Skip to main content ↑ ↓

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: How good is AAC VBR in iTunes 5? (Read 100893 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

How good is AAC VBR in iTunes 5?

Reply #50 –
Indeed a listening test would be a good idea. Let's not forget that the only really 'telling' tests we have (afaik anyway) are those by Roberto...And I sometimes get the impression people have blind faith in their results. If you look at the software involved:

* lame 3.97 has reached beta stage and as such should be usable. The general concensus also seems to be that the new vbr algorythm is now better, quality and speed-wise, than the old one.
* nero AAC has had a major version update
* quicktime went from 6 to 7 too.

(yes, okay, so I'm only interested in aac or mp3 for lossy codecs, probably other codecs underwent changes too in all this time).

I think the results today for such a test could be really surprising. lame certainly underwent some serious changes. The general assumption is that both nero aac and quicktime aac will sound better than lame at an equal bitrate and that quicktime is the better encoder of the two. Seeing as guruboulez has shown that nero aacplus (just an example) is in fact an inferior encoder compared to winamp (of all things...) and that in the past we have seen that even between minor revisions of itunes/quicktime (remember the 4.5 to 4.6 upgrade with all the discussions about deterioration in quality) there can be significant quality changes, I cannot help but wonder what a statistically valid comparison between these codecs would reveal these days...

What actually is the tool people use to measure bitrate distribution of aac files? Foobar's mentioned in this thread but I cannot find any function in it that does this...

How good is AAC VBR in iTunes 5?

Reply #51 –
I too would be interested in an updated listening test as I am starting to rip my music to the Apple Lossless format for backup purposes.  I would then be able to convert between a lossless and lossy file (ie lossless to lossy) very easily so choosing a format would not be so much of a pain.  Lame 3.97 has changed and we have now seen the introduction of QuickTime 7, it would be nice to know how the updated versions of the popular mp3 and mpeg-4 AAC have changed.  Indeed a listening test at 128kbps or higher with both codecs, in the previous test the QuickTime mpeg-4 AAC encoder was ommitied because it couldn't produce a file with a 180kbps bitrate, would be much appreciated.

So far it seems like Apple is limiting the VBR functionality of their mpeg-4 AAC encoder as they have similarly done with their mp3 encoder.  Hopefully they will add greater VBR performance to their encoder.  I doubt they will ever get the VBR range of the Lame mp3 encoder but they could approach the VBR capabilities of the Lame mp3 encoder if they wanted (ie the ability to go fully from 32kbps to 320kbps with none of this minimum bitrate crap, instead they should do different VBR levels as Lame and OGG do).

How good is AAC VBR in iTunes 5?

Reply #52 –
I think iTunes just writes the mp4 tags a bit differently - if you remux the AAC stream into a new .m4a using mp4creator and add it to itunes the actual bitrate is displayed (as opposed to what option you selected in the prefs) - a bit annoying behaviour for us but it's probably a usability thing. (why do songs come out as x kbps when I have y selected in the settings, i dont understand etc...) They'd probably make mp3 do it too if they could (or if anybody used it).

How good is AAC VBR in iTunes 5?

Reply #53 –
Hi! I encoded(iTunes5.0.1) a couple songs using the new VBR mode at 128 kbps(prg trance) and run a check on Winamp5.1 with "in_mp4" decoder from RareWares.org... The bitrate fluctuates in the range: 113kbit/s-150kbit/s... So  is it ABR or VBR actually???

(sorry for my grammar - english isn't my native language)
hdd: flac -8\wavpack -h  iPod: lame3.98.2 -V2\NeroAAC -q 0.43

How good is AAC VBR in iTunes 5?

Reply #54 –
Quote
Hi! I encoded(iTunes5.0.1) a couple songs using the new VBR mode at 128 kbps(prg trance) and run ? check on Winamp5.1 with "in_mp4" decoder from RareWares.om... The bitrate fluctuates in the range: 113kbit/s-150kbit/s... So  is it ABR or VBR actually???

(sorry for my grammar - english isn't my native language)
[a href="index.php?act=findpost&pid=331260"][{POST_SNAPBACK}][/a]


By definition, it's VBR. As long as the bitrate isn't constant, it's variable 

How good is AAC VBR in iTunes 5?

Reply #55 –
Quote
By definition, it's VBR. As long as the bitrate isn't constant, it's variable 
[a href="index.php?act=findpost&pid=331262"][{POST_SNAPBACK}][/a]

i see...
hdd: flac -8\wavpack -h  iPod: lame3.98.2 -V2\NeroAAC -q 0.43

How good is AAC VBR in iTunes 5?

Reply #56 –
This is false.

The difference between ABR and VBR is that ABR aims at a target average bitrate and VBR aims at a target quality level.

A file with a varying bitrate can still be ABR.

How good is AAC VBR in iTunes 5?

Reply #57 –
Quote
This is false.

The difference between ABR and VBR is that ABR aims at a target average bitrate and VBR aims at a target quality level.

A file with a varying bitrate can still be ABR.
[a href="index.php?act=findpost&pid=331267"][{POST_SNAPBACK}][/a]

I think so...  but what  can you say about iTunes encoder??? is it VBR or ABR???
hdd: flac -8\wavpack -h  iPod: lame3.98.2 -V2\NeroAAC -q 0.43

How good is AAC VBR in iTunes 5?

Reply #58 –
For the iTunes encoder specifically, if you encode a difficult sample such as fatboy, the bitrate will deviate wildly from the target. It's just that unlike LAME, the encoder doesn't use a much lower bitrate on mono stuff, and because it's AAC there is no need for any bitrate bloat on most HF material. Hence, the target bitrate will usually be met pretty closely, but if the music requires it, it will go much higher.

In the end, what matters is how it sounds IMHO, but it seems there's not enough interest in the encoder for the golden ears around here to do listening tests of it. 

How good is AAC VBR in iTunes 5?

Reply #59 –
Quote
For the iTunes encoder specifically, if you encode a difficult sample such as castanets, the bitrate will deviate wildly from the target. It's just that unlike LAME, the encoder doesn't use a lower bitrate on mono stuff, and because it's AAC there is no need for bitrate bloat on most HF material. Hence, the target bitrate will usually be met pretty closely, but if the music requires it, it will go much higher.
[a href="index.php?act=findpost&pid=331274"][{POST_SNAPBACK}][/a]

ok! i will try to encode using CBR mode and check for the bitrate fluctuation.... may be there is no difference between CBR and VBR mode....
hdd: flac -8\wavpack -h  iPod: lame3.98.2 -V2\NeroAAC -q 0.43

How good is AAC VBR in iTunes 5?

Reply #60 –
Yesterday i encoded Pink Floyd's Album "The Dark Side Of The Moon" at 128 kbps VBR(iTunes5.0.1) and checked for the bitrate fluctuation: for example - Money (124kbps-159kbps)... hmm...
hdd: flac -8\wavpack -h  iPod: lame3.98.2 -V2\NeroAAC -q 0.43

How good is AAC VBR in iTunes 5?

Reply #61 –
three pages of discussion and no abx results, do i see this right?
I know, that I know nothing (Socrates)

How good is AAC VBR in iTunes 5?

Reply #62 –
Quote
Indeed a listening test would be a good idea. Let's not forget that the only really 'telling' tests we have (afaik anyway) are those by Roberto...And I sometimes get the impression people have blind faith in their results. If you look at the software involved:

* nero AAC has had a major version update
[a href="index.php?act=findpost&pid=329420"][{POST_SNAPBACK}][/a]


Since when?  I'm pretty sure the aac encoder used is still the same one that's been out for some time.  Check the properties of aacenc32.dll.  Although the file version shown may be different to previous version, the Product Version still shows up as 2.9.9.999, the same as with older Nero Versions.

How good is AAC VBR in iTunes 5?

Reply #63 –
Test done by Guruboolez show different results with the AAC encoder from the last QT version and QT7.
we was young an' full of beans

How good is AAC VBR in iTunes 5?

Reply #64 –
Instead of looking at bitrates fluctuating, perhaps someone should try to ABX between an ABR and VBR iTunes-encoded file.  I mean, you can vary the bitrate all you want but if it leads to no increase in quality, then there's no point really.  So an ABX should be able to tell us whether there are any perceptible differences.

How good is AAC VBR in iTunes 5?

Reply #65 –
has somebody encoded a file with a silence and then secret track, where the bitrate should drop to it's lowest possible rate allowed.  If its VBR, it will go down all the way.  My guessing it will not as iTunes VBR is actually ABR

I would check it out but I really don't want to install iTunes
Chaintech AV-710

How good is AAC VBR in iTunes 5?

Reply #66 –
Quote
has somebody encoded a file with a silence and then secret track, where the bitrate should drop to it's lowest possible rate allowed.  If its VBR, it will go down all the way.  My guessing it will not as iTunes VBR is actually ABR

1. yup, it goes down to 2kbps for the silence according to foobar

2. fatboy does 116kbps(87KB) at the 128kbps VBR AAC setting.

3. I dunno if this is interesting or not, never tried this on a track with LAME.

I took the last track off Green Day's Dookie and split it into 4 parts(intro, fastpart, silence, secret track) and encoded them(129, 142, 3, 126)kbps respectively.  The full track comes out to 104kbps.

total filesizes: 4,612,542 bytes(4 tracks), 4,679,101 bytes(full track)

the size discrepency might be because EAC's wave editor isn't so good, the times in an itunes playlist add up to 5:45, whereas the full track is 5:47. edit: another thing i forgot to add is that each of the 4 sections is between 1:15 and 1:39, so the times are fairly similar, I think I will try splitting it into equal parts tonight.

anyhow, the reason I decided to do this was to see if each section tried to stick close to an ABR rather than really going VBR.  Hopefully there's something to learn from this, maybe I'll do a pink floyd album and see what happens.

