|
|
|
@ -75,20 +75,22 @@ typedef struct AlacEncodeContext { |
|
|
|
|
} AlacEncodeContext; |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
static void init_sample_buffers(AlacEncodeContext *s, const int16_t *input_samples) |
|
|
|
|
static void init_sample_buffers(AlacEncodeContext *s, |
|
|
|
|
const int16_t *input_samples) |
|
|
|
|
{ |
|
|
|
|
int ch, i; |
|
|
|
|
|
|
|
|
|
for(ch=0;ch<s->avctx->channels;ch++) { |
|
|
|
|
for (ch = 0; ch < s->avctx->channels; ch++) { |
|
|
|
|
const int16_t *sptr = input_samples + ch; |
|
|
|
|
for(i=0;i<s->avctx->frame_size;i++) { |
|
|
|
|
for (i = 0; i < s->avctx->frame_size; i++) { |
|
|
|
|
s->sample_buf[ch][i] = *sptr; |
|
|
|
|
sptr += s->avctx->channels; |
|
|
|
|
} |
|
|
|
|
} |
|
|
|
|
} |
|
|
|
|
|
|
|
|
|
static void encode_scalar(AlacEncodeContext *s, int x, int k, int write_sample_size) |
|
|
|
|
static void encode_scalar(AlacEncodeContext *s, int x, |
|
|
|
|
int k, int write_sample_size) |
|
|
|
|
{ |
|
|
|
|
int divisor, q, r; |
|
|
|
|
|
|
|
|
@ -97,17 +99,17 @@ static void encode_scalar(AlacEncodeContext *s, int x, int k, int write_sample_s |
|
|
|
|
q = x / divisor; |
|
|
|
|
r = x % divisor; |
|
|
|
|
|
|
|
|
|
if(q > 8) { |
|
|
|
|
if (q > 8) { |
|
|
|
|
// write escape code and sample value directly
|
|
|
|
|
put_bits(&s->pbctx, 9, ALAC_ESCAPE_CODE); |
|
|
|
|
put_bits(&s->pbctx, write_sample_size, x); |
|
|
|
|
} else { |
|
|
|
|
if(q) |
|
|
|
|
if (q) |
|
|
|
|
put_bits(&s->pbctx, q, (1<<q) - 1); |
|
|
|
|
put_bits(&s->pbctx, 1, 0); |
|
|
|
|
|
|
|
|
|
if(k != 1) { |
|
|
|
|
if(r > 0) |
|
|
|
|
if (k != 1) { |
|
|
|
|
if (r > 0) |
|
|
|
|
put_bits(&s->pbctx, k, r+1); |
|
|
|
|
else |
|
|
|
|
put_bits(&s->pbctx, k-1, 0); |
|
|
|
@ -164,7 +166,7 @@ static int estimate_stereo_mode(int32_t *left_ch, int32_t *right_ch, int n) |
|
|
|
|
|
|
|
|
|
/* calculate sum of 2nd order residual for each channel */ |
|
|
|
|
sum[0] = sum[1] = sum[2] = sum[3] = 0; |
|
|
|
|
for(i=2; i<n; i++) { |
|
|
|
|
for (i = 2; i < n; i++) { |
|
|
|
|
lt = left_ch[i] - 2*left_ch[i-1] + left_ch[i-2]; |
|
|
|
|
rt = right_ch[i] - 2*right_ch[i-1] + right_ch[i-2]; |
|
|
|
|
sum[2] += FFABS((lt + rt) >> 1); |
|
|
|
@ -181,8 +183,8 @@ static int estimate_stereo_mode(int32_t *left_ch, int32_t *right_ch, int n) |
|
|
|
|
|
|
|
|
|
/* return mode with lowest score */ |
|
|
|
|
best = 0; |
|
|
|
|
for(i=1; i<4; i++) { |
|
|
|
|
if(score[i] < score[best]) { |
|
|
|
|
for (i = 1; i < 4; i++) { |
|
|
|
|
if (score[i] < score[best]) { |
|
|
|
|
best = i; |
|
|
|
|
} |
|
|
|
|
} |
|
|
|
