[FFmpeg-devel] [PATCH] Long-term prediction for ALS

[Thilo Borgmann]

Thilo Borgmann schrieb:
> Ronald S. Bultje schrieb:
>> Hi,
>>
>> 2009/11/12 M?ns Rullg?rd <mans at mansr.com>:
>>> "Ronald S. Bultje" <rsbultje at gmail.com> writes:
>>>> int idx = get_unary(gb, 0, 4) << 2 + get_bits(gb, 2);
>>>> ltp_gain[2] = ltp_gain_values[idx];
>>>>
>>>> Or even:
>>>>
>>>> int a = get_unary(gb, 0, 4), b = get_bits(gb, 2);
>>>> ltp_gain[2] = ltp_gain_values[a][b];
>>> Is that a 1D or 2D array.  Please try to index arrays as they are
>>> declared, not in some other random fashion.  It *DOES* break
>>> sometimes.
>> It's declared in this very patch, he could easily make it 2D instead
>> of 1D (I didn't say it explicitely, but obviously the intention is for
>> him to make the array 2D then).
> 
> Ah yes, this makes more sense then.
> 
> And thinking about the clipping, which would break this thing
> nevertheless - I think I've put in the clipping too fast because I think
> it is not necessary here.
> 
> So base line is I would be fine to make it 2D and clipping has to be
> removed again, so rev. 3 comes for sure.
> 
> So if no one interrupts, I would send in rev. 3 using a 2D array shortly.

Here it is, revision 3.

-Thilo
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: als_ltp.rev3.patch
URL: <http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/attachments/20091112/1a6f2f38/attachment.txt>