colorspace_rgb_mmx.asm

从FFMPEG转换而来的H264解码程序,VC下编译..

;/*****************************************************************************
; *
; *  XVID MPEG-4 VIDEO CODEC
; *  - RGB colorspace conversions -
; *
; *  Copyright(C) 2002-2003 Michael Militzer <[e-mail]>
; *               2002-2003 Peter Ross <[e-mail]>
; *
; *  This program is free software ; you can redistribute it and/or modify
; *  it under the terms of the GNU General Public License as published by
; *  the Free Software Foundation ; either version 2 of the License, or
; *  (at your option) any later version.
; *
; *  This program 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 General Public License for more details.
; *
; *  You should have received a copy of the GNU General Public License
; *  along with this program ; if not, write to the Free Software
; *  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307 USA
; *
; ****************************************************************************/

BITS 32

%macro cglobal 1
	%ifdef PREFIX
		%ifdef MARK_FUNCS
			global _%1:function %1.endfunc-%1
			%define %1 _%1:function %1.endfunc-%1
		%else
			global _%1
			%define %1 _%1
		%endif
	%else
		%ifdef MARK_FUNCS
			global %1:function %1.endfunc-%1
		%else
			global %1
		%endif
	%endif
%endmacro

;=============================================================================
; Some constants
;=============================================================================

;-----------------------------------------------------------------------------
; RGB->YV12 yuv constants
;-----------------------------------------------------------------------------

%define Y_R		0.257
%define Y_G		0.504
%define Y_B		0.098
%define Y_ADD	16

%define U_R		0.148
%define U_G		0.291
%define U_B		0.439
%define U_ADD	128

%define V_R		0.439
%define V_G		0.368
%define V_B		0.071
%define V_ADD	128

; Scaling used during conversion
%define SCALEBITS 6

;=============================================================================
; Read only data
;=============================================================================

%ifdef FORMAT_COFF
SECTION .rodata
%else
SECTION .rodata align=16
%endif
ALIGN 16

;-----------------------------------------------------------------------------
; RGB->YV12 multiplication matrices
;-----------------------------------------------------------------------------
;         FIX(Y_B)	FIX(Y_G)	FIX(Y_R) Ignored

bgr_y_mul: dw    25,      129,        66,      0
bgr_u_mul: dw   112,      -74,       -38,      0
bgr_v_mul: dw   -18,      -94,       112,      0

;-----------------------------------------------------------------------------
; BGR->YV12 multiplication matrices
;-----------------------------------------------------------------------------
;         FIX(Y_R)	FIX(Y_G)	FIX(Y_B) Ignored

rgb_y_mul: dw    66,      129,        25,      0
rgb_u_mul: dw   -38,      -74,       112,      0
rgb_v_mul: dw   112,      -94,       -18,      0

;-----------------------------------------------------------------------------
; YV12->RGB data
;-----------------------------------------------------------------------------

Y_SUB: dw  16,  16,  16,  16
U_SUB: dw 128, 128, 128, 128
V_SUB: dw 128, 128, 128, 128

Y_MUL: dw  74,  74,  74,  74

UG_MUL: dw  25,  25,  25,  25
VG_MUL: dw  52,  52,  52,  52

UB_MUL: dw 129, 129, 129, 129
VR_MUL: dw 102, 102, 102, 102

BRIGHT: db 128, 128, 128, 128, 128, 128, 128, 128

;=============================================================================
; Helper macros used with the colorspace_mmx.inc file
;=============================================================================

;------------------------------------------------------------------------------
; BGR_TO_YV12( BYTES )
;
; BYTES		3=bgr(24bit), 4=bgra(32-bit)
;
; bytes=3/4, pixels = 2, vpixels=2
;------------------------------------------------------------------------------

%macro BGR_TO_YV12_INIT		2
  movq mm7, [bgr_y_mul]
%endmacro


%macro BGR_TO_YV12			2
    ; y_out
  pxor mm4, mm4
  pxor mm5, mm5
  movd mm0, [edi]               ; x_ptr[0...]
  movd mm2, [edi+edx]           ; x_ptr[x_stride...]
  punpcklbw mm0, mm4            ; [  |b |g |r ]
  punpcklbw mm2, mm5            ; [  |b |g |r ]
  movq mm6, mm0                 ; = [  |b4|g4|r4]
  paddw mm6, mm2                ; +[  |b4|g4|r4]
  pmaddwd mm0, mm7              ; *= Y_MUL
  pmaddwd mm2, mm7              ; *= Y_MUL
  movq mm4, mm0                 ; [r]
  movq mm5, mm2                 ; [r]
  psrlq mm4, 32                 ; +[g]
  psrlq mm5, 32                 ; +[g]
  paddd mm0, mm4                ; +[b]
  paddd mm2, mm5                ; +[b]

