mirror of
https://github.com/FFmpeg/FFmpeg.git
synced 2024-12-23 12:43:46 +02:00
avcodec/cfhdenc: refactor DSP code for CFHD encoder
This is needed to implement x86 SIMD.
This commit is contained in:
parent
767d3f7dcf
commit
d69f4be1c6
@ -258,7 +258,7 @@ OBJS-$(CONFIG_CDGRAPHICS_DECODER) += cdgraphics.o
|
|||||||
OBJS-$(CONFIG_CDTOONS_DECODER) += cdtoons.o
|
OBJS-$(CONFIG_CDTOONS_DECODER) += cdtoons.o
|
||||||
OBJS-$(CONFIG_CDXL_DECODER) += cdxl.o
|
OBJS-$(CONFIG_CDXL_DECODER) += cdxl.o
|
||||||
OBJS-$(CONFIG_CFHD_DECODER) += cfhd.o cfhddata.o cfhddsp.o
|
OBJS-$(CONFIG_CFHD_DECODER) += cfhd.o cfhddata.o cfhddsp.o
|
||||||
OBJS-$(CONFIG_CFHD_ENCODER) += cfhdenc.o cfhddata.o
|
OBJS-$(CONFIG_CFHD_ENCODER) += cfhdenc.o cfhddata.o cfhdencdsp.o
|
||||||
OBJS-$(CONFIG_CINEPAK_DECODER) += cinepak.o
|
OBJS-$(CONFIG_CINEPAK_DECODER) += cinepak.o
|
||||||
OBJS-$(CONFIG_CINEPAK_ENCODER) += cinepakenc.o elbg.o
|
OBJS-$(CONFIG_CINEPAK_ENCODER) += cinepakenc.o elbg.o
|
||||||
OBJS-$(CONFIG_CLEARVIDEO_DECODER) += clearvideo.o
|
OBJS-$(CONFIG_CLEARVIDEO_DECODER) += clearvideo.o
|
||||||
|
@ -33,6 +33,7 @@
|
|||||||
#include "avcodec.h"
|
#include "avcodec.h"
|
||||||
#include "bytestream.h"
|
#include "bytestream.h"
|
||||||
#include "cfhd.h"
|
#include "cfhd.h"
|
||||||
|
#include "cfhdencdsp.h"
|
||||||
#include "put_bits.h"
|
#include "put_bits.h"
|
||||||
#include "internal.h"
|
#include "internal.h"
|
||||||
#include "thread.h"
|
#include "thread.h"
|
||||||
@ -239,6 +240,8 @@ typedef struct CFHDEncContext {
|
|||||||
Runbook rb[321];
|
Runbook rb[321];
|
||||||
Codebook cb[513];
|
Codebook cb[513];
|
||||||
int16_t *alpha;
|
int16_t *alpha;
|
||||||
|
|
||||||
|
CFHDEncDSPContext dsp;
|
||||||
} CFHDEncContext;
|
} CFHDEncContext;
|
||||||
|
|
||||||
static av_cold int cfhd_encode_init(AVCodecContext *avctx)
|
static av_cold int cfhd_encode_init(AVCodecContext *avctx)
|
||||||
@ -359,6 +362,8 @@ static av_cold int cfhd_encode_init(AVCodecContext *avctx)
|
|||||||
s->lut[i] = last;
|
s->lut[i] = last;
|
||||||
}
|
}
|
||||||
|
|
||||||
|
ff_cfhdencdsp_init(&s->dsp);
|
||||||
|
|
||||||
if (s->planes != 4)
|
if (s->planes != 4)
|
||||||
return 0;
|
return 0;
|
||||||
|
|
||||||
@ -369,42 +374,6 @@ static av_cold int cfhd_encode_init(AVCodecContext *avctx)
|
|||||||
return 0;
|
return 0;
|
||||||
}
|
}
|
||||||
|
|
||||||
static av_always_inline void filter(int16_t *input, ptrdiff_t in_stride,
|
|
||||||
int16_t *low, ptrdiff_t low_stride,
|
|
||||||
int16_t *high, ptrdiff_t high_stride,
|
|
||||||
int len)
|
|
||||||
{
|
|
||||||
low[(0>>1) * low_stride] = av_clip_int16(input[0*in_stride] + input[1*in_stride]);
|
|
||||||
high[(0>>1) * high_stride] = av_clip_int16((5 * input[0*in_stride] - 11 * input[1*in_stride] +
|
|
||||||
4 * input[2*in_stride] + 4 * input[3*in_stride] -
|
|
||||||
1 * input[4*in_stride] - 1 * input[5*in_stride] + 4) >> 3);
|
|
||||||
|
|
||||||
for (int i = 2; i < len - 2; i += 2) {
|
|
||||||
low[(i>>1) * low_stride] = av_clip_int16(input[i*in_stride] + input[(i+1)*in_stride]);
|
|
||||||
high[(i>>1) * high_stride] = av_clip_int16(((-input[(i-2)*in_stride] - input[(i-1)*in_stride] +
|
|
||||||
input[(i+2)*in_stride] + input[(i+3)*in_stride] + 4) >> 3) +
|
|
||||||
