yuv2rgb.c

PocketMVP V0.8082503 source for Pocket 的程序源代码

 
  "add $9,$10,$12;" 
  "sra $8,$9,31;" 
  "srlv $9,$9,$8;"); 
 
__asm( "andi $9,$9,0x00F8;" 
  "lbu $10,1($4);"//cargamos aqui para no tener interbloqueo 
 
  "sll $9,$9,8;" 
  "or  $24,$24,$9;" 
 
  //primer punto ,ahora procesamos segundo. 
 
  //r2 
 
  "add $9,$10,$11;" 
  "sra $8,$9,31;" 
  "srlv $9,$9,$8;"
  "andi $9,$9,0x00F8;"
  "srl $19,$9,3;" 
 
  //g2 
 
  "addi $9,$10,-16;" 
  "sra $8,$9,31;" 
  "srlv $9,$9,$8;" 
  "andi $9,$9,0x00FC;" 
  "sll $9,$9,3;" 
  "or  $19,$19,$9;" 
 
  //b2 
 
  "add $9,$10,$12;" 
  "lbu $10,0($15);"//cargamos aqui para no tener interbloqueo 
  "sra $8,$9,31;" 
  "srlv $9,$9,$8;" 
  "andi $9,$9,0x00F8;" 
  "sll $9,$9,8;" 
  "or  $19,$19,$9;" 
 
  //presentamos el punto 
 
); 
 
__asm(//ahora procesamos la segunda l韓ea 
 
  //f2r1 
 
  "add $9,$10,$11;" 
  "sra $8,$9,31;" 
  "srlv $9,$9,$8;" 
  "andi $9,$9,0x00F8;" 
  "sll $9,$9,13;" 
  "or  $24,$24,$9;" 
 
  //f2g1 
 
  "addi $9,$10,-16;" 
  "sra $8,$9,31;" 
  "srlv $9,$9,$8;" 
  "andi $9,$9,0x00FC;" 
  "sll $9,$9,19;" 
  "or  $24,$24,$9;" 

  //f2b1 
 
  "add $9,$10,$12;" 
  "sra $8,$9,31;" 
  "srlv $9,$9,$8;" 
  "andi $9,$9,0x00F8;" 
  "lbu $10,1($15);"//cargamos aqui para no tener interbloqueo 
 
  "sll $9,$9,24;" 
  "or  $24,$24,$9;" 
); 
 
  //primer punto ,ahora procesamos segundo. 
 
  //f2r2 
 
__asm( "add $9,$10,$11;" 
  "lbu $11,1($6);" 
  "sra $8,$9,31;" 
  "srlv $9,$9,$8;"
  "andi $9,$9,0x00F8;"
  "sll $9,$9,13;" 
  "or  $19,$19,$9;" 
 
  //f2g2 
 
  "addi $9,$10,-16;" 
  "sra $8,$9,31;" 
  "srlv $9,$9,$8;" 
  "andi $9,$9,0x00FC;" 
  "sll $9,$9,19;" 
  "or  $19,$19,$9;" 
 
  //f2b2 
 
  "add $9,$10,$12;" 
  "lbu $12,1($7);" 
  "sra $8,$9,31;" 
  "srlv $9,$9,$8;" 
  "andi $9,$9,0x00F8;" 
 
  "sll $9,$9,24;" 
  "lbu $10,2($4);" 
  "or  $19,$19,$9;" 
 
  //presentamos el punto 
 
  "sw $24,0($25);"
  "sub $8,$25,$17;" 
  "sw $19,0($8);" 
);
 
  //incrementamos todos los punteros y volvemos al bucle 
__asm(
  "addi $4,$4,+2;" 
  "addi $15,$15,+2;" 
  "addi $6,$6,+1;" 
  "addi $7,$7,+1;" 
  "addi $3,$3,-1;" 
  "sub $25,$25,$17;" 
  "sub $25,$25,$17;" 

  "bgtz $3,bucle_xR;" 
 
  "add $4,$15,$16;" 
 
  "addu $6,$6,$14;" 
  "addu $7,$7,$14;" 
 
  //cargamos aqui para no tener interbloqueo 
  "lbu $11,0($6);" 
  "lbu $10,0($4);" 
  "lbu $12,0($7);" 
 
  "mul $8,$13,$17;"
  "addu $25,$25,$8;" //puntero_video+desplazamiento 
  "addiu $25,$25,+4;"

  "addi $2,$2,-1;" 

  "bgtz $2,bucle_yR;" 
 
  //restauramos registros, la pila y finalizamos 
  "lw  $16,12(sp);" 
  "lw  $17,16(sp);" 
  "lw  $18,20(sp);" 
  "lw $19,24(sp);"

  "addiu sp,sp,+36;" 
 
);} 

 
void yuv2rgb_565Z(uint8_t *puc_y, int stride_y, 
uint8_t *puc_u, uint8_t *puc_v, int stride_uv, 
uint8_t *puc_out, int width_y, int height_y, int stride_dest, int Dither) 
{ 
// yuv2rgb 4xzoom.32bits 
// v0.2->some optimizations+horizontal crop 
// v0.3->precise bilinear+some optimizations
// v0.4a->removed precise filtering+save word opt.
// v0.5a->fixed big bug. Precise bilinear only on VR4122.
 
