FFmpeg
lossless_videoencdsp_init.c
Go to the documentation of this file.
1 /*
2  * SIMD-optimized lossless video encoding functions
3  * Copyright (c) 2000, 2001 Fabrice Bellard
4  * Copyright (c) 2002-2004 Michael Niedermayer <michaelni@gmx.at>
5  *
6  * MMX optimization by Nick Kurshev <nickols_k@mail.ru>
7  *
8  * This file is part of FFmpeg.
9  *
10  * FFmpeg is free software; you can redistribute it and/or
11  * modify it under the terms of the GNU Lesser General Public
12  * License as published by the Free Software Foundation; either
13  * version 2.1 of the License, or (at your option) any later version.
14  *
15  * FFmpeg is distributed in the hope that it will be useful,
16  * but WITHOUT ANY WARRANTY; without even the implied warranty of
17  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
18  * Lesser General Public License for more details.
19  *
20  * You should have received a copy of the GNU Lesser General Public
21  * License along with FFmpeg; if not, write to the Free Software
22  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
23  */
24 
25 #include "libavutil/attributes.h"
26 #include "libavutil/cpu.h"
27 #include "libavutil/x86/asm.h"
28 #include "libavutil/x86/cpu.h"
30 #include "libavcodec/mathops.h"
31 
32 void ff_diff_bytes_sse2(uint8_t *dst, const uint8_t *src1, const uint8_t *src2,
33  intptr_t w);
34 void ff_diff_bytes_avx2(uint8_t *dst, const uint8_t *src1, const uint8_t *src2,
35  intptr_t w);
36 
37 void ff_sub_left_predict_avx(uint8_t *dst, const uint8_t *src,
38  ptrdiff_t stride, ptrdiff_t width, int height);
39 
40 #if HAVE_INLINE_ASM
41 
42 static void sub_median_pred_mmxext(uint8_t *dst, const uint8_t *src1,
43  const uint8_t *src2, intptr_t w,
44  int *left, int *left_top)
45 {
46  x86_reg i = 0;
47  uint8_t l, lt;
48 
49  __asm__ volatile (
50  "movq (%1, %0), %%mm0 \n\t" // LT
51  "psllq $8, %%mm0 \n\t"
52  "1: \n\t"
53  "movq (%1, %0), %%mm1 \n\t" // T
54  "movq -1(%2, %0), %%mm2 \n\t" // L
55  "movq (%2, %0), %%mm3 \n\t" // X
56  "movq %%mm2, %%mm4 \n\t" // L
57  "psubb %%mm0, %%mm2 \n\t"
58  "paddb %%mm1, %%mm2 \n\t" // L + T - LT
59  "movq %%mm4, %%mm5 \n\t" // L
60  "pmaxub %%mm1, %%mm4 \n\t" // max(T, L)
61  "pminub %%mm5, %%mm1 \n\t" // min(T, L)
62  "pminub %%mm2, %%mm4 \n\t"
63  "pmaxub %%mm1, %%mm4 \n\t"
64  "psubb %%mm4, %%mm3 \n\t" // dst - pred
65  "movq %%mm3, (%3, %0) \n\t"
66  "add $8, %0 \n\t"
67  "movq -1(%1, %0), %%mm0 \n\t" // LT
68  "cmp %4, %0 \n\t"
69  " jb 1b \n\t"
70  : "+r" (i)
71  : "r" (src1), "r" (src2), "r" (dst), "r" ((x86_reg) w));
72 
73  l = *left;
74  lt = *left_top;
75 
76  dst[0] = src2[0] - mid_pred(l, src1[0], (l + src1[0] - lt) & 0xFF);
77 
78  *left_top = src1[w - 1];
79  *left = src2[w - 1];
80 }
81 
82 #endif /* HAVE_INLINE_ASM */
83 
85 {
87 
88 #if HAVE_INLINE_ASM
89  if (INLINE_MMXEXT(cpu_flags)) {
90  c->sub_median_pred = sub_median_pred_mmxext;
91  }
92 #endif /* HAVE_INLINE_ASM */
93 
94  if (EXTERNAL_SSE2(cpu_flags)) {
95  c->diff_bytes = ff_diff_bytes_sse2;
96  }
97 
98  if (EXTERNAL_AVX(cpu_flags)) {
99  c->sub_left_predict = ff_sub_left_predict_avx;
100  }
101 
103  c->diff_bytes = ff_diff_bytes_avx2;
104  }
105 }
cpu.h
src1
const pixel * src1
Definition: h264pred_template.c:421
x86_reg
int x86_reg
Definition: asm.h:72
av_unused
#define av_unused
Definition: attributes.h:131
EXTERNAL_AVX2_FAST
#define EXTERNAL_AVX2_FAST(flags)
Definition: cpu.h:79
w
uint8_t w
Definition: llviddspenc.c:38
ff_sub_left_predict_avx
void ff_sub_left_predict_avx(uint8_t *dst, const uint8_t *src, ptrdiff_t stride, ptrdiff_t width, int height)
av_get_cpu_flags
int av_get_cpu_flags(void)
Return the flags which specify extensions supported by the CPU.
