[FFmpeg-devel] [PATCH 1/3] aacdec: move the LTP and TNS tables out of aacdectab.h

Rostislav Pehlivanov atomnuker at gmail.com
Mon Jul 20 21:15:33 CEST 2015


Yep, with this patch compute_lpc_coefs() (referenced in aacdec_template.c
when applying TNS) in ltp.c prints a warning when compiling the fixed
decoder, which is fine since it changed a type from INTFLOAT to float.

>And it will probably be accepted first anyhow :)
I'm fine either way too, it's not hard to substitute a few words :)
The encoder has no problem with INTFLOATs (as long as its 'float' when
compiling the encoder) so feel free to change any shared tables later when
you fully implement the fixed point decoder.

Looking at it I was wrong, compute_lpc_coefs() is used in ra288.c as well,
so disregard what I said about moving that in the decoder, it's fine as it
is.

On 20 July 2015 at 16:52, Nedeljko Babic <Nedeljko.Babic at imgtec.com> wrote:

>  >Thanks for your response.
> >If you are okay with it then it would make sense to push this patch to
> >master first then because you have to rebase aactab.* to INTFLOATs anyway.
> >
>
> I am ok with that.
>
> As I said, I don't have a problem to make necessary changes to my patch
> set if
> this patch is accepted first.
>
> And it will probably be accepted first anyhow :)
>
> >Could you consider moving ltp.c into the aacdec_template.c file? It's
> >relatively short (compared to the decoder source) and it's only used a
> >single time in the entire libavcodec by the aac decoder? Perhaps you can
> >just inline the whole function in the decoder. When compiling aacdec_fixed
> >it prints a warning saying it expects float * rather than an INTFLOAT *.
>
> ltp.c? Are you sure that is the name of the file that is causing a problem?
>
> The only warning of that type that I am having while building is in
> aacdec_template.c
> and that is for ff_aac_eld_window_480 table.
>
> The reason for this warning I explained in one of my previous mails.
>
> -Nedeljko
>


More information about the ffmpeg-devel mailing list