[Speex-dev] Experimental psycho-acoustic model

[Thorvald Natvig]

>> When enabling and compiling this with mingw on Win32, my debugger
>> complains about heap overruns when calling speex_encoder_destroy(). This
>> could be a mingw issue though, as I also found a stackalignment bug which
>> prevents me from _USING_SSE (apparantly -mpreferred-stack-size is just
>> "prefered" and therefore ignored... *Sigh*).
>
> Just so I understand, the debugger complains, but the code still runs?
> What's the message?

That's right. Ah, I don't have the code here right now, but it was "heap 
overrun of 200 bytes", meaning it wrote 200 bytes beond the allocated 
block.

If I am to guess, the debugger adds safety markers at the start and end of 
each heap block and checks those during deallocation, so the actual 
overrun probably happens sometime earlier.

>> But, before that point, there seems to be.. well.. I seem to think it's
>> better, but that could be just because I expect it to be better.
>
> Well, that's a start... what bit-rate are you using?

VBR, quality 6.0, wideband, with preprocessor.

>> If the heap problem can be fixed, I can ninja-include this in the next
>> version of Mumble and then poll users' opinions :)
>
> That would indeed give nice testing. Just be aware that it's very
> experimental and I don't know all the side effects of that.

I guess we'll find out.