 of 1952 0.79
tigregotico-hey_chatterbox hey_chatterbox en-US TigreGotico/synthetic-wakeword-hey_chatterbox 74b7355b cc-by-4.0 400 of 1888 0.79
tigregotico-hey_christopher hey_christopher en-US TigreGotico/synthetic-wakeword-hey_christopher 6f2718b0 cc-by-4.0 168 of 168 0.33
tigregotico-hey_computer hey_computer en-US TigreGotico/synthetic-wakeword-hey_computer 3c3c6e1d cc-by-4.0 400 of 14092 0.76
tigregotico-hey_eggbert hey_eggbert en-US TigreGotico/synthetic-wakeword-hey_eggbert 5365931e cc-by-4.0 101 of 101 0.19
tigregotico-hey_firefox hey_firefox en-US TigreGotico/synthetic-wakeword-hey_firefox dfb510ac cc-by-4.0 400 of 975 0.79
tigregotico-hey_floyd hey_floyd en-US TigreGotico/synthetic-wakeword-hey_floyd c3c3d933 cc-by-4.0 400 of 1767 0.77
tigregotico-hey_jarbas hey_jarbas en-US TigreGotico/synthetic-wakeword-hey_jarbas 4d2f80eb cc-by-4.0 266 of 266 0.51
tigregotico-hey_jarvis hey_jarvis en-US TigreGotico/synthetic-wakeword-hey_jarvis 5783361d cc-by-4.0 400 of 12900 0.76
tigregotico-hey_k9 hey_k9 en-US TigreGotico/synthetic-wakeword-hey_k9 a0c191f4 cc-by-4.0 400 of 581 0.78
tigregotico-hey_lazarus hey_lazarus en-US TigreGotico/synthetic-wakeword-hey_lazarus 2df68e7c cc-by-4.0 100 of 100 0.19
tigregotico-hey_marvin hey_marvin en-US TigreGotico/synthetic-wakeword-hey_marvin 62004493 cc-by-4.0 400 of 482 0.76
tigregotico-hey_mike hey_mike en-US TigreGotico/synthetic-wakeword-hey_mike 878e296a cc-by-4.0 100 of 100 0.19
tigregotico-hey_mycroft hey_mycroft en-US TigreGotico/synthetic-wakeword-hey_mycroft 026ecd10 cc-by-4.0 400 of 24614 0.78
tigregotico-hey_neon hey_neon en-US TigreGotico/synthetic-wakeword-hey_neon cf3c6016 cc-by-4.0 400 of 899 0.75
tigregotico-hey_potato hey_potato en-US TigreGotico/synthetic-wakeword-hey_potato b669d0c5 cc-by-4.0 400 of 3514 0.76
tigregotico-hey_rhasspy hey_rhasspy en-US TigreGotico/synthetic-wakeword-hey_rhasspy ebf720d5 cc-by-4.0 400 of 1676 0.77
tigregotico-hey_robin hey_robin en-US TigreGotico/synthetic-wakeword-hey_robin 3281ed8e cc-by-4.0 400 of 21731 0.75
tigregotico-hey_scout hey_scout en-US TigreGotico/synthetic-wakeword-hey_scout d99db22a cc-by-4.0 400 of 648 0.76
tigregotico-hey_siri hey_siri en-US TigreGotico/synthetic-wakeword-hey_siri 17ad73df cc-by-4.0 296 of 296 0.55
tigregotico-hey_snapdragon hey_snapdragon en-US TigreGotico/synthetic-wakeword-hey_snapdragon f6f4f272 cc-by-4.0 400 of 1000 0.79
tigregotico-hey_ziggy hey_ziggy en-US TigreGotico/synthetic-wakeword-hey_ziggy 41c0124e cc-by-4.0 400 of 809 0.75
tigregotico-hi_galaxy hi_galaxy en-US TigreGotico/synthetic-wakeword-hi_galaxy 895e8ee5 cc-by-4.0 400 of 2309 0.78
tigregotico-hi_lumina hi_lumina en-US TigreGotico/synthetic-wakeword-hi_lumina 83b8f98e cc-by-4.0 400 of 1000 0.76
tigregotico-home_assistant home_assistant en-US TigreGotico/synthetic-wakeword-home_assistant e89416c9 cc-by-4.0 400 of 10304 0.76
tigregotico-jarvis jarvis en-US TigreGotico/synthetic-wakeword-jarvis c4365e23 cc-by-4.0 400 of 10924 0.74
tigregotico-joao-pt joao pt-PT TigreGotico/synthetic-wakeword-joao f6557ba5 cc-by-4.0 40 of 40 0.07
tigregotico-julio-pt julio pt-PT TigreGotico/synthetic-wakeword-julio 49745816 cc-by-4.0 40 of 40 0.08
