speech_recognition/FunASR
1
0
Fork 0
mirror of https://github.com/modelscope/FunASR synced 2025-09-15 14:48:36 +08:00
Code Issues Actions Packages Projects Releases Wiki Activity

sort frame_queue

Browse Source
This commit is contained in:
雾聪 2024-04-17 19:14:20 +08:00
parent 05f5021864
commit 200bba87f0
1 changed files with 19 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

22
runtime/onnxruntime/src/audio.cpp
View File

@ -1111,7 +1111,7 @@ void Audio::Split(OfflineStream* offline_stream)
}
}
void Audio::CutSplit(OfflineStream* offline_stream)
void Audio::CutSplit(OfflineStream* offline_stream, std::vector<int> &index_vector)
{
std::unique_ptr<VadModel> vad_online_handle = make_unique<FsmnVadOnline>((FsmnVad*)(offline_stream->vad_handle).get());
AudioFrame *frame;
@ -1138,6 +1138,7 @@ void Audio::CutSplit(OfflineStream* offline_stream)
}
int speech_start_i = -1, speech_end_i =-1;
std::vector<AudioFrame*> vad_frames;
for(vector<int> vad_segment:vad_segments)
{
if(vad_segment.size() != 2){
@ -1152,16 +1153,31 @@ void Audio::CutSplit(OfflineStream* offline_stream)
}
if(speech_start_i!=-1 && speech_end_i!=-1){
frame = new AudioFrame();
int start = speech_start_i*seg_sample;
int end = speech_end_i*seg_sample;
frame = new AudioFrame(end-start);
frame->SetStart(start);
frame->SetEnd(end);
frame_queue.push(frame);
vad_frames.push_back(frame);
frame = nullptr;
speech_start_i=-1;
speech_end_i=-1;
}
}
// sort
{
index_vector.clear();
index_vector.resize(vad_frames.size());
for (int i = 0; i < index_vector.size(); ++i) {
index_vector[i] = i;
}
std::sort(index_vector.begin(), index_vector.end(), [&vad_frames](const int a, const int b) {
return vad_frames[a]->len < vad_frames[b]->len;
});
for (int idx : index_vector) {
frame_queue.push(vad_frames[idx]);
}
}
}