bool SpeexFile::Open(const ACE_TString& filename,
int bandmode,
bool vbr)
{
switch(bandmode)
{
case SPEEX_MODEID_NB :
speex_init_header(&m_spx_header, 8000, 1,
speex_lib_get_mode(SPEEX_MODEID_NB));
break;
case SPEEX_MODEID_WB :
speex_init_header(&m_spx_header, 16000, 1,
speex_lib_get_mode(SPEEX_MODEID_WB));
break;
case SPEEX_MODEID_UWB :
speex_init_header(&m_spx_header, 32000, 1,
speex_lib_get_mode(SPEEX_MODEID_UWB));
break;
default:
assert(0);
return false;
}
m_spx_header.frames_per_packet = 1;
m_spx_header.nb_channels = 1;
m_spx_header.vbr = vbr;
SpeexEncoder enc;
if(!enc.Initialize(bandmode, 2, 5))
{
Close();
return false;
}
spx_int32_t lookahead = 0;
enc.GetOption(SPEEX_GET_LOOKAHEAD, lookahead);
if(!m_speex.Open(m_spx_header, lookahead))
{
Close();
return false;
}
if(!m_ogg.Open(filename))
{
Close();
return false;
}
//write remaining audio headers
int ret;
while(m_speex.FlushPageOut(m_aud_page)>0)
{
ret = m_ogg.WriteOggPage(m_aud_page);
assert(ret>=0);
}
return true;
}
static int CreateDefaultHeader( decoder_t *p_dec )
{
ogg_packet oggpacket;
SpeexHeader *p_header = malloc( sizeof(SpeexHeader) );
if( !p_header )
return VLC_ENOMEM;
const int rate = p_dec->fmt_in.audio.i_rate;
const unsigned i_mode = (rate / 8000) >> 1;
const SpeexMode *mode;
int ret = VLC_SUCCESS;
oggpacket.packet = NULL;
switch( rate )
{
case 8000:
case 16000:
case 32000:
mode = speex_lib_get_mode( i_mode );
break;
default:
msg_Err( p_dec, "Unexpected rate %d", rate );
ret = VLC_EGENERIC;
goto cleanup;
}
speex_init_header( p_header, rate, p_dec->fmt_in.audio.i_channels, mode );
p_header->frames_per_packet = 160 << i_mode;
oggpacket.packet = (unsigned char *) speex_header_to_packet( p_header,
(int *) &oggpacket.bytes );
if( !oggpacket.packet )
{
ret = VLC_ENOMEM;
goto cleanup;
}
oggpacket.b_o_s = 1;
oggpacket.e_o_s = 0;
oggpacket.granulepos = -1;
oggpacket.packetno = 0;
ret = ProcessInitialHeader( p_dec, &oggpacket );
if( ret != VLC_SUCCESS )
{
msg_Err( p_dec, "default Speex header is corrupted" );
}
cleanup:
free( oggpacket.packet );
free( p_header );
return ret;
}
void Ihu2Spx::setup(QString filename, const SpeexMode *mode, int rate, int fsize)
{
if (fout)
fclose(fout);
filename.append(IHU_SPX_EXT);
fout = fopen(filename.ascii(), "wb");
if (fout)
{
total_samples = 0;
bytes_written = 0;
id = -1;
frame_size = fsize;
if (ogg_stream_init(&os, rand())==-1)
throw Error(QString("stream init failed"));
speex_init_header(&header, rate, 1, mode);
header.frames_per_packet = 1;
op.packet = (unsigned char *)speex_header_to_packet(&header, (int*)&(op.bytes));
op.b_o_s = 1;
op.e_o_s = 0;
op.granulepos = 0;
op.packetno = 0;
ogg_stream_packetin(&os, &op);
free(op.packet);
op.packet = (unsigned char *)comments;
op.bytes = comments_length;
op.b_o_s = 0;
op.e_o_s = 0;
op.granulepos = 0;
op.packetno = 1;
ogg_stream_packetin(&os, &op);
while((result = ogg_stream_flush(&os, &og)))
{
if(!result) break;
ret = oe_write_page(&og, fout);
if(ret != og.header_len + og.body_len)
throw Error(QString("failed writing header to output stream"));
else
bytes_written += ret;
}
}
else
{
throw Error(QString("%1: %2").arg(filename).arg(strerror(errno)));
}
}
static int speex_output_open(const char *fname, const char *comment)
{
Speex_ctx *ctx;
int fd;
if (!(speex_ctx = (Speex_ctx *)calloc(sizeof(Speex_ctx), 1))) {
ctl->cmsg(CMSG_ERROR, VERB_NORMAL, "%s", strerror(errno));
return -1;
}
ctx = speex_ctx;
/* init id */
ctx->in_bytes = ctx->out_bytes = 0;
ctx->ogg_packetid = -1;
ctx->channels = (dpm.encoding & PE_MONO) ? 1 : 2;
if(strcmp(fname, "-") == 0) {
fd = 1; /* data to stdout */
if(comment == NULL)
comment = "(stdout)";
} else {
/* Open the audio file */
fd = open(fname, FILE_OUTPUT_MODE);
if (fd < 0) {
ctl->cmsg(CMSG_ERROR, VERB_NORMAL, "%s: %s",
fname, strerror(errno));
return -1;
}
}
/*Initialize Ogg stream struct*/
if (ogg_seqnum == 0) {
srand(time(NULL));
ogg_seqnum = rand();
}
if (ogg_stream_init(&ctx->os, ogg_seqnum++) == -1) {
ctl->cmsg(CMSG_ERROR, VERB_NORMAL, "Ogg stream init failed\n");
return -1;
}
if (!speex_mode_preset())
return -1;
speex_init_header(&ctx->header, dpm.rate, 1, ctx->mode);
ctx->header.nb_channels = ctx->channels;