input[(i+0)*in_stride] - input[(i+1)*in_stride]);
|
|
||||||
}
|
|
||||||
|
|
||||||
low[((len-2)>>1) * low_stride] = av_clip_int16(input[((len-2)+0)*in_stride] + input[((len-2)+1)*in_stride]);
|
|
||||||
high[((len-2)>>1) * high_stride] = av_clip_int16((11* input[((len-2)+0)*in_stride] - 5 * input[((len-2)+1)*in_stride] -
|
|
||||||
4 * input[((len-2)-1)*in_stride] - 4 * input[((len-2)-2)*in_stride] +
|
|
||||||
1 * input[((len-2)-3)*in_stride] + 1 * input[((len-2)-4)*in_stride] + 4) >> 3);
|
|
||||||
}
|
|
||||||
|
|
||||||
static void horiz_filter(int16_t *input, int16_t *low, int16_t *high,
|
|
||||||
int width)
|
|
||||||
{
|
|
||||||
filter(input, 1, low, 1, high, 1, width);
|
|
||||||
}
|
|
||||||
|
|
||||||
static void vert_filter(int16_t *input, ptrdiff_t in_stride,
|
|
||||||
int16_t *low, ptrdiff_t low_stride,
|
|
||||||
int16_t *high, ptrdiff_t high_stride, int len)
|
|
||||||
{
|
|
||||||
filter(input, in_stride, low, low_stride, high, high_stride, len);
|
|
||||||
}
|
|
||||||
|
|
||||||
static void quantize_band(int16_t *input, int width, int a_width,
|
static void quantize_band(int16_t *input, int width, int a_width,
|
||||||
int height, unsigned quantization)
|
int height, unsigned quantization)
|
||||||
{
|
{
|
||||||
@ -454,6 +423,7 @@ static int cfhd_encode_frame(AVCodecContext *avctx, AVPacket *pkt,
|
|||||||
const AVFrame *frame, int *got_packet)
|
const AVFrame *frame, int *got_packet)
|
||||||
{
|
{
|
||||||
CFHDEncContext *s = avctx->priv_data;
|
CFHDEncContext *s = avctx->priv_data;
|
||||||
|
CFHDEncDSPContext *dsp = &s->dsp;
|
||||||
PutByteContext *pby = &s->pby;
|
PutByteContext *pby = &s->pby;
|
||||||
PutBitContext *pb = &s->pb;
|
PutBitContext *pb = &s->pb;
|
||||||
const Codebook *const cb = s->cb;
|
const Codebook *const cb = s->cb;
|
||||||
@ -480,12 +450,9 @@ static int cfhd_encode_frame(AVCodecContext *avctx, AVPacket *pkt,
|
|||||||
in_stride = avctx->width;
|
in_stride = avctx->width;
|
||||||
}
|
}
|
||||||
|
|
||||||
for (int i = 0; i < height * 2; i++) {
|
dsp->horiz_filter(input, low, high,
|
||||||
horiz_filter(input, low, high, width * 2);
|
in_stride, a_width, a_width,
|
||||||
input += in_stride;
|
width * 2, height * 2);
|
||||||
low += a_width;
|
|
||||||
high += a_width;
|
|
||||||
}
|
|
||||||
|
|
||||||
input = s->plane[plane].l_h[7];
|
input = s->plane[plane].l_h[7];
|
||||||
low = s->plane[plane].subband[7];
|
low = s->plane[plane].subband[7];
|
||||||
@ -493,23 +460,17 @@ static int cfhd_encode_frame(AVCodecContext *avctx, AVPacket *pkt,
|
|||||||
high = s->plane[plane].subband[9];
|
high = s->plane[plane].subband[9];
|
||||||
high_stride = s->plane[plane].band[2][0].a_width;
|
high_stride = s->plane[plane].band[2][0].a_width;
|
||||||
|
|
||||||
for (int i = 0; i < width; i++) {
|
dsp->vert_filter(input, low, high,
|
||||||
vert_filter(input, a_width, low, low_stride, high, high_stride, height * 2);
|
a_width, low_stride, high_stride,
|
||||||
input++;
|
width, height * 2);
|
||||||
low++;
|
|
||||||
high++;
|
|
||||||
}
|
|
||||||
|
|
||||||
input = s->plane[plane].l_h[6];
|
input = s->plane[plane].l_h[6];
|
||||||
low = s->plane[plane].l_h[7];
|
low = s->plane[plane].l_h[7];
|
||||||
high = s->plane[plane].subband[8];
|
high = s->plane[plane].subband[8];