How good is AAC VBR in iTunes 5?

Reply #67 –
pacohaas,
Thanks for posting that info! 

How good is AAC VBR in iTunes 5?

Reply #68 –
Quote
Quote
has somebody encoded a file with a silence and then secret track, where the bitrate should drop to it's lowest possible rate allowed.  If its VBR, it will go down all the way.  My guessing it will not as iTunes VBR is actually ABR

1. yup, it goes down to 2kbps for the silence according to foobar

2. fatboy does 116kbps(87KB) at the 128kbps VBR AAC setting.

3. I dunno if this is interesting or not, never tried this on a track with LAME.

I took the last track off Green Day's Dookie and split it into 4 parts(intro, fastpart, silence, secret track) and encoded them(129, 142, 3, 126)kbps respectively.  The full track comes out to 104kbps.

total filesizes: 4,612,542 bytes(4 tracks), 4,679,101 bytes(full track)

the size discrepency might be because EAC's wave editor isn't so good, the times in an itunes playlist add up to 5:45, whereas the full track is 5:47. edit: another thing i forgot to add is that each of the 4 sections is between 1:15 and 1:39, so the times are fairly similar, I think I will try splitting it into equal parts tonight.

anyhow, the reason I decided to do this was to see if each section tried to stick close to an ABR rather than really going VBR.  Hopefully there's something to learn from this, maybe I'll do a pink floyd album and see what happens.
[a href="index.php?act=findpost&pid=331760"][{POST_SNAPBACK}][/a]


hey, thanks for that info, maybe I will FINALLY start using AAC now that it seems to be officially VBR, it would be nice if there was a "transparent" like -aps setting as in LAME
Chaintech AV-710

How good is AAC VBR in iTunes 5?

Reply #69 –
Quote
Quote
has somebody encoded a file with a silence and then secret track, where the bitrate should drop to it's lowest possible rate allowed.  If its VBR, it will go down all the way.  My guessing it will not as iTunes VBR is actually ABR

1. yup, it goes down to 2kbps for the silence according to foobar


iTunes-encoded AAC files behaved this way even before the addition of "VBR" capability, so this doesn't really demonstrate anything.

How good is AAC VBR in iTunes 5?

Reply #70 –
so it has something hardcoded for silence, even in CBR?
Chaintech AV-710

How good is AAC VBR in iTunes 5?

Reply #71 –
Quote
so it has something hardcoded for silence, even in CBR?
[a href="index.php?act=findpost&pid=331944"][{POST_SNAPBACK}][/a]
no, i think the main point people keep forgetting is that AAC is not a CBR codec, and unless it's possible for an encoder to mindlessly encode each frame the same, you will always exhibit some variability in the bitrates.

I've heard it explained as similar to LAME(and other mp3 encoders)'s bitrate resevoir, except the resevoir is much deeper.

To add to this discussion, I will encode the same track in CBR(128) as well when I get home and report the actual bitrate.

How good is AAC VBR in iTunes 5?

Reply #72 –
i am lazy, i just want somebody to tell me what settings to encode with... playing around now withe the different VBR settings...

to me it seems that it is ABR, though bitrate resevoir allows the bitrate to go down to 2kbps on silence...

I am playing around with the same track too, Green Day - 14 - F.O.D. 
Chaintech AV-710

How good is AAC VBR in iTunes 5?

Reply #73 –
nothing too interesting to report, the CBR 128 version of FOD came out to 100kbps because it tried to stick closer to 128kbps, but when the silence comes, it doesn't need it and the average goes down.

Splitting the track into 4 equal parts yielded predictable(from my earlier results) results: VBR(129,139,22,127)kbps CBR(127,125,22,127)kbps.

It would be nice if I could find a program similar to EncSpot that would show a graph of bitrate over time, and an overall min/max bitrate....anyone know of one that I haven't been able to find?

also...as far as abx tests go, you have to get the attention of people with better ears than me.  I'm a numbers guy, and these numbers are telling me that simply adding the VBR checkbox next to your favorite AAC setting in iTunes will increase bitrate on more difficult areas(ie part 2 here) and "theoretically" increase the quality on those parts, which is what I'm looking for.  I'm sure we all wish apple would make a quality based VBR setting, but for now this is a start.

How good is AAC VBR in iTunes 5?

Reply #74 –
Thanks again pacohaas!  With me being a numbers guy too, I like your posts.  The data you have provided implies checking the VBR box in iTunes will allow the encoder to use more bits if needed but won't really affect the minimum number of bits it will use (compared to "CBR" mode).  I was a bit surprised by that, but that is because I didn't realize how low iTunes' "CBR" would go.  I agree what we really need (to better analyze this new VBR mode) is a program that shows a graph of bitrate over time.