ctx->header.frames_per_packet = ctx->nframes = speex_options.nframes;
ctx->state = speex_encoder_init(ctx->mode);
/*Set the quality to 8 (15 kbps)*/
speex_encoder_ctl(ctx->state, SPEEX_SET_QUALITY, &speex_options.quality);
if(strcmp(fname, "-") == 0) {
fd = 1; /* data to stdout */
if(comment == NULL)
comment = "(stdout)";
} else {
/* Open the audio file */
fd = open(fname, FILE_OUTPUT_MODE);
if(fd < 0) {
ctl->cmsg(CMSG_ERROR, VERB_NORMAL, "%s: %s",
fname, strerror(errno));
return -1;
}
if(comment == NULL)
comment = fname;
}
write_ogg_header(ctx, fd, (char *)comment);
speex_encoder_ctl(ctx->state, SPEEX_GET_FRAME_SIZE, &ctx->frame_size);
speex_encoder_ctl(ctx->state, SPEEX_SET_COMPLEXITY, &speex_options.complexity);
speex_encoder_ctl(ctx->state, SPEEX_SET_SAMPLING_RATE, &dpm.rate);
if (speex_options.vbr) {
speex_encoder_ctl(ctx->state, SPEEX_SET_VBR, &speex_options.vbr);
}
else if (speex_options.vad) {
speex_encoder_ctl(ctx->state, SPEEX_SET_VAD, &speex_options.vad);
}
if (speex_options.dtx)
speex_encoder_ctl(ctx->state, SPEEX_SET_DTX, &speex_options.dtx);
if (speex_options.dtx && !(speex_options.vbr ||
speex_options.abr ||
speex_options.vad)) {
ctl->cmsg(CMSG_ERROR, VERB_NORMAL,
"--speex-dtx is useless without --speex-vad, --speex-vbr or --speex-abr");
}
else if ((speex_options.vbr || speex_options.abr) && (speex_options.vad)) {
ctl->cmsg(CMSG_ERROR, VERB_NORMAL,
"--speex-vad is already implied by --speex-vbr or --speex-abr");
}
if (speex_options.abr) {
speex_encoder_ctl(ctx->state, SPEEX_SET_ABR, &speex_options.abr);
}
speex_bits_init(&ctx->bits);
ctx->input = (float *)safe_malloc(sizeof(float) * MAX_FRAME_SIZE * ctx->channels);
ctx->input_idx = 0;
return fd;
}
static block_t *DecodeRtpSpeexPacket( decoder_t *p_dec, block_t **pp_block )
{
block_t *p_speex_bit_block = *pp_block;
decoder_sys_t *p_sys = p_dec->p_sys;
block_t *p_aout_buffer;
int i_decode_ret;
unsigned int i_speex_frame_size;
if ( !p_speex_bit_block || p_speex_bit_block->i_pts <= VLC_TS_INVALID )
return NULL;
/*
If the SpeexBits buffer size is 0 (a default value),
we know that a proper initialization has not yet been done.
*/
if ( p_sys->bits.buf_size==0 )
{
p_sys->p_header = malloc(sizeof(SpeexHeader));
if ( !p_sys->p_header )
{
msg_Err( p_dec, "Could not allocate a Speex header.");
return NULL;
}
const SpeexMode *mode = speex_lib_get_mode((p_sys->rtp_rate / 8000) >> 1);
speex_init_header( p_sys->p_header,p_sys->rtp_rate, 1, mode );
speex_bits_init( &p_sys->bits );
p_sys->p_state = speex_decoder_init( mode );
if ( !p_sys->p_state )
{
msg_Err( p_dec, "Could not allocate a Speex decoder." );
free( p_sys->p_header );
return NULL;
}
/*
Assume that variable bit rate is enabled. Also assume
that there is only one frame per packet.
*/
p_sys->p_header->vbr = 1;
p_sys->p_header->frames_per_packet = 1;
p_dec->fmt_out.audio.i_channels = p_sys->p_header->nb_channels;
p_dec->fmt_out.audio.i_physical_channels =
p_dec->fmt_out.audio.i_original_channels =
pi_channels_maps[p_sys->p_header->nb_channels];
p_dec->fmt_out.audio.i_rate = p_sys->p_header->rate;
if ( speex_mode_query( &speex_nb_mode,
SPEEX_MODE_FRAME_SIZE,
&i_speex_frame_size ) )
{
msg_Err( p_dec, "Could not determine the frame size." );
speex_decoder_destroy( p_sys->p_state );
free( p_sys->p_header );
return NULL;
}
p_dec->fmt_out.audio.i_bytes_per_frame = i_speex_frame_size;
date_Init(&p_sys->end_date, p_sys->p_header->rate, 1);
}
static gboolean
gst_speex_enc_setup (GstSpeexEnc * enc)
{
enc->setup = FALSE;
switch (enc->mode) {
case GST_SPEEX_ENC_MODE_UWB:
GST_LOG_OBJECT (enc, "configuring for requested UWB mode");
enc->speex_mode = (SpeexMode *) & speex_uwb_mode;
break;
case GST_SPEEX_ENC_MODE_WB:
GST_LOG_OBJECT (enc, "configuring for requested WB mode");
enc->speex_mode = (SpeexMode *) & speex_wb_mode;
break;
case GST_SPEEX_ENC_MODE_NB:
GST_LOG_OBJECT (enc, "configuring for requested NB mode");
enc->speex_mode = (SpeexMode *) & speex_nb_mode;
break;
case GST_SPEEX_ENC_MODE_AUTO:
/* fall through */
GST_LOG_OBJECT (enc, "finding best mode");
default:
break;
}
if (enc->rate > 25000) {
if (enc->mode == GST_SPEEX_ENC_MODE_AUTO) {
GST_LOG_OBJECT (enc, "selected UWB mode for samplerate %d", enc->rate);
