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: Itunes AAC (Read 10815 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Itunes AAC

I know I could test it myself but I was just curious to see if anyone has already compared itunes at 160vbr vs 192vbr vs 224 cbr for example. Where does it achieve transparency while maintaining a good file size. What is the most efficient setting? I am testing it myself right now but having more people to test it would be better.

Itunes AAC

Reply #1 –
I conducted my own test about two weeks ago using iTunes and QuickTime to encode 20 samples at the mpeg-4 AAC file standard.  I tested the 128kbps CBR/VBR, 160kbps CBR/VBR, and 192kbps CBR/VBR bitrates.  I didn't go above 192kbps because I already have my library in the -V 2 --vbr-new Lame 3.97b1 format and 224/256kbps vbr/cbr mpeg-4 AAC's don't offer large file size differences if any.

I don't have my results graphed or even saved on my computer anymore.  However, I still remember the details of my results as I keep telling people them.  I found that the VBR bitrates for the 128kbps and 160kbps bitrates were better than the CBR bitrates.  I found it hard to ABX between a source and a 192kbps CBR mpeg-4 AAC but I scored better for 192kbps VBR mpeg-4 AAC.  So, even though the iTunes mpeg-4 VBR setting is not yet matured, I would still use it.  Besides, VBR is still better than ABR.  By default, the mpeg-4 AAC format is ABR and this is true for iTunes mpeg-4 AAC.  The VBR setting would allow the bitrate to increase even more than the standard CBR (actually ABR) setting.  I recomend using VBR instead of the CBR option in iTunes.  As for the bitrate, I found 128kbps VBR to be quite listenable but there were still some tracks that posed problems (mainly off of Lamb Of God's New American Gospel album).  I found 160kbps VBR to be my recomended bitrate for iTunes mpeg-4 AAC.  192kbps VBR was indistinguishable from the CD but I could not hear a difference between a 192kbps VBR and -V 2 --vbr-new mp3 and the file size differences are minimal.  So, I recomend going with 160kbps VBR mpeg-4 AAC as I could not readily hear the difference between that and the source CD.  I had to try really hard to ABX the difference and my ears felt exhausted after testing 160kbps VBR mpeg-4 AAC.  I could not ABX the difference between a -V 2 --vbr-new Lame mp3 and a 160kbps VBR mpeg-4 AAC (iTunes).

I hope that helps.  I ended up sticking with Lame mp3 simply because it is still considered the universal file format (mp3) for digital audio and the Lame mp3 encoder has a lot of development into it.  Additionally, I only have three devices which can play mpeg-4 AAC (PocketPC, iPod, and PSP) while I have 5 devices that play mp3 (PocketPC, iPod, car CD deck, PSP, and Xbox 360).  You can see guruboolez's 130kbps multiformat listening test to see how 130kbps VBR iTunes mpeg-4 AAC compared against other formats including -V 2 --vbr-new Lame 3.97b1.  Just a FYI - -V 2 --vbr-new Lame 3.97b1 is still of higher quality than 130kbps mpeg-4 iTunes AAC.

Itunes AAC

Reply #2 –
Thanks a lot, I read guruboolezs thread. awesome! it's what got me so interested in itunes encoding in the first place.
Now, anyone here that can tell the difference between cd and high bitrate aacs? if so, pls share your experience

Itunes AAC

Reply #3 –
Kornchild, when you said that you scored better using 192vbr vs cbr, were you saying that the 192vbr file was inferior and could be abxed successfully with the original source file???

As for my tests, I found 160vbr to be pretty impressive for the cds i tested so  far... 192vbr is better for some songs. 160vbr is a little duller than the original for one track i thoroughly tested but i could only abx it successfully at the beggining of my test. When fatigue set in, I couldn't tell the difference between the 160vbr and the wav file. The wav file was ripped with eac separately though and the aacs directly from the cd on itunes. This might make a difference but probably not that much.

Itunes AAC

Reply #4 –
I have used 160K CBR, and immediately switched to 160K (sort of) VBR as soon as that setting was added to iTunes (version 5.0?). It is a perfect setting for me, and considering iTunes encodes in that mode at ~25X on my PC it has made batch transcoding my FLAC files a breeze.

I have found one issue with the VBR mode. There is significant bitrate inflation for CDs that have a significant amount of tape hiss (high frequency distortion?).

For example, several tracks from the Miles Davis concert Black Beauty Live at the Filmore West reach 180K using the 160K VBR setting. So, it seems that Apple need to tweak the VBR mode somewhat, or add better filtering with the VBR mode. It seems that the encoder is trying to hard to replicate the hiss, when it could really just be filtered to produce a better sounding file (if not a file that is perfectly accurate to the original CD).

Anyway, I love 160K VBR, for my bad hearing it is an excellent compromise between file size and sound quality.

I still hope that Apple will open up all the encoder settings to tweaking, so that the community could develop a system of presets.

Itunes AAC

Reply #5 –
Quote from: ShowsOn,Nov 18 2005, 01:50 PM

I have used 160K CBR, and immediately switched to 160K (sort of) VBR as soon as that setting was added to iTunes (version 5.0?). It is a perfect setting for me, and considering iTunes encodes in that mode at ~25X on my PC it has made batch transcoding my FLAC files a breeze.

