interpolate8x8.c

这是一个压缩解压包,用C语言进行编程的,里面有详细的源代码.

				dst[j + 2] = (uint8_t)((src[j + 2] + src[j + stride + 2] )>>1);
				dst[j + 3] = (uint8_t)((src[j + 3] + src[j + stride + 3] )>>1);
				dst[j + 4] = (uint8_t)((src[j + 4] + src[j + stride + 4] )>>1);
				dst[j + 5] = (uint8_t)((src[j + 5] + src[j + stride + 5] )>>1);
				dst[j + 6] = (uint8_t)((src[j + 6] + src[j + stride + 6] )>>1);
				dst[j + 7] = (uint8_t)((src[j + 7] + src[j + stride + 7] )>>1);
		}
	} else {
		for (j = 0; j < 8*stride; j+=stride) {
				dst[j + 0] = (uint8_t)((src[j + 0] + src[j + stride + 0] + 1)>>1);
				dst[j + 1] = (uint8_t)((src[j + 1] + src[j + stride + 1] + 1)>>1);
				dst[j + 2] = (uint8_t)((src[j + 2] + src[j + stride + 2] + 1)>>1);
				dst[j + 3] = (uint8_t)((src[j + 3] + src[j + stride + 3] + 1)>>1);
				dst[j + 4] = (uint8_t)((src[j + 4] + src[j + stride + 4] + 1)>>1);
				dst[j + 5] = (uint8_t)((src[j + 5] + src[j + stride + 5] + 1)>>1);
				dst[j + 6] = (uint8_t)((src[j + 6] + src[j + stride + 6] + 1)>>1);
				dst[j + 7] = (uint8_t)((src[j + 7] + src[j + stride + 7] + 1)>>1);
		}
	}
}

/* dst = interpolate(src) */

void
interpolate8x4_halfpel_v_c(uint8_t * const dst,
						   const uint8_t * const src,
						   const uint32_t stride,
						   const uint32_t rounding)
{
	uintptr_t j;


	if (rounding) {
		for (j = 0; j < 4*stride; j+=stride) {
				dst[j + 0] = (uint8_t)((src[j + 0] + src[j + stride + 0] )>>1);
				dst[j + 1] = (uint8_t)((src[j + 1] + src[j + stride + 1] )>>1);
				dst[j + 2] = (uint8_t)((src[j + 2] + src[j + stride + 2] )>>1);
				dst[j + 3] = (uint8_t)((src[j + 3] + src[j + stride + 3] )>>1);
				dst[j + 4] = (uint8_t)((src[j + 4] + src[j + stride + 4] )>>1);
				dst[j + 5] = (uint8_t)((src[j + 5] + src[j + stride + 5] )>>1);
				dst[j + 6] = (uint8_t)((src[j + 6] + src[j + stride + 6] )>>1);
				dst[j + 7] = (uint8_t)((src[j + 7] + src[j + stride + 7] )>>1);
		}
	} else {
		for (j = 0; j < 4*stride; j+=stride) {
				dst[j + 0] = (uint8_t)((src[j + 0] + src[j + stride + 0] + 1)>>1);
				dst[j + 1] = (uint8_t)((src[j + 1] + src[j + stride + 1] + 1)>>1);
				dst[j + 2] = (uint8_t)((src[j + 2] + src[j + stride + 2] + 1)>>1);
				dst[j + 3] = (uint8_t)((src[j + 3] + src[j + stride + 3] + 1)>>1);
				dst[j + 4] = (uint8_t)((src[j + 4] + src[j + stride + 4] + 1)>>1);
				dst[j + 5] = (uint8_t)((src[j + 5] + src[j + stride + 5] + 1)>>1);
				dst[j + 6] = (uint8_t)((src[j + 6] + src[j + stride + 6] + 1)>>1);
				dst[j + 7] = (uint8_t)((src[j + 7] + src[j + stride + 7] + 1)>>1);
		}
	}
}