enc->speex_mode = (SpeexMode *) & speex_uwb_mode;
} else {
if (enc->speex_mode != &speex_uwb_mode) {
gst_speex_enc_set_last_msg (enc,
"Warning: suggest to use ultra wide band mode for this rate");
}
}
} else if (enc->rate > 12500) {
if (enc->mode == GST_SPEEX_ENC_MODE_AUTO) {
GST_LOG_OBJECT (enc, "selected WB mode for samplerate %d", enc->rate);
enc->speex_mode = (SpeexMode *) & speex_wb_mode;
} else {
if (enc->speex_mode != &speex_wb_mode) {
gst_speex_enc_set_last_msg (enc,
"Warning: suggest to use wide band mode for this rate");
}
}
} else {
if (enc->mode == GST_SPEEX_ENC_MODE_AUTO) {
GST_LOG_OBJECT (enc, "selected NB mode for samplerate %d", enc->rate);
enc->speex_mode = (SpeexMode *) & speex_nb_mode;
} else {
if (enc->speex_mode != &speex_nb_mode) {
gst_speex_enc_set_last_msg (enc,
"Warning: suggest to use narrow band mode for this rate");
}
}
}
if (enc->rate != 8000 && enc->rate != 16000 && enc->rate != 32000) {
gst_speex_enc_set_last_msg (enc,
"Warning: speex is optimized for 8, 16 and 32 KHz");
}
speex_init_header (&enc->header, enc->rate, 1, enc->speex_mode);
enc->header.frames_per_packet = enc->nframes;
enc->header.vbr = enc->vbr;
enc->header.nb_channels = enc->channels;
/*Initialize Speex encoder */
enc->state = speex_encoder_init (enc->speex_mode);
speex_encoder_ctl (enc->state, SPEEX_GET_FRAME_SIZE, &enc->frame_size);
speex_encoder_ctl (enc->state, SPEEX_SET_COMPLEXITY, &enc->complexity);
speex_encoder_ctl (enc->state, SPEEX_SET_SAMPLING_RATE, &enc->rate);
if (enc->vbr)
speex_encoder_ctl (enc->state, SPEEX_SET_VBR_QUALITY, &enc->quality);
else {
gint tmp = floor (enc->quality);
speex_encoder_ctl (enc->state, SPEEX_SET_QUALITY, &tmp);
}
if (enc->bitrate) {
if (enc->quality >= 0.0 && enc->vbr) {
gst_speex_enc_set_last_msg (enc,
"Warning: bitrate option is overriding quality");
}
speex_encoder_ctl (enc->state, SPEEX_SET_BITRATE, &enc->bitrate);
}
if (enc->vbr) {
gint tmp = 1;
speex_encoder_ctl (enc->state, SPEEX_SET_VBR, &tmp);
} else if (enc->vad) {
gint tmp = 1;
speex_encoder_ctl (enc->state, SPEEX_SET_VAD, &tmp);
}
if (enc->dtx) {
gint tmp = 1;
speex_encoder_ctl (enc->state, SPEEX_SET_DTX, &tmp);
}
if (enc->dtx && !(enc->vbr || enc->abr || enc->vad)) {
gst_speex_enc_set_last_msg (enc,
"Warning: dtx is useless without vad, vbr or abr");
} else if ((enc->vbr || enc->abr) && (enc->vad)) {
gst_speex_enc_set_last_msg (enc,
"Warning: vad is already implied by vbr or abr");
}
if (enc->abr) {
speex_encoder_ctl (enc->state, SPEEX_SET_ABR, &enc->abr);
}
speex_encoder_ctl (enc->state, SPEEX_GET_LOOKAHEAD, &enc->lookahead);
GST_LOG_OBJECT (enc, "we have frame size %d, lookahead %d", enc->frame_size,
enc->lookahead);
enc->setup = TRUE;
return TRUE;
}
int main(int argc, char **argv)
{
int nb_samples, total_samples=0, nb_encoded;
int c;
int option_index = 0;
char *inFile, *outFile;
FILE *fin, *fout;
short input[MAX_FRAME_SIZE];
spx_int32_t frame_size;
int quiet=0;
spx_int32_t vbr_enabled=0;
spx_int32_t vbr_max=0;
int abr_enabled=0;
spx_int32_t vad_enabled=0;
spx_int32_t dtx_enabled=0;
int nbBytes;
const SpeexMode *mode=NULL;
int modeID = -1;
void *st;
SpeexBits bits;
char cbits[MAX_FRAME_BYTES];
int with_skeleton = 0;
struct option long_options[] =
{
{"wideband", no_argument, NULL, 0},
{"ultra-wideband", no_argument, NULL, 0},
{"narrowband", no_argument, NULL, 0},
{"vbr", no_argument, NULL, 0},
{"vbr-max-bitrate", required_argument, NULL, 0},
{"abr", required_argument, NULL, 0},
{"vad", no_argument, NULL, 0},
{"dtx", no_argument, NULL, 0},
{"quality", required_argument, NULL, 0},
{"bitrate", required_argument, NULL, 0},
{"nframes", required_argument, NULL, 0},
{"comp", required_argument, NULL, 0},
#ifdef USE_SPEEXDSP
{"denoise", no_argument, NULL, 0},
{"agc", no_argument, NULL, 0},
#endif
{"no-highpass", no_argument, NULL, 0},
{"skeleton",no_argument,NULL, 0},
{"help", no_argument, NULL, 0},
{"quiet", no_argument, NULL, 0},
{"le", no_argument, NULL, 0},
{"be", no_argument, NULL, 0},
{"8bit", no_argument, NULL, 0},
{"16bit", no_argument, NULL, 0},
{"stereo", no_argument, NULL, 0},
{"rate", required_argument, NULL, 0},
{"version", no_argument, NULL, 0},
{"version-short", no_argument, NULL, 0},
