OGG vs AAC

[xeo]

Ok, so newbie here, and I would like to know which people prefer. I make divx 3 dbc 2CD rips, and would like to create digital 5.1 audio tracks to go with new films and dolby 2 channel ones for old ones. I rip a lot of japanese films, and therefore usually end up with 2-3 audio streams (eng, jap & directors comments) as well as subtitles. I am interested in putting subtitles into an ogg stream as described by doom9 in his ogg stream creation guide.

So what will it be? Is OGG the best, or is AAC? Can only one succeed? Or is one better in 5.1 and the other better in normal 2 channel? And can they both handle subtitle streams?

Give me your thoughts, and even though their both still in their primary stages, let's try and settle this once and for all.

[MaTTeR]

Well this issue isn't going to be settled very soon.

IMO AAC is not a viable solution for the community yet mainly because we have no muxing tools ATM. Nic is also working on the DSF (Right Nic?;) ) AAC is certainly showing promise but it seems premature for the ripping needs now.

Stick with AC3 for 2 CD rips and Vorbis for 1 CD. That's my stance anyways.

[ReferenceDivx]

At 64 ogg is def. better than AAC. I wouldn't try AAC any lower than 96.

Thats just my two cents.

[ChristianHJW]

The biggest point to consider here is that AAC is developing pretty slow now, as the specs have been fixed long ago, while Vorbis is alive and developing incredibly fast overall .... but the perspective to have hardware support in future MPEG4 players is a strong argument for using AAC i admit ....

[DSPguru]

Originally posted by ChristianHJW
The biggest point to consider here is that AAC is developing pretty slow now, as the specs have been fixed hi Chris :)
didn't we hear the same statement from the vorbis-devs ? ("spec closed").


the biggest advantage of Ogg-vorbis, imho, is the fact that it's free of patent.
and yes, the quality is very good :D.

[Nic]

With AAC thats not strictly true, the reason development is slow is because only Ivan is developing it. The spec is closed...but there are more tools that Ivan could develop & introduce (such as IS (Intensity Stereo) which would help at low bitrates (also he could develop PNS more, also helping)

But they are both very good formats :)

Cheers,
-Nic

[DSPguru]

pay attention that the mp3 spec was closed before aac spec had started, but still we see the quality improving, and i believe lame didn't reach its limits.

1. Alexander Leidinger had now inroduced a new nonlinear psymodel to lame.
2. there are talkings about implementing Frank Baumgartes master thesis (psycho-physiologic model)
3. Project L.A.M.E.) Please add option in lame to redirect all kilobits
above 160 to encode 16000hz-22050hz frequencies. After several tests
with Spectralab (www.SoundTechnology.com)i saw, that Lame 3.92 MMX
with options: -b 32 -m s --abr 160 -B 320 -q 0 -d -k produce sound better
than mp3PRO at 96khz (max quality) and ogg with 124kbps all in range
20hz-16000hz and almost identical with original. There is no need to spend
any additional kilobits to freq 20hz-16000hz. But above 16000 (especially
above 18500) sound differs from original.

NOTE: Lame with options -b 32 -m s --abr 160 -B
320 -q 0 -d -k give also better results than Lame with options -b 32 -m s
--abr 160 -B 320 - q 0