[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6XDRMS3APNG1850VBFMQ4KG":3,"skill-same:01M6XDRMS3APNG1850VBFMQ4KG":63},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":15,"categories":16,"keywords":17,"publishedAt":37,"linkOnly":38,"markdownContent":39,"markdownBody":40,"frontMatter":41,"refName":42,"commitSha":43,"skillDirectory":44,"forksCount":13,"repositoryHtmlUrl":45,"repositoryHomepageUrl":6,"ownerHtmlUrl":46,"ownerBlogUrl":6,"ownerDescription":47,"updatedAt":48,"install":49,"resources":62},"01M6XDRMS3APNG1850VBFMQ4KG","Voice Activity Detection (VAD)",null,"Detect speech segments in audio using VAD tools like Silero VAD, SpeechBrain VAD, or WebRTC VAD. Use when preprocessing audio for speaker diarization, filtering silence, or segmenting audio into speech chunks. Choose Silero VAD for short segments, SpeechBrain VAD for general purpose, or WebRTC VAD for lightweight applications.","Waaaalnut\u002Fskillsbench","Waaaalnut","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F103995656?v=4","01M6XCS9K1V51EY3JMB7DGMX24","01M6XCS9Q4WPW88W62C55G1H44",0,"Apache-2.0","https:\u002F\u002Fgithub.com\u002FWaaaalnut\u002Fskillsbench\u002Fblob\u002Fmain\u002Ftasks-no-skills\u002Fspeaker-diarization-subtitles\u002Fenvironment\u002Fskills\u002Fvoice-activity-detection\u002FSKILL.md",[],[18,5,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36],"detection","activity","vad","voice","choose","short","filtering","use","general","webrtc","preprocessing","Audio","detect","applications","segments","for","chunks","when","speech","2026-11-10 10:12:30",false,"---\nname: Voice Activity Detection (VAD)\ndescription: Detect speech segments in audio using VAD tools like Silero VAD, SpeechBrain VAD, or WebRTC VAD. Use when preprocessing audio for speaker diarization, filtering silence, or segmenting audio into speech chunks. Choose Silero VAD for short segments, SpeechBrain VAD for general purpose, or WebRTC VAD for lightweight applications.\n---\n\n# Voice Activity Detection (VAD)\n\n## Overview\n\nVoice Activity Detection identifies which parts of an audio signal contain speech versus silence or background noise. This is a critical first step in speaker diarization pipelines.\n\n## When to Use\n\n- Preprocessing audio before speaker diarization\n- Filtering out silence and noise\n- Segmenting audio into speech chunks\n- Improving diarization accuracy by focusing on speech regions\n\n## Available VAD Tools\n\n### 1. Silero VAD (Recommended for Short Segments)\n\n**Best for:** Short audio segments, real-time applications, better detection of brief speech\n\n```python\nimport torch\n\n# Load Silero VAD model\nmodel, utils = torch.hub.load(\n    repo_or_dir='snakers4\u002Fsilero-vad',\n    model='silero_vad',\n    force_reload=False,\n    onnx=False\n)\nget_speech_timestamps = utils[0]\n\n# Run VAD\nspeech_timestamps = get_speech_timestamps(\n    waveform[0],  # mono audio waveform\n    model,\n    threshold=0.6,  # speech probability threshold\n    min_speech_duration_ms=350,  # minimum speech segment length\n    min_silence_duration_ms=400,  # minimum silence between segments\n    sampling_rate=sample_rate\n)\n\n# Convert to boundaries format\nboundaries = [[ts['start'] \u002F sample_rate, ts['end'] \u002F sample_rate]\n              for ts in speech_timestamps]\n```\n\n**Advantages:**\n- Better at detecting short speech segments\n- Lower false alarm rate\n- Optimized for real-time processing\n\n### 2. SpeechBrain VAD\n\n**Best for:** General-purpose VAD, longer audio files\n\n```python\nfrom speechbrain.inference.VAD import VAD\n\nVAD_model = VAD.from_hparams(\n    source=\"speechbrain\u002Fvad-crdnn-libriparty\",\n    savedir=\"\u002Ftmp\u002Fspeechbrain_vad\"\n)\n\n# Get speech segments\nboundaries = VAD_model.get_speech_segments(audio_path)\n```\n\n**Advantages:**\n- Well-tested and reliable\n- Good for longer audio files\n- Part of comprehensive SpeechBrain toolkit\n\n### 3. WebRTC VAD\n\n**Best for:** Lightweight applications, real-time processing\n\n```python\nimport webrtcvad\n\nvad = webrtcvad.Vad(2)  # Aggressiveness: 0-3 (higher = more aggressive)\n\n# Process audio frames (must be 10ms, 20ms, or 