__asm ( "addiu sp,sp,-8;" 
  "sw  $16,0(sp);" 
  "sw  $17,4(sp);"
 
  //cargamos todas las variables 
 
  "lw  $9,40(sp);" 
  "lw  $2,36(sp);" 
  "lw  $3,32(sp);" 
  "lw  $25,28(sp);" 
 
  //recortamos video x_crop
  
  "addi $14,$3,-120;" 
  "blez $14,no_recorte_x;" 
  
  "li $3,120;" 
  "srl	$14,$14,1;"
  "add	$4,$4,$14;"
  "srl	$14,$14,1;"
  "add	$6,$6,$14;"
  "add	$7,$7,$14;"

  "no_recorte_x:"
  
  //recortamos video y_crop
  
  "addi $14,$2,-160;" 
  "blez $14,no_recorte_y;" 
  
  "li	$2,160;" 
  "srl	$14,$14,1;"
  "mul	$14,$14,$5;"
  "add	$4,$4,$14;"
  "srl	$14,$14,2;"
  "add	$6,$6,$14;"
  "add	$7,$7,$14;"
    
  //diferencias de strides: 
  
  "no_recorte_y:srl $8,$5,1;" 
  "sub $5,$5,$3;" 
  
  //
 
  "sll $9,$9,1;" 
  
  //preparando segunda linea
 
  "sll $14,$3,1;"
  "add $17,$25,$9;"
  "add $17,$17,$14;"
 
  //
 
  "bucle_y1:lbu $12,0($7);" 
  "lbu $11,0($6);" 
  "lbu $10,0($4);" 
  "add $16,$3,$0;" 
  "addi $11,$11,-128;" 
  "addi $12,$12,-128;"
  "addi $10,$10,-16;"
  
  //r1 
  
  "add $14,$10,$11;" 
  "sra $15,$14,31;"
  "srlv $15,$14,$15;"
  "srl $13,$15,3;" 
  
  //g1 
  
  "sra $15,$14,31;"
  "srlv $15,$14,$15;"
  "andi $14,$15,0x00FC;" 
  "sll $14,$14,3;" 
  "or  $13,$13,$14;" 
  
  //b1 
  
  "add $14,$10,$12;" 
  "sra $15,$14,31;"
  "srlv $15,$14,$15;"
  "andi $14,$15,0x00F8;" 
  "sll $14,$14,8;" 
  "or  $13,$13,$14;" 
  
  "lbu $10,1($4);" 
  "addi $4,$4,+1;" 
//  "addi $16,$16,-1;"
  "addi $10,$10,-16;"
  ); 
 
__asm( "bucle_x1:add $14,$10,$11;"  
  "sra $15,$14,31;"
  "srlv $24,$14,$15;"
  "srl $24,$24,3;" 
 
  //g3 
 
  "sra $15,$10,31;" 
  "srlv $15,$10,$15;"
  "andi $15,$15,0x00FC;" 
  "sll $14,$15,3;" 
  "or  $24,$24,$14;" 
 
  //b3 
 
  "add $14,$10,$12;" 
  "sra $15,$14,31;"
  "srlv $15,$14,$15;"
  "andi $15,$15,0x00F8;" 
  "sll $14,$15,8;" 
  "or  $24,$24,$14;" 
#if (_WIN32_WCE < 300) 
  //calculamos el punto interpolado 
 
  "srl $14,$13,1;"  //div por 2 p1 
  "srl $15,$24,1;"  //y p2 
 
  "andi $14,$14,0x7BEF;" //los enmascaramos 
  "andi $15,$15,0x7BEF;" 
 
  "addu $13,$14,$15;"  //y los sumamos 
 
#else
  //calculamos el punto interpolado 
 
  "andi $10,$13,0x0821;"//
  "and $10,$10,$24;"//calcula un valor medio perfecto
 
  "srl $14,$13,1;"//div por 2 p1 
  "srl $15,$24,1;"//y p2 
 
  "andi $14,$14,0x7BEF;"//los enmascaramos 
  "andi $15,$15,0x7BEF;" 
 
  "addu $13,$14,$15;"//y los sumamos 
  "addu $13,$13,$10;"
#endif  
  //guardamos puntos 
    
  "sll $15,$24,16;"
  "or  $13,$13,$15;"
 
  "sw  $13,0($25);"
  "sw  $13,0($17);"
     
  //cambiamos de lugar ultimo punto 
  
  "add $13,$24,$0;" 
  
  //incrementamos punteros 
  
  "lbu $10,0($4);"  //cargamos nuevo puc_y 
  "addi $4,$4,+1;" 
  "addi $16,$16,-1;" 
  "sll $14,$16,31;" 
  "addi $25,$25,+4;" 
  "addi $17,$17,+4;"
 
  "addi $10,$10,-16;" //*
     
  "beq $14,$0,par_x;" 
  "addi $6,$6,+1;" 
  "addi $7,$7,+1;" 
  "lbu $12,0($7);" 
  "lbu $11,0($6);" 
  "addi $12,$12,-128;" 
  "addi $11,$11,-128;"
  "par_x:bgtz $16,bucle_x1;"); 
 