  pxor mm4, mm4
  pxor mm5, mm5
  movd mm1, [edi+%1]            ; src[%1...]
  movd mm3, [edi+edx+%1]        ; src[x_stride+%1...]
  punpcklbw mm1, mm4            ; [  |b |g |r ]
  punpcklbw mm3, mm5            ; [  |b |g |r ]
  paddw mm6, mm1                ; +[  |b4|g4|r4]
  paddw mm6, mm3                ; +[  |b4|g4|r4]
  pmaddwd mm1, mm7              ; *= Y_MUL
  pmaddwd mm3, mm7              ; *= Y_MUL
  movq mm4, mm1                 ; [r]
  movq mm5, mm3                 ; [r]
  psrlq mm4, 32                 ; +[g]
  psrlq mm5, 32                 ; +[g]
  paddd mm1, mm4                ; +[b]
  paddd mm3, mm5                ; +[b]

  push edx

  movd edx, mm0
  shr edx, 8
  add edx, Y_ADD
  mov [esi], dl                 ; y_ptr[0]

  movd edx, mm1
  shr edx, 8
  add edx, Y_ADD
  mov [esi + 1], dl             ; y_ptr[1]

  movd edx, mm2
  shr edx, 8
  add edx, Y_ADD
  mov [esi + eax + 0], dl       ; y_ptr[y_stride + 0]

  movd edx, mm3
  shr edx, 8
  add edx, Y_ADD
  mov [esi + eax + 1], dl       ; y_ptr[y_stride + 1]

  ; u_ptr, v_ptr
  movq mm0, mm6                 ; = [  |b4|g4|r4]
  pmaddwd mm6, [bgr_v_mul]          ; *= V_MUL
  pmaddwd mm0, [bgr_u_mul]          ; *= U_MUL
  movq mm1, mm0
  movq mm2, mm6
  psrlq mm1, 32
  psrlq mm2, 32
  paddd mm0, mm1
  paddd mm2, mm6

  movd edx, mm0
  shr edx, 10
  add edx, U_ADD
  mov [ebx], dl

  movd edx, mm2
  shr edx, 10
  add edx, V_ADD
  mov [ecx], dl

  pop edx
%endmacro

;------------------------------------------------------------------------------
; RGB_TO_YV12( BYTES )
;
; BYTES		3=rgb(24bit), 4=rgba(32-bit)
;
; bytes=3/4, pixels = 2, vpixels=2
;------------------------------------------------------------------------------

%macro RGB_TO_YV12_INIT		2
  movq mm7, [rgb_y_mul]
%endmacro


%macro RGB_TO_YV12			2
    ; y_out
  pxor mm4, mm4
  pxor mm5, mm5
  movd mm0, [edi]               ; x_ptr[0...]
  movd mm2, [edi+edx]           ; x_ptr[x_stride...]
  punpcklbw mm0, mm4            ; [  |b |g |r ]
  punpcklbw mm2, mm5            ; [  |b |g |r ]
  movq mm6, mm0                 ; = [  |b4|g4|r4]
  paddw mm6, mm2                ; +[  |b4|g4|r4]
  pmaddwd mm0, mm7              ; *= Y_MUL
  pmaddwd mm2, mm7              ; *= Y_MUL
  movq mm4, mm0                 ; [r]
  movq mm5, mm2                 ; [r]
  psrlq mm4, 32                 ; +[g]
  psrlq mm5, 32                 ; +[g]
  paddd mm0, mm4                ; +[b]
  paddd mm2, mm5                ; +[b]

  pxor mm4, mm4
  pxor mm5, mm5
  movd mm1, [edi+%1]            ; src[%1...]
  movd mm3, [edi+edx+%1]        ; src[x_stride+%1...]
  punpcklbw mm1, mm4            ; [  |b |g |r ]
  punpcklbw mm3, mm5            ; [  |b |g |r ]
  paddw mm6, mm1                ; +[  |b4|g4|r4]
  paddw mm6, mm3                ; +[  |b4|g4|r4]
  pmaddwd mm1, mm7              ; *= Y_MUL
  pmaddwd mm3, mm7              ; *= Y_MUL
  movq mm4, mm1                 ; [r]
  movq mm5, mm3                 ; [r]
  psrlq mm4, 32                 ; +[g]
  psrlq mm5, 32                 ; +[g]
  paddd mm1, mm4                ; +[b]
  paddd mm3, mm5                ; +[b]

  push edx

  movd edx, mm0
  shr edx, 8