diff --git a/src/opus.h b/src/opus.h
index 37c1129e77997d5c6380dac8e1581de6b9b5ba37..e62bedd3151c6279da4983d206001882063008f0 100644
--- a/src/opus.h
+++ b/src/opus.h
@@ -65,13 +65,12 @@ extern "C" {
#define OPUS_BITRATE_AUTO -1
-#define OPUS_MODE_AUTO 2000
-#define OPUS_MODE_VOICE 2001
-#define OPUS_MODE_AUDIO 2002
+#define OPUS_MODE_VOICE 2000
+#define OPUS_MODE_AUDIO 2001
-#define MODE_SILK_ONLY 1000
-#define MODE_HYBRID 1001
-#define MODE_CELT_ONLY 1002
+#define MODE_SILK_ONLY 1000
+#define MODE_HYBRID 1001
+#define MODE_CELT_ONLY 1002
#define BANDWIDTH_AUTO 1100
#define BANDWIDTH_NARROWBAND 1101
@@ -140,25 +139,65 @@ extern "C" {
typedef struct OpusEncoder OpusEncoder;
typedef struct OpusDecoder OpusDecoder;
-OPUS_EXPORT OpusEncoder *opus_encoder_init(OpusEncoder* st, int Fs, int channels);
-
-OPUS_EXPORT OpusEncoder *opus_encoder_create(int Fs, int channels);
+/*
+ * There are two coding modes:
+ * OPUS_MODE_VOICE gives best quality at a given bitrate for voice signals. It enhances the
+ * input signal by high-pass filtering and emphasizing formants and harmonics. Optionally
+ * it includes in-band forward error correction to protect against packet loss. Use this
+ * mode for typical VoIP applications. Because of the enhancement, even at high bitrates
+ * the output may sound different from the input.
+ * OPUS_MODE_AUDIO gives best quality at a given bitrate for most non-voice signals like music.
+ * Use this mode for music and mixed (music/voice) content, and applications requiring less
+ * than 15 ms of coding delay.
+ */
+
+/* Returns initialized encoder state */
+OPUS_EXPORT OpusEncoder *opus_encoder_create(
+ int Fs, /* Sampling rate of input signal (Hz) */
+ int channels, /* Number of channels (1/2) in input signal */
+ int mode /* Coding mode (OPUS_MODE_VOICE/OPUS_MODE_AUDIO) */
+);
+
+OPUS_EXPORT OpusEncoder *opus_encoder_init(
+ OpusEncoder *st, /* Encoder state */
+ int Fs, /* Sampling rate of input signal (Hz) */
+ int channels, /* Number of channels (1/2) in input signal */
+ int mode /* Coding mode (OPUS_MODE_VOICE/OPUS_MODE_AUDIO) */
+);
/* returns length of data payload (in bytes) */
-OPUS_EXPORT int opus_encode(OpusEncoder *st, const short *pcm, int frame_size,
- unsigned char *data, int max_data_bytes);
+OPUS_EXPORT int opus_encode(
+ OpusEncoder *st, /* Encoder state */
+ const short *pcm, /* Input signal (interleaved if 2 channels). length is frame_size*channels */
+ int frame_size, /* Number of samples per frame of input signal */
+ unsigned char *data, /* Output payload (no more than max_data_bytes long) */
+ int max_data_bytes /* Allocated memory for payload; don't use for controlling bitrate */
+);
OPUS_EXPORT void opus_encoder_destroy(OpusEncoder *st);
OPUS_EXPORT int opus_encoder_ctl(OpusEncoder *st, int request, ...);
-OPUS_EXPORT OpusDecoder *opus_decoder_init(OpusDecoder *st, int Fs, int channels);
+OPUS_EXPORT OpusDecoder *opus_decoder_create(
+ int Fs, /* Sampling rate of output signal (Hz) */
+ int channels /* Number of channels (1/2) in output signal */
+);
-OPUS_EXPORT OpusDecoder *opus_decoder_create(int Fs, int channels);
+OPUS_EXPORT OpusDecoder *opus_decoder_init(OpusDecoder *st,
+ int Fs, /* Sampling rate of output signal (Hz) */
+ int channels /* Number of channels (1/2) in output signal */
+);
/* returns (CELT) error code */
-OPUS_EXPORT int opus_decode(OpusDecoder *st, const unsigned char *data, int len,
- short *pcm, int frame_size, int decode_fec);
+OPUS_EXPORT int opus_decode(
+ OpusDecoder *st, /* Decoder state */
+ const unsigned char *data, /* Input payload. Use a NULL pointer to indicate packet loss */
+ int len, /* Number of bytes in payload */
+ short *pcm, /* Output signal (interleaved if 2 channels). length is frame_size*channels */