{"comment", required_argument, NULL, 0},
{"author", required_argument, NULL, 0},
{"title", required_argument, NULL, 0},
{"print-rate", no_argument, NULL, 0},
{0, 0, 0, 0}
};
int print_bitrate=0;
spx_int32_t rate=0;
spx_int32_t size;
int chan=1;
int fmt=16;
spx_int32_t quality=-1;
float vbr_quality=-1;
int lsb=1;
ogg_stream_state os;
ogg_stream_state so; /* ogg stream for skeleton bitstream */
ogg_page og;
ogg_packet op;
int bytes_written=0, ret, result;
int id=-1;
SpeexHeader header;
int nframes=1;
spx_int32_t complexity=3;
const char* speex_version;
char vendor_string[64];
char *comments;
int comments_length;
int close_in=0, close_out=0;
int eos=0;
spx_int32_t bitrate=0;
double cumul_bits=0, enc_frames=0;
char first_bytes[12];
int wave_input=0;
spx_int32_t tmp;
#ifdef USE_SPEEXDSP
SpeexPreprocessState *preprocess = NULL;
int denoise_enabled=0, agc_enabled=0;
#endif
int highpass_enabled=1;
int output_rate=0;
spx_int32_t lookahead = 0;
speex_lib_ctl(SPEEX_LIB_GET_VERSION_STRING, (void*)&speex_version);
snprintf(vendor_string, sizeof(vendor_string), "Encoded with Speex %s", speex_version);
comment_init(&comments, &comments_length, vendor_string);
/*Process command-line options*/
while(1)
{
c = getopt_long (argc, argv, "nwuhvV",
long_options, &option_index);
if (c==-1)
break;
switch(c)
{
case 0:
if (strcmp(long_options[option_index].name,"narrowband")==0)
{
modeID = SPEEX_MODEID_NB;
} else if (strcmp(long_options[option_index].name,"wideband")==0)
{
modeID = SPEEX_MODEID_WB;
} else if (strcmp(long_options[option_index].name,"ultra-wideband")==0)
{
modeID = SPEEX_MODEID_UWB;
} else if (strcmp(long_options[option_index].name,"vbr")==0)
{
vbr_enabled=1;
} else if (strcmp(long_options[option_index].name,"vbr-max-bitrate")==0)
{
vbr_max=atoi(optarg);
if (vbr_max<1)
{
fprintf (stderr, "Invalid VBR max bit-rate value: %d\n", vbr_max);
exit(1);
}
} else if (strcmp(long_options[option_index].name,"abr")==0)
{
abr_enabled=atoi(optarg);
if (!abr_enabled)
{
fprintf (stderr, "Invalid ABR value: %d\n", abr_enabled);
exit(1);
}
} else if (strcmp(long_options[option_index].name,"vad")==0)
{
vad_enabled=1;
} else if (strcmp(long_options[option_index].name,"dtx")==0)
{
dtx_enabled=1;
} else if (strcmp(long_options[option_index].name,"quality")==0)
{
quality = atoi (optarg);
vbr_quality=atof(optarg);
} else if (strcmp(long_options[option_index].name,"bitrate")==0)
{
bitrate = atoi (optarg);
} else if (strcmp(long_options[option_index].name,"nframes")==0)
{
nframes = atoi (optarg);
if (nframes<1)
nframes=1;
if (nframes>10)
nframes=10;
} else if (strcmp(long_options[option_index].name,"comp")==0)
{
complexity = atoi (optarg);
#ifdef USE_SPEEXDSP
} else if (strcmp(long_options[option_index].name,"denoise")==0)
{
denoise_enabled=1;
} else if (strcmp(long_options[option_index].name,"agc")==0)
{
agc_enabled=1;
#endif
} else if (strcmp(long_options[option_index].name,"no-highpass")==0)
{
highpass_enabled=0;
} else if (strcmp(long_options[option_index].name,"skeleton")==0)
{
with_skeleton=1;
} else if (strcmp(long_options[option_index].name,"help")==0)
{
usage();
exit(0);
} else if (strcmp(long_options[option_index].name,"quiet")==0)
{
quiet = 1;
} else if (strcmp(long_options[option_index].name,"version")==0)
{
version();
exit(0);
} else if (strcmp(long_options[option_index].name,"version-short")==0)
{
version_short();
exit(0);
} else if (strcmp(long_options[option_index].name,"print-rate")==0)
{
output_rate=1;
} else if (strcmp(long_options[option_index].name,"le")==0)
{
lsb=1;
} else if (strcmp(long_options[option_index].name,"be")==0)
{
lsb=0;
} else if (strcmp(long_options[option_index].name,"8bit")==0)
{
fmt=8;
} else if (strcmp(long_options[option_index].name,"16bit")==0)
{
fmt=16;
} else if (strcmp(long_options[option_index].name,"stereo")==0)
{
chan=2;
} else if (strcmp(long_options[option_index].name,"rate")==0)
{
rate=atoi (optarg);
} else if (strcmp(long_options[option_index].name,"comment")==0)
{
if (!strchr(optarg, '='))
{
fprintf (stderr, "Invalid comment: %s\n", optarg);
fprintf (stderr, "Comments must be of the form name=value\n");
exit(1);
}
comment_add(&comments, &comments_length, NULL, optarg);