30ms)\nis_speech = vad.is_speech(frame_bytes, sample_rate)\n```\n\n**Advantages:**\n- Very lightweight\n- Fast processing\n- Good for real-time applications\n\n## Postprocessing VAD Boundaries\n\nAfter VAD, you should postprocess boundaries to:\n- Merge close segments\n- Remove very short segments\n- Smooth boundaries\n\n```python\ndef postprocess_boundaries(boundaries, min_dur=0.30, merge_gap=0.25):\n    \"\"\"\n    boundaries: list of [start_sec, end_sec]\n    min_dur: drop segments shorter than this (sec)\n    merge_gap: merge segments if silence gap \u003C= this (sec)\n    \"\"\"\n    # Sort by start time\n    boundaries = sorted(boundaries, key=lambda x: x[0])\n\n    # Remove short segments\n    boundaries = [(s, e) for s, e in boundaries if (e - s) >= min_dur]\n\n    # Merge close segments\n    merged = [list(boundaries[0])]\n    for s, e in boundaries[1:]:\n        prev_s, prev_e = merged[-1]\n        if s - prev_e \u003C= merge_gap:\n            merged[-1][1] = max(prev_e, e)\n        else:\n            merged.append([s, e])\n\n    return merged\n```\n\n## Choosing the Right VAD\n\n| Tool | Best For | Pros | Cons |\n|------|----------|------|------|\n| Silero VAD | Short segments, real-time | Better short-segment detection | Requires PyTorch |\n| SpeechBrain VAD | General purpose | Reliable, well-tested | May miss short segments |\n| WebRTC VAD | Lightweight apps | Fast, lightweight | Less accurate, requires specific frame sizes |\n\n## Common Issues and Solutions\n\n1. **Too many false alarms**: Increase threshold or min_speech_duration_ms\n2. **Missing short segments**: Use Silero VAD or decrease threshold\n3. **Over-segmentation**: Increase merge_gap in postprocessing\n4. **Missing speech at boundaries**: Decrease min_silence_duration_ms\n\n## Integration with Speaker Diarization\n\nVAD boundaries are used to:\n1. Extract speech segments for speaker embedding extraction\n2. Filter out non-speech regions\n3. Improve clustering by focusing on actual speech\n\n```python\n# After VAD, extract embeddings only for speech segments\nfor start, end in vad_boundaries:\n    segment_audio = waveform[:, int(start*sr):int(end*sr)]\n    embedding = speaker_model.encode_batch(segment_audio)\n    # ... continue with clustering\n```\n","# Voice Activity Detection (VAD)\n\n## Overview\n\nVoice Activity Detection identifies which parts of an audio signal contain speech versus silence or background noise. This is a critical first step in speaker diarization pipelines.\n\n## When to Use\n\n- Preprocessing audio before speaker diarization\n- Filtering out silence and noise\n- Segmenting audio into speech chunks\n- Improving diarization accuracy by focusing on speech regions\n\n## Available VAD Tools\n\n### 1. Silero VAD (Recommended for Short Segments)\n\n**Best for:** Short audio segments, real-time applications, better detection of brief speech\n\n```python\nimport torch\n\n# Load Silero VAD model\nmodel, utils = torch.hub.load(\n    repo_or_dir='snakers4\u002Fsilero-vad',\n    model='silero_vad',\n    force_reload=False,\n    onnx=False\n)\nget_speech_timestamps = utils[0]\n\n# Run VAD\nspeech_timestamps = get_speech_timestamps(\n    waveform[0],  # mono audio waveform\n    model,\n    threshold=0.6,  # speech probability threshold\n    min_speech_duration_ms=350,  # minimum speech segment length\n    min_silence_duration_ms=400,  # minimum silence between segments\n    sampling_rate=sample_rate\n)\n\n# Convert to boundaries format\nboundaries = [[ts['start'] \u002F sample_rate, ts['end'] \u002F sample_rate]\n              for ts in speech_timestamps]\n```\n\n**Advantages:**\n- Better at detecting short speech segments\n- Lower false alarm rate\n- Optimized for real-time processing\n\n### 2. SpeechBrain VAD\n\n**Best for:** General-purpose VAD, longer audio files\n\n```python\nfrom speechbrain.inference.VAD import VAD\n\nVAD_model = VAD.from_hparams(\n    source=\"speechbrain\u002Fvad-crdnn-libriparty\",\n    savedir=\"\u002Ftmp\u002Fspeechbrain_vad\"\n)\n\n# Get speech segments\nboundaries = VAD_model.get_speech_segments(audio_path)\n```\n\n**Advantages:**\n- Well-tested and reliable\n- Good for longer audio files\n- Part of comprehensive SpeechBrain toolkit\n\n### 3. WebRTC VAD\n\n**Best for:** Lightweight applications, real-time processing\n\n```python\nimport webrtcvad\n\nvad = webrtcvad.Vad(2)  # Aggressiveness: 0-3 (higher = more aggressive)\n\n# Process audio frames (must be 10ms, 20ms, or 30ms)\nis_speech = vad.is_speech(frame_bytes, sample_rate)\n```\n\n**Advantages:**\n- Very lightweight\n- Fast processing\n- Good for real-time applications\n\n## Postprocessing VAD Boundaries\n\nAfter VAD, you should postprocess boundaries to:\n- Merge close segments\n- Remove very short segments\n- Smooth boundaries\n\n```python\ndef postprocess_boundaries(boundaries, min_dur=0.30, merge_gap=0.25):\n    \"\"\"\n    boundaries: list of [start_sec, end_sec]\n    min_dur: drop segments shorter than this (sec)\n    merge_gap: merge segments if silence gap \u003C= this (sec)\n    \"\"\"\n    # Sort by start time\n    boundaries = sorted(boundaries, key=lambda x: x[0])\n\n    # Remove short segments\n    boundaries = [(s, e) for s, e in boundaries if (e - s) >= min_dur]\n\n    # Merge close segments\n    merged = [list(boundaries[0])]\n    for s, e in boundaries[1:]:\n        prev_s, prev_e = merged[-1]\n        if s - prev_e \u003C= merge_gap:\n            merged[-1][1] = max(prev_e, e)\n        else:\n            merged.append([s, e])\n\n    return merged\n```\n\n## Choosing the Right VAD\n\n| Tool | Best For | Pros | Cons |\n|------|----------|------|------|\n| Silero VAD | Short segments, real-time | Better short-segment detection | Requires PyTorch |\n| SpeechBrain VAD | General purpose | Reliable, well-tested | May miss short segments |\n| WebRTC VAD | Lightweight apps | Fast, lightweight | Less accurate, requires specific frame sizes |\n\n## Common Issues and Solutions\n\n1. **Too many false alarms**: Increase threshold or min_speech_duration_ms\n2. **Missing short segments**: Use Silero VAD or decrease threshold\n3. **Over-segmentation**: Increase merge_gap in postprocessing\n4. **Missing speech at boundaries**: Decrease min_silence_duration_ms\n\n## Integration with Speaker Diarization\n\nVAD boundaries are used to:\n1. Extract speech segments for speaker embedding extraction\n2. Filter out non-speech regions\n3. Improve clustering by focusing on actual speech\n\n```python\n# After VAD, extract embeddings only for speech segments\nfor start, end in vad_boundaries:\n    segment_audio = waveform[:, int(start*sr):int(end*sr)]\n    embedding = speaker_model.encode_batch(segment_audio)\n    # ... continue with clustering\n```\n",{"name":5,"description":7},"main","7eeb258ada72b7f58119b0e7a2c00ea89d9beedd","tasks-no-skills\u002Fspeaker-diarization-subtitles\u002Fenvironment\u002Fskills\u002Fvoice-activity-detection","https:\u002F\u002Fgithub.com\u002FWaaaalnut\u002Fskillsbench","https:\u002F\u002Fgithub.com\u002FWaaaalnut","student of Zhejiang University","2026-04-21 23:16:23",{"source":8,"skillName":5,"commands":50,"prompt":60,"zipAvailable":61},[51,54,57],{"tool":52,"command":53},"npx","npx skills add Waaaalnut\u002Fskillsbench --skill \"Voice Activity Detection (VAD)\"",{"tool":55,"command":56},"bunx","bunx skills add Waaaalnut\u002Fskillsbench --skill \"Voice Activity Detection (VAD)\"",{"tool":58,"command":59},"pnpm","pnpm dlx skills add Waaaalnut\u002Fskillsbench --skill \"Voice Activity Detection (VAD)\"","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FWaaaalnut\u002Fskillsbench\u002Fblob\u002Fmain\u002Ftasks-no-skills\u002Fspeaker-diarization-subtitles\u002Fenvironment\u002Fskills\u002Fvoice-activity-detection\u002FSKILL.md",true,[],[64,78,92,117,137,154],{"skillNo":65,"name":66,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":67,"categories":68,"keywords":74,"publishedAt":37},"01M6XEWMJ6VY3ZC53MPDCPG3GH","data-reconciliation","https:\u002F\u002Fgithub.com\u002FWaaaalnut\u002Fskillsbench\u002Fblob\u002Fmain\u002Ftasks\u002Fxlsx-recover-data\u002Fenvironment\u002Fskills\u002Fdata-reconciliation\u002FSKILL.md",[69],{"categoryNo":70,"categoryName":71,"taxonomyType":72,"taxonomyTypeDesc":73},"01JBSEEDCAT000000000000033","数据分析师","PROFESSION","适用职业",[75,76,77],"data