@ -205,7 +207,7 @@ static void alac_stereo_decorrelation(AlacEncodeContext *s) |
|
|
|
|
break; |
|
|
|
|
|
|
|
|
|
case ALAC_CHMODE_LEFT_SIDE: |
|
|
|
|
for(i=0; i<n; i++) { |
|
|
|
|
for (i = 0; i < n; i++) { |
|
|
|
|
right[i] = left[i] - right[i]; |
|
|
|
|
} |
|
|
|
|
s->interlacing_leftweight = 1; |
|
|
|
@ -213,7 +215,7 @@ static void alac_stereo_decorrelation(AlacEncodeContext *s) |
|
|
|
|
break; |
|
|
|
|
|
|
|
|
|
case ALAC_CHMODE_RIGHT_SIDE: |
|
|
|
|
for(i=0; i<n; i++) { |
|
|
|
|
for (i = 0; i < n; i++) { |
|
|
|
|
tmp = right[i]; |
|
|
|
|
right[i] = left[i] - right[i]; |
|
|
|
|
left[i] = tmp + (right[i] >> 31); |
|
|
|
@ -223,7 +225,7 @@ static void alac_stereo_decorrelation(AlacEncodeContext *s) |
|
|
|
|
break; |
|
|
|
|
|
|
|
|
|
default: |
|
|
|
|
for(i=0; i<n; i++) { |
|
|
|
|
for (i = 0; i < n; i++) { |
|
|
|
|
tmp = left[i]; |
|
|
|
|
left[i] = (tmp + right[i]) >> 1; |
|
|
|
|
right[i] = tmp - right[i]; |
|
|
|
@ -239,10 +241,10 @@ static void alac_linear_predictor(AlacEncodeContext *s, int ch) |
|
|
|
|
int i; |
|
|
|
|
AlacLPCContext lpc = s->lpc[ch]; |
|
|
|
|
|
|
|
|
|
if(lpc.lpc_order == 31) { |
|
|
|
|
if (lpc.lpc_order == 31) { |
|
|
|
|
s->predictor_buf[0] = s->sample_buf[ch][0]; |
|
|
|
|
|
|
|
|
|
for(i=1; i<s->avctx->frame_size; i++) |
|
|
|
|
for (i = 1; i < s->avctx->frame_size; i++) |
|
|
|
|
s->predictor_buf[i] = s->sample_buf[ch][i] - s->sample_buf[ch][i-1]; |
|
|
|
|
|
|
|
|
|
return; |
|
|
|
@ -250,17 +252,17 @@ static void alac_linear_predictor(AlacEncodeContext *s, int ch) |
|
|
|
|
|
|
|
|
|
// generalised linear predictor
|
|
|
|
|
|
|
|
|
|
if(lpc.lpc_order > 0) { |
|
|
|
|
if (lpc.lpc_order > 0) { |
|
|
|
|
int32_t *samples = s->sample_buf[ch]; |
|
|
|
|
int32_t *residual = s->predictor_buf; |
|
|
|
|
|
|
|
|
|
// generate warm-up samples
|
|
|
|
|
residual[0] = samples[0]; |
|
|
|
|
for(i=1;i<=lpc.lpc_order;i++) |
|
|
|
|
for (i = 1; i <= lpc.lpc_order; i++) |
|
|
|
|
residual[i] = samples[i] - samples[i-1]; |
|
|
|
|
|
|
|
|
|
// perform lpc on remaining samples
|
|
|
|
|
for(i = lpc.lpc_order + 1; i < s->avctx->frame_size; i++) { |
|
|
|
|
for (i = lpc.lpc_order + 1; i < s->avctx->frame_size; i++) { |
|
|
|
|
int sum = 1 << (lpc.lpc_quant - 1), res_val, j; |
|
|
|
|
|
|
|
|
|
for (j = 0; j < lpc.lpc_order; j++) { |
|
|
|
@ -303,7 +305,7 @@ static void alac_entropy_coder(AlacEncodeContext *s) |
|
|
|
|
int sign_modifier = 0, i, k; |
|
|
|
|
int32_t *samples = s->predictor_buf; |
|
|
|
|
|
|
|
|
|
for(i=0;i < s->avctx->frame_size;) { |
|
|
|
|
for (i = 0; i < s->avctx->frame_size;) { |
|
|
