ffmpeg/libavcodec/x86/simple_idct10.asm
Christophe Gisquet 7ece8b50b1 x86: simple_idct: 12bits versions
On 12 frames of a 444p 12 bits DNxHR sequence, _put function:
C:         78902 decicycles in idct,  262071 runs,     73 skips
avx:       32478 decicycles in idct,  262045 runs,     99 skips

Difference between the 2:
stddev:    0.39 PSNR:104.47 MAXDIFF:    2

This is unavoidable and due to the scale factors used in the x86
version, which cannot match the C ones.

In addition, the trick of adding an initial bias to the input of a
pass can overflow, as the input coefficients are already 15bits,
which is the maximum this function can handle.

Overall, however, the omse on 12 bits samples goes from 0.16916 to
0.16883. Reducing rowshift by 1 improves to 0.0908, but causes
overflows.

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
2015-10-13 15:34:32 +02:00

75 lines
2.1 KiB
NASM

;******************************************************************************
;* x86-SIMD-optimized IDCT for prores
;* this is identical to "simple" IDCT written by Michael Niedermayer
;* except for the clip range
;*
;* Copyright (c) 2011 Ronald S. Bultje <rsbultje@gmail.com>
;* Copyright (c) 2015 Christophe Gisquet
;*
;* 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
;* 51, Inc., Foundation Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
;******************************************************************************
%include "libavutil/x86/x86util.asm"
%if ARCH_X86_64
SECTION_RODATA
cextern pw_2
cextern pw_16
cextern pw_1023
cextern pw_4095
pd_round_12: times 4 dd 1<<(12-1)
pd_round_15: times 4 dd 1<<(15-1)
pd_round_19: times 4 dd 1<<(19-1)
%include "libavcodec/x86/simple_idct10_template.asm"
section .text align=16
%macro idct_fn 0
cglobal simple_idct10, 1, 1, 16
IDCT_FN "", 12, "", 19
RET
cglobal simple_idct10_put, 3, 3, 16
IDCT_FN "", 12, "", 19, 0, pw_1023
RET
cglobal simple_idct12, 1, 1, 16
; coeffs are already 15bits, adding the offset would cause
; overflow in the input
IDCT_FN "", 15, pw_2, 16
RET
cglobal simple_idct12_put, 3, 3, 16
; range isn't known, so the C simple_idct range is used
; Also, using a bias on input overflows, so use the bias
; on output of the first butterfly instead
IDCT_FN "", 15, pw_2, 16, 0, pw_4095
RET
%endmacro
INIT_XMM sse2
idct_fn
%if HAVE_AVX_EXTERNAL
INIT_XMM avx
idct_fn
%endif
%endif