reconciliation","reconciliation","data",{"skillNo":79,"name":80,"nameZh":6,"description":81,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":82,"categories":83,"keywords":84,"publishedAt":37},"01M6XEW8MNPMF1KRXV6ATGFRQF","speech-to-text","Transcribe video to timestamped text using Whisper tiny model (pre-installed).","https:\u002F\u002Fgithub.com\u002FWaaaalnut\u002Fskillsbench\u002Fblob\u002Fmain\u002Ftasks\u002Fvideo-tutorial-indexer\u002Fenvironment\u002Fskills\u002Fspeech-to-text\u002FSKILL.md",[],[36,85,86,87,88,89,90,91],"text","speech to text","transcribe video","whisper","model","Video","transcribe",{"skillNo":93,"name":94,"nameZh":6,"description":95,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":96,"categories":97,"keywords":103,"publishedAt":37},"01M6XEW3FQZP0T1RKQQJFXVY9T","video-processor","Process videos by removing segments and concatenating remaining parts. Use when you need to remove detected pauses\u002Fopenings from videos, create highlight reels, or batch process segment removals using ffmpeg filter_complex.","https:\u002F\u002Fgithub.com\u002FWaaaalnut\u002Fskillsbench\u002Fblob\u002Fmain\u002Ftasks\u002Fvideo-silence-remover\u002Fenvironment\u002Fskills\u002Fvideo-processor\u002FSKILL.md",[98],{"categoryNo":99,"categoryName":100,"taxonomyType":101,"taxonomyTypeDesc":102},"01JBSEEDCAT000000000000016","图像与多媒体","SCENARIO","使用场景",[90,104,105,106,107,108,35,109,32,110,111,112,113,114,115,25,116],"processor","video processor","batch","create","reels","parts","process","videos","you","segment","filter","ffmpeg","remove",{"skillNo":118,"name":119,"nameZh":6,"description":120,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":121,"categories":122,"keywords":124,"publishedAt":37},"01M6XEVXJ5KZ0NTY3XZKRY4AS1","silence-detector","Detect initial silence segments in audio\u002Fvideo using energy-based analysis. Use when you need to find low-energy periods at the start of recordings (title slides, setup time, pre-roll silence).","https:\u002F\u002Fgithub.com\u002FWaaaalnut\u002Fskillsbench\u002Fblob\u002Fmain\u002Ftasks\u002Fvideo-silence-remover\u002Fenvironment\u002Fskills\u002Fsilence-detector\u002FSKILL.md",[123],{"categoryNo":99,"categoryName":100,"taxonomyType":101,"taxonomyTypeDesc":102},[125,126,127,128,35,112,129,130,131,132,30,32,133,29,90,134,135,136,25],"detector","silence detector","title","slides","time","energy","at","recordings","find","analysis","start","setup",{"skillNo":138,"name":139,"nameZh":6,"description":140,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":141,"categories":142,"keywords":143,"publishedAt":37},"01M6XEVT3NGRN1FT7JPMAPXS6M","segment-combiner","Combine multiple segment detection results into a unified list. Use when you need to merge segments from different detectors, prepare removal lists for video processing, or consolidate detection outputs.","https:\u002F\u002Fgithub.com\u002FWaaaalnut\u002Fskillsbench\u002Fblob\u002Fmain\u002Ftasks\u002Fvideo-silence-remover\u002Fenvironment\u002Fskills\u002Fsegment-combiner\u002FSKILL.md",[],[113,144,90,18,145,146,147,33,35,148,149,32,112,150,151,25,152,153],"segment combiner","prepare","merge","video processing","outputs","combine","processing","consolidate","results","list",{"skillNo":155,"name":156,"nameZh":6,"description":157,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":9,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":158,"categories":159,"keywords":163,"publishedAt":37},"01M6XEVMPVASGS2KX86QK4FMT7","report-generator","Generate compression reports for video processing. Use when you need to create structured JSON reports with duration statistics, compression ratios, and segment details after video processing.","https:\u002F\u002Fgithub.com\u002FWaaaalnut\u002Fskillsbench\u002Fblob\u002Fmain\u002Ftasks\u002Fvideo-silence-remover\u002Fenvironment\u002Fskills\u002Freport-generator\u002FSKILL.md",[160],{"categoryNo":161,"categoryName":162,"taxonomyType":72,"taxonomyTypeDesc":73},"01JBSEEDCAT000000000000034","行政与办公人员",[164,165,166,113,35,33,112,167,150,168,169,170,25,147,171,90,172,173,107],"generator","report generator","report","with","statistics","structured","reports","json","compression","generate"]