EngineX-Iluvatar/enginex_bi_series-sherpa-onnx
Archived
16
0
Fork 0
Code Issues Pull Requests Actions Projects Releases Wiki Activity

Play generated audio as it is generating. (#457)

Browse Source
This commit is contained in:
Fangjun Kuang
2023-12-02 15:35:11 +08:00
committed by GitHub
parent 539b27e575
commit 99ff6a834c
20 changed files with 876 additions and 79 deletions
Download Patch File Download Diff File Expand all files Collapse all files

25
sherpa-onnx/csrc/offline-tts-vits-impl.h
View File

@@ -69,8 +69,11 @@ class OfflineTtsVitsImpl : public OfflineTtsImpl {
}
#endif
GeneratedAudio Generate(const std::string &_text, int64_t sid = 0,
float speed = 1.0) const override {
int32_t SampleRate() const override { return model_->SampleRate(); }
GeneratedAudio Generate(
const std::string &_text, int64_t sid = 0, float speed = 1.0,
GeneratedAudioCallback callback = nullptr) const override {
int32_t num_speakers = model_->NumSpeakers();
if (num_speakers == 0 && sid != 0) {
SHERPA_ONNX_LOGE(
@@ -118,7 +121,11 @@ class OfflineTtsVitsImpl : public OfflineTtsImpl {
int32_t x_size = static_cast<int32_t>(x.size());
if (config_.max_num_sentences <= 0 || x_size <= config_.max_num_sentences) {
return Process(x, sid, speed);
auto ans = Process(x, sid, speed);
if (callback) {
callback(ans.samples.data(), ans.samples.size());
}
return ans;
}
// the input text is too long, we process sentences within it in batches
@@ -149,6 +156,12 @@ class OfflineTtsVitsImpl : public OfflineTtsImpl {
ans.sample_rate = audio.sample_rate;
ans.samples.insert(ans.samples.end(), audio.samples.begin(),
audio.samples.end());
if (callback) {
callback(audio.samples.data(), audio.samples.size());
// Caution(fangjun): audio is freed when the callback returns, so users
// should copy the data if they want to access the data after
// the callback returns to avoid segmentation fault.
}
}
batch.clear();
@@ -162,6 +175,12 @@ class OfflineTtsVitsImpl : public OfflineTtsImpl {
ans.sample_rate = audio.sample_rate;
ans.samples.insert(ans.samples.end(), audio.samples.begin(),
audio.samples.end());
if (callback) {
callback(audio.samples.data(), audio.samples.size());
// Caution(fangjun): audio is freed when the callback returns, so users
// should copy the data if they want to access the data after
// the callback returns to avoid segmentation fault.
}
}
return ans;