Definition: cpu.c:109
cpu_flags
static atomic_int cpu_flags
Definition: cpu.c:56
ff_diff_bytes_avx2
void ff_diff_bytes_avx2(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, intptr_t w)
av_cold
#define av_cold
Definition: attributes.h:90
asm.h
mathops.h
c
Undefined Behavior In the C some operations are like signed integer dereferencing freed accessing outside allocated Undefined Behavior must not occur in a C it is not safe even if the output of undefined operations is unused The unsafety may seem nit picking but Optimizing compilers have in fact optimized code on the assumption that no undefined Behavior occurs Optimizing code based on wrong assumptions can and has in some cases lead to effects beyond the output of computations The signed integer overflow problem in speed critical code Code which is highly optimized and works with signed integers sometimes has the problem that often the output of the computation does not c
Definition: undefined.txt:32
height
#define height
Definition: dsp.h:85
dst
uint8_t ptrdiff_t const uint8_t ptrdiff_t int intptr_t intptr_t int int16_t * dst
Definition: dsp.h:83
cpu.h
attributes.h
EXTERNAL_SSE2
#define EXTERNAL_SSE2(flags)
Definition: cpu.h:59
lossless_videoencdsp.h
i
#define i(width, name, range_min, range_max)
Definition: cbs_h2645.c:256
src2
const pixel * src2
Definition: h264pred_template.c:422
LLVidEncDSPContext
Definition: lossless_videoencdsp.h:25
stride
#define stride
Definition: h264pred_template.c:537
mid_pred
#define mid_pred
Definition: mathops.h:96
__asm__
__asm__(".macro parse_r var r\n\t" "\\var = -1\n\t" _IFC_REG(0) _IFC_REG(1) _IFC_REG(2) _IFC_REG(3) _IFC_REG(4) _IFC_REG(5) _IFC_REG(6) _IFC_REG(7) _IFC_REG(8) _IFC_REG(9) _IFC_REG(10) _IFC_REG(11) _IFC_REG(12) _IFC_REG(13) _IFC_REG(14) _IFC_REG(15) _IFC_REG(16) _IFC_REG(17) _IFC_REG(18) _IFC_REG(19) _IFC_REG(20) _IFC_REG(21) _IFC_REG(22) _IFC_REG(23) _IFC_REG(24) _IFC_REG(25) _IFC_REG(26) _IFC_REG(27) _IFC_REG(28) _IFC_REG(29) _IFC_REG(30) _IFC_REG(31) ".iflt \\var\n\t" ".error \"Unable to parse register name \\r\"\n\t" ".endif\n\t" ".endm")
EXTERNAL_AVX
#define EXTERNAL_AVX(flags)
Definition: cpu.h:70
left
Tag MUST be and< 10hcoeff half pel interpolation filter coefficients, hcoeff[0] are the 2 middle coefficients[1] are the next outer ones and so on, resulting in a filter like:...eff[2], hcoeff[1], hcoeff[0], hcoeff[0], hcoeff[1], hcoeff[2] ... the sign of the coefficients is not explicitly stored but alternates after each coeff and coeff[0] is positive, so ...,+,-,+,-,+,+,-,+,-,+,... hcoeff[0] is not explicitly stored but found by subtracting the sum of all stored coefficients with signs from 32 hcoeff[0]=32 - hcoeff[1] - hcoeff[2] - ... a good choice for hcoeff and htaps is htaps=6 hcoeff={40,-10, 2} an alternative which requires more computations at both encoder and decoder side and may or may not be better is htaps=8 hcoeff={42,-14, 6,-2}ref_frames minimum of the number of available reference frames and max_ref_frames for example the first frame after a key frame always has ref_frames=1spatial_decomposition_type wavelet type 0 is a 9/7 symmetric compact integer wavelet 1 is a 5/3 symmetric compact integer wavelet others are reserved stored as delta from last, last is reset to 0 if always_reset||keyframeqlog quality(logarithmic quantizer scale) stored as delta from last, last is reset to 0 if always_reset||keyframemv_scale stored as delta from last, last is reset to 0 if always_reset||keyframe FIXME check that everything works fine if this changes between framesqbias dequantization bias stored as delta from last, last is reset to 0 if always_reset||keyframeblock_max_depth maximum depth of the block tree stored as delta from last, last is reset to 0 if always_reset||keyframequant_table quantization tableHighlevel bitstream structure:==============================--------------------------------------------|Header|--------------------------------------------|------------------------------------|||Block0||||split?