/* dst = (dst + interpolate(src))/2 */

void
interpolate8x8_halfpel_v_add_c(uint8_t * const dst,
						   const uint8_t * const src,
						   const uint32_t stride,
						   const uint32_t rounding)
{
	uintptr_t j;


	if (rounding) {
		for (j = 0; j < 8*stride; j+=stride) {
				dst[j + 0] = (uint8_t)((((src[j + 0] + src[j + stride + 0] )>>1) + dst[j+0] + 1)>>1);
				dst[j + 1] = (uint8_t)((((src[j + 1] + src[j + stride + 1] )>>1) + dst[j+1] + 1)>>1);
				dst[j + 2] = (uint8_t)((((src[j + 2] + src[j + stride + 2] )>>1) + dst[j+2] + 1)>>1);
				dst[j + 3] = (uint8_t)((((src[j + 3] + src[j + stride + 3] )>>1) + dst[j+3] + 1)>>1);
				dst[j + 4] = (uint8_t)((((src[j + 4] + src[j + stride + 4] )>>1) + dst[j+4] + 1)>>1);
				dst[j + 5] = (uint8_t)((((src[j + 5] + src[j + stride + 5] )>>1) + dst[j+5] + 1)>>1);
				dst[j + 6] = (uint8_t)((((src[j + 6] + src[j + stride + 6] )>>1) + dst[j+6] + 1)>>1);
				dst[j + 7] = (uint8_t)((((src[j + 7] + src[j + stride + 7] )>>1) + dst[j+7] + 1)>>1);
		}
	} else {
		for (j = 0; j < 8*stride; j+=stride) {
				dst[j + 0] = (uint8_t)((((src[j + 0] + src[j + stride + 0] + 1)>>1) + dst[j+0] + 1)>>1);
				dst[j + 1] = (uint8_t)((((src[j + 1] + src[j + stride + 1] + 1)>>1) + dst[j+1] + 1)>>1);
				dst[j + 2] = (uint8_t)((((src[j + 2] + src[j + stride + 2] + 1)>>1) + dst[j+2] + 1)>>1);
				dst[j + 3] = (uint8_t)((((src[j + 3] + src[j + stride + 3] + 1)>>1) + dst[j+3] + 1)>>1);
				dst[j + 4] = (uint8_t)((((src[j + 4] + src[j + stride + 4] + 1)>>1) + dst[j+4] + 1)>>1);
				dst[j + 5] = (uint8_t)((((src[j + 5] + src[j + stride + 5] + 1)>>1) + dst[j+5] + 1)>>1);
				dst[j + 6] = (uint8_t)((((src[j + 6] + src[j + stride + 6] + 1)>>1) + dst[j+6] + 1)>>1);
				dst[j + 7] = (uint8_t)((((src[j + 7] + src[j + stride + 7] + 1)>>1) + dst[j+7] + 1)>>1);
		}
	}
}

/* dst = interpolate(src) */

void
interpolate8x8_halfpel_hv_c(uint8_t * const dst,
							const uint8_t * const src,
							const uint32_t stride,
							const uint32_t rounding)
{
	uintptr_t j;

	if (rounding) {
		for (j = 0; j < 8*stride; j+=stride) {
				dst[j + 0] = (uint8_t)((src[j+0] + src[j+1] + src[j+stride+0] + src[j+stride+1] +1)>>2);
				dst[j + 1] = (uint8_t)((src[j+1] + src[j+2] + src[j+stride+1] + src[j+stride+2] +1)>>2);
				dst[j + 2] = (uint8_t)((src[j+2] + src[j+3] + src[j+stride+2] + src[j+stride+3] +1)>>2);
				dst[j + 3] = (uint8_t)((src[j+3] + src[j+4] + src[j+stride+3] + src[j+stride+4] +1)>>2);
				dst[j + 4] = (uint8_t)((src[j+4] + src[j+5] + src[j+stride+4] + src[j+stride+5] +1)>>2);
				dst[j + 5] = (uint8_t)((src[j+5] + src[j+6] + src[j+stride+5] + src[j+stride+6] +1)>>2);
				dst[j + 6] = (uint8_t)((src[j+6] + src[j+7] + src[j+stride+6] + src[j+stride+7] +1)>>2);
				dst[j + 7] = (uint8_t)((src[j+7] + src[j+8] + src[j+stride+7] + src[j+stride+8] +1)>>2);
		}
	} else {
		for (j = 0; j < 8*stride; j+=stride) {
				dst[j + 0] = (uint8_t)((src[j+0] + src[j+1] + src[j+stride+0] + src[j+stride+1] +2)>>2);
				dst[j + 1] = (uint8_t)((src[j+1] + src[j+2] + src[j+stride+1] + src[j+stride+2] +2)>>2);
				dst[j + 2] = (uint8_t)((src[j+2] + src[j+3] + src[j+stride+2] + src[j+stride+3] +2)>>2);
				dst[j + 3] = (uint8_t)((src[j+3] + src[j+4] + src[j+stride+3] + src[j+stride+4] +2)>>2);
				dst[j + 4] = (uint8_t)((src[j+4] + src[j+5] + src[j+stride+4] + src[j+stride+5] +2)>>2);
				dst[j + 5] = (uint8_t)((src[j+5] + src[j+6] + src[j+stride+5] + src[j+stride+6] +2)>>2);
				dst[j + 6] = (uint8_t)((src[j+6] + src[j+7] + src[j+stride+6] + src[j+stride+7] +2)>>2);
				dst[j + 7] = (uint8_t)((src[j+7] + src[j+8] + src[j+stride+7] + src[j+stride+8] +2)>>2);
		}
	}
}

