ExternalVendorCode/whisper.cpp
1
0
Fork 0
mirror of https://github.com/ggerganov/whisper.cpp.git synced 2025-06-13 04:28:07 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

examples : use miniaudio for direct decoding flac, mp3, ogg and wav (#2759)

Browse Source
This commit is contained in:
Dmitry Atamanov
2025-02-27 12:06:54 +05:00
committed by GitHub
parent b5d21359c1
commit 7d3da68f79
10 changed files with 99150 additions and 8930 deletions
Download Patch File Download Diff File Expand all files Collapse all files

9
examples/cli/cli.cpp
View File

@ -199,7 +199,8 @@ static bool whisper_params_parse(int argc, char ** argv, whisper_params & params
static void whisper_print_usage(int /*argc*/, char ** argv, const whisper_params & params) {
fprintf(stderr, "\n");
fprintf(stderr, "usage: %s [options] file0.wav file1.wav ...\n", argv[0]);
fprintf(stderr, "usage: %s [options] file0 file1 ...\n", argv[0]);
fprintf(stderr, "supported audio formats: flac, mp3, ogg, wav\n");
fprintf(stderr, "\n");
fprintf(stderr, "options:\n");
fprintf(stderr, " -h, --help [default] show this help message and exit\n");
@ -244,7 +245,7 @@ static void whisper_print_usage(int /*argc*/, char ** argv, const whisper_params
fprintf(stderr, " -dl, --detect-language [%-7s] exit after automatically detecting language\n", params.detect_language ? "true" : "false");
fprintf(stderr, " --prompt PROMPT [%-7s] initial prompt (max n_text_ctx/2 tokens)\n", params.prompt.c_str());
fprintf(stderr, " -m FNAME, --model FNAME [%-7s] model path\n", params.model.c_str());
fprintf(stderr, " -f FNAME, --file FNAME [%-7s] input WAV file path\n", "");
fprintf(stderr, " -f FNAME, --file FNAME [%-7s] input audio file path\n", "");
fprintf(stderr, " -oved D, --ov-e-device DNAME [%-7s] the OpenVINO device used for encode inference\n", params.openvino_encode_device.c_str());
fprintf(stderr, " -dtw MODEL --dtw MODEL [%-7s] compute token-level timestamps\n", params.dtw.c_str());
fprintf(stderr, " -ls, --log-score [%-7s] log best decoder scores of tokens\n", params.log_score?"true":"false");
@ -1069,8 +1070,8 @@ int main(int argc, char ** argv) {
std::vector<float> pcmf32; // mono-channel F32 PCM
std::vector<std::vector<float>> pcmf32s; // stereo-channel F32 PCM
if (!::read_wav(fname_inp, pcmf32, pcmf32s, params.diarize)) {
fprintf(stderr, "error: failed to read WAV file '%s'\n", fname_inp.c_str());
if (!::read_audio_data(fname_inp, pcmf32, pcmf32s, params.diarize)) {
fprintf(stderr, "error: failed to read audio file '%s'\n", fname_inp.c_str());
continue;
}