/*
* Open codec.
*/
static pj_status_t amr_codec_open( pjmedia_codec *codec,
pjmedia_codec_param *attr )
{
struct amr_data *amr_data = (struct amr_data*) codec->codec_data;
pjmedia_codec_amr_pack_setting *setting;
unsigned i;
pj_uint8_t octet_align = 0;
pj_int8_t enc_mode;
const pj_str_t STR_FMTP_OCTET_ALIGN = {"octet-align", 11};
unsigned idx;
PJ_ASSERT_RETURN(codec && attr, PJ_EINVAL);
PJ_ASSERT_RETURN(amr_data != NULL, PJ_EINVALIDOP);
idx = (attr->info.clock_rate <= 8000? IDX_AMR_NB: IDX_AMR_WB);
enc_mode = pjmedia_codec_amr_get_mode(attr->info.avg_bps);
pj_assert(enc_mode >= 0 && enc_mode < amr_bitrates_size[idx]);
/* Check octet-align */
for (i = 0; i < attr->setting.dec_fmtp.cnt; ++i) {
if (pj_stricmp(&attr->setting.dec_fmtp.param[i].name,
&STR_FMTP_OCTET_ALIGN) == 0)
{
octet_align = (pj_uint8_t)
(pj_strtoul(&attr->setting.dec_fmtp.param[i].val));
break;
}
}
/* Check mode-set */
for (i = 0; i < attr->setting.enc_fmtp.cnt; ++i) {
const pj_str_t STR_FMTP_MODE_SET = {"mode-set", 8};
if (pj_stricmp(&attr->setting.enc_fmtp.param[i].name,
&STR_FMTP_MODE_SET) == 0)
{
const char *p;
pj_size_t l;
pj_int8_t diff = 99;
/* Encoding mode is chosen based on local default mode setting:
* - if local default mode is included in the mode-set, use it
* - otherwise, find the closest mode to local default mode;
* if there are two closest modes, prefer to use the higher
* one, e.g: local default mode is 4, the mode-set param
* contains '2,3,5,6', then 5 will be chosen.
*/
p = pj_strbuf(&attr->setting.enc_fmtp.param[i].val);
l = pj_strlen(&attr->setting.enc_fmtp.param[i].val);
while (l--) {
if (*p>='0' && *p<=('0'+amr_bitrates_size[idx]-1)) {
pj_int8_t tmp = *p - '0' - enc_mode;
if (PJ_ABS(diff) > PJ_ABS(tmp) ||
(PJ_ABS(diff) == PJ_ABS(tmp) && tmp > diff))
{
diff = tmp;
if (diff == 0) break;
}
}
++p;
}
PJ_ASSERT_RETURN(diff != 99, PJMEDIA_CODEC_EFAILED);
enc_mode = enc_mode + diff;
break;
}
}
amr_data->clock_rate = attr->info.clock_rate;
amr_data->vad_enabled = (attr->setting.vad != 0);
amr_data->plc_enabled = (attr->setting.plc != 0);
amr_data->enc_mode = enc_mode;
if (idx == IDX_AMR_NB) {
#ifdef USE_AMRNB
amr_data->encoder = Encoder_Interface_init(amr_data->vad_enabled);
#endif
} else {
#ifdef USE_AMRWB
amr_data->encoder = E_IF_init();
#endif
}
if (amr_data->encoder == NULL) {
TRACE_((THIS_FILE, "Encoder initialization failed"));
amr_codec_close(codec);
return PJMEDIA_CODEC_EFAILED;
}
setting = &amr_data->enc_setting;
pj_bzero(setting, sizeof(pjmedia_codec_amr_pack_setting));
setting->amr_nb = (idx == IDX_AMR_NB? 1: 0);
setting->reorder = 0;
setting->octet_aligned = octet_align;
setting->cmr = 15;
if (idx == IDX_AMR_NB) {
#ifdef USE_AMRNB
amr_data->decoder = Decoder_Interface_init();
#endif
} else {
#ifdef USE_AMRWB
amr_data->decoder = D_IF_init();
#endif
}
if (amr_data->decoder == NULL) {
TRACE_((THIS_FILE, "Decoder initialization failed"));
amr_codec_close(codec);
return PJMEDIA_CODEC_EFAILED;
}
setting = &amr_data->dec_setting;
pj_bzero(setting, sizeof(pjmedia_codec_amr_pack_setting));
setting->amr_nb = (idx == IDX_AMR_NB? 1: 0);
setting->reorder = 0;
setting->octet_aligned = octet_align;
TRACE_((THIS_FILE, "AMR codec allocated: clockrate=%d vad=%d, plc=%d,"
" bitrate=%d", amr_data->clock_rate,
amr_data->vad_enabled, amr_data->plc_enabled,
amr_bitrates[idx][amr_data->enc_mode]));
return PJ_SUCCESS;
}
int encoder_write(unsigned long id, void* rbuffer, unsigned long bsize)
{
short lspeech[160];
unsigned int i = 0, j, z, byte_counter, k;
if(pestreams[id].firstwrite)
