virtualenv/FFmpeg
1
0
Fork 0
mirror of https://github.com/FFmpeg/FFmpeg.git synced 2024-12-23 12:43:46 +02:00
Code Issues Releases Activity

mpegaudiodec: interleave iMDCT buffer to simplify future SIMD implementations

Browse Source 
Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>
This commit is contained in:
Vitor Sessak 2012-01-03 21:40:57 +01:00 committed by Ronald S. Bultje
parent 2170a0e6ad
commit 421c99a4a7
1 changed files with 25 additions and 25 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

50
libavcodec/mpegaudiodec.c
View File

@ -621,17 +621,17 @@ static void imdct36(INTFLOAT *out, INTFLOAT *buf, INTFLOAT *in, INTFLOAT *win)
t0 = s0 + s1; t0 = s0 + s1;
t1 = s0 - s1; t1 = s0 - s1;
out[(9 + j) * SBLIMIT] = MULH3(t1, win[ 9 + j], 1) + buf[9 + j]; out[(9 + j) * SBLIMIT] = MULH3(t1, win[ 9 + j], 1) + buf[4*(9 + j)];
out[(8 - j) * SBLIMIT] = MULH3(t1, win[ 8 - j], 1) + buf[8 - j]; out[(8 - j) * SBLIMIT] = MULH3(t1, win[ 8 - j], 1) + buf[4*(8 - j)];
buf[ 9 + j ] = MULH3(t0, win[18 + 9 + j], 1); buf[4 * ( 9 + j )] = MULH3(t0, win[18 + 9 + j], 1);
buf[ 8 - j ] = MULH3(t0, win[18 + 8 - j], 1); buf[4 * ( 8 - j )] = MULH3(t0, win[18 + 8 - j], 1);
t0 = s2 + s3; t0 = s2 + s3;
t1 = s2 - s3; t1 = s2 - s3;
out[(9 + 8 - j) * SBLIMIT] = MULH3(t1, win[ 9 + 8 - j], 1) + buf[9 + 8 - j]; out[(9 + 8 - j) * SBLIMIT] = MULH3(t1, win[ 9 + 8 - j], 1) + buf[4*(9 + 8 - j)];
out[ j * SBLIMIT] = MULH3(t1, win[ j], 1) + buf[ j]; out[ j * SBLIMIT] = MULH3(t1, win[ j], 1) + buf[4*( j)];
buf[ 9 + 8 - j ] = MULH3(t0, win[18 + 9 + 8 - j], 1); buf[4 * ( 9 + 8 - j )] = MULH3(t0, win[18 + 9 + 8 - j], 1);
buf[ j ] = MULH3(t0, win[18 + j], 1); buf[4 * ( j )] = MULH3(t0, win[18 + j], 1);
i += 4; i += 4;
} }
@ -639,10 +639,10 @@ static void imdct36(INTFLOAT *out, INTFLOAT *buf, INTFLOAT *in, INTFLOAT *win)
s1 = MULH3(tmp[17], icos36h[4], 2); s1 = MULH3(tmp[17], icos36h[4], 2);
t0 = s0 + s1; t0 = s0 + s1;
t1 = s0 - s1; t1 = s0 - s1;
out[(9 + 4) * SBLIMIT] = MULH3(t1, win[ 9 + 4], 1) + buf[9 + 4]; out[(9 + 4) * SBLIMIT] = MULH3(t1, win[ 9 + 4], 1) + buf[4*(9 + 4)];
out[(8 - 4) * SBLIMIT] = MULH3(t1, win[ 8 - 4], 1) + buf[8 - 4]; out[(8 - 4) * SBLIMIT] = MULH3(t1, win[ 8 - 4], 1) + buf[4*(8 - 4)];
buf[ 9 + 4 ] = MULH3(t0, win[18 + 9 + 4], 1); buf[4 * ( 9 + 4 )] = MULH3(t0, win[18 + 9 + 4], 1);
buf[ 8 - 4 ] = MULH3(t0, win[18 + 8 - 4], 1); buf[4 * ( 8 - 4 )] = MULH3(t0, win[18 + 8 - 4], 1);