|
||||||
|
|
||||||
for (int i = 0; i < width; i++) {
|
dsp->vert_filter(input, low, high,
|
||||||
vert_filter(input, a_width, low, low_stride, high, high_stride, height * 2);
|
a_width, low_stride, high_stride,
|
||||||
input++;
|
width, height * 2);
|
||||||
low++;
|
|
||||||
high++;
|
|
||||||
}
|
|
||||||
|
|
||||||
a_width = s->plane[plane].band[1][0].a_width;
|
a_width = s->plane[plane].band[1][0].a_width;
|
||||||
width = s->plane[plane].band[1][0].width;
|
width = s->plane[plane].band[1][0].width;
|
||||||
@ -527,34 +488,25 @@ static int cfhd_encode_frame(AVCodecContext *avctx, AVPacket *pkt,
|
|||||||
}
|
}
|
||||||
|
|
||||||
input = s->plane[plane].l_h[7];
|
input = s->plane[plane].l_h[7];
|
||||||
for (int i = 0; i < height * 2; i++) {
|
dsp->horiz_filter(input, low, high,
|
||||||
horiz_filter(input, low, high, width * 2);
|
a_width * 2, low_stride, high_stride,
|
||||||
input += a_width * 2;
|
width * 2, height * 2);
|
||||||
low += low_stride;
|
|
||||||
high += high_stride;
|
|
||||||
}
|
|
||||||
|
|
||||||
input = s->plane[plane].l_h[4];
|
input = s->plane[plane].l_h[4];
|
||||||
low = s->plane[plane].subband[4];
|
low = s->plane[plane].subband[4];
|
||||||
high = s->plane[plane].subband[6];
|
high = s->plane[plane].subband[6];
|
||||||
|
|
||||||
for (int i = 0; i < width; i++) {
|
dsp->vert_filter(input, low, high,
|
||||||
vert_filter(input, a_width, low, low_stride, high, high_stride, height * 2);
|
a_width, low_stride, high_stride,
|
||||||
input++;
|
width, height * 2);
|
||||||
low++;
|
|
||||||
high++;
|
|
||||||
}
|
|
||||||
|
|
||||||
input = s->plane[plane].l_h[3];
|
input = s->plane[plane].l_h[3];
|
||||||
low = s->plane[plane].l_h[4];
|
low = s->plane[plane].l_h[4];
|
||||||
high = s->plane[plane].subband[5];
|
high = s->plane[plane].subband[5];
|
||||||
|
|
||||||
for (int i = 0; i < width; i++) {
|
dsp->vert_filter(input, low, high,
|
||||||
vert_filter(input, a_width, low, low_stride, high, high_stride, height * 2);
|
a_width, low_stride, high_stride,
|
||||||
input++;
|
width, height * 2);
|
||||||
low++;
|
|
||||||
high++;
|
|
||||||
}
|
|
||||||
|
|
||||||
a_width = s->plane[plane].band[0][0].a_width;
|
a_width = s->plane[plane].band[0][0].a_width;
|
||||||
width = s->plane[plane].band[0][0].width;
|
width = s->plane[plane].band[0][0].width;
|
||||||
@ -574,34 +526,25 @@ static int cfhd_encode_frame(AVCodecContext *avctx, AVPacket *pkt,
|
|||||||
}
|
}
|
||||||
|
|
||||||
input = s->plane[plane].l_h[4];
|
input = s->plane[plane].l_h[4];
|
||||||
for (int i = 0; i < height * 2; i++) {
|
dsp->horiz_filter(input, low, high,
|
||||||
horiz_filter(input, low, high, width * 2);
|
a_width * 2, low_stride, high_stride,
|
||||||
input += a_width * 2;
|
width * 2, height * 2);
|
||||||
low += low_stride;
|
|
||||||
high += high_stride;
|
|
||||||
}
|
|
||||||
|
|
||||||
low = s->plane[plane].subband[1];
|
low = s->plane[plane].subband[1];
|
||||||
high = s->plane[plane].subband[3];
|
high = s->plane[plane].subband[3];
|
||||||
input = s->plane[plane].l_h[1];
|
input = s->plane[plane].l_h[1];
|
||||||
|
|
||||||
for (int i = 0; i < width; i++) {
|
dsp->vert_filter(input, low, high,
|
||||||
vert_filter(input, a_width, low, low_stride, high, high_stride, height * 2);
