FFmpeg
 All Data Structures Namespaces Files Functions Variables Typedefs Enumerations Enumerator Macros Groups Pages
mdct15_init.c
Go to the documentation of this file.
1 /*
2  * SIMD optimized non-power-of-two MDCT functions
3  *
4  * Copyright (C) 2017 Rostislav Pehlivanov <atomnuker@gmail.com>
5  *
6  * This file is part of FFmpeg.
7  *
8  * FFmpeg is free software; you can redistribute it and/or
9  * modify it under the terms of the GNU Lesser General Public
10  * License as published by the Free Software Foundation; either
11  * version 2.1 of the License, or (at your option) any later version.
12  *
13  * FFmpeg is distributed in the hope that it will be useful,
14  * but WITHOUT ANY WARRANTY; without even the implied warranty of
15  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
16  * Lesser General Public License for more details.
17  *
18  * You should have received a copy of the GNU Lesser General Public
19  * License along with FFmpeg; if not, write to the Free Software
20  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
21  */
22 
23 #include "config.h"
24 
25 #include "libavutil/x86/cpu.h"
26 #include "libavcodec/mdct15.h"
27 
28 void ff_mdct15_postreindex_sse3(FFTComplex *out, FFTComplex *in, FFTComplex *exp, int *lut, ptrdiff_t len8);
29 void ff_mdct15_postreindex_avx2(FFTComplex *out, FFTComplex *in, FFTComplex *exp, int *lut, ptrdiff_t len8);
30 
32 
34 {
35  int k;
36  FFTComplex exp_5point[4];
37 
38  FFTComplex tmp[21], tmp2[30];
39  memcpy(tmp, s->exptab, sizeof(FFTComplex)*21);
40 
41  /* 15-point FFT twiddles */
42  for (k = 0; k < 5; k++) {
43  tmp2[6*k + 0] = tmp[k + 0];
44  tmp2[6*k + 2] = tmp[k + 5];
45  tmp2[6*k + 4] = tmp[k + 10];
46 
47  tmp2[6*k + 1] = tmp[2 * (k + 0)];
48  tmp2[6*k + 3] = tmp[2 * (k + 5)];
49  tmp2[6*k + 5] = tmp[2 * k + 5 ];
50  }
51 
52  for (k = 0; k < 6; k++) {
53  FFTComplex ac_exp[] = {
54  { tmp2[6*1 + k].re, tmp2[6*1 + k].re },
55  { tmp2[6*2 + k].re, tmp2[6*2 + k].re },
56  { tmp2[6*3 + k].re, tmp2[6*3 + k].re },
57  { tmp2[6*4 + k].re, tmp2[6*4 + k].re },
58  { tmp2[6*1 + k].im, -tmp2[6*1 + k].im },
59  { tmp2[6*2 + k].im, -tmp2[6*2 + k].im },
60  { tmp2[6*3 + k].im, -tmp2[6*3 + k].im },
61  { tmp2[6*4 + k].im, -tmp2[6*4 + k].im },
62  };
63  memcpy(s->exptab + 8*k, ac_exp, 8*sizeof(FFTComplex));
64  }
65 
66  /* Specialcase when k = 0 */
67  for (k = 0; k < 3; k++) {
68  FFTComplex dc_exp[] = {
69  { tmp2[2*k + 0].re, -tmp2[2*k + 0].im },
70  { tmp2[2*k + 0].im, tmp2[2*k + 0].re },
71  { tmp2[2*k + 1].re, -tmp2[2*k + 1].im },
72  { tmp2[2*k + 1].im, tmp2[2*k + 1].re },
73  };
74  memcpy(s->exptab + 8*6 + 4*k, dc_exp, 4*sizeof(FFTComplex));
75  }
76 
77  /* 5-point FFT twiddles */
78  exp_5point[0].re = exp_5point[0].im = tmp[19].re;
79  exp_5point[1].re = exp_5point[1].im = tmp[19].im;
80  exp_5point[2].re = exp_5point[2].im = tmp[20].re;
81  exp_5point[3].re = exp_5point[3].im = tmp[20].im;
82 
83  memcpy(s->exptab + 8*6 + 4*3, exp_5point, 4*sizeof(FFTComplex));
84 }
85 
87 {
88  int adjust_twiddles = 0;
90 
91  if (EXTERNAL_SSE3(cpu_flags))
93 
94  if (ARCH_X86_64 && EXTERNAL_AVX(cpu_flags)) {
95  s->fft15 = ff_fft15_avx;
96  adjust_twiddles = 1;
97  }
98 
99  if (ARCH_X86_64 && EXTERNAL_AVX2_FAST(cpu_flags))
101 
102  if (adjust_twiddles)
103  perm_twiddles(s);
104 }
const char * s
Definition: avisynth_c.h:768
void(* fft15)(FFTComplex *out, FFTComplex *in, FFTComplex *exptab, ptrdiff_t stride)
Definition: mdct15.h:43
av_cold void ff_mdct15_init_x86(MDCT15Context *s)
Definition: mdct15_init.c:86
FFTSample re
Definition: avfft.h:38
static atomic_int cpu_flags
Definition: cpu.c:50
void ff_mdct15_postreindex_sse3(FFTComplex *out, FFTComplex *in, FFTComplex *exp, int *lut, ptrdiff_t len8)
#define av_cold
Definition: attributes.h:82
void ff_fft15_avx(FFTComplex *out, FFTComplex *in, FFTComplex *exptab, ptrdiff_t stride)
#define EXTERNAL_SSE3(flags)
Definition: cpu.h:62
#define EXTERNAL_AVX2_FAST(flags)
Definition: cpu.h:79
int8_t exp
Definition: eval.c:65
void(* postreindex)(FFTComplex *out, FFTComplex *in, FFTComplex *exp, int *lut, ptrdiff_t len8)
