1 ;******************************************************************************
2 ;* SIMD optimized MPEG-4 Parametric Stereo decoding functions
4 ;* Copyright (C) 2015 James Almer
6 ;* This file is part of FFmpeg.
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.
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.
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 ;******************************************************************************
23 %include "libavutil/x86/x86util.asm"
27 ps_p1m1p1m1: dd 0, 0x80000000, 0, 0x80000000
31 ;*************************************************************************
32 ;void ff_ps_add_squares_<opt>(float *dst, const float (*src)[2], int n);
33 ;*************************************************************************
34 %macro PS_ADD_SQUARES 1
35 cglobal ps_add_squares, 3, 3, %1, dst, src, n
43 movaps m1, [srcq+nq+mmsize]
60 ;*******************************************************************
61 ;void ff_ps_mul_pair_single_sse(float (*dst)[2], float (*src0)[2],
62 ; float *src1, int n);
63 ;*******************************************************************
65 cglobal ps_mul_pair_single, 4, 4, 4, dst, src1, src2, n
74 movu m1, [src1q+nq+mmsize]
82 mova [dstq+nq+mmsize], m1
88 ;***********************************************************************
89 ;void ff_ps_stereo_interpolate_sse3(float (*l)[2], float (*r)[2],
90 ; float h[2][4], float h_step[2][4],
92 ;***********************************************************************
94 cglobal ps_stereo_interpolate, 5, 5, 6, l, r, h, h_step, n
121 ;***************************************************************************
122 ;void ps_stereo_interpolate_ipdopd_sse3(float (*l)[2], float (*r)[2],
123 ; float h[2][4], float h_step[2][4],
125 ;***************************************************************************
127 cglobal ps_stereo_interpolate_ipdopd, 5, 5, 10, l, r, h, h_step, n
129 movaps m1, [hq+mmsize]
132 movaps m9, [h_stepq+mmsize]
136 %define H_STEP0 [h_stepq]
137 %define H_STEP1 [h_stepq+mmsize]
150 shufps m4, m2, m2, q2301
151 shufps m5, m3, m3, q2301
169 ;**********************************************************
170 ;void ps_hybrid_analysis_ileave_sse(float out[2][38][64],
171 ; float (*in)[32][2],
173 ;**********************************************************
175 cglobal ps_hybrid_analysis_ileave, 3, 7, 5, out, in, i, len, in0, in1, tmp
201 movaps m2, [in0q+lenq]
202 movaps m3, [in1q+lenq]
203 TRANSPOSE4x4PS 0, 1, 2, 3, 4
205 movaps [outq+lenq], m1
206 movaps [outq+lenq*2], m2
207 movaps [outq+3*32*2*4], m3
208 lea in0q, [in0q+lenq*2]
209 lea in1q, [in1q+lenq*2]
229 movhps m0, [in0q+lenq]
230 movhps m1, [in1q+lenq]
234 movaps [outq+lenq], m1
235 lea in0q, [in0q+lenq*2]
236 lea in1q, [in1q+lenq*2]
256 movss m2, [in0q+lenq]
257 movss m3, [in1q+lenq]
262 lea in0q, [in0q+lenq*2]
263 lea in1q, [in1q+lenq*2]
275 ;***********************************************************
276 ;void ps_hybrid_synthesis_deint_sse4(float out[2][38][64],
277 ; float (*in)[32][2],
279 ;***********************************************************
280 %macro HYBRID_SYNTHESIS_DEINT 0
281 cglobal ps_hybrid_synthesis_deint, 3, 7, 5, out, in, i, len, out0, out1, tmp
289 lea outq, [outq+iq*4]
311 movaps m1, [inq+lenq]
312 movaps m2, [inq+lenq*2]
313 movaps m3, [inq+3*32*2*4]
314 TRANSPOSE4x4PS 0, 1, 2, 3, 4
317 movaps [out0q+lenq], m2
318 movaps [out1q+lenq], m3
319 lea out0q, [out0q+lenq*2]
320 lea out1q, [out1q+lenq*2]
339 movaps m1, [inq+lenq]
344 movhps [out0q+lenq], m0
345 movhps [out1q+lenq], m1
346 lea out0q, [out0q+lenq*2]
347 lea out1q, [out1q+lenq*2]
368 extractps [out1q], m0, 1
369 extractps [out0q+lenq], m0, 2
370 extractps [out1q+lenq], m0, 3
373 movss [out0q+lenq], m1
377 movss [out1q+lenq], m1
379 lea out0q, [out0q+lenq*2]
380 lea out1q, [out1q+lenq*2]
394 HYBRID_SYNTHESIS_DEINT
396 HYBRID_SYNTHESIS_DEINT
398 ;*******************************************************************
399 ;void ff_ps_hybrid_analysis_<opt>(float (*out)[2], float (*in)[2],
400 ; const float (*filter)[8][2],
401 ; ptrdiff_t stride, int n);
402 ;*******************************************************************
403 %macro PS_HYBRID_ANALYSIS_LOOP 3
404 movu %1, [inq+mmsize*%3]
405 movu m1, [inq+mmsize*(5-%3)+8]
410 pshufd m2, [filterq+nq+mmsize*%3], q2301
414 mova m2, [filterq+nq+mmsize*%3]
434 %macro PS_HYBRID_ANALYSIS 0
435 cglobal ps_hybrid_analysis, 5, 5, 8, out, in, filter, stride, n
445 mova m7, [ps_p1m1p1m1]
449 PS_HYBRID_ANALYSIS_LOOP m0, m3, 0
450 PS_HYBRID_ANALYSIS_LOOP m5, m6, 1
451 PS_HYBRID_ANALYSIS_LOOP m5, m6, 2
471 movu m2, [inq+6*8] ; faster than movlps and no risk of overread
473 movss m3, [filterq+nq+8*6]