|
|
int x; |
|
|
|
|
|
|
|
|
|
k = av_log2((history >> 9) + 3); |
|
|
|
@ -320,15 +322,15 @@ static void alac_entropy_coder(AlacEncodeContext *s) |
|
|
|
|
- ((history * s->rc.history_mult) >> 9); |
|
|
|
|
|
|
|
|
|
sign_modifier = 0; |
|
|
|
|
if(x > 0xFFFF) |
|
|
|
|
if (x > 0xFFFF) |
|
|
|
|
history = 0xFFFF; |
|
|
|
|
|
|
|
|
|
if((history < 128) && (i < s->avctx->frame_size)) { |
|
|
|
|
if (history < 128 && i < s->avctx->frame_size) { |
|
|
|
|
unsigned int block_size = 0; |
|
|
|
|
|
|
|
|
|
k = 7 - av_log2(history) + ((history + 16) >> 6); |
|
|
|
|
|
|
|
|
|
while((*samples == 0) && (i < s->avctx->frame_size)) { |
|
|
|
|
while (*samples == 0 && i < s->avctx->frame_size) { |
|
|
|
|
samples++; |
|
|
|
|
i++; |
|
|
|
|
block_size++; |
|
|
|
@ -347,12 +349,12 @@ static void write_compressed_frame(AlacEncodeContext *s) |
|
|
|
|
{ |
|
|
|
|
int i, j; |
|
|
|
|
|
|
|
|
|
if(s->avctx->channels == 2) |
|
|
|
|
if (s->avctx->channels == 2) |
|
|
|
|
alac_stereo_decorrelation(s); |
|
|
|
|
put_bits(&s->pbctx, 8, s->interlacing_shift); |
|
|
|
|
put_bits(&s->pbctx, 8, s->interlacing_leftweight); |
|
|
|
|
|
|
|
|
|
for(i=0;i<s->avctx->channels;i++) { |
|
|
|
|
for (i = 0; i < s->avctx->channels; i++) { |
|
|
|
|
|
|
|
|
|
calc_predictor_params(s, i); |
|
|
|
|
|
|
|
|
@ -362,14 +364,14 @@ static void write_compressed_frame(AlacEncodeContext *s) |
|
|
|
|
put_bits(&s->pbctx, 3, s->rc.rice_modifier); |
|
|
|
|
put_bits(&s->pbctx, 5, s->lpc[i].lpc_order); |
|
|
|
|
// predictor coeff. table
|
|
|
|
|
for(j=0;j<s->lpc[i].lpc_order;j++) { |
|
|
|
|
for (j = 0; j < s->lpc[i].lpc_order; j++) { |
|
|
|
|
put_sbits(&s->pbctx, 16, s->lpc[i].lpc_coeff[j]); |
|
|
|
|
} |
|
|
|
|
} |
|
|
|
|
|
|
|
|
|
// apply lpc and entropy coding to audio samples
|
|
|
|
|
|
|
|
|
|
for(i=0;i<s->avctx->channels;i++) { |
|
|
|
|
for (i = 0; i < s->avctx->channels; i++) { |
|
|
|
|
alac_linear_predictor(s, i); |
|
|
|
|
alac_entropy_coder(s); |
|
|
|
|
} |
|
|
|
@ -384,13 +386,13 @@ static av_cold int alac_encode_init(AVCodecContext *avctx) |
|
|
|
|
avctx->frame_size = DEFAULT_FRAME_SIZE; |
|
|
|
|
avctx->bits_per_coded_sample = DEFAULT_SAMPLE_SIZE; |
|
|
|
|
|
|
|
|
|
if(avctx->sample_fmt != AV_SAMPLE_FMT_S16) { |
|
|
|
|
if (avctx->sample_fmt != AV_SAMPLE_FMT_S16) { |
|
|
|
|
av_log(avctx, AV_LOG_ERROR, "only pcm_s16 input samples are supported\n"); |
|
|
|
|
return -1; |
|
|
|
|
} |
|
|
|
|
|
|
|
|
|
// Set default compression level
|
|
|
|
|
if(avctx->compression_level == FF_COMPRESSION_DEFAULT) |
|
|
|
|
if (avctx->compression_level == FF_COMPRESSION_DEFAULT) |