} else if (strcmp(long_options[option_index].name,"author")==0)
{
comment_add(&comments, &comments_length, "author=", optarg);
} else if (strcmp(long_options[option_index].name,"title")==0)
{
comment_add(&comments, &comments_length, "title=", optarg);
}
break;
case 'n':
modeID = SPEEX_MODEID_NB;
break;
case 'h':
usage();
exit(0);
break;
case 'v':
version();
exit(0);
break;
case 'V':
print_bitrate=1;
break;
case 'w':
modeID = SPEEX_MODEID_WB;
break;
case 'u':
modeID = SPEEX_MODEID_UWB;
break;
case '?':
usage();
exit(1);
break;
}
}
if (argc-optind!=2)
{
usage();
exit(1);
}
inFile=argv[optind];
outFile=argv[optind+1];
/*Initialize Ogg stream struct*/
srand(time(NULL));
if (ogg_stream_init(&os, rand())==-1)
{
fprintf(stderr,"Error: stream init failed\n");
exit(1);
}
if (with_skeleton && ogg_stream_init(&so, rand())==-1)
{
fprintf(stderr,"Error: stream init failed\n");
exit(1);
}
if (strcmp(inFile, "-")==0)
{
#if defined WIN32 || defined _WIN32
_setmode(_fileno(stdin), _O_BINARY);
#elif defined OS2
_fsetmode(stdin,"b");
#endif
fin=stdin;
}
else
{
fin = fopen(inFile, "rb");
if (!fin)
{
perror(inFile);
exit(1);
}
close_in=1;
}
{
if (fread(first_bytes, 1, 12, fin) != 12)
{
perror("short file");
exit(1);
}
if (strncmp(first_bytes,"RIFF",4)==0 || strncmp(first_bytes,"riff",4)==0)
{
if (read_wav_header(fin, &rate, &chan, &fmt, &size)==-1)
exit(1);
wave_input=1;
lsb=1; /* CHECK: exists big-endian .wav ?? */
}
}
if (modeID==-1 && !rate)
{
/* By default, use narrowband/8 kHz */
modeID = SPEEX_MODEID_NB;
rate=8000;
} else if (modeID!=-1 && rate)
{
mode = speex_lib_get_mode (modeID);
if (rate>48000)
{
fprintf (stderr, "Error: sampling rate too high: %d Hz, try down-sampling\n", rate);
exit(1);
} else if (rate>25000)
{
if (modeID != SPEEX_MODEID_UWB)
{
fprintf (stderr, "Warning: Trying to encode in %s at %d Hz. I'll do it but I suggest you try ultra-wideband instead\n", mode->modeName , rate);
}
} else if (rate>12500)
{
if (modeID != SPEEX_MODEID_WB)
{
fprintf (stderr, "Warning: Trying to encode in %s at %d Hz. I'll do it but I suggest you try wideband instead\n", mode->modeName , rate);
}
} else if (rate>=6000)
{
if (modeID != SPEEX_MODEID_NB)
{
fprintf (stderr, "Warning: Trying to encode in %s at %d Hz. I'll do it but I suggest you try narrowband instead\n", mode->modeName , rate);
}
} else {
fprintf (stderr, "Error: sampling rate too low: %d Hz\n", rate);
exit(1);
}
} else if (modeID==-1)
{
if (rate>48000)
{
fprintf (stderr, "Error: sampling rate too high: %d Hz, try down-sampling\n", rate);
exit(1);
} else if (rate>25000)
{
modeID = SPEEX_MODEID_UWB;
} else if (rate>12500)
{
modeID = SPEEX_MODEID_WB;
} else if (rate>=6000)
{
modeID = SPEEX_MODEID_NB;
} else {
fprintf (stderr, "Error: Sampling rate too low: %d Hz\n", rate);
exit(1);
}
} else if (!rate)
{
if (modeID == SPEEX_MODEID_NB)
rate=8000;
else if (modeID == SPEEX_MODEID_WB)
rate=16000;
else if (modeID == SPEEX_MODEID_UWB)
rate=32000;
}
if (!quiet)
if (rate!=8000 && rate!=16000 && rate!=32000)
fprintf (stderr, "Warning: Speex is only optimized for 8, 16 and 32 kHz. It will still work at %d Hz but your mileage may vary\n", rate);
if (!mode)
mode = speex_lib_get_mode (modeID);
speex_init_header(&header, rate, 1, mode);
header.frames_per_packet=nframes;
header.vbr=vbr_enabled;
header.nb_channels = chan;
{
char *st_string="mono";
if (chan==2)
st_string="stereo";
if (!quiet)
fprintf (stderr, "Encoding %d Hz audio using %s mode (%s)\n",
header.rate, mode->modeName, st_string);
}
/*fprintf (stderr, "Encoding %d Hz audio at %d bps using %s mode\n",
header.rate, mode->bitrate, mode->modeName);*/
/*Initialize Speex encoder*/
st = speex_encoder_init(mode);
if (strcmp(outFile,"-")==0)
{
#if defined WIN32 || defined _WIN32
_setmode(_fileno(stdout), _O_BINARY);
#endif
fout=stdout;
}
else
{
fout = fopen(outFile, "wb");
if (!fout)
{
perror(outFile);
exit(1);
}
close_out=1;
}
speex_encoder_ctl(st, SPEEX_GET_FRAME_SIZE, &frame_size);
speex_encoder_ctl(st, SPEEX_SET_COMPLEXITY, &complexity);
speex_encoder_ctl(st, SPEEX_SET_SAMPLING_RATE, &rate);
if (quality >= 0)
{
if (vbr_enabled)
{
if (vbr_max>0)
speex_encoder_ctl(st, SPEEX_SET_VBR_MAX_BITRATE, &vbr_max);
speex_encoder_ctl(st, SPEEX_SET_VBR_QUALITY, &vbr_quality);
