Hugging Face's logo

Spaces:
nsfwalex
/
audio-edit
Paused

App Files Community
audio-edit

Commit History

  1. revert

    liuyang
    a7c5fd6

  1. Update VAD parameters and enhance silence detection logic in AudioJobRunner. This commit modifies the VAD aggressiveness setting and adjusts silence detection thresholds, improving the accuracy of silence spans and fallback mechanisms. Additionally, it refines the handling of stereo audio channels during VAD processing.

    liuyang
    46f053a

  2. Refactor VAD alignment logic in AudioJobRunner to improve silence detection and expand search window. This update introduces a new method for attempting silence detection with adjustable seek windows, enhancing the flexibility and accuracy of audio chunk processing.

    liuyang
    1739f44

  3. Enhance logging in AudioJobRunner for VAD alignment and chunk processing. This update adds detailed logging statements to track VAD alignment status, chunk durations, and silence detection, improving traceability and debugging capabilities during audio processing.

    liuyang
    37570e3

  4. Refactor chunk processing in AudioJobRunner to enhance VAD alignment and overlap handling. This update improves the logic for determining chunk durations and ensures proper progress through audio segments, while also adding alignment parameters when VAD is enabled.

    liuyang
    370f42a

  5. Add VAD-based features to AudioJobRunner for improved silence detection and alignment⁠

    liuyang
    570e908

  1. fix value

    liuyang
    f69d26d

  1. Enhance dual-mono detection in _stereo_decision_quick method of AudioJobRunner by implementing VAD-based similarity metrics, improving audio processing accuracy and robustness with fallback to previous methods if VAD is unavailable.

    liuyang
    00d0703

  2. Refactor comments and improve readability in audiojob.py, clarifying thresholds and processing logic in the AudioJobRunner class for better maintainability.

    liuyang
    2c2b916

  3. Refactor dual-mono detection in _stereo_decision_quick method of AudioJobRunner, adjusting thresholds and improving correlation metrics for enhanced audio processing accuracy.

    liuyang
    f3ec09a

  4. Add dual-mono detection logic to _stereo_decision_quick method in AudioJobRunner, enhancing audio processing accuracy with new thresholds and correlation metrics.

    liuyang
    d03f859

  5. Update chunk_target_ms in DEFAULT_PRESETS to improve audio processing window size configuration.

    liuyang
    74c0262

  1. Add chunk_target_ms configuration to S3 upload settings in app.py for improved audio processing control.

    liuyang
    c854aa9

  2. Remove LocalStorageAdapter import from app.py, simplifying AudioJobRunner integration.

    liuyang
    7b72138

  3. Refactor AudioJobRunner to remove LocalStorageAdapter dependency; update manifest handling and improve chunk processing logic for fixed window splits with configurable overlap.

    liuyang
    fbedc7c

  4. Update app.py to include S3 prefix input in audio job UI; enhance logging in audiojob.py for command execution, upload stages, and VAD processing, improving traceability and error handling.

    liuyang
    e8c330f

  5. Refactor audio processing command in AudioJobRunner to utilize filter_complex for channel splitting, improving clarity and efficiency of audio output generation.

    liuyang
    3d01174

  1. Fix audio processing command in AudioJobRunner by adding input stream reference for channel splitting in filter_complex.

    liuyang
    89dc159

  2. Implement detailed logging in AudioJobRunner for upload, probe, and preprocess stages; enhance error handling for VAD pipeline fallback.

    liuyang
    4c2c2a3

  3. Add AudioJob integration to app.py with UI for running audio jobs and handling manifests. Updated requirements.txt to include webrtcvad and boto3.

    liuyang
    a4ab88e

  1. Enhance cut_audio function to include error handling for ffmpeg stream copy. Added fallback to re-encode using pydub if the stream copy fails, with improved logging of errors.

    liuyang
    a095ed4

  2. Fix temporary file handling in cut_audio function by removing unnecessary fd closure.

    liuyang
    52cf8f9

  3. Refactor cut_audio function to use ffmpeg for audio cutting, preserving original codec and file extension. Improved logging and temporary file handling.

    liuyang
    5c87856

  1. Specify temporary directory for audio export in cut_audio function

    liuyang
    d4de52b

  1. Remove info text from audio output component in app.py

    liuyang
    50f7140

  2. init

    liuyang
    4771e75

  3. init

    liuyang
    e1500f1

  4. initial commit

    nsfwalex
    Verified signature
    d0d1c87