__asm( //aumentamos strides: 
  "sll $14,$2,31;" 
  "add $4,$4,$5;" 
  "addi $4,$4,-1;"
  "srl $15,$3,1;" 
  "sub $6,$6,$15;" 
  "sub $7,$7,$15;" 
    
  "beq $14,$0,par_y;" 
  "add $6,$6,$8;" 
  "add $7,$7,$8;" 
 
  //aumentamos stride_dest 
 
  "par_y:addi  $2,$2,-1;" 
  
  "sll $14,$3,1;"
  "sub $17,$17,$14;"
  "add $25,$17,$9;"
 
  "add $17,$25,$9;"
  "add $17,$17,$14;"
  
  "bgtz $2,bucle_y1;" 
 
  "lw  $16,0(sp);" 
  "lw  $17,4(sp);"
 
  "addiu sp,sp,+8;" ); 
} 
 
 
 

void yuv2rgb_565ZPP(uint8_t *puc_y, int stride_y, 
uint8_t *puc_u, uint8_t *puc_v, int stride_uv, 
uint8_t *puc_out, int width_y, int height_y, int stride_dest, int Dither) 
{ 
//yuv2rgb 4xzoom BILINEAR FILTERING. 32bit access 
// v0.1=>base code of 0.2 linear zoom asm +bilinear filtering 
// v0.2=>finally 32bits code
 
__asm ( "addiu sp,sp,-36;" 
  "sw  $16,0(sp);" 
  "sw  $17,4(sp);" 
  "sw  $18,8(sp);" 
  "sw  $19,12(sp);" 
  "sw  $20,16(sp);" 
  "sw  $21,20(sp);"
 
  //cargamos todas las variables 
 
  "lw  $9,68(sp);" 
  "lw  $2,64(sp);" 
  "lw  $3,60(sp);" 
  "lw  $25,56(sp);" 
 
  //recortamos video x_crop
  
  "addi $14,$3,-120;" 
  "blez $14,no_recorte_xzpp;" 
  
  "li $3,120;" 
  "srl	$14,$14,1;"
  "add	$4,$4,$14;"
  "srl	$14,$14,1;"
  "add	$6,$6,$14;"
  "add	$7,$7,$14;"

  "no_recorte_xzpp:"
  
  //recortamos video y_crop
  
  "addi $14,$2,-160;" 
  "blez $14,no_recorte_yzpp;" 
  
  "li	$2,160;" 
  "srl	$14,$14,1;"
  "mul	$14,$14,$5;"
  "add	$4,$4,$14;"
  "srl	$14,$14,2;"
  "add	$6,$6,$14;"
  "add	$7,$7,$14;"
 
  //diferencias de strides: 
 
  "no_recorte_yzpp:srl $8,$5,1;" 
 
  "add $17,$5,$4;"//y tomamos el primer puntero 
 
  "sub $5,$5,$3;" 
  
  //
 
  "sll $9,$9,1;" 
  
  //preparando segunda linea
 
  "sll $14,$3,1;"
  "add $21,$25,$9;"
  "add $21,$21,$14;"
   
  //
 
  "bucle_y2:lbu $12,0($7);" 
  "lbu $11,0($6);" 
  "lbu $10,0($4);" 
  "lbu $18,0($17);"//tomamos puc_y de la siguiente linea 
  "add $16,$3,$0;" 
  "addi $11,$11,-144;" 
  "addi $12,$12,-144;"); 
 
  //r1 
 
__asm( "add $14,$10,$11;" 
  "sra $15,$14,31;"
  "srlv $15,$14,$15;"
  "srl $13,$15,3;" 
  
  //g1 
 
  "addi $14,$10,-16;" 
  "sra $15,$14,31;"
  "srlv $15,$14,$15;"
  "andi $14,$15,0x00FC;" 
  "sll $14,$14,3;" 
  "or  $13,$13,$14;" 
 
  //b1 
  
  "add $14,$10,$12;" 
  "sra $15,$14,31;"
  "srlv $15,$14,$15;"
  "andi $14,$15,0x00F8;" 
  "sll $14,$14,8;" 
  "or  $13,$13,$14;" 
 
  //r2 
 
  "srl $10,$10,1;" 
  "srl $18,$18,1;" 
  
  "add $18,$18,$10;" 
 
  "add $14,$18,$11;" 
  "sra $15,$14,31;"
  "srlv $15,$14,$15;"
  "srl $19,$15,3;" 
 
  //g2 
 
  "addi $14,$18,-16;" 
  "sra $15,$14,31;"
  "srlv $15,$14,$15;" 
  "andi $14,$15,0x00FC;" 
  "sll $14,$14,3;" 
  "or  $19,$19,$14;" 
 
  //b2
  
  "add $14,$18,$12;" 
  "sra $15,$14,31;"
  "srlv $15,$14,$15;" 
  "andi $14,$15,0x00F8;" 
  "sll $14,$14,8;" 
  "or  $19,$19,$14;"