}
else
speex_encoder_ctl(st, SPEEX_SET_QUALITY, &quality);
}
if (bitrate)
{
if (quality >= 0 && vbr_enabled)
fprintf (stderr, "Warning: --bitrate option is overriding --quality\n");
speex_encoder_ctl(st, SPEEX_SET_BITRATE, &bitrate);
}
if (vbr_enabled)
{
tmp=1;
speex_encoder_ctl(st, SPEEX_SET_VBR, &tmp);
} else if (vad_enabled)
{
tmp=1;
speex_encoder_ctl(st, SPEEX_SET_VAD, &tmp);
}
if (dtx_enabled)
speex_encoder_ctl(st, SPEEX_SET_DTX, &tmp);
if (dtx_enabled && !(vbr_enabled || abr_enabled || vad_enabled))
{
fprintf (stderr, "Warning: --dtx is useless without --vad, --vbr or --abr\n");
} else if ((vbr_enabled || abr_enabled) && (vad_enabled))
{
fprintf (stderr, "Warning: --vad is already implied by --vbr or --abr\n");
}
if (with_skeleton) {
fprintf (stderr, "Warning: Enabling skeleton output may cause some decoders to fail.\n");
}
if (abr_enabled)
{
speex_encoder_ctl(st, SPEEX_SET_ABR, &abr_enabled);
}
speex_encoder_ctl(st, SPEEX_SET_HIGHPASS, &highpass_enabled);
speex_encoder_ctl(st, SPEEX_GET_LOOKAHEAD, &lookahead);
#ifdef USE_SPEEXDSP
if (denoise_enabled || agc_enabled)
{
preprocess = speex_preprocess_state_init(frame_size, rate);
speex_preprocess_ctl(preprocess, SPEEX_PREPROCESS_SET_DENOISE, &denoise_enabled);
speex_preprocess_ctl(preprocess, SPEEX_PREPROCESS_SET_AGC, &agc_enabled);
lookahead += frame_size;
}
#endif
/* first packet should be the skeleton header. */
if (with_skeleton) {
add_fishead_packet(&so);
if ((ret = flush_ogg_stream_to_file(&so, fout))) {
fprintf (stderr,"Error: failed skeleton (fishead) header to output stream\n");
exit(1);
} else
bytes_written += ret;
}
/*Write header*/
{
int packet_size;
op.packet = (unsigned char *)speex_header_to_packet(&header, &packet_size);
op.bytes = packet_size;
op.b_o_s = 1;
op.e_o_s = 0;
op.granulepos = 0;
op.packetno = 0;
ogg_stream_packetin(&os, &op);
free(op.packet);
while((result = ogg_stream_flush(&os, &og)))
{
if(!result) break;
ret = oe_write_page(&og, fout);
if(ret != og.header_len + og.body_len)
{
fprintf (stderr,"Error: failed writing header to output stream\n");
exit(1);
}
else
bytes_written += ret;
}
op.packet = (unsigned char *)comments;
op.bytes = comments_length;
op.b_o_s = 0;
op.e_o_s = 0;
op.granulepos = 0;
op.packetno = 1;
ogg_stream_packetin(&os, &op);
}
/* fisbone packet should be write after all bos pages */
if (with_skeleton) {
add_fisbone_packet(&so, os.serialno, &header);
if ((ret = flush_ogg_stream_to_file(&so, fout))) {
fprintf (stderr,"Error: failed writing skeleton (fisbone )header to output stream\n");
exit(1);
} else
bytes_written += ret;
}
/* writing the rest of the speex header packets */
while((result = ogg_stream_flush(&os, &og)))
{
if(!result) break;
ret = oe_write_page(&og, fout);
if(ret != og.header_len + og.body_len)
{
fprintf (stderr,"Error: failed writing header to output stream\n");
exit(1);
}
else
bytes_written += ret;
}
free(comments);
/* write the skeleton eos packet */
if (with_skeleton) {
add_eos_packet_to_stream(&so);
if ((ret = flush_ogg_stream_to_file(&so, fout))) {
fprintf (stderr,"Error: failed writing skeleton header to output stream\n");
exit(1);
} else
bytes_written += ret;
}
speex_bits_init(&bits);
if (!wave_input)
{
nb_samples = read_samples(fin,frame_size,fmt,chan,lsb,input, first_bytes, NULL);
} else {
nb_samples = read_samples(fin,frame_size,fmt,chan,lsb,input, NULL, &size);
}
if (nb_samples==0)
eos=1;
total_samples += nb_samples;
nb_encoded = -lookahead;
/*Main encoding loop (one frame per iteration)*/
while (!eos || total_samples>nb_encoded)
{
id++;
/*Encode current frame*/
if (chan==2)
speex_encode_stereo_int(input, frame_size, &bits);
#ifdef USE_SPEEXDSP
if (preprocess)
speex_preprocess(preprocess, input, NULL);
#endif
speex_encode_int(st, input, &bits);
nb_encoded += frame_size;
if (print_bitrate) {
int tmp;
char ch=13;
speex_encoder_ctl(st, SPEEX_GET_BITRATE, &tmp);
fputc (ch, stderr);
cumul_bits += tmp;
enc_frames += 1;
if (!quiet)
{
if (vad_enabled || vbr_enabled || abr_enabled)
fprintf (stderr, "Bitrate is use: %d bps (average %d bps) ", tmp, (int)(cumul_bits/enc_frames));
else
fprintf (stderr, "Bitrate is use: %d bps ", tmp);
if (output_rate)