|
a_width, low_stride, high_stride,
|
||||||
input++;
|
width, height * 2);
|
||||||
low++;
|
|
||||||
high++;
|
|
||||||
}
|
|
||||||
|
|
||||||
low = s->plane[plane].subband[0];
|
low = s->plane[plane].subband[0];
|
||||||
high = s->plane[plane].subband[2];
|
high = s->plane[plane].subband[2];
|
||||||
input = s->plane[plane].l_h[0];
|
input = s->plane[plane].l_h[0];
|
||||||
|
|
||||||
for (int i = 0; i < width; i++) {
|
dsp->vert_filter(input, low, high,
|
||||||
vert_filter(input, a_width, low, low_stride, high, high_stride, height * 2);
|
a_width, low_stride, high_stride,
|
||||||
input++;
|
width, height * 2);
|
||||||
low++;
|
|
||||||
high++;
|
|
||||||
}
|
|
||||||
}
|
}
|
||||||
|
|
||||||
ret = ff_alloc_packet2(avctx, pkt, 64LL + s->planes * (2LL * avctx->width * avctx->height + 1000LL), 0);
|
ret = ff_alloc_packet2(avctx, pkt, 64LL + s->planes * (2LL * avctx->width * avctx->height + 1000LL), 0);
|
||||||
|
76
libavcodec/cfhdencdsp.c
Normal file
76
libavcodec/cfhdencdsp.c
Normal file
@ -0,0 +1,76 @@
|
|||||||
|
/*
|
||||||
|
* Copyright (c) 2015-2016 Kieran Kunhya <kieran@kunhya.com>
|
||||||
|
*
|
||||||
|
* This file is part of FFmpeg.
|
||||||
|
*
|
||||||
|
* FFmpeg is free software; you can redistribute it and/or
|
||||||
|
* modify it under the terms of the GNU Lesser General Public
|
||||||
|
* License as published by the Free Software Foundation; either
|
||||||
|
* version 2.1 of the License, or (at your option) any later version.
|
||||||
|
*
|
||||||
|
* FFmpeg is distributed in the hope that it will be useful,
|
||||||
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
||||||
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
||||||
|
* Lesser General Public License for more details.
|
||||||
|
*
|
||||||
|
* You should have received a copy of the GNU Lesser General Public
|
||||||
|
* License along with FFmpeg; if not, write to the Free Software
|
||||||
|
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
|
||||||
|
*/
|
||||||
|
|
||||||
|
#include "libavutil/attributes.h"
|
||||||
|
#include "libavutil/common.h"
|
||||||
|
#include "libavutil/avassert.h"
|
||||||
|
|
||||||
|
#include "cfhdencdsp.h"
|
||||||
|
|
||||||
|
static av_always_inline void filter(int16_t *input, ptrdiff_t in_stride,
|
||||||
|
int16_t *low, ptrdiff_t low_stride,
|
||||||
|
int16_t *high, ptrdiff_t high_stride,
|
||||||
|
int len)
|
||||||
|
{
|
||||||
|
low[(0>>1) * low_stride] = av_clip_int16(input[0*in_stride] + input[1*in_stride]);
|
||||||
|
high[(0>>1) * high_stride] = av_clip_int16((5 * input[0*in_stride] - 11 * input[1*in_stride] +
|
||||||
|
4 * input[2*in_stride] + 4 * input[3*in_stride] -
|
||||||
|
1 * input[4*in_stride] - 1 * input[5*in_stride] + 4) >> 3);
|
||||||
|
|
||||||
|
for (int i = 2; i < len - 2; i += 2) {
|
||||||
|
low[(i>>1) * low_stride] = av_clip_int16(input[i*in_stride] + input[(i+1)*in_stride]);
|
||||||
|
high[(i>>1) * high_stride] = av_clip_int16(((-input[(i-2)*in_stride] - input[(i-1)*in_stride] +
|
||||||
|
input[(i+2)*in_stride] + input[(i+3)*in_stride] + 4) >> 3) +
|
||||||
|
input[(i+0)*in_stride] - input[(i+1)*in_stride]);
|
||||||
|
}
|
||||||
|
|
||||||
|
low[((len-2)>>1) * low_stride] = av_clip_int16(input[((len-2)+0)*in_stride] + input[((len-2)+1)*in_stride]);
