auracaster/piper
3
0
Fork 0
mirror of https://github.com/pstrueb/piper.git synced 2026-04-23 00:04:49 +00:00
Code Issues Packages Projects Releases Wiki Activity

Fix synthesis speaker id bug

Browse Source
This commit is contained in:
Michael Hansen
2023-01-10 14:01:47 -06:00
parent dfa03d80e6
commit a34b2d4f3b
3 changed files with 7 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
src/cpp/main.cpp
View File

@@ -188,7 +188,7 @@ void parseArgs(int argc, char *argv[], RunConfig &runConfig) {
runConfig.outputPath = filesystem::path(argv[++i]);
} else if (arg == "-s" || arg == "--speaker") {
ensureArg(argc, argv, i);
runConfig.speakerId = (larynx::SpeakerId)stoll(argv[++i]);
runConfig.speakerId = (larynx::SpeakerId)stoi(argv[++i]);
} else if (arg == "-h" || arg == "--help") {
printUsage(argv);
exit(0);

8
src/cpp/synthesize.hpp
View File

@@ -53,10 +53,12 @@ void synthesize(SynthesisConfig &synthesisConfig, ModelSession &session,
Ort::Value::CreateTensor<float>(memoryInfo, scales.data(), scales.size(),
scalesShape.data(), scalesShape.size()));
// Add speaker id.
// NOTE: These must be kept outside the "if" below to avoid being deallocated.
vector<int64_t> speakerId{(int64_t)synthesisConfig.speakerId.value_or(0)};
vector<int64_t> speakerIdShape{(int64_t)speakerId.size()};
if (synthesisConfig.speakerId) {
// Add speaker id
vector<int64_t> speakerId{(int64_t)synthesisConfig.speakerId.value()};
vector<int64_t> speakerIdShape{(int64_t)speakerId.size()};
inputTensors.push_back(Ort::Value::CreateTensor<int64_t>(
memoryInfo, speakerId.data(), speakerId.size(), speakerIdShape.data(),
speakerIdShape.size()));

2
src/cpp/wavfile.hpp
View File

@@ -27,7 +27,7 @@ struct WavHeader {
// Write WAV file header only
void writeWavHeader(int sampleRate, int sampleWidth, int channels,
uint32_t numSamples, ostream &audioFile) {
uint32_t numSamples, std::ostream &audioFile) {
WavHeader header;
header.chunkSize = numSamples + sizeof(WavHeader) - 8;
header.sampleRate = sampleRate;