+ int frame_size, /* Number of samples per frame of input signal */
+ int decode_fec /* Flag (0/1) to request that any in-band forward error correction data be */
+ /* decoded. If no such data is available the frame is decoded as if it were lost. */
+);
OPUS_EXPORT int opus_decoder_ctl(OpusDecoder *st, int request, ...);
@@ -170,7 +209,7 @@ OPUS_EXPORT int opus_packet_get_nb_channels(const unsigned char *data);
OPUS_EXPORT int opus_packet_get_nb_frames(const unsigned char packet[], int len);
OPUS_EXPORT int opus_decoder_get_nb_samples(const OpusDecoder *dec, const unsigned char packet[], int len);
-
+/* For testing purposes: the encoder and decoder state should always be identical after coding a payload */
#if OPUS_TEST_RANGE_CODER_STATE
OPUS_EXPORT int opus_encoder_get_final_range(OpusEncoder *st);
OPUS_EXPORT int opus_decoder_get_final_range(OpusDecoder *st);
diff --git a/src/opus_decoder.c b/src/opus_decoder.c
index c784233685f2cd9cacb40358756b8fd02254c911..add9226ebf9d69f3d8ee14c118d37f8c36666d32 100644
--- a/src/opus_decoder.c
+++ b/src/opus_decoder.c
@@ -103,7 +103,7 @@ failure:
OpusDecoder *opus_decoder_create(int Fs, int channels)
{
- char *raw_state = malloc(opus_decoder_get_size(channels));
+ char *raw_state = (char*)malloc(opus_decoder_get_size(channels));
if (raw_state == NULL)
return NULL;
return opus_decoder_init((OpusDecoder*)raw_state, Fs, channels);
diff --git a/src/opus_encoder.c b/src/opus_encoder.c
index f306606a74e14281028804339d91835d658db535..b4fcd93ebd64994add528b7a91150a7d041c789d 100644
--- a/src/opus_encoder.c
+++ b/src/opus_encoder.c
@@ -75,7 +75,7 @@ int opus_encoder_get_size(int channels)
}
-OpusEncoder *opus_encoder_init(OpusEncoder* st, int Fs, int channels)
+OpusEncoder *opus_encoder_init(OpusEncoder* st, int Fs, int channels, int mode)
{
void *silk_enc;
CELTEncoder *celt_enc;
@@ -115,6 +115,7 @@ OpusEncoder *opus_encoder_init(OpusEncoder* st, int Fs, int channels)
st->silk_mode.useInBandFEC = 0;
st->silk_mode.useDTX = 0;
st->silk_mode.useCBR = 0;
+ st->silk_mode.HP_cutoff_Hz = 0;
st->hybrid_stereo_width_Q14 = 1 << 14;
@@ -130,7 +131,7 @@ OpusEncoder *opus_encoder_init(OpusEncoder* st, int Fs, int channels)
st->use_vbr = 0;
st->user_bitrate_bps = OPUS_BITRATE_AUTO;
st->bitrate_bps = 3000+Fs*channels;
- st->user_mode = OPUS_MODE_AUTO;
+ st->user_mode = mode;
st->user_bandwidth = BANDWIDTH_AUTO;
st->voice_ratio = 90;
st->first = 1;
@@ -145,12 +146,12 @@ failure:
return NULL;
}
-OpusEncoder *opus_encoder_create(int Fs, int channels)
+OpusEncoder *opus_encoder_create(int Fs, int channels, int mode)
{
char *raw_state = (char *)malloc(opus_encoder_get_size(channels));
if (raw_state == NULL)
return NULL;
- return opus_encoder_init((OpusEncoder*)raw_state, Fs, channels);
+ return opus_encoder_init((OpusEncoder*)raw_state, Fs, channels, mode);
}
int opus_encode(OpusEncoder *st, const short *pcm, int frame_size,
@@ -160,7 +161,7 @@ int opus_encode(OpusEncoder *st, const short *pcm, int frame_size,
CELTEncoder *celt_enc;
int i;
int ret=0;
- SKP_int32 nBytes;
+ int nBytes;
ec_enc enc;
int framerate, period;
int silk_internal_bandwidth=-1;
@@ -187,8 +188,9 @@ int opus_encode(OpusEncoder *st, const short *pcm, int frame_size,
/* Rate-dependent mono-stereo decision */
if (st->force_mono)
+ {
st->stream_channels = 1;
- if (st->mode == MODE_CELT_ONLY && st->channels == 2)
+ } else if (st->mode == MODE_CELT_ONLY && st->channels == 2)
{
celt_int32 decision_rate;
decision_rate = st->bitrate_bps + st->voice_ratio*st->voice_ratio;
@@ -213,21 +215,7 @@ int opus_encode(OpusEncoder *st, const short *pcm, int frame_size,
mono_rate -= 60*(st->Fs/frame_size - 50);
/* Mode selection */
- if (st->user_mode==OPUS_MODE_AUTO)
- {
- celt_int32 decision_rate;