printf ("%d\n", tmp);
}
}
if (wave_input)
{
nb_samples = read_samples(fin,frame_size,fmt,chan,lsb,input, NULL, &size);
} else {
nb_samples = read_samples(fin,frame_size,fmt,chan,lsb,input, NULL, NULL);
}
if (nb_samples==0)
{
eos=1;
}
if (eos && total_samples<=nb_encoded)
op.e_o_s = 1;
else
op.e_o_s = 0;
total_samples += nb_samples;
if ((id+1)%nframes!=0)
continue;
speex_bits_insert_terminator(&bits);
nbBytes = speex_bits_write(&bits, cbits, MAX_FRAME_BYTES);
speex_bits_reset(&bits);
op.packet = (unsigned char *)cbits;
op.bytes = nbBytes;
op.b_o_s = 0;
/*Is this redundent?*/
if (eos && total_samples<=nb_encoded)
op.e_o_s = 1;
else
op.e_o_s = 0;
op.granulepos = (id+1)*frame_size-lookahead;
if (op.granulepos>total_samples)
op.granulepos = total_samples;
/*printf ("granulepos: %d %d %d %d %d %d\n", (int)op.granulepos, id, nframes, lookahead, 5, 6);*/
op.packetno = 2+id/nframes;
ogg_stream_packetin(&os, &op);
/*Write all new pages (most likely 0 or 1)*/
while (ogg_stream_pageout(&os,&og))
{
ret = oe_write_page(&og, fout);
if(ret != og.header_len + og.body_len)
{
fprintf (stderr,"Error: failed writing header to output stream\n");
exit(1);
}
else
bytes_written += ret;
}
}
if ((id+1)%nframes!=0)
{
while ((id+1)%nframes!=0)
{
id++;
speex_bits_pack(&bits, 15, 5);
}
nbBytes = speex_bits_write(&bits, cbits, MAX_FRAME_BYTES);
op.packet = (unsigned char *)cbits;
op.bytes = nbBytes;
op.b_o_s = 0;
op.e_o_s = 1;
op.granulepos = (id+1)*frame_size-lookahead;
if (op.granulepos>total_samples)
op.granulepos = total_samples;
op.packetno = 2+id/nframes;
ogg_stream_packetin(&os, &op);
}
/*Flush all pages left to be written*/
while (ogg_stream_flush(&os, &og))
{
ret = oe_write_page(&og, fout);
if(ret != og.header_len + og.body_len)
{
fprintf (stderr,"Error: failed writing header to output stream\n");
exit(1);
}
else
bytes_written += ret;
}
speex_encoder_destroy(st);
speex_bits_destroy(&bits);
ogg_stream_clear(&os);
if (close_in)
fclose(fin);
if (close_out)
fclose(fout);
return 0;
}
int SpeexEncoder::Initialize(const char* filename, char* modeInput, int channels, int pcmRate)
{
char *comments;
int comments_length;
fprintf(stderr, "SpeexEncoder: Initialize\n");
SpeexHeader header;
fprintf(stderr, "modeInput: %s\n", modeInput);
if (modeInput != NULL){
if (strcmp(modeInput, "narrowband") == 0)
{
modeID = SPEEX_MODEID_NB;
}
else if (strcmp(modeInput, "wideband") == 0)
{
modeID = SPEEX_MODEID_WB;
}
else if (strcmp(modeInput, "ultra-wideband") == 0)
{
modeID = SPEEX_MODEID_UWB;
}
}
if (channels > 0){
chan = channels;
}
speex_lib_ctl(SPEEX_LIB_GET_VERSION_STRING, (void*)&speex_version);
snprintf(vendor_string, sizeof(vendor_string), "Encoded with Speex %s", speex_version);
comment_init(&comments, &comments_length, vendor_string);
comment_add(&comments, &comments_length, "TITLE=", "saa1");
comment_add(&comments, &comments_length, "AUTHOR=", "uptivity");
comment_add(&comments, &comments_length, "FORMATTYPE=", (char*)speex_version);
comment_add(&comments, &comments_length, "DURATION=", "00000");
/*Initialize Ogg stream struct*/
srand(time(NULL));
if (ogg_stream_init(&os, rand()) == -1)
{
fprintf(stderr, "Error: stream init failed\n");
exit(1);
}
rate = pcmRate;
if (modeID == -1 )
{
/* By default, use narrowband/8 kHz */
modeID = SPEEX_MODEID_NB;
if (pcmRate < 0){
rate = 8000;
}
}
if (pcmRate < 0){
if (modeID == SPEEX_MODEID_NB)
rate = 8000;
else if (modeID == SPEEX_MODEID_WB)
rate = 16000;
else if (modeID == SPEEX_MODEID_UWB)
rate = 32000;
}
if (!mode)
mode = speex_lib_get_mode(modeID);
speex_init_header(&header, rate, 1, mode);
header.frames_per_packet = 1;
header.vbr = 0;
header.nb_channels = chan;
{
char *st_string = "mono";
if (chan == 2)
st_string = "stereo";
fprintf(stderr, "Encoding %d Hz audio using %s mode (%s)\n",
header.rate, mode->modeName, st_string);
}
/*Initialize Speex encoder*/
st = speex_encoder_init(mode);
fout = fopen(filename, "wb");
if (!fout)
{
perror(filename);
exit(1);
}
speex_encoder_ctl(st, SPEEX_GET_FRAME_SIZE, &frame_size);
speex_encoder_ctl(st, SPEEX_SET_COMPLEXITY, &complexity);
speex_encoder_ctl(st, SPEEX_SET_SAMPLING_RATE, &rate);
if (quality >= 0)
{
speex_encoder_ctl(st, SPEEX_SET_QUALITY, &quality);
}
