sherpa-onnx C API 1.0
Public C API and C++ wrapper for sherpa-onnx
Loading...
Searching...
No Matches
SherpaOnnxOfflineSpeechDenoiserVoxLockModelConfig Struct Reference

VoxLock target-speaker voice isolation model configuration. More...

#include <c-api.h>

Public Attributes

const char * enroll_model
const char * chunk_model
const char * voice_sample

Detailed Description

VoxLock keeps the voice of one speaker (the caller) and removes other voices. The model is from https://github.com/arbazsiddiqui/voxlock

Definition at line 4570 of file c-api.h.

Member Data Documentation

◆ chunk_model

const char* SherpaOnnxOfflineSpeechDenoiserVoxLockModelConfig::chunk_model

Path to the VoxLock chunk model, e.g., chunk.onnx or chunk_int8.onnx.

It runs on each chunk of the audio.

Definition at line 4585 of file c-api.h.

◆ enroll_model

const char* SherpaOnnxOfflineSpeechDenoiserVoxLockModelConfig::enroll_model

Path to the VoxLock enroll model, e.g., enroll.onnx.

It runs once on a voice sample of the speaker to keep. Download the models from https://github.com/arbazsiddiqui/voxlock/releases/tag/sherpa-onnx-models

Definition at line 4578 of file c-api.h.

◆ voice_sample

const char* SherpaOnnxOfflineSpeechDenoiserVoxLockModelConfig::voice_sample

Optional. Path to a wave file with at least 3 seconds of the caller's voice. VoxLock keeps this voice and removes the others. From a longer file, it uses the loudest 3 seconds within the first 30 seconds. A file that is shorter or cannot be read is ignored, as if it were empty.

If empty or ignored, the voice sample is taken from the input audio:

  • Offline denoiser: the loudest 3 seconds in the first 15 seconds of the input are used.
  • Online denoiser: the audio is passed through unchanged until 3 seconds of speech have been heard; that speech is used as the voice sample.

Definition at line 4599 of file c-api.h.


The documentation for this struct was generated from the following file: