I was just wondering if aac variable is better quality than regular aac. i have done this test with wma variable compared to standard wma, and the variable bit rate file had nearly half the size of the regular, but it also sounded worse. however i could use a higher bit general rate for the variable wma as the file size was so low.

do you know if variable aac is better than regular in terms of filesize and quality.

I haven't used Neo's AAC encoder, but it's settings are based on quality not bitrate, anyway.

In general, with iTunes, a nominal bitrate plus VBR will give you a larger file than the same bitrate without VBR checked. Whether that would be better is another question, since there's a subjective element there.

I believe Ivan dimkovich has said that all AAC is VBR.

any so called CBR should just be VBR with bitrate constraints

but i might be wrong

You are correct. Even when AAC is listed as CBR, the bitrate will still change. I like to think of it as quasi-ABR in that you pick a 128kbps CBR bitrate but the bitrate in the song can change +-3-4kbps while the overall average bitrate will remain at about 128kbps.

In general, with any lossy encoder, VBR is always a better choice than CBR. AAC is no different here. Nero's AAC encoder performs much better in a VBR environment. The iTunes AAC encoder dosn't benefit too much from VBR implementations but that is because Apple restricted it. With iTunes AAC, you pick a minimum bitrate and the encoder can change the bitrate +10-15kbps throughout the song. The overall bitrate will still remain near the target you selected. Nero AAC works with different quality levels (like Lame mp3) while iTunes AAC works with a known bitrate. Either way, enabling VBR will only help with the quality of the file.

Variable bitrate has different goals than constant bitrate. Variable bitrate algorithms generally aim for "constant quality".

Imagine this scenario:

For the sake of argument, let's assume that 96kbps is the bitrate where, on average, AAC becomes good quality (the obvious audio distortions start to disappear). If you were to encode at 96kbps CBR, you would likely end up with good quality, but with frequent artifacts (moments of low quality). If you were to encode at 96kbps VBR, the median bitrate might be 80kbps and you would likely end up with less-than-good quality, but it would be consistent (no moments of low quality).

However, you might also end up with a scenario where CBR is often producing too-good quality at 96kbps, because song is simple to encode. Good average quality, moments of high quality. With VBR, it is likely that the median bitrate ends up higher (as many segments of a song can be encoded at good quality with a very low bitrate) and so you end up with high overall quality.

That being said, generally speaking you should use VBR unless your application demands otherwise (streaming audio, target filesize, etc). Used correctly, VBR will give you your desired quality at the smallest file size.

I ran into a similar situation the other day when I was trying to decide if I should use iTunes "CBR" or "VBR". I've ripped a few albums in VBR, and the overall average bitrate is slightly higher than the target I had set (130-135). Then I compared it to the iTunes "CBR" version. By listening to the tracks, I could not tell the difference between the two, and they both sound great to me. So in the end, I decided to use CBR since I would end up having more space available on my portable and that I could not tell the difference between CBR and the VBR encodings anyway.

iTunes CBR isn't exactly CBR as stated before, so you still have some of the benefits from VBR encoding anyway. For example, if you had some digital silence on a track, that segment would get encoded at 2kbps in iTunes "CBR" AAC, while an MP3 encoder (and I think WMA does it the same way as MP3) would still put it at the CBR (128kbps maybe?) that you had it set to. Even in VBR, LAME encodes digital silence at 32kbps since it can't make the bit rate any lower than that.

I'm not surprised. I always use VBR myself, and, theoretically, there are reasons why that's a good idea; but as the O.P. had asked which is "better" I didn't like to say "VBR", because, obviously, a real improvement may be imperceptible to any given listener.

I can definitely hear a difference between 160 aac variable and constant. the variable does sound a lot better, however average albums are about 4 MB bigger than constant, I have reripped a few CD's to the computer at variable using ITUNES however. i cannot find out which ones are constant and which ones are variable once i have ripped them.

There's an option for showing bitrate as one of the columns in iTunes under "View options". That might help track down any you might prefer to re-rip. It's not an infallible method, as AAC VBR files aren't specifically shown as VBR, and, obviously, the final average bitrate might work out to 160kbps anyway. But it may help.

Same thing for MP3, it's just that you have no easy tool to check it with mp3.
By using something like Mp3X, you could see that the actual amount of bits for a given audio frame also changes in an CBR mp3.

iTunes will always show the target bitrate (ie: 160kbps) for an AAC file encoded by itself, VBR or not. It will not show the real average bitrate.

However, if you use other tools like Max (which uses the same encoder as iTunes) or Nero AAC, it will show the correct average bitrate in iTunes. I think those tools just write the proper value to the bitrate tag, while iTunes just writes the target bitrate. Probably for simplicity's sake.

I do use Max to encode, so if what you say in the next paragraph is so, maybe that's why I can see them.