{
if(pestreams[id].firstwrite == -1) return 0;
if(pestreams[id].cfrequency != 8000 || pestreams[id].cchannels != 1)
{
MessageBox(0, uni("Sorry, input file needs to be in following format:\n8000Hz, Mono.\n\nPlease use \"Sound File Encoder\" to convert between formats (example: convert your file(s) into Wave format (with above settings), then convert those (.wav) file(s) into AMR)."), uni("AMR Conversion"), MB_ICONEXCLAMATION);
pestreams[id].firstwrite = -1;
return 0;
}
pestreams[id].fhandle = sys_file_createforcestream(pestreams[id].filepath, v_sys_file_forwrite);
if(pestreams[id].fhandle == v_error_sys_file_create)return 0;
memset(pestreams[id].serial_data, 0, sizeof(pestreams[id].serial_data));
pestreams[id].req_mode = MR122;
pestreams[id].dtx = 0;
pestreams[id].c = 0;
pestreams[id].enstate = Encoder_Interface_init(pestreams[id].dtx);
sys_file_write(pestreams[id].fhandle, AMR_MAGIC_NUMBER, strlen(AMR_MAGIC_NUMBER));
pestreams[id].firstwrite = 0;
}
j = pestreams[id].c;
for(;;)
{
z = min(bsize - i, sizeof(pestreams[id].speech));
memcpy(((char*)pestreams[id].speech) + j, ((char*)rbuffer) + i, z - j);
pestreams[id].c = z - j;
if(pestreams[id].c == sizeof(pestreams[id].speech))
{
for(k=0; k<160; k++)
lspeech[k] = (short)(pestreams[id].speech[k] * 32767.0);
byte_counter = Encoder_Interface_Encode(pestreams[id].enstate, pestreams[id].req_mode, lspeech, pestreams[id].serial_data, 0);
sys_file_write(pestreams[id].fhandle, pestreams[id].serial_data, byte_counter);
}
i += z - j;
j = 0;
if(i >= bsize)break;
}
return 1;
}
static void enc_preprocess(MSFilter *f) {
EncState *s = (EncState*) f->data;
s->enc = Encoder_Interface_init(s->dtx);
}
static void enc_preprocess(AmrEncState *s) {
s->amr_enc = Encoder_Interface_init(s->dtx);
}
/*
* main
*
*
* Function:
* Speech encoder main program
*
* Usage: encoder speech_file bitstream_file mode dtx mode_file
*
* Format for speech_file:
* Speech is read from a binary file of 16 bits data.
*
* Format for ETSI bitstream file:
* 1 word (2-byte) for the TX frame type
* 244 words (2-byte) containing 244 bits.
* Bit 0 = 0x0000 and Bit 1 = 0x0001
* 1 word (2-byte) for the mode indication
* 4 words for future use, currently written as zero
*
* Format for 3GPP bitstream file:
* Holds mode information and bits packed to octets.
* Size is from 1 byte to 31 bytes.
*
* ETSI bitstream file format is defined using ETSI as preprocessor
* definition
*
* mode : MR475, MR515, MR59, MR67, MR74, MR795, MR102, MR122
* mode_file : reads mode information from a file
* Returns:
* 0
*/
int main (int argc, char * argv[]){
/* file strucrures */
FILE * file_speech = NULL;
FILE * file_encoded = NULL;
FILE * file_mode = NULL;
/* input speech vector */
short speech[160];
/* counters */
int byte_counter, frames = 0, bytes = 0;
/* pointer to encoder state structure */
int *enstate;
/* requested mode */
enum Mode req_mode = MR122;
int dtx = 0;
/* temporary variables */
char mode_string[9];
long mode_tmp;
/* bitstream filetype */
#ifndef ETSI
unsigned char serial_data[32];
#else
short serial_data[250] = {0};
#endif
/* Process command line options */
if ((argc == 5) || (argc == 4)){
file_encoded = fopen(argv[argc - 1], "wb");
if (file_encoded == NULL){
Usage(argv);
return 1;
}
file_speech = fopen(argv[argc - 2], "rb");
if (file_speech == NULL){
fclose(file_encoded);
Usage(argv);
return 1;
}
if (strncmp(argv[argc - 3], "-modefile=", 10) == 0){
file_mode = fopen(&argv[argc - 3][10], "rt");
if (file_mode == NULL){
Usage(argv);
fclose(file_speech);
