Commit cf0f2642 authored by Oded Shimon's avatar Oded Shimon

Original Commit: r61 | ods15 | 2006-09-25 13:10:56 +0300 (Mon, 25 Sep 2006) | 2 lines

step 2, floor_fit, decides on the values to be encoded

Originally committed as revision 6466 to svn://svn.ffmpeg.org/ffmpeg/trunk
parent 62ca51b3
...@@ -643,6 +643,27 @@ static int put_main_header(venc_context_t * venc, uint8_t ** out) { ...@@ -643,6 +643,27 @@ static int put_main_header(venc_context_t * venc, uint8_t ** out) {
return p - *out; return p - *out;
} }
static void floor_fit(venc_context_t * venc, floor_t * fc, float * coeffs, int * posts, int samples) {
int range = 255 / fc->multiplier + 1;
int i;
for (i = 0; i < fc->values; i++) {
int position = fc->list[fc->list[i].sort].x;
int begin = fc->list[fc->list[FFMAX(i-1, 0)].sort].x;
int end = fc->list[fc->list[FFMIN(i+1, fc->values - 1)].sort].x;
int j;
float average = 0;
begin = (position + begin) / 2;
end = (position + end ) / 2;
assert(end <= samples);
for (j = begin; j < end; j++) average += fabs(coeffs[j]);
average /= end - begin;
average /= 64; // MAGIC!
for (j = 0; j < range; j++) if (floor1_inverse_db_table[j * fc->multiplier] > average) break;
posts[fc->list[i].sort] = j;
}
}
static void floor_encode(venc_context_t * venc, floor_t * fc, PutBitContext * pb, float * floor, int samples) { static void floor_encode(venc_context_t * venc, floor_t * fc, PutBitContext * pb, float * floor, int samples) {
int range = 255 / fc->multiplier + 1; int range = 255 / fc->multiplier + 1;
int j; int j;
...@@ -805,6 +826,7 @@ static int vorbis_encode_frame(AVCodecContext * avccontext, unsigned char * pack ...@@ -805,6 +826,7 @@ static int vorbis_encode_frame(AVCodecContext * avccontext, unsigned char * pack
int i; int i;
if (!window(venc, audio, samples)) return 0; if (!window(venc, audio, samples)) return 0;
samples = 1 << (venc->blocksize[0] - 1);
init_put_bits(&pb, packets, buf_size); init_put_bits(&pb, packets, buf_size);
...@@ -821,6 +843,8 @@ static int vorbis_encode_frame(AVCodecContext * avccontext, unsigned char * pack ...@@ -821,6 +843,8 @@ static int vorbis_encode_frame(AVCodecContext * avccontext, unsigned char * pack
for (i = 0; i < venc->channels; i++) { for (i = 0; i < venc->channels; i++) {
floor_t * fc = &venc->floors[mapping->floor[mapping->mux[i]]]; floor_t * fc = &venc->floors[mapping->floor[mapping->mux[i]]];
int posts[fc->values];
floor_fit(venc, fc, &venc->coeffs[i * samples], posts, samples);
floor_encode(venc, fc, &pb, &venc->floor[i * samples], samples); floor_encode(venc, fc, &pb, &venc->floor[i * samples], samples);
} }
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment