[Vorbis] AACplus

[Borphee]

Ross;

> I'm seeing more and more streaming stations using AACplus, with many 
> listeners being amazed at the sound quality.  Most say that a 48kb/s sounds 
> better than a 128kb/s MP3, which would put Ogg Vorbis at around 96kb/s IMO. 

I haven't done any listening tests comparing those rates, but I would *seriously* doubt that claim.

It would depend a lot on what mp3 encoder you are using...  There are a lot of bad encoders, and many more recommendations for using poor encoder settings.  Not every mp3 encoder is created equal.  And not every encoder setting is a good one.

(Side rant:  The p2p networks are flooded with poor quality 128k mp3's.  The reality is those files are brainwashing a generation of music listeners to accept poor quality encodings.  They have accepted that a 128k mp3 has to sound bad.  I have some p2p mp3's of songs from some cd's I have.  And I can assure you, the quality between those mp3's and the ones I made are substantial.)

Roberto's public listening test showed that at 128k, mp3 was only slightly worse than the best full AAC at 128k.  (And other tests have shown there really are some very poor quality mp3 encoders.)  (Technically, it's possible mp3 might even be better, although the odds are against that.)

http://www.rjamorim.com/test/

Methods such as AACPlus, mp3pro, and so on do pretty well at producing tolerable quality at lower bit rates, but they do it by giving up any chance to sound 'correct' by trying to be 'good enough'.  And by the higher bit rates (above 96k), they can't match the quality of their own full, non 'pro' / 'plus' encoders (AAC, mp3, etc.)

In other words, it may sound okay (very listenable), but you wont mistake it for the original in a listening test.

Considering that at 128k, mp3 is only slightly worse than AAC, and you are saying people are saying that 48k aacplus is better than 128k mp3, then that would be like saying that 48k aacplus is as good as 128k aac.  That just isn't even close to being true.

So even though I haven't done any listening tests comparing 48k aacplus against 128k mp3 against ~96k vorbis, I find it very very hard to believe that 48k aacplus would actually sound closest to the original cd.

For streaming, 48k aacplus may indeed be more than good enough.  But I sure wouldn't call it "near cd quality"....


You might want to go over to HydrogenAudio.org and look through their messages, their FAQ, and maybe post a question or two.  (Be prepared to get flamed, though.  Some of the older members aren incredibly intolerant of any newbie.)

> That means only half the bitrate is required in AACplus compared to Ogg 
> Vorbis for the same sound quality.  Up until this codec was available, Ogg 
> Vorbis compared favourably with all the others.

At normal listening rates, OggVorbis is still one of the best.  Especially with the tunings that were done outside of Xiph.

But at lower 'streaming radio' rates, no, Vorbis isn't the best.  It wasn't tuned / designed for those rates.
 
Roberto's 32k rate listening test didn't show Vorbis in a good light...  I doubt a 48k test would be much different.