} }
/* return the number of decoded frames */ /* return the number of decoded frames */
@ -1407,7 +1407,7 @@ static void compute_imdct(MPADecodeContext *s, GranuleDef *g,
imdct36(out_ptr, buf, ptr, win); imdct36(out_ptr, buf, ptr, win);
out_ptr += 18 * SBLIMIT; out_ptr += 18 * SBLIMIT;
ptr += 18; ptr += 18;
buf += 18; buf += (j&3) != 3 ? 1 : (4*18-3);
} }
for (j = mdct_long_end; j < sblimit; j++) { for (j = mdct_long_end; j < sblimit; j++) {
/* select frequency inversion */ /* select frequency inversion */
@ -1415,40 +1415,40 @@ static void compute_imdct(MPADecodeContext *s, GranuleDef *g,
out_ptr = sb_samples + j; out_ptr = sb_samples + j;
for (i = 0; i < 6; i++) { for (i = 0; i < 6; i++) {
*out_ptr = buf[i]; *out_ptr = buf[4*i];
out_ptr += SBLIMIT; out_ptr += SBLIMIT;
} }
imdct12(out2, ptr + 0); imdct12(out2, ptr + 0);
for (i = 0; i < 6; i++) { for (i = 0; i < 6; i++) {
*out_ptr = MULH3(out2[i ], win[i ], 1) + buf[i + 6*1]; *out_ptr = MULH3(out2[i ], win[i ], 1) + buf[4*(i + 6*1)];
buf[i + 6*2] = MULH3(out2[i + 6], win[i + 6], 1); buf[4*(i + 6*2)] = MULH3(out2[i + 6], win[i + 6], 1);
out_ptr += SBLIMIT; out_ptr += SBLIMIT;
} }
imdct12(out2, ptr + 1); imdct12(out2, ptr + 1);
for (i = 0; i < 6; i++) { for (i = 0; i < 6; i++) {
*out_ptr = MULH3(out2[i ], win[i ], 1) + buf[i + 6*2]; *out_ptr = MULH3(out2[i ], win[i ], 1) + buf[4*(i + 6*2)];
buf[i + 6*0] = MULH3(out2[i + 6], win[i + 6], 1); buf[4*(i + 6*0)] = MULH3(out2[i + 6], win[i + 6], 1);
out_ptr += SBLIMIT; out_ptr += SBLIMIT;
} }
imdct12(out2, ptr + 2); imdct12(out2, ptr + 2);
for (i = 0; i < 6; i++) { for (i = 0; i < 6; i++) {
buf[i + 6*0] = MULH3(out2[i ], win[i ], 1) + buf[i + 6*0]; buf[4*(i + 6*0)] = MULH3(out2[i ], win[i ], 1) + buf[4*(i + 6*0)];
buf[i + 6*1] = MULH3(out2[i + 6], win[i + 6], 1); buf[4*(i + 6*1)] = MULH3(out2[i + 6], win[i + 6], 1);
buf[i + 6*2] = 0; buf[4*(i + 6*2)] = 0;
} }
ptr += 18; ptr += 18;
buf += 18; buf += (j&3) != 3 ? 1 : (4*18-3);
} }
/* zero bands */ /* zero bands */
for (j = sblimit; j < SBLIMIT; j++) { for (j = sblimit; j < SBLIMIT; j++) {
/* overlap */ /* overlap */
out_ptr = sb_samples + j; out_ptr = sb_samples + j;
for (i = 0; i < 18; i++) { for (i = 0; i < 18; i++) {
*out_ptr = buf[i]; *out_ptr = buf[4*i];
buf[i] = 0; buf[4*i] = 0;
out_ptr += SBLIMIT; out_ptr += SBLIMIT;
} }
buf += 18; buf += (j&3) != 3 ? 1 : (4*18-3);
} }
} }