- /* SILK/CELT threshold is higher for voice than for music */
- decision_rate = mono_rate - 3*st->voice_ratio*st->voice_ratio;
- /* Hysteresis */
- if (st->prev_mode == MODE_CELT_ONLY)
- decision_rate += 4000;
- else if (st->prev_mode>0)
- decision_rate -= 4000;
- if (decision_rate>24000)
- st->mode = MODE_CELT_ONLY;
- else
- st->mode = MODE_SILK_ONLY;
- } else if (st->user_mode==OPUS_MODE_VOICE)
+ if (st->user_mode==OPUS_MODE_VOICE)
{
st->mode = MODE_SILK_ONLY;
} else {/* OPUS_AUDIO_MODE */
@@ -525,7 +513,6 @@ int opus_encode(OpusEncoder *st, const short *pcm, int frame_size,
if (st->mode != MODE_SILK_ONLY)
{
- /* Encode high band with CELT */
ret = celt_encode_with_ec(celt_enc, pcm_buf, frame_size, NULL, nb_compr_bytes, &enc);
}
diff --git a/src/opus_encoder.h b/src/opus_encoder.h
index ca1e28aedc6f546b27139cb4dec195bec4c9be80..87910b4376c4b0349bbd1a5e3aba91b6c4e84944 100644
--- a/src/opus_encoder.h
+++ b/src/opus_encoder.h
@@ -36,20 +36,20 @@
#define MAX_ENCODER_BUFFER 480
struct OpusEncoder {
- int celt_enc_offset;
- int silk_enc_offset;
- silk_EncControlStruct silk_mode;
+ int celt_enc_offset;
+ int silk_enc_offset;
+ silk_EncControlStruct silk_mode;
int hybrid_stereo_width_Q14;
- int channels;
- int stream_channels;
+ int channels;
+ int stream_channels;
+ int force_mono;
int mode;
int user_mode;
- int force_mono;
int prev_mode;
- int bandwidth;
- int user_bandwidth;
- int voice_ratio;
+ int bandwidth;
+ int user_bandwidth;
+ int voice_ratio;
/* Sampling rate (at the API level) */
int Fs;
int use_vbr;
diff --git a/src/test_opus.c b/src/test_opus.c
index 19165840ab3f6c607c8dccd30daf2a3ee5158550..f2d3db576d94baa336dd8d182b92ea6fa5bcb98f 100644
--- a/src/test_opus.c
+++ b/src/test_opus.c
@@ -42,11 +42,12 @@
void print_usage( char* argv[] )
{
- fprintf(stderr, "Usage: %s <mode (0/1/2)> <sampling rate (Hz)> <channels> "
+ fprintf(stderr, "Usage: %s <mode (0/1)> <sampling rate (Hz)> <channels (1/2)> "
"<bits per second> [options] <input> <output>\n\n", argv[0]);
- fprintf(stderr, "mode: 0 for auto, 1 for voice, 2 for audio:\n" );
+ fprintf(stderr, "mode: 0 for voice, 1 for audio:\n" );
fprintf(stderr, "options:\n" );
- fprintf(stderr, "-cbr : enable constant bitrate; default: VBR\n" );
+ fprintf(stderr, "-cbr : enable constant bitrate; default: variable bitrate\n" );
+ fprintf(stderr, "-cvbr : enable constraint variable bitrate; default: unconstraint\n" );
fprintf(stderr, "-bandwidth <NB|MB|WB|SWB|FB> : audio bandwidth (from narrowband to fullband); default: sampling rate\n" );
fprintf(stderr, "-framesize <2.5|5|10|20|40|60> : frame size in ms; default: 20 \n" );
fprintf(stderr, "-max_payload <bytes> : maximum payload size in bytes, default: 1024\n" );
@@ -106,7 +107,7 @@ int main(int argc, char *argv[])
return 1;
}
- mode = atoi(argv[1]) + OPUS_MODE_AUTO;
+ mode = atoi(argv[1]) + OPUS_MODE_VOICE;
sampling_rate = atoi(argv[2]);
channels = atoi(argv[3]);
bitrate_bps = atoi(argv[4]);
@@ -199,8 +200,8 @@ int main(int argc, char *argv[])
}
}
- if( mode < OPUS_MODE_AUTO || mode > OPUS_MODE_AUDIO) {
- fprintf (stderr, "mode must be: 0, 1 or 2\n");
+ if( mode < OPUS_MODE_VOICE || mode > OPUS_MODE_AUDIO) {
+ fprintf (stderr, "mode must be: 0 or 1\n");
return 1;
}
@@ -232,7 +233,7 @@ int main(int argc, char *argv[])
return 1;
}
- enc = opus_encoder_create(sampling_rate, channels);
+ enc = opus_encoder_create(sampling_rate, channels, mode);
dec = opus_decoder_create(sampling_rate, channels);
if (enc==NULL)
@@ -246,7 +247,6 @@ int main(int argc, char *argv[])
exit(1);
}
- opus_encoder_ctl(enc, OPUS_SET_MODE(mode));
opus_encoder_ctl(enc, OPUS_SET_BITRATE(bitrate_bps));
opus_encoder_ctl(enc, OPUS_SET_BANDWIDTH(bandwidth));
opus_encoder_ctl(enc, OPUS_SET_VBR_FLAG(use_vbr));