/* dst = interpolate(src) */

void
interpolate8x4_halfpel_hv_c(uint8_t * const dst,
							const uint8_t * const src,
							const uint32_t stride,
							const uint32_t rounding)
{
	uintptr_t j;

	if (rounding) {
		for (j = 0; j < 4*stride; j+=stride) {
				dst[j + 0] = (uint8_t)((src[j+0] + src[j+1] + src[j+stride+0] + src[j+stride+1] +1)>>2);
				dst[j + 1] = (uint8_t)((src[j+1] + src[j+2] + src[j+stride+1] + src[j+stride+2] +1)>>2);
				dst[j + 2] = (uint8_t)((src[j+2] + src[j+3] + src[j+stride+2] + src[j+stride+3] +1)>>2);
				dst[j + 3] = (uint8_t)((src[j+3] + src[j+4] + src[j+stride+3] + src[j+stride+4] +1)>>2);
				dst[j + 4] = (uint8_t)((src[j+4] + src[j+5] + src[j+stride+4] + src[j+stride+5] +1)>>2);
				dst[j + 5] = (uint8_t)((src[j+5] + src[j+6] + src[j+stride+5] + src[j+stride+6] +1)>>2);
				dst[j + 6] = (uint8_t)((src[j+6] + src[j+7] + src[j+stride+6] + src[j+stride+7] +1)>>2);
				dst[j + 7] = (uint8_t)((src[j+7] + src[j+8] + src[j+stride+7] + src[j+stride+8] +1)>>2);
		}
	} else {
		for (j = 0; j < 4*stride; j+=stride) {
				dst[j + 0] = (uint8_t)((src[j+0] + src[j+1] + src[j+stride+0] + src[j+stride+1] +2)>>2);
				dst[j + 1] = (uint8_t)((src[j+1] + src[j+2] + src[j+stride+1] + src[j+stride+2] +2)>>2);
				dst[j + 2] = (uint8_t)((src[j+2] + src[j+3] + src[j+stride+2] + src[j+stride+3] +2)>>2);
				dst[j + 3] = (uint8_t)((src[j+3] + src[j+4] + src[j+stride+3] + src[j+stride+4] +2)>>2);
				dst[j + 4] = (uint8_t)((src[j+4] + src[j+5] + src[j+stride+4] + src[j+stride+5] +2)>>2);
				dst[j + 5] = (uint8_t)((src[j+5] + src[j+6] + src[j+stride+5] + src[j+stride+6] +2)>>2);
				dst[j + 6] = (uint8_t)((src[j+6] + src[j+7] + src[j+stride+6] + src[j+stride+7] +2)>>2);
				dst[j + 7] = (uint8_t)((src[j+7] + src[j+8] + src[j+stride+7] + src[j+stride+8] +2)>>2);
		}
	}
}

/* dst = (interpolate(src) + dst)/2 */

void
interpolate8x8_halfpel_hv_add_c(uint8_t * const dst,
							const uint8_t * const src,
							const uint32_t stride,
							const uint32_t rounding)
{
	uintptr_t j;

