experiments/STT
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Remove sample rate parameter usage from client.cc

Browse Source
This commit is contained in:
Reuben Morais 2019-10-09 16:40:30 +02:00
parent 2f4116695f
commit abb11f040d
1 changed files with 5 additions and 10 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

15
native_client/client.cc
View File

@ -54,23 +54,23 @@ char* JSONOutput(Metadata* metadata);
ds_result ds_result
LocalDsSTT(ModelState* aCtx, const short* aBuffer, size_t aBufferSize, LocalDsSTT(ModelState* aCtx, const short* aBuffer, size_t aBufferSize,
int aSampleRate, bool extended_output, bool json_output) bool extended_output, bool json_output)
{ {
ds_result res = {0}; ds_result res = {0};
clock_t ds_start_time = clock(); clock_t ds_start_time = clock();
if (extended_output) { if (extended_output) {
Metadata *metadata = DS_SpeechToTextWithMetadata(aCtx, aBuffer, aBufferSize, aSampleRate); Metadata *metadata = DS_SpeechToTextWithMetadata(aCtx, aBuffer, aBufferSize);
res.string = metadataToString(metadata); res.string = metadataToString(metadata);
DS_FreeMetadata(metadata); DS_FreeMetadata(metadata);
} else if (json_output) { } else if (json_output) {
Metadata *metadata = DS_SpeechToTextWithMetadata(aCtx, aBuffer, aBufferSize, aSampleRate); Metadata *metadata = DS_SpeechToTextWithMetadata(aCtx, aBuffer, aBufferSize);
res.string = JSONOutput(metadata); res.string = JSONOutput(metadata);
DS_FreeMetadata(metadata); DS_FreeMetadata(metadata);
} else if (stream_size > 0) { } else if (stream_size > 0) {
StreamingState* ctx; StreamingState* ctx;
int status = DS_CreateStream(aCtx, aSampleRate, &ctx); int status = DS_CreateStream(aCtx, &ctx);
if (status != DS_ERR_OK) { if (status != DS_ERR_OK) {
res.string = strdup(""); res.string = strdup("");
return res; return res;
@ -94,7 +94,7 @@ LocalDsSTT(ModelState* aCtx, const short* aBuffer, size_t aBufferSize,
} }
res.string = DS_FinishStream(ctx); res.string = DS_FinishStream(ctx);
} else { } else {
res.string = DS_SpeechToText(aCtx, aBuffer, aBufferSize, aSampleRate); res.string = DS_SpeechToText(aCtx, aBuffer, aBufferSize);
} }
clock_t ds_end_infer = clock(); clock_t ds_end_infer = clock();
@ -108,7 +108,6 @@ LocalDsSTT(ModelState* aCtx, const short* aBuffer, size_t aBufferSize,
typedef struct { typedef struct {
char* buffer; char* buffer;
size_t buffer_size; size_t buffer_size;
int sample_rate;
} ds_audio_buffer; } ds_audio_buffer;
ds_audio_buffer ds_audio_buffer
@ -159,8 +158,6 @@ GetAudioBuffer(const char* path)
assert(output); assert(output);
res.sample_rate = (int)output->signal.rate;
if ((int)input->signal.rate < 16000) { if ((int)input->signal.rate < 16000) {
fprintf(stderr, "Warning: original sample rate (%d) is lower than 16kHz. Up-sampling might produce erratic speech recognition.\n", (int)input->signal.rate); fprintf(stderr, "Warning: original sample rate (%d) is lower than 16kHz. Up-sampling might produce erratic speech recognition.\n", (int)input->signal.rate);
} }
@ -221,7 +218,6 @@ GetAudioBuffer(const char* path)
unsigned int sample_rate; unsigned int sample_rate;
fseek(wave, 24, SEEK_SET); rv = fread(&sample_rate, 4, 1, wave); fseek(wave, 24, SEEK_SET); rv = fread(&sample_rate, 4, 1, wave);
res.sample_rate = (int)sample_rate;
unsigned short bits_per_sample; unsigned short bits_per_sample;
fseek(wave, 34, SEEK_SET); rv = fread(&bits_per_sample, 2, 1, wave); fseek(wave, 34, SEEK_SET); rv = fread(&bits_per_sample, 2, 1, wave);
@ -269,7 +265,6 @@ ProcessFile(ModelState* context, const char* path, bool show_times)
ds_result result = LocalDsSTT(context, ds_result result = LocalDsSTT(context,
(const short*)audio.buffer, (const short*)audio.buffer,
audio.buffer_size / 2, audio.buffer_size / 2,
audio.sample_rate,
extended_metadata, extended_metadata,
json_output); json_output);
free(audio.buffer); free(audio.buffer);