[FFmpeg-devel] [PATCH 2/2] Add hflip filter.

[Frank Barchard]

On Wed, Aug 4, 2010 at 5:23 AM, Michael Niedermayer <michaelni at gmx.at>wrote:

> variable length memcpy on a per pixel base is slow
>

Ya, thats what I was thinking.  It supports many pixel formats, but wouldnt
be fast enough for realtime.

Check the pixel size and write a function that deals with that size, for at
least the common cases.

The ideal is something like bswap or pshufb.  pshuf can handle byte, word or
dword, but not 3 bytes (RGB) naturally... you'd have to unroll more.
in libavcodec/dsputil.c theres an endian swapping function that is close to
what you need

static void bswap_buf(uint32_t *dst, const uint32_t *src, int w){    int i;
    for(i=0; i+8<=w; i+=8){        dst[i+0]= av_bswap32(src[i-0]);
   dst[i+1]= av_bswap32(src[i-1]);        dst[i+2]=
av_bswap32(src[i-2]);        dst[i+3]= av_bswap32(src[i-3]);
dst[i+4]= av_bswap32(src[i-4]);        dst[i+5]= av_bswap32(src[i-5]);
       dst[i+6]= av_bswap32(src[i-6]);        dst[i+7]=
av_bswap32(src[i-7]);    }    for(;i<w; i++){        dst[i+0]=
av_bswap32(src[i-0]);    }}

Then figure out how to do something like bswap but for other pixel types.

Looking at pixfmt.h your current code could handle more of the formats.