	if (rounding) {
		for (j = 0; j < 8*stride; j+=stride) {
				dst[j + 0] = (uint8_t)((((src[j+0] + src[j+1] + src[j+stride+0] + src[j+stride+1] +1)>>2) + dst[j+0])>>1);
				dst[j + 1] = (uint8_t)((((src[j+1] + src[j+2] + src[j+stride+1] + src[j+stride+2] +1)>>2) + dst[j+1])>>1);
				dst[j + 2] = (uint8_t)((((src[j+2] + src[j+3] + src[j+stride+2] + src[j+stride+3] +1)>>2) + dst[j+2])>>1);
				dst[j + 3] = (uint8_t)((((src[j+3] + src[j+4] + src[j+stride+3] + src[j+stride+4] +1)>>2) + dst[j+3])>>1);
				dst[j + 4] = (uint8_t)((((src[j+4] + src[j+5] + src[j+stride+4] + src[j+stride+5] +1)>>2) + dst[j+4])>>1);
				dst[j + 5] = (uint8_t)((((src[j+5] + src[j+6] + src[j+stride+5] + src[j+stride+6] +1)>>2) + dst[j+5])>>1);
				dst[j + 6] = (uint8_t)((((src[j+6] + src[j+7] + src[j+stride+6] + src[j+stride+7] +1)>>2) + dst[j+6])>>1);
				dst[j + 7] = (uint8_t)((((src[j+7] + src[j+8] + src[j+stride+7] + src[j+stride+8] +1)>>2) + dst[j+7])>>1);
		}
	} else {
		for (j = 0; j < 8*stride; j+=stride) {
				dst[j + 0] = (uint8_t)((((src[j+0] + src[j+1] + src[j+stride+0] + src[j+stride+1] +2)>>2) + dst[j+0] + 1)>>1);
				dst[j + 1] = (uint8_t)((((src[j+1] + src[j+2] + src[j+stride+1] + src[j+stride+2] +2)>>2) + dst[j+1] + 1)>>1);
				dst[j + 2] = (uint8_t)((((src[j+2] + src[j+3] + src[j+stride+2] + src[j+stride+3] +2)>>2) + dst[j+2] + 1)>>1);
				dst[j + 3] = (uint8_t)((((src[j+3] + src[j+4] + src[j+stride+3] + src[j+stride+4] +2)>>2) + dst[j+3] + 1)>>1);
				dst[j + 4] = (uint8_t)((((src[j+4] + src[j+5] + src[j+stride+4] + src[j+stride+5] +2)>>2) + dst[j+4] + 1)>>1);
				dst[j + 5] = (uint8_t)((((src[j+5] + src[j+6] + src[j+stride+5] + src[j+stride+6] +2)>>2) + dst[j+5] + 1)>>1);
				dst[j + 6] = (uint8_t)((((src[j+6] + src[j+7] + src[j+stride+6] + src[j+stride+7] +2)>>2) + dst[j+6] + 1)>>1);
				dst[j + 7] = (uint8_t)((((src[j+7] + src[j+8] + src[j+stride+7] + src[j+stride+8] +2)>>2) + dst[j+7] + 1)>>1);
		}
	}
}

/*************************************************************
 * QPEL STUFF STARTS HERE                                    *
 *************************************************************/

void interpolate8x8_6tap_lowpass_h_c(uint8_t *dst, uint8_t *src, int32_t stride, int32_t rounding)
{
    int32_t i;
	uint8_t round_add = 16 - rounding;

    for(i = 0; i < 8; i++)
    {

        dst[0] = CLIP((((src[-2] + src[3]) + 5 * (((src[0] + src[1])<<2) - (src[-1] + src[2])) + round_add) >> 5), 0, 255);
        dst[1] = CLIP((((src[-1] + src[4]) + 5 * (((src[1] + src[2])<<2) - (src[0] + src[3])) + round_add) >> 5), 0, 255);
        dst[2] = CLIP((((src[0] + src[5]) + 5 * (((src[2] + src[3])<<2) - (src[1] + src[4])) + round_add) >> 5), 0, 255);
        dst[3] = CLIP((((src[1] + src[6]) + 5 * (((src[3] + src[4])<<2) - (src[2] + src[5])) + round_add) >> 5), 0, 255);
        dst[4] = CLIP((((src[2] + src[7]) + 5 * (((src[4] + src[5])<<2) - (src[3] + src[6])) + round_add) >> 5), 0, 255);
        dst[5] = CLIP((((src[3] + src[8]) + 5 * (((src[5] + src[6])<<2) - (src[4] + src[7])) + round_add) >> 5), 0, 255);
        dst[6] = CLIP((((src[4] + src[9]) + 5 * (((src[6] + src[7])<<2) - (src[5] + src[8])) + round_add) >> 5), 0, 255);
        dst[7] = CLIP((((src[5] + src[10]) + 5 * (((src[7] + src[8])<<2) - (src[6] + src[9])) + round_add) >> 5), 0, 255);

        dst += stride;
        src += stride;
    }
}

void interpolate16x16_lowpass_h_c(uint8_t *dst, uint8_t *src, int32_t stride, int32_t rounding)
{
    int32_t i;
	uint8_t round_add = 16 - rounding;