tigregotico-liga_a_luz-pt liga_a_luz pt-PT TigreGotico/synthetic-wakeword-liga_a_luz a48e306d cc-by-4.0 80 of 80 0.15
tigregotico-lights_off lights_off en-US TigreGotico/synthetic-wakeword-lights_off 3b4b93ee cc-by-4.0 168 of 168 0.32
tigregotico-lights_on lights_on en-US TigreGotico/synthetic-wakeword-lights_on fdd1883e cc-by-4.0 168 of 168 0.32
tigregotico-lurdes-pt lurdes pt-PT TigreGotico/synthetic-wakeword-lurdes 23e60547 cc-by-4.0 40 of 40 0.08
tigregotico-maria-pt maria pt-PT TigreGotico/synthetic-wakeword-maria 896e11b8 cc-by-4.0 43 of 43 0.08
tigregotico-marvin marvin en-US TigreGotico/synthetic-wakeword-marvin 19c1fdec cc-by-4.0 400 of 3844 0.73
tigregotico-no no en-US TigreGotico/synthetic-wakeword-no 0d55eb31 cc-by-4.0 400 of 920 0.72
tigregotico-o_computador-pt o_computador pt-PT TigreGotico/synthetic-wakeword-computador d825d1ae cc-by-4.0 80 of 80 0.15
tigregotico-o_flor-pt o_flor pt-PT TigreGotico/synthetic-wakeword-o_flor d1e265b5 cc-by-4.0 80 of 80 0.15
tigregotico-o_ines-pt o_ines pt-PT TigreGotico/synthetic-wakeword-o_ines 2b0a9232 cc-by-4.0 80 of 80 0.15
tigregotico-o_jarbas-pt o_jarbas pt-PT TigreGotico/synthetic-wakeword-jarbas 1d75ca8e cc-by-4.0 80 of 80 0.15
tigregotico-o_maria-pt o_maria pt-PT TigreGotico/synthetic-wakeword-o_maria 66f88760 cc-by-4.0 80 of 80 0.15
tigregotico-o_princesa-pt o_princesa pt-PT TigreGotico/synthetic-wakeword-o_princesa e9216aa6 cc-by-4.0 80 of 80 0.15
tigregotico-oi_computador-pt oi_computador pt-PT TigreGotico/synthetic-wakeword-computador d825d1ae cc-by-4.0 80 of 80 0.15
tigregotico-oi_jarbas-pt oi_jarbas pt-PT TigreGotico/synthetic-wakeword-jarbas 1d75ca8e cc-by-4.0 80 of 80 0.15
tigregotico-ok_nabu okay_nabu en-US TigreGotico/synthetic-wakeword-ok_nabu ccec0206 cc-by-4.0 400 of 16149 0.78
tigregotico-reveille_toi-fr reveille_toi fr-FR TigreGotico/synthetic-wakeword-reveille_toi 056b6302 cc-by-4.0 77 of 77 0.14
tigregotico-sheila sheila en-US TigreGotico/synthetic-wakeword-sheila 6e900aaa cc-by-4.0 400 of 3284 0.73
tigregotico-snowboy snowboy en-US TigreGotico/synthetic-wakeword-snowboy 05894d18 cc-by-4.0 400 of 1000 0.76
tigregotico-stop stop en-US TigreGotico/synthetic-wakeword-stop 0cfddb2d cc-by-4.0 400 of 2670 0.72
tigregotico-sveglia-it sveglia it-IT TigreGotico/synthetic-wakeword-sveglia e2f0b628 cc-by-4.0 222 of 222 0.42
tigregotico-vagn_op-da vagn_op da-DK TigreGotico/synthetic-wakeword-vagn_op 53d0dc1b cc-by-4.0 264 of 264 0.49
tigregotico-view_glass view_glass en-US TigreGotico/synthetic-wakeword-view_glass 26dea652 cc-by-4.0 400 of 1000 0.76
tigregotico-voice_assistant voice_assistant en-US TigreGotico/synthetic-wakeword-voice_assistant 834b30bb cc-by-4.0 400 of 10431 0.78
tigregotico-wake_up wake_up en-US TigreGotico/synthetic-wakeword-wake_up 06d453fc cc-by-4.0 400 of 22335 0.73
tigregotico-wakker_worden-nl wakker_worden nl-NL TigreGotico/synthetic-wakeword-wakker_worden eb5d7adb cc-by-4.0 390 of 390 0.76
tigregotico-yes yes en-US TigreGotico/synthetic-wakeword-yes ee433a8f cc-by-4.0 400 of 1000 0.74

These clips hold no sound and are left out: positive/positive_00440.wav of sam-wake-word.

Funded by the NGI0 Commons Fund / NLnet under grant agreement No 101135429, through the European Commission's Next Generation Internet programme.

Downloads last month
174