speex_encoder_ctl(st, SPEEX_GET_LOOKAHEAD, &lookahead);
/*Write header*/
{
int packet_size;
op.packet = (unsigned char *)speex_header_to_packet(&header, &packet_size);
op.bytes = packet_size;
op.b_o_s = 1;
op.e_o_s = 0;
op.granulepos = 0;
op.packetno = 0;
// submit the packet to the ogg streaming layer
ogg_stream_packetin(&os, &op);
free(op.packet);
while ((result = ogg_stream_flush(&os, &og)))
{
if (!result) break;
ret = oe_write_page(&og, fout);
if (ret != og.header_len + og.body_len)
{
fprintf(stderr, "Error: failed writing header to output stream\n");
exit(1);
}
else
bytes_written += ret;
}
op.packet = (unsigned char *)comments;
op.bytes = comments_length;
op.b_o_s = 0;
op.e_o_s = 0;
op.granulepos = 0;
op.packetno = 1;
ogg_stream_packetin(&os, &op);
durationIndex = bytes_written + os.header_fill + os.body_fill - 5;
}
/* writing the rest of the speex header packets */
while ((result = ogg_stream_flush(&os, &og)))
{
if (!result) break;
ret = oe_write_page(&og, fout);
if (ret != og.header_len + og.body_len)
{
fprintf(stderr, "Error: failed writing header to output stream\n");
exit(1);
}
else
bytes_written += ret;
}
free(comments);
speex_bits_init(&bits);
return 0;
}
static block_t *DecodeRtpSpeexPacket( decoder_t *p_dec, block_t **pp_block )
{
block_t *p_speex_bit_block = *pp_block;
decoder_sys_t *p_sys = p_dec->p_sys;
block_t *p_aout_buffer;
int i_decode_ret;
unsigned int i_speex_frame_size;
if ( !p_speex_bit_block || p_speex_bit_block->i_pts <= VLC_TS_INVALID )
return NULL;
/*
If the SpeexBits buffer size is 0 (a default value),
we know that a proper initialization has not yet been done.
*/
if ( p_sys->bits.buf_size==0 )
{
p_sys->p_header = (SpeexHeader *)malloc(sizeof(SpeexHeader));
if ( !p_sys->p_header )
{
msg_Err( p_dec, "Could not allocate a Speex header.");
return NULL;
}
speex_init_header( p_sys->p_header,p_sys->rtp_rate,1,&speex_nb_mode );
speex_bits_init( &p_sys->bits );
p_sys->p_state = speex_decoder_init( &speex_nb_mode );
if ( !p_sys->p_state )
{
msg_Err( p_dec, "Could not allocate a Speex decoder." );
free( p_sys->p_header );
return NULL;
}
/*
Assume that variable bit rate is enabled. Also assume
that there is only one frame per packet.
*/
p_sys->p_header->vbr = 1;
p_sys->p_header->frames_per_packet = 1;
p_dec->fmt_out.audio.i_channels = p_sys->p_header->nb_channels;
p_dec->fmt_out.audio.i_physical_channels =
p_dec->fmt_out.audio.i_original_channels =
pi_channels_maps[p_sys->p_header->nb_channels];
p_dec->fmt_out.audio.i_rate = p_sys->p_header->rate;
if ( speex_mode_query( &speex_nb_mode,
SPEEX_MODE_FRAME_SIZE,
&i_speex_frame_size ) )
{
msg_Err( p_dec, "Could not determine the frame size." );
speex_decoder_destroy( p_sys->p_state );
free( p_sys->p_header );
return NULL;
}
p_dec->fmt_out.audio.i_bytes_per_frame = i_speex_frame_size;
date_Init(&p_sys->end_date, p_sys->p_header->rate, 1);
}
/*
If the SpeexBits are initialized but there is
still no header, an error must be thrown.
*/
if ( !p_sys->p_header )
{
msg_Err( p_dec, "There is no valid Speex header found." );
return NULL;
}
*pp_block = NULL;
if ( !date_Get( &p_sys->end_date ) )
date_Set( &p_sys->end_date, p_speex_bit_block->i_dts );
/*
Ask for a new audio output buffer and make sure
we get one.
*/
p_aout_buffer = decoder_NewAudioBuffer( p_dec,
p_sys->p_header->frame_size );
if ( !p_aout_buffer || p_aout_buffer->i_buffer == 0 )
{
msg_Err(p_dec, "Oops: No new buffer was returned!");
return NULL;
}
/*
Read the Speex payload into the SpeexBits buffer.
*/
speex_bits_read_from( &p_sys->bits,
(char*)p_speex_bit_block->p_buffer,
p_speex_bit_block->i_buffer );
/*
Decode the input and ensure that no errors
were encountered.
*/
i_decode_ret = speex_decode_int( p_sys->p_state, &p_sys->bits,
(int16_t*)p_aout_buffer->p_buffer );
if ( i_decode_ret < 0 )
{
msg_Err( p_dec, "Decoding failed. Perhaps we have a bad stream?" );
return NULL;
}
/*
Handle date management on the audio output buffer.
*/
p_aout_buffer->i_pts = date_Get( &p_sys->end_date );
p_aout_buffer->i_length = date_Increment( &p_sys->end_date,
p_sys->p_header->frame_size ) - p_aout_buffer->i_pts;
p_sys->i_frame_in_packet++;
block_Release( p_speex_bit_block );
return p_aout_buffer;
}
