[FFmpeg-devel] [PATCH] V210 decoder and encoder

[Ramiro Polla]

On Mon, May 11, 2009 at 6:41 PM, Baptiste Coudurier
<baptiste.coudurier at gmail.com> wrote:
> On 5/11/2009 1:59 PM, Reimar D?ffinger wrote:
>> On Mon, May 11, 2009 at 12:35:49PM -0700, Baptiste Coudurier wrote:
>>> On 5/11/2009 12:27 PM, Reimar D?ffinger wrote:
>>>>> + ? ? ? ? ? ?*udst++ = (v & 0x3FF) ? ? ?<< ?6;
>>>>> + ? ? ? ? ? ?*ydst++ = (v & 0xFFC00) ? ?>> ?4;
>>>>> + ? ? ? ? ? ?*vdst++ = (v & 0x3FF00000) >> 14;
>>>> Both
>>>>
>>>>> + ? ? ? ? ? ?*udst++ = (v & 0x000003FF) << ?6;
>>>>> + ? ? ? ? ? ?*ydst++ = (v & 0x000FFC00) >> ?4;
>>>>> + ? ? ? ? ? ?*vdst++ = (v & 0x3FF00000) >> 14;
>>>> and
>>>>
>>>>> + ? ? ? ? ? ?*udst++ = (v << ?6) & 0xFFC0;
>>>>> + ? ? ? ? ? ?*ydst++ = (v >> ?4) & 0xFFC0;
>>>>> + ? ? ? ? ? ?*vdst++ = (v >> 14) & 0xFFC0;
>>>> Seem nicer to me.
>>> I don't get what you mean.
>>
>> That I consider both alternatives more readable.
>>
>>>> I think the later one might have a speed advantage due to needing only one
>>>> constant.
>>>> Also like in the encoder you don't really need one of the ands.
>>> You mean ydst >> 14 ? What if the 2 bits are not zero ? Doesn't the bit
>>> gets replicated ?
>>
>> No, the << 6 one, Upper bits are dropped because udst is only 16 bits, lower
>> bits 0 is shifted in. Thus the & is pointless.
>
> I see what you mean now.
>
> Update attached.

> +            v= le2me_32(*src++);
> +            *udst++ =  v <<  6;
> +            *ydst++ = (v >>  4) & 0xFFC0;
> +            *vdst++ = (v >> 14) & 0xFFC0;
> +
> +            v= le2me_32(*src++);
> +            *ydst++ =  v <<  6;
> +            *udst++ = (v >>  4) & 0xFFC0;
> +            *ydst++ = (v >> 14) & 0xFFC0;
> +
> +            v= le2me_32(*src++);
> +            *vdst++ =  v <<  6;
> +            *ydst++ = (v >>  4) & 0xFFC0;
> +            *udst++ = (v >> 14) & 0xFFC0;
> +
> +            v= le2me_32(*src++);
> +            *ydst++ =  v <<  6;
> +            *vdst++ = (v >>  4) & 0xFFC0;
> +            *ydst++ = (v >> 14) & 0xFFC0;

Maybe:
#define FOOBAR(a, b, c) \
[...]

FOOBAR(u, v, y)
FOOBAR(y, u, y)
FOOBAR(v, y, u)
FOOBAR(y, v, y)

And there are some more further down the patch.

Ramiro Polla