Definition: mdct15.h:46
void ff_mdct15_postreindex_avx2(FFTComplex *out, FFTComplex *in, FFTComplex *exp, int *lut, ptrdiff_t len8)
static struct @119 * exptab
uint8_t pi<< 24) CONV_FUNC_GROUP(AV_SAMPLE_FMT_FLT, float, AV_SAMPLE_FMT_U8, uint8_t,(*(constuint8_t *) pi-0x80)*(1.0f/(1<< 7))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_DBL, double, AV_SAMPLE_FMT_U8, uint8_t,(*(constuint8_t *) pi-0x80)*(1.0/(1<< 7))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_U8, uint8_t, AV_SAMPLE_FMT_S16, int16_t,(*(constint16_t *) pi >>8)+0x80) CONV_FUNC_GROUP(AV_SAMPLE_FMT_FLT, float, AV_SAMPLE_FMT_S16, int16_t,*(constint16_t *) pi *(1.0f/(1<< 15))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_DBL, double, AV_SAMPLE_FMT_S16, int16_t,*(constint16_t *) pi *(1.0/(1<< 15))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_U8, uint8_t, AV_SAMPLE_FMT_S32, int32_t,(*(constint32_t *) pi >>24)+0x80) CONV_FUNC_GROUP(AV_SAMPLE_FMT_FLT, float, AV_SAMPLE_FMT_S32, int32_t,*(constint32_t *) pi *(1.0f/(1U<< 31))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_DBL, double, AV_SAMPLE_FMT_S32, int32_t,*(constint32_t *) pi *(1.0/(1U<< 31))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_U8, uint8_t, AV_SAMPLE_FMT_FLT, float, av_clip_uint8(lrintf(*(constfloat *) pi *(1<< 7))+0x80)) CONV_FUNC_GROUP(AV_SAMPLE_FMT_S16, int16_t, AV_SAMPLE_FMT_FLT, float, av_clip_int16(lrintf(*(constfloat *) pi *(1<< 15)))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_S32, int32_t, AV_SAMPLE_FMT_FLT, float, av_clipl_int32(llrintf(*(constfloat *) pi *(1U<< 31)))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_U8, uint8_t, AV_SAMPLE_FMT_DBL, double, av_clip_uint8(lrint(*(constdouble *) pi *(1<< 7))+0x80)) CONV_FUNC_GROUP(AV_SAMPLE_FMT_S16, int16_t, AV_SAMPLE_FMT_DBL, double, av_clip_int16(lrint(*(constdouble *) pi *(1<< 15)))) CONV_FUNC_GROUP(AV_SAMPLE_FMT_S32, int32_t, AV_SAMPLE_FMT_DBL, double, av_clipl_int32(llrint(*(constdouble *) pi *(1U<< 31))))#defineSET_CONV_FUNC_GROUP(ofmt, ifmt) staticvoidset_generic_function(AudioConvert *ac){}voidff_audio_convert_free(AudioConvert **ac){if(!*ac) return;ff_dither_free(&(*ac) ->dc);av_freep(ac);}AudioConvert *ff_audio_convert_alloc(AVAudioResampleContext *avr, enumAVSampleFormatout_fmt, enumAVSampleFormatin_fmt, intchannels, intsample_rate, intapply_map){AudioConvert *ac;intin_planar, out_planar;ac=av_mallocz(sizeof(*ac));if(!ac) returnNULL;ac->avr=avr;ac->out_fmt=out_fmt;ac->in_fmt=in_fmt;ac->channels=channels;ac->apply_map=apply_map;if(avr->dither_method!=AV_RESAMPLE_DITHER_NONE &&av_get_packed_sample_fmt(out_fmt)==AV_SAMPLE_FMT_S16 &&av_get_bytes_per_sample(in_fmt)>2){ac->dc=ff_dither_alloc(avr, out_fmt, in_fmt, channels, sample_rate, apply_map);if(!ac->dc){av_free(ac);returnNULL;}returnac;}in_planar=ff_sample_fmt_is_planar(in_fmt, channels);out_planar=ff_sample_fmt_is_planar(out_fmt, channels);if(in_planar==out_planar){ac->func_type=CONV_FUNC_TYPE_FLAT;ac->planes=in_planar?ac->channels:1;}elseif(in_planar) ac->func_type=CONV_FUNC_TYPE_INTERLEAVE;elseac->func_type=CONV_FUNC_TYPE_DEINTERLEAVE;set_generic_function(ac);if(ARCH_AARCH64) ff_audio_convert_init_aarch64(ac);if(ARCH_ARM) ff_audio_convert_init_arm(ac);if(ARCH_X86) ff_audio_convert_init_x86(ac);returnac;}intff_audio_convert(AudioConvert *ac, AudioData *out, AudioData *in){intuse_generic=1;intlen=in->nb_samples;intp;if(ac->dc){av_log(ac->avr, AV_LOG_TRACE,"%dsamples-audio_convert:%sto%s(dithered)\n", len, av_get_sample_fmt_name(ac->in_fmt), av_get_sample_fmt_name(ac->out_fmt));returnff_convert_dither(ac-> in
static void perm_twiddles(MDCT15Context *s)
Definition: mdct15_init.c:33
int av_get_cpu_flags(void)
Return the flags which specify extensions supported by the CPU.
Definition: cpu.c:92
GLint GLenum GLboolean GLsizei stride
Definition: opengl_enc.c:105
FFTComplex exptab[64]
Definition: mdct15.h:40
FFTSample im
Definition: avfft.h:38
FILE * out
Definition: movenc.c:54
#define EXTERNAL_AVX(flags)
Definition: cpu.h:70
static uint8_t tmp[11]
Definition: aes_ctr.c:26