static double fetch_and_process_audio( ucil_theora_video_file_object_t *vobj, double audiopos )
{
#if HAVE_ALSA
int n_samples;
signed char *buf;
if( !vobj->audio )
{
return audiopos;
}
while( ucil_alsa_fill_audio_buffer( vobj->audio_data ) )
;
n_samples = ucil_alsa_get_audio_buffer( vobj->audio_data, (short**)&buf );
if( !n_samples )
{
return audiopos;
}
if( vobj->async_audio_encoding )
{
audiopos = write_pcm( vobj, audiopos, buf, n_samples );
}
else
{
audiopos = encode_vorbis( vobj->f, &vobj->vd, &vobj->vo, &vobj->vb, vobj->audio_rate, audiopos, buf, n_samples );
}
#endif
return audiopos;
}
int main(int argc, char *argv[]) {
if (argc == 3) {
video_filename = argv[1];
pcm_filaname = argv[2];
} else {
printf("Missing input video file or output file.\n");
return 1;
}
int err;
av_register_all();
AVFormatContext * fmt_ctx = NULL;
err = avformat_open_input( &fmt_ctx, video_filename, NULL, NULL );
assert( err == 0 );
err = avformat_find_stream_info( fmt_ctx, NULL );
assert( err >= 0 );
av_dump_format( fmt_ctx, 0, argv[1], 0 );
// int const video_idx = get_stream_idx( fmt_ctx, AVMEDIA_TYPE_VIDEO );
int const audio_idx = get_stream_idx( fmt_ctx, AVMEDIA_TYPE_AUDIO );
assert(audio_idx >= 0);
// AVCodecContext * video_codec_ctx = new_codec_ctx( fmt_ctx, video_idx );
AVCodecContext * audio_codec_ctx = new_codec_ctx( fmt_ctx, audio_idx );
int is_frame_finish;
static AVPacket pkt;
static uint8_t *pkt_data = NULL;
static int pkt_size = 0;
static AVFrame frame;
AVFrame *decoded_frame = av_frame_alloc();
assert( decoded_frame );
while ( true ) {
if ( av_read_frame( fmt_ctx, &pkt ) < 0 ) {
break;
}
if ( pkt.stream_index == audio_idx )
{
pkt_data = pkt.data;
pkt_size = pkt.size;
while ( pkt_size > 0 )
{
int const byte_consumed = avcodec_decode_audio4(
audio_codec_ctx, decoded_frame, &is_frame_finish, &pkt );
ASSERT_OR_ERRMSG(byte_consumed >= 0, "Errors when decode audio.\n");
pkt_data += byte_consumed;
pkt_size -= byte_consumed;
assert(pkt_size >= 0);
if ( is_frame_finish ) {
int const data_size = av_samples_get_buffer_size(
NULL, audio_codec_ctx->channels,
decoded_frame->nb_samples,
audio_codec_ctx->sample_fmt,
1);
if ( av_sample_fmt_is_planar(audio_codec_ctx->sample_fmt)) {
uint8_t *buf = malloc(data_size);
interleave(decoded_frame->data, buf,
audio_codec_ctx->channels, audio_codec_ctx->sample_fmt, data_size);
write_pcm( buf, data_size, pcm_filaname );
free(buf);
} else {
write_pcm( decoded_frame->data[0], data_size, pcm_filaname );
}
}
}
if (pkt.data) {
av_free_packet( &pkt );
}
}
else
{
av_free_packet( &pkt );
}
}
avcodec_close(audio_codec_ctx);
avformat_close_input(&fmt_ctx);
printf("audio decode done\n");
return 0;
}
static void ci_pcmbuf_insert(const void *ch1, const void *ch2, int count)
{
num_output_samples += count;
if (use_dsp) {
struct dsp_buffer src;
src.remcount = count;
src.pin[0] = ch1;
src.pin[1] = ch2;
src.proc_mask = 0;
while (1) {
int out_count = MAX(count, 512);
int16_t buf[2 * out_count];
struct dsp_buffer dst;
dst.remcount = 0;
dst.p16out = buf;
dst.bufcount = out_count;
dsp_process(ci.dsp, &src, &dst);
if (dst.remcount > 0) {
if (mode == MODE_WRITE)
write_pcm(buf, dst.remcount);
else if (mode == MODE_PLAY)
playback_pcm(buf, dst.remcount);
} else if (src.remcount <= 0) {
break;
}
}
} else {
/* Convert to 32-bit interleaved. */
count *= format.channels;
int i;
int32_t buf[count];
if (format.depth > 16) {
if (format.stereo_mode == STEREO_NONINTERLEAVED) {
for (i = 0; i < count; i += 2) {
buf[i+0] = ((int32_t*)ch1)[i/2];
buf[i+1] = ((int32_t*)ch2)[i/2];
}
} else {
memcpy(buf, ch1, sizeof(buf));
}
} else {
if (format.stereo_mode == STEREO_NONINTERLEAVED) {
for (i = 0; i < count; i += 2) {
buf[i+0] = ((int16_t*)ch1)[i/2];
buf[i+1] = ((int16_t*)ch2)[i/2];
}
} else {
for (i = 0; i < count; i++) {
buf[i] = ((int16_t*)ch1)[i];
}
}
}
if (mode == MODE_WRITE)
write_pcm_raw(buf, count);
}
perform_config();
}
void get_word(int flag)
{
unsigned int size = 0, n = 0, i = 0;
char ans[500];
int frame_n;
void *buffer = NULL;
frame *frames = NULL;
mfcc_frame *mfcc_frames = NULL;
voice_signal *signals = NULL;
printf("\nFale agora e pressione enter\n");
capture_start(NULL);
getchar();
getchar();
capture_stop(&size, &buffer);
n = split(buffer, size / 2, &signals);
for (i = 0; i < n; i++)
{
frame_n = make_frames_hamming(signals[i].buffer, signals[i].number, &frames);
mfcc_frames = malloc(sizeof(mfcc_frame) * frame_n);
mfcc_features(frames, frame_n, mfcc_frames);
if (!flag)
{
record(NULL, signals[i].buffer, signals[i].number * 2);
printf("Escreva a palavra: ");
scanf("%s", (char *)ans);
if (!(ans[0] == 'x' && ans[1] == '\0'))
{
new_word(mfcc_frames, frame_n, (char *)ans);
chdir("waves");
char *path = malloc(strlen(ans) + 5);
char *ext = ".wav";
memcpy(path, ans, strlen(ans));
memcpy(path + strlen(ans), ext, 5);
write_pcm(signals[i].buffer, signals[i].number * 2, path);
free(path);
chdir("..");
}
}
else
{
/*laço de reconhecimento*/
puts("analisando dados...");
word *words = malloc(sizeof(word));
double best = 1e10;
char *name = NULL;
void **word_adresses;
unsigned int n = 0, i = 0, count = 0;
if ((n = get_list(words)))
{
word_adresses = malloc(n * sizeof(void *));
while (words != NULL)
{
double now = compare(mfcc_frames, frame_n, words->frames, words->n);
word_adresses[count++] = words;
if (now < best)
{
best = now;
name = words->name;
}
words = words->next;
}
for (i = 0; i < count; i++)
free(word_adresses[i]);
free(word_adresses);
}
if (best < 3.5)
printf("%f %s", best, name);
}
free(mfcc_frames);
free(frames);
}
if (flag)
printf("\n");
free(buffer);
}
