Converting framerate of an aac - Reducing depth from 64 to 32 bits...

[maudit]

hi, I'm trying to convert the framerate of an AAC with useac3to, with the following command:

 %_25p.aac -24.000 -changeTo25.000

eac3to says:

AAC, 2.0 channels, 24kHz
Decoding with DirectShow (Nero Audio Decoder 2)...
DirectShow reports 2.0 channels, 24 bits, 48kHz
Changing FPS from 24.000 to 25.000...
Reducing depth from 64 to 32 bits...
Encoding AAC <0.50> with NeroAacEnc...

- why it is saying from 64 to 32bits?
what parameter do I have to add in the line in order to keep the bitdepth at its original 24bits?

edit: ok I think the parameter -down24 is working..


but now I have some other doubts:
- which "Output format" I have to use: aac or qaac-ext?

the eac3to manual talks about:

-quality=0.35 any audio format M4A / MP4 Nero AAC encoding quality (0.00=lowest; 1.00=highest) (default: 0.5)

- how to deal with the quality? do I keep it to default or maximize it?

[ProWo]

I would first encode aac to pcm, then use the pcm to make the fps change, and then encode the new pcm back to aac.
This way you avoid timing problems that often occur with aac.

[Cornucopia]

@ProWo, the eac3to code is saying that it is decoding to LPCM.

Am betting that the 64 to 32bits has to do with internal resampling calculations and reverting back to standard stored-in-RAM bit depths, but since 32bits is over and above 24bit LPCM, you are still doing fine in that respect.
Other that me having a laugh at a piece of software that thinks audio has any framerate to begin with (it doesn't), I expect this is just using those terms as a reference for how much to time stretch/shrink resample.

Note: any time you go from uncompressed to compressed format, the meaning of bitdepth is not the same. Bit depth in the compressed realm is dependent upon the compression format's limits, but also upon the capabilities of the compression app and how it was coded. But they are internal to the compressed format and bear no real relation to what was in the uncompressed source. And when you decode & decompress it, it should revert to the nearest equivalent that the decoder/decompressor can emulate of the settings that were stored about the source.
So, 24bit LPCM --> ??bit aac --> 24bit LPCM. But since you are resampling, it is kind of doing this: 24bit LPCM (maybe, do you really know this for a fact?) --> ??bit aac --> 32bit float or LPCM --> 64bit float (internal calculation) or LPCM --> 32bit float or LPCM --> maybe 24bit LPCM. And if you have it going back to aac, you have to tell it what samplerate, and bitrate to choose but you may or may not be able to tell the encoder the bitdepth.

If you want to lose less (aka maximize the remaining quality), you should choose highest quality. But you should try multiple options and let your ears be your guide.


Scott

[maudit]

thank you both guys, sry for the delay, I tried with qaac instead of aac nd the bithdepth was ok, it said from 64 to 24.
anyway, the file has lost 20MB over 90MB, even at the "quality 99", with ffmpeg instead the size was working better,
at the end I decided to no longer change its fps and use it at 24p.

thanks.