fclose(file_encoded);
return 1;
}
}
else {
mode_tmp = strtol(&argv[argc - 3][2], NULL, 0);
for (req_mode = 0; req_mode < 8; req_mode++){
if (mode_tmp == modeConv[req_mode])
break;
}
if (req_mode == 8){
Usage(argv);
fclose(file_speech);
fclose(file_encoded);
if (file_mode != NULL)
fclose(file_mode);
return 1;
}
}
if (argc == 5){
if ((strcmp(argv[1], "-dtx") != 0)){
Usage(argv);
fclose(file_speech);
fclose(file_encoded);
if (file_mode != NULL){
fclose(file_mode);
}
return 1;
}
else {
dtx = 1;
}
}
}
else {
Usage(argv);
return 1;
}
enstate = Encoder_Interface_init(dtx);
Copyright();
#ifndef VAD2
fprintf( stderr, "%s\n", "Code compiled with VAD option: VAD1");
#else
fprintf( stderr, "%s\n", "Code compiled with VAD option: VAD2");
#endif
#ifndef ETSI
#ifndef IF2
/* write magic number to indicate single channel AMR file storage format */
bytes = fwrite(AMR_MAGIC_NUMBER, sizeof(char), strlen(AMR_MAGIC_NUMBER), file_encoded);
#endif
#endif
/* read file */
while (fread( speech, sizeof (Word16), 160, file_speech ) > 0)
{
/* read mode */
if (file_mode != NULL){
req_mode = 8;
if (fscanf(file_mode, "%9s\n", mode_string) != EOF) {
mode_tmp = strtol(&mode_string[2], NULL, 0);
for (req_mode = 0; req_mode < 8; req_mode++){
if (mode_tmp == modeConv[req_mode]){
break;
}
}
}
if (req_mode == 8){
break;
}
}
frames ++;
/* call encoder */
byte_counter = Encoder_Interface_Encode(enstate, req_mode, speech, serial_data, 0);
bytes += byte_counter;
fwrite(serial_data, sizeof (UWord8), byte_counter, file_encoded );
fflush(file_encoded);
}
Encoder_Interface_exit(enstate);
#ifndef ETSI
#ifdef IF2
fprintf ( stderr, "\n%s%i%s%i%s\n", "Frame structure AMR IF2: ", frames, " frames, ", bytes, " bytes.");
#else
fprintf ( stderr, "\n%s%i%s%i%s\n", "Frame structure AMR MIME file storage format: ", frames, " frames, ", bytes, " bytes.");
#endif
#else
fprintf ( stderr, "\n%s%i%s\n", "Frame structure AMR ETSI: ", frames, " frames. ");
#endif
fclose(file_speech);
fclose(file_encoded);
if (file_mode != NULL)
fclose(file_mode);
return 0;
}
int encode_amr(const char* infile, const char* outfile) {
enum Mode mode = MR122;
FILE *out;
void *wav, *amr;
int format, sampleRate, channels, bitsPerSample;
int inputSize;
uint8_t* inputBuf;
wav = wav_read_open(infile);
if (!wav) {
fprintf(stderr, "Unable to open wav file %s\n", infile);
return 1;
}
if (!wav_get_header(wav, &format, &channels, &sampleRate, &bitsPerSample, NULL)) {
fprintf(stderr, "Bad wav file %s\n", infile);
return 1;
}
if (format != 1) {
fprintf(stderr, "Unsupported WAV format %d\n", format);
return 1;
}
if (bitsPerSample != 16) {
fprintf(stderr, "Unsupported WAV sample depth %d\n", bitsPerSample);
return 1;
}
if (channels != 1)
fprintf(stderr, "Warning, only compressing one audio channel\n");
if (sampleRate != 8000)
fprintf(stderr, "Warning, AMR-NB uses 8000 Hz sample rate (WAV file has %d Hz)\n", sampleRate);
inputSize = channels*2*160;
inputBuf = (uint8_t*) malloc(inputSize);
amr = Encoder_Interface_init(0);
out = fopen(outfile, "wb");
if (!out) {
perror(outfile);
return 1;
}
fwrite("#!AMR\n", 1, 6, out);
while (1) {
short buf[160];
uint8_t outbuf[500];
int read, i, n;
read = wav_read_data(wav, inputBuf, inputSize);
read /= channels;
read /= 2;
if (read < 160)
break;
for (i = 0; i < 160; i++) {
const uint8_t* in = &inputBuf[2*channels*i];
buf[i] = in[0] | (in[1] << 8);
}
n = Encoder_Interface_Encode(amr, mode, buf, outbuf, 0);
fwrite(outbuf, 1, n, out);
}
free(inputBuf);
fclose(out);
Encoder_Interface_exit(amr);
wav_read_close(wav);