|
||||||
|
high[((len-2)>>1) * high_stride] = av_clip_int16((11* input[((len-2)+0)*in_stride] - 5 * input[((len-2)+1)*in_stride] -
|
||||||
|
4 * input[((len-2)-1)*in_stride] - 4 * input[((len-2)-2)*in_stride] +
|
||||||
|
1 * input[((len-2)-3)*in_stride] + 1 * input[((len-2)-4)*in_stride] + 4) >> 3);
|
||||||
|
}
|
||||||
|
|
||||||
|
static void horiz_filter(int16_t *input, int16_t *low, int16_t *high,
|
||||||
|
ptrdiff_t in_stride, ptrdiff_t low_stride,
|
||||||
|
ptrdiff_t high_stride,
|
||||||
|
int width, int height)
|
||||||
|
{
|
||||||
|
for (int i = 0; i < height; i++) {
|
||||||
|
filter(input, 1, low, 1, high, 1, width);
|
||||||
|
input += in_stride;
|
||||||
|
low += low_stride;
|
||||||
|
high += high_stride;
|
||||||
|
}
|
||||||
|
}
|
||||||
|
|
||||||
|
static void vert_filter(int16_t *input, int16_t *low, int16_t *high,
|
||||||
|
ptrdiff_t in_stride, ptrdiff_t low_stride,
|
||||||
|
ptrdiff_t high_stride,
|
||||||
|
int width, int height)
|
||||||
|
{
|
||||||
|
for (int i = 0; i < width; i++)
|
||||||
|
filter(&input[i], in_stride, &low[i], low_stride, &high[i], high_stride, height);
|
||||||
|
}
|
||||||
|
|
||||||
|
av_cold void ff_cfhdencdsp_init(CFHDEncDSPContext *c)
|
||||||
|
{
|
||||||
|
c->horiz_filter = horiz_filter;
|
||||||
|
c->vert_filter = vert_filter;
|
||||||
|
}
|
41
libavcodec/cfhdencdsp.h
Normal file
41
libavcodec/cfhdencdsp.h
Normal file
@ -0,0 +1,41 @@
|
|||||||
|
/*
|
||||||
|
* This file is part of FFmpeg.
|
||||||
|
*
|
||||||
|
* FFmpeg is free software; you can redistribute it and/or
|
||||||
|
* modify it under the terms of the GNU Lesser General Public
|
||||||
|
* License as published by the Free Software Foundation; either
|
||||||
|
* version 2.1 of the License, or (at your option) any later version.
|
||||||
|
*
|
||||||
|
* FFmpeg is distributed in the hope that it will be useful,
|
||||||
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
||||||
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
||||||
|
* Lesser General Public License for more details.
|
||||||
|
*
|
||||||
|
* You should have received a copy of the GNU Lesser General Public
|
||||||
|
* License along with FFmpeg; if not, write to the Free Software
|
||||||
|
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
|
||||||
|
*/
|
||||||
|
|
||||||
|
#ifndef AVCODEC_CFHDENCDSP_H
|
||||||
|
#define AVCODEC_CFHDENCDSP_H
|
||||||
|
|
||||||
|
#include <stddef.h>
|
||||||
|
#include <stdint.h>
|
||||||
|
|
||||||
|
typedef struct CFHDEncDSPContext {
|
||||||
|
void (*horiz_filter)(int16_t *input, int16_t *low, int16_t *high,
|
||||||
|
ptrdiff_t in_stride, ptrdiff_t low_stride,
|
||||||
|
ptrdiff_t high_stride,
|
||||||
|
int width, int height);
|
||||||
|
|
||||||
|
void (*vert_filter)(int16_t *input, int16_t *low, int16_t *high,
|
||||||
|
ptrdiff_t in_stride, ptrdiff_t low_stride,
|
||||||
|
ptrdiff_t high_stride,
|
||||||
|
int width, int height);
|
||||||
|
} CFHDEncDSPContext;
|
||||||
|
|
||||||
|
void ff_cfhdencdsp_init(CFHDEncDSPContext *c);
|
||||||
|
|
||||||
|
void ff_cfhdencdsp_init_x86(CFHDEncDSPContext *c);
|
||||||
|
|
||||||
|
#endif /* AVCODEC_CFHDENCDSP_H */
|
Loading…
Reference in New Issue
Block a user