|
|
|
|
s->compression_level = 2; |
|
|
|
|
else |
|
|
|
|
s->compression_level = av_clip(avctx->compression_level, 0, 2); |
|
|
|
@ -411,21 +413,23 @@ static av_cold int alac_encode_init(AVCodecContext *avctx) |
|
|
|
|
AV_WB8 (alac_extradata+17, avctx->bits_per_coded_sample); |
|
|
|
|
AV_WB8 (alac_extradata+21, avctx->channels); |
|
|
|
|
AV_WB32(alac_extradata+24, s->max_coded_frame_size); |
|
|
|
|
AV_WB32(alac_extradata+28, avctx->sample_rate*avctx->channels*avctx->bits_per_coded_sample); // average bitrate
|
|
|
|
|
AV_WB32(alac_extradata+28, |
|
|
|
|
avctx->sample_rate * avctx->channels * avctx->bits_per_coded_sample); // average bitrate
|
|
|
|
|
AV_WB32(alac_extradata+32, avctx->sample_rate); |
|
|
|
|
|
|
|
|
|
// Set relevant extradata fields
|
|
|
|
|
if(s->compression_level > 0) { |
|
|
|
|
if (s->compression_level > 0) { |
|
|
|
|
AV_WB8(alac_extradata+18, s->rc.history_mult); |
|
|
|
|
AV_WB8(alac_extradata+19, s->rc.initial_history); |
|
|
|
|
AV_WB8(alac_extradata+20, s->rc.k_modifier); |
|
|
|
|
} |
|
|
|
|
|
|
|
|
|
s->min_prediction_order = DEFAULT_MIN_PRED_ORDER; |
|
|
|
|
if(avctx->min_prediction_order >= 0) { |
|
|
|
|
if(avctx->min_prediction_order < MIN_LPC_ORDER || |
|
|
|
|
if (avctx->min_prediction_order >= 0) { |
|
|
|
|
if (avctx->min_prediction_order < MIN_LPC_ORDER || |
|
|
|
|
avctx->min_prediction_order > ALAC_MAX_LPC_ORDER) { |
|
|
|
|
av_log(avctx, AV_LOG_ERROR, "invalid min prediction order: %d\n", avctx->min_prediction_order); |
|
|
|
|
av_log(avctx, AV_LOG_ERROR, "invalid min prediction order: %d\n", |
|
|
|
|
avctx->min_prediction_order); |
|
|
|
|
return -1; |
|
|
|
|
} |
|
|
|
|
|
|
|
|
@ -433,18 +437,20 @@ static av_cold int alac_encode_init(AVCodecContext *avctx) |
|
|
|
|
} |
|
|
|
|
|
|
|
|
|
s->max_prediction_order = DEFAULT_MAX_PRED_ORDER; |
|
|
|
|
if(avctx->max_prediction_order >= 0) { |
|
|
|
|
if(avctx->max_prediction_order < MIN_LPC_ORDER || |
|
|
|
|
if (avctx->max_prediction_order >= 0) { |
|
|
|
|
if (avctx->max_prediction_order < MIN_LPC_ORDER || |
|
|
|
|
avctx->max_prediction_order > ALAC_MAX_LPC_ORDER) { |
|
|
|
|
av_log(avctx, AV_LOG_ERROR, "invalid max prediction order: %d\n", avctx->max_prediction_order); |
|
|
|
|
av_log(avctx, AV_LOG_ERROR, "invalid max prediction order: %d\n", |
|
|
|
|
avctx->max_prediction_order); |
|
|
|
|
return -1; |
|
|
|
|
} |
|
|
|
|
|
|
|
|
|
s->max_prediction_order = avctx->max_prediction_order; |
|
|
|
|
} |
|
|
|
|
|
|
|
|
|
if(s->max_prediction_order < s->min_prediction_order) { |
|
|
|
|
av_log(avctx, AV_LOG_ERROR, "invalid prediction orders: min=%d max=%d\n", |
|
|
|
|
if (s->max_prediction_order < s->min_prediction_order) { |
|
|
|
|
av_log(avctx, AV_LOG_ERROR, |
|
|
|
|
"invalid prediction orders: min=%d max=%d\n", |
|
|
|
|
s->min_prediction_order, s->max_prediction_order); |
|
|
|
|
return -1; |
|
|
|
|
} |
|
|
|
@ -469,12 +475,12 @@ static int alac_encode_frame(AVCodecContext *avctx, uint8_t *frame, |
|
|
|
|
PutBitContext *pb = &s->pbctx; |
|
|
|
|
int i, out_bytes, verbatim_flag = 0; |
|
|
|
|
|
|
|
|
|
if(avctx->frame_size > DEFAULT_FRAME_SIZE) { |
|
|
|
|
if (avctx->frame_size > DEFAULT_FRAME_SIZE) { |
|
|
|
|
av_log(avctx, AV_LOG_ERROR, "input frame size exceeded\n"); |
|
|
|
|
return -1; |
|
|
|
|
} |
|
|
|
|
|
|
|
|
|
if(buf_size < 2*s->max_coded_frame_size) { |
|
|
|
|
if (buf_size < 2 * s->max_coded_frame_size) { |
|
|
|
|
av_log(avctx, AV_LOG_ERROR, "buffer size is too small\n"); |
|
|
|
|
return -1; |
|
|
|
|
} |
|
|
|
@ -482,11 +488,11 @@ static int alac_encode_frame(AVCodecContext *avctx, uint8_t *frame, |
|
|
|
|
verbatim: |
|
|
|
|
init_put_bits(pb, frame, buf_size); |
|
|
|
|
|
|
|
|
|
if((s->compression_level == 0) || verbatim_flag) { |
|
|
|
|
if (s->compression_level == 0 || verbatim_flag) { |
|
|
|
|
// Verbatim mode
|
|
|
|
|
const int16_t *samples = data; |
|
|
|
|
write_frame_header(s, 1); |
|
|
|
|
for(i=0; i<avctx->frame_size*avctx->channels; i++) { |
|
|
|
|
for (i = 0; i < avctx->frame_size * avctx->channels; i++) { |
|
|
|
|
put_sbits(pb, 16, *samples++); |
|
|
|
|
} |
|
|
|
|
} else { |
|
|
|
@ -499,9 +505,9 @@ verbatim: |
|
|
|
|
flush_put_bits(pb); |
|
|
|
|
out_bytes = put_bits_count(pb) >> 3; |
|
|
|
|
|
|
|
|
|
if(out_bytes > s->max_coded_frame_size) { |
|
|
|
|
if (out_bytes > s->max_coded_frame_size) { |
|
|
|
|
/* frame too large. use verbatim mode */ |
|
|
|
|
if(verbatim_flag || (s->compression_level == 0)) { |
|
|
|
|
if (verbatim_flag || s->compression_level == 0) { |
|
|
|
|
/* still too large. must be an error. */ |
|
|
|
|
av_log(avctx, AV_LOG_ERROR, "error encoding frame\n"); |
|
|
|
|
return -1; |
|
|
|
@ -532,6 +538,7 @@ AVCodec ff_alac_encoder = { |
|
|
|
|
.encode = alac_encode_frame, |
|
|
|
|
.close = alac_encode_close, |
|
|
|
|
.capabilities = CODEC_CAP_SMALL_LAST_FRAME, |
|
|
|
|
.sample_fmts = (const enum AVSampleFormat[]){ AV_SAMPLE_FMT_S16, AV_SAMPLE_FMT_NONE}, |
|
|
|
|
.sample_fmts = (const enum AVSampleFormat[]){ AV_SAMPLE_FMT_S16, |
|
|
|
|
AV_SAMPLE_FMT_NONE }, |
|
|
|
|
.long_name = NULL_IF_CONFIG_SMALL("ALAC (Apple Lossless Audio Codec)"), |
|
|
|
|
}; |
|
|
|
|