Spaces:
Running
Running
stream : add "--capture" option to select capture device (ref #10)
Browse files
examples/stream/stream.cpp
CHANGED
|
@@ -35,10 +35,11 @@ std::string to_timestamp(int64_t t) {
|
|
| 35 |
|
| 36 |
// command-line parameters
|
| 37 |
struct whisper_params {
|
| 38 |
-
int32_t seed
|
| 39 |
-
int32_t n_threads
|
| 40 |
-
int32_t step_ms
|
| 41 |
-
int32_t length_ms
|
|
|
|
| 42 |
|
| 43 |
bool verbose = false;
|
| 44 |
bool translate = false;
|
|
@@ -65,6 +66,8 @@ bool whisper_params_parse(int argc, char ** argv, whisper_params & params) {
|
|
| 65 |
params.step_ms = std::stoi(argv[++i]);
|
| 66 |
} else if (arg == "--length") {
|
| 67 |
params.length_ms = std::stoi(argv[++i]);
|
|
|
|
|
|
|
| 68 |
} else if (arg == "-v" || arg == "--verbose") {
|
| 69 |
params.verbose = true;
|
| 70 |
} else if (arg == "--translate") {
|
|
@@ -109,6 +112,7 @@ void whisper_print_usage(int argc, char ** argv, const whisper_params & params)
|
|
| 109 |
fprintf(stderr, " -t N, --threads N number of threads to use during computation (default: %d)\n", params.n_threads);
|
| 110 |
fprintf(stderr, " --step N audio step size in milliseconds (default: %d)\n", params.step_ms);
|
| 111 |
fprintf(stderr, " --length N audio length in milliseconds (default: %d)\n", params.length_ms);
|
|
|
|
| 112 |
fprintf(stderr, " -v, --verbose verbose output\n");
|
| 113 |
fprintf(stderr, " --translate translate from source language to english\n");
|
| 114 |
fprintf(stderr, " -kc, --keep-context keep text context from earlier audio (default: false)\n");
|
|
@@ -201,7 +205,7 @@ int main(int argc, char ** argv) {
|
|
| 201 |
|
| 202 |
// init audio
|
| 203 |
|
| 204 |
-
if (!audio_sdl_init(
|
| 205 |
fprintf(stderr, "%s: audio_sdl_init() failed!\n", __func__);
|
| 206 |
return 1;
|
| 207 |
}
|
|
|
|
| 35 |
|
| 36 |
// command-line parameters
|
| 37 |
struct whisper_params {
|
| 38 |
+
int32_t seed = -1; // RNG seed, not used currently
|
| 39 |
+
int32_t n_threads = std::min(4, (int32_t) std::thread::hardware_concurrency());
|
| 40 |
+
int32_t step_ms = 3000;
|
| 41 |
+
int32_t length_ms = 10000;
|
| 42 |
+
int32_t capture_id = -1;
|
| 43 |
|
| 44 |
bool verbose = false;
|
| 45 |
bool translate = false;
|
|
|
|
| 66 |
params.step_ms = std::stoi(argv[++i]);
|
| 67 |
} else if (arg == "--length") {
|
| 68 |
params.length_ms = std::stoi(argv[++i]);
|
| 69 |
+
} else if (arg == "-c" || arg == "--capture") {
|
| 70 |
+
params.capture_id = std::stoi(argv[++i]);
|
| 71 |
} else if (arg == "-v" || arg == "--verbose") {
|
| 72 |
params.verbose = true;
|
| 73 |
} else if (arg == "--translate") {
|
|
|
|
| 112 |
fprintf(stderr, " -t N, --threads N number of threads to use during computation (default: %d)\n", params.n_threads);
|
| 113 |
fprintf(stderr, " --step N audio step size in milliseconds (default: %d)\n", params.step_ms);
|
| 114 |
fprintf(stderr, " --length N audio length in milliseconds (default: %d)\n", params.length_ms);
|
| 115 |
+
fprintf(stderr, " -c ID, --capture ID capture device ID (default: -1)\n");
|
| 116 |
fprintf(stderr, " -v, --verbose verbose output\n");
|
| 117 |
fprintf(stderr, " --translate translate from source language to english\n");
|
| 118 |
fprintf(stderr, " -kc, --keep-context keep text context from earlier audio (default: false)\n");
|
|
|
|
| 205 |
|
| 206 |
// init audio
|
| 207 |
|
| 208 |
+
if (!audio_sdl_init(params.capture_id)) {
|
| 209 |
fprintf(stderr, "%s: audio_sdl_init() failed!\n", __func__);
|
| 210 |
return 1;
|
| 211 |
}
|