mirror of
https://github.com/FFmpeg/FFmpeg.git
synced 2024-11-21 10:55:51 +02:00
downmix before the IMDCT if no block switching is used
Originally committed as revision 11228 to svn://svn.ffmpeg.org/ffmpeg/trunk
This commit is contained in:
parent
15c57ced2f
commit
b40211ff67
@ -750,27 +750,28 @@ static inline void do_imdct(AC3DecodeContext *ctx)
|
||||
/**
|
||||
* Downmix the output to mono or stereo.
|
||||
*/
|
||||
static void ac3_downmix(float samples[AC3_MAX_CHANNELS][256], int fbw_channels,
|
||||
int output_mode, float coef[AC3_MAX_CHANNELS][2])
|
||||
static void ac3_downmix(float samples[][256], int fbw_channels,
|
||||
int output_mode, float coef[AC3_MAX_CHANNELS][2],
|
||||
int ch_offset)
|
||||
{
|
||||
int i, j;
|
||||
float v0, v1, s0, s1;
|
||||
|
||||
for(i=0; i<256; i++) {
|
||||
v0 = v1 = s0 = s1 = 0.0f;
|
||||
for(j=0; j<fbw_channels; j++) {
|
||||
v0 += samples[j][i] * coef[j][0];
|
||||
v1 += samples[j][i] * coef[j][1];
|
||||
s0 += coef[j][0];
|
||||
s1 += coef[j][1];
|
||||
for(j=ch_offset; j<fbw_channels+ch_offset; j++) {
|
||||
v0 += samples[j][i] * coef[j-ch_offset][0];
|
||||
v1 += samples[j][i] * coef[j-ch_offset][1];
|
||||
s0 += coef[j-ch_offset][0];
|
||||
s1 += coef[j-ch_offset][1];
|
||||
}
|
||||
v0 /= s0;
|
||||
v1 /= s1;
|
||||
if(output_mode == AC3_CHMODE_MONO) {
|
||||
samples[0][i] = (v0 + v1) * LEVEL_MINUS_3DB;
|
||||
samples[ch_offset][i] = (v0 + v1) * LEVEL_MINUS_3DB;
|
||||
} else if(output_mode == AC3_CHMODE_STEREO) {
|
||||
samples[0][i] = v0;
|
||||
samples[1][i] = v1;
|
||||
samples[ch_offset][i] = v0;
|
||||
samples[ch_offset+1][i] = v1;
|
||||
}
|
||||
}
|
||||
}
|
||||
@ -785,12 +786,16 @@ static int ac3_parse_audio_block(AC3DecodeContext *ctx, int blk)
|
||||
int i, bnd, seg, ch;
|
||||
GetBitContext *gb = &ctx->gb;
|
||||
uint8_t bit_alloc_stages[AC3_MAX_CHANNELS];
|
||||
int any_block_switching = 0;
|
||||
int num_channels_bak, fbw_channels_bak;
|
||||
|
||||
memset(bit_alloc_stages, 0, AC3_MAX_CHANNELS);
|
||||
|
||||
/* block switch flags */
|
||||
for (ch = 1; ch <= fbw_channels; ch++)
|
||||
for (ch = 1; ch <= fbw_channels; ch++) {
|
||||
ctx->block_switch[ch] = get_bits1(gb);
|
||||
any_block_switching |= ctx->block_switch[ch];
|
||||
}
|
||||
|
||||
/* dithering flags */
|
||||
ctx->dither_all = 1;
|
||||
@ -1062,13 +1067,26 @@ static int ac3_parse_audio_block(AC3DecodeContext *ctx, int blk)
|
||||
}
|
||||
}
|
||||
|
||||
/* if no block switching is used, downmixing can be done before IMDCT */
|
||||
num_channels_bak = ctx->channels;
|
||||
fbw_channels_bak = ctx->fbw_channels;
|
||||
if(!any_block_switching) {
|
||||
if(ctx->channels != ctx->out_channels && !((ctx->output_mode & AC3_OUTPUT_LFEON) &&
|
||||
ctx->fbw_channels == ctx->out_channels)) {
|
||||
ac3_downmix(ctx->transform_coeffs, ctx->fbw_channels,
|
||||
ctx->output_mode, ctx->downmix_coeffs, 1);
|
||||
ctx->channels = ctx->out_channels;
|
||||
ctx->fbw_channels = ctx->channels - (ctx->output_mode & AC3_OUTPUT_LFEON);
|
||||
}
|
||||
}
|
||||
|
||||
do_imdct(ctx);
|
||||
|
||||
/* downmix output if needed */
|
||||
/* downmix output now if it wasn't done before IMDCT */
|
||||
if(ctx->channels != ctx->out_channels && !((ctx->output_mode & AC3_OUTPUT_LFEON) &&
|
||||
ctx->fbw_channels == ctx->out_channels)) {
|
||||
ac3_downmix(ctx->output, ctx->fbw_channels, ctx->output_mode,
|
||||
ctx->downmix_coeffs);
|
||||
ctx->downmix_coeffs, 0);
|
||||
}
|
||||
|
||||
/* convert float to 16-bit integer */
|
||||
@ -1079,6 +1097,9 @@ static int ac3_parse_audio_block(AC3DecodeContext *ctx, int blk)
|
||||
ctx->dsp.float_to_int16(ctx->int_output[ch], ctx->output[ch], 256);
|
||||
}
|
||||
|
||||
ctx->channels = num_channels_bak;
|
||||
ctx->fbw_channels = fbw_channels_bak;
|
||||
|
||||
return 0;
|
||||
}
|
||||
|
||||
|
Loading…
Reference in New Issue
Block a user