mirror of
https://github.com/k2-fsa/sherpa-onnx.git
synced 2026-11-10 21:11:03 +08:00
Fangjun Kuang
f2e01fe8e3
Fixes #3635 It is for v2.0.0 and breaks the ABI. CircularBuffer::head_ and tail_ are absolute sample counters that grow forever. As int32_t they overflow after 2^31 samples (~37 h at 16 kHz); tail_ % capacity then goes negative and std::copy() writes to an invalid iterator, crashing long-running VAD streams. Widen every absolute stream position to int64_t. Counts and capacities (Push/Pop/Get's n, Size(), capacity) stay int32_t — they are bounded by buffer memory, not stream duration.
Introduction
This directory contains examples for how to use the Object Pascal APIs of sherpa-onnx.
Documentation for this directory: https://k2-fsa.github.io/sherpa/onnx/pascal-api/index.html
| Directory | Description |
|---|---|
| read-wav | It shows how to read a wave file. |
| speaker-diarization | It shows how to use Pascal API for speaker diarization. |
| speech-enhancement-gtcrn | It shows how to use the offline speech denoiser API with GTCRN. |
| speech-enhancement-dpdfnet | It shows how to use the offline speech denoiser API with DPDFNet. Use dpdfnet_baseline.onnx, dpdfnet2.onnx, dpdfnet4.onnx, or dpdfnet8.onnx for 16 kHz downstream ASR and dpdfnet2_48khz_hr.onnx for 48 kHz enhancement output. |
| streaming-speech-enhancement-gtcrn | It shows how to use the streaming speech denoiser API with GTCRN. |
| streaming-speech-enhancement-dpdfnet | It shows how to use the streaming speech denoiser API with DPDFNet. |
| streaming-asr | It shows how to use streaming models for speech recognition. |
| non-streaming-asr | It shows how to use non-streaming models for speech recognition. |
| vad | It shows how to use the voice activity detection API. |
| vad-with-non-streaming-asr | It shows how to use the voice activity detection API with non-streaming models for speech recognition. |
| portaudio-test | It shows how to use PortAudio for recording and playing. |
| tts | It shows how to use the text-to-speech API. |