return 0;
}
static switch_status_t switch_amr_init(switch_codec_t *codec, switch_codec_flag_t flags, const switch_codec_settings_t *codec_settings)
{
#ifdef AMR_PASSTHROUGH
codec->flags |= SWITCH_CODEC_FLAG_PASSTHROUGH;
if (codec->fmtp_in) {
codec->fmtp_out = switch_core_strdup(codec->memory_pool, codec->fmtp_in);
}
return SWITCH_STATUS_SUCCESS;
#else
struct amr_context *context = NULL;
switch_codec_fmtp_t codec_fmtp;
amr_codec_settings_t amr_codec_settings;
int encoding, decoding;
int x, i, argc;
char *argv[10];
char fmtptmp[128];
encoding = (flags & SWITCH_CODEC_FLAG_ENCODE);
decoding = (flags & SWITCH_CODEC_FLAG_DECODE);
if (!(encoding || decoding) || (!(context = switch_core_alloc(codec->memory_pool, sizeof(struct amr_context))))) {
return SWITCH_STATUS_FALSE;
} else {
memset(&codec_fmtp, '\0', sizeof(struct switch_codec_fmtp));
codec_fmtp.private_info = &amr_codec_settings;
switch_amr_fmtp_parse(codec->fmtp_in, &codec_fmtp);
if (context->enc_modes) {
for (i = 7; i > -1; i++) {
if (context->enc_modes & (1 << i)) {
context->enc_mode = (switch_byte_t) i;
break;
}
}
}
if (!context->enc_mode) {
context->enc_mode = globals.default_bitrate;
}
switch_snprintf(fmtptmp, sizeof(fmtptmp), "octet-align=%d; mode-set=%d", switch_test_flag(context, AMR_OPT_OCTET_ALIGN) ? 1 : 0,
context->enc_mode);
codec->fmtp_out = switch_core_strdup(codec->memory_pool, fmtptmp);
context->enc_mode = AMR_DEFAULT_BITRATE;
context->encoder_state = NULL;
context->decoder_state = NULL;
if (encoding) {
context->encoder_state = Encoder_Interface_init(context->dtx_mode);
}
if (decoding) {
context->decoder_state = Decoder_Interface_init();
}
codec->private_info = context;
return SWITCH_STATUS_SUCCESS;
}
#endif
}
int main(int argc, char *argv[]) {
enum Mode mode = MR122;
int ch, dtx = 0;
const char *infile, *outfile;
FILE *out;
void *wav, *amr;
int format, sampleRate, channels, bitsPerSample;
int inputSize;
uint8_t* inputBuf;
while ((ch = getopt(argc, argv, "r:d")) != -1) {
switch (ch) {
case 'r':
mode = findMode(optarg);
break;
case 'd':
dtx = 1;
break;
case '?':
default:
usage(argv[0]);
return 1;
}
}
if (argc - optind < 2) {
usage(argv[0]);
return 1;
}
infile = argv[optind];
outfile = argv[optind + 1];
wav = wav_read_open(infile);
if (!wav) {
fprintf(stderr, "Unable to open wav file %s\n", infile);
return 1;
}
if (!wav_get_header(wav, &format, &channels, &sampleRate, &bitsPerSample, NULL)) {
fprintf(stderr, "Bad wav file %s\n", infile);
return 1;
}
if (format != 1) {
fprintf(stderr, "Unsupported WAV format %d\n", format);
return 1;
}
if (bitsPerSample != 16) {
fprintf(stderr, "Unsupported WAV sample depth %d\n", bitsPerSample);
return 1;
}
if (channels != 1)
fprintf(stderr, "Warning, only compressing one audio channel\n");
if (sampleRate != 8000)
fprintf(stderr, "Warning, AMR-NB uses 8000 Hz sample rate (WAV file has %d Hz)\n", sampleRate);
inputSize = channels*2*160;
inputBuf = (uint8_t*) malloc(inputSize);
amr = Encoder_Interface_init(dtx);
out = fopen(outfile, "wb");
if (!out) {
perror(outfile);
return 1;
}
fwrite("#!AMR\n", 1, 6, out);
while (1) {
short buf[160];
uint8_t outbuf[500];
int read, i, n;
read = wav_read_data(wav, inputBuf, inputSize);
read /= channels;
read /= 2;
if (read < 160)
break;
for (i = 0; i < 160; i++) {
const uint8_t* in = &inputBuf[2*channels*i];
buf[i] = in[0] | (in[1] << 8);
}
n = Encoder_Interface_Encode(amr, mode, buf, outbuf, 0);
fwrite(outbuf, 1, n, out);
}
free(inputBuf);
fclose(out);
Encoder_Interface_exit(amr);
wav_read_close(wav);
return 0;
}