[opus] PLC oddity

[John Ridges]

The current git shows ALLOC(etmp, overlap, opus_val32);, where "overlap" 
is st->mode->overlap. I don't know what the difference is between 
st->mode->overlap and st->overlap (maybe none), but it looked odd.

John Ridges


On 11/30/2012 4:44 PM, Jean-Marc Valin wrote:
> That's already what's at 429. I'm not sure I see what's the issue.
>
> 	Jean-Marc
>
> On 12-11-30 06:16 PM, John Ridges wrote:
>> Hi Jean-Marc
>>
>> Shouldn't line 429 in celt_decoder.c be ALLOC(etmp, st->overlap,
>> opus_val32); ? It seems that's how much memory is being written by
>> comb_filter(). Or at least there seems to be some kind of disconnect
>> somewhere.
>>
>> Cheers,
>> John Ridges
>>
>>
>> _______________________________________________
>> opus mailing list
>> opus at xiph.org
>> http://lists.xiph.org/mailman/listinfo/opus
>>
>>
>>
>