How do you batch encode from FLAC using iTunes please ?

Itunes AAC

Reply #6 –
I use Foobar, and the iTunes Encoder from here:
http://otto.homedns.org:8888/iTunes/iTunesEncode46.zip

You just set up iTunesEncode.exe as an "output preset" under Diskwriter.

Use M4A as the extension, and I use the following "parameters":
"-e "AAC Encoder" -a "%artist%" -l "%album%" -t "%title%" -g "%genre%" -y %date% -n %tracknumber% -i %s -o %d"

Then make the Display Name "iTunes" or whatever. Then to convert just drag flac files into Foobar, then right click on them, and select CONVERT.

Then wait...

Itunes AAC

Reply #9 –
So Can we use itunes aac with eac? where do i get the latest version for download and how is it used properly with eac?

Itunes AAC

Reply #10 –
Donnus: That is the last version, and the readme tells you how to set it up with EAC.

Itunes AAC

Reply #11 –
Quote
Donnus: That is the last version, and the readme tells you how to set it up with EAC.
[a href="index.php?act=findpost&pid=343098"][{POST_SNAPBACK}][/a]


Thanks

Now any comparisons between 160vbr and 192 vbr? Can anyone tell the difference in sound?

Itunes AAC

Reply #12 –
I did comparisons between 192 vbr and 160 vbr then threw in lame 3.97b1 --preset fast standard. 192 vbr wins. lame is second then 160 vbr aac third. The 2 aac files sound very close with the same characteristics but the 192 has a brighter sound and attacks are more defined. The mp3 file has as much highs as the 192 vbr file but with less body. The 160 vbr has the same rich sound as the 192 file just with less definition. For all types of music I tested, 192vbr AAC wins in all tests by a slight margin. 160vbr AAC and the mp3 file is a tossup. music that needs richness benefit from the ever so slightly darker sound of the 160 aac file and ones that need detail will benefit from the mp3( the transients are quicker). The mp3 is second to the 192 vbr aac though cause it sounds closer to the original cd than the 160 vbr aac.

Whats nice about the 160vbr aac is that if it is not compared to the original cd, It sounds nice and rounded and better for a lot of music especially electronic. 192vbr has it all compared to the others here and is therefore my choice so far.

Just to addon, I feel that the overall distortion of aac errs to the smooth and sort of chalky side while mp3 errs to be a little grainy.

On my quick tests of 128 aac vs 128 mp3, the mp3s noticeably roll off the treble more than the aac and when complex passages come, the aac sounds like it clips the sound a bit(sounding like an amp clipping) while the mp3 gives its more digital sounding mess of an artifact. With the aac, it fooled me into thinking that my headphone amp clipped while the mp3s distortion was unmistakeably mp3s own thing. stepping aac up to 160 got rid of the clipping for the specific passage.

I'll keep posting my findings...

Itunes AAC

Reply #13 –
I would also like to add that the clipping sound i heard with the aac really reminds me of how tube amps clip. It starts by softening the sound then breaks when really overloaded. A weird thing since 128 clips 160 sounds a little soft then upping the bitrate more fixes it. hmm. weird that it has this kind of distortion since its not actually clipping but just having inadequate bitrate for the sample. Any AAC developers here able to explain this??? I like it much better than how mp3 acts when it chokes(when the bitrate not enough for the sample).

Itunes AAC

Reply #14 –
Quote
I did comparisons between 192 vbr and 160 vbr then threw in lame 3.97b1 --preset fast standard. 192 vbr wins. lame is second then 160 vbr aac third. The 2 aac files sound very close with the same characteristics but the 192 has a brighter sound and attacks are more defined. The mp3 file has as much highs as the 192 vbr file but with less body. The 160 vbr has the same rich sound as the 192 file just with less definition. For all types of music I tested, 192vbr AAC wins in all tests by a slight margin. 160vbr AAC and the mp3 file is a tossup. music that needs richness benefit from the ever so slightly darker sound of the 160 aac file and ones that need detail will benefit from the mp3( the transients are quicker). The mp3 is second to the 192 vbr aac though cause it sounds closer to the original cd than the 160 vbr aac.

Whats nice about the 160vbr aac is that if it is not compared to the original cd, It sounds nice and rounded and better for a lot of music especially electronic. 192vbr has it all compared to the others here and is therefore my choice so far.

Just to addon, I feel that the overall distortion of aac errs to the smooth and sort of chalky side while mp3 errs to be a little grainy.

On my quick tests of 128 aac vs 128 mp3, the mp3s noticeably roll off the treble more than the aac and when complex passages come, the aac sounds like it clips the sound a bit(sounding like an amp clipping) while the mp3 gives its more digital sounding mess of an artifact. With the aac, it fooled me into thinking that my headphone amp clipped while the mp3s distortion was unmistakeably mp3s own thing. stepping aac up to 160 got rid of the clipping for the specific passage.

I'll keep posting my findings...
[a