||||yes no||||......... intra?||||:Block01 :yes no||||:Block02 :....... ..........||||:Block03 ::y DC ::ref index:||||:Block04 ::cb DC ::motion x :||||......... :cr DC ::motion y :||||....... ..........|||------------------------------------||------------------------------------|||Block1|||...|--------------------------------------------|------------ ------------ ------------|||Y subbands||Cb subbands||Cr subbands||||--- ---||--- ---||--- ---|||||LL0||HL0||||LL0||HL0||||LL0||HL0|||||--- ---||--- ---||--- ---||||--- ---||--- ---||--- ---|||||LH0||HH0||||LH0||HH0||||LH0||HH0|||||--- ---||--- ---||--- ---||||--- ---||--- ---||--- ---|||||HL1||LH1||||HL1||LH1||||HL1||LH1|||||--- ---||--- ---||--- ---||||--- ---||--- ---||--- ---|||||HH1||HL2||||HH1||HL2||||HH1||HL2|||||...||...||...|||------------ ------------ ------------|--------------------------------------------Decoding process:=================------------|||Subbands|------------||||------------|Intra DC||||LL0 subband prediction ------------|\ Dequantization ------------------- \||Reference frames|\ IDWT|------- -------|Motion \|||Frame 0||Frame 1||Compensation . OBMC v -------|------- -------|--------------. \------> Frame n output Frame Frame<----------------------------------/|...|------------------- Range Coder:============Binary Range Coder:------------------- The implemented range coder is an adapted version based upon "Range encoding: an algorithm for removing redundancy from a digitised message." by G. N. N. Martin. The symbols encoded by the Snow range coder are bits(0|1). The associated probabilities are not fix but change depending on the symbol mix seen so far. bit seen|new state ---------+----------------------------------------------- 0|256 - state_transition_table[256 - old_state];1|state_transition_table[old_state];state_transition_table={ 0, 0, 0, 0, 0, 0, 0, 0, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 94, 95, 96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 128, 129, 130, 131, 132, 133, 133, 134, 135, 136, 137, 138, 139, 140, 141, 142, 143, 144, 145, 146, 147, 148, 149, 150, 151, 152, 152, 153, 154, 155, 156, 157, 158, 159, 160, 161, 162, 163, 164, 165, 166, 167, 168, 169, 170, 171, 171, 172, 173, 174, 175, 176, 177, 178, 179, 180, 181, 182, 183, 184, 185, 186, 187, 188, 189, 190, 190, 191, 192, 194, 194, 195, 196, 197, 198, 199, 200, 201, 202, 202, 204, 205, 206, 207, 208, 209, 209, 210, 211, 212, 213, 215, 215, 216, 217, 218, 219, 220, 220, 222, 223, 224, 225, 226, 227, 227, 229, 229, 230, 231, 232, 234, 234, 235, 236, 237, 238, 239, 240, 241, 242, 243, 244, 245, 246, 247, 248, 248, 0, 0, 0, 0, 0, 0, 0};FIXME Range Coding of integers:------------------------- FIXME Neighboring Blocks:===================left and top are set to the respective blocks unless they are outside of the image in which case they are set to the Null block top-left is set to the top left block unless it is outside of the image in which case it is set to the left block if this block has no larger parent block or it is at the left side of its parent block and the top right block is not outside of the image then the top right block is used for top-right else the top-left block is used Null block y, cb, cr are 128 level, ref, mx and my are 0 Motion Vector Prediction:=========================1. the motion vectors of all the neighboring blocks are scaled to compensate for the difference of reference frames scaled_mv=(mv *(256 *(current_reference+1)/(mv.reference+1))+128)> the median of the scaled left
Definition: snow.txt:386
ff_llvidencdsp_init_x86
av_cold void ff_llvidencdsp_init_x86(LLVidEncDSPContext *c)
Definition: lossless_videoencdsp_init.c:84
INLINE_MMXEXT
#define INLINE_MMXEXT(flags)
Definition: cpu.h:88
width
#define width
Definition: dsp.h:85
src
#define src
Definition: vp8dsp.c:248
ff_diff_bytes_sse2
void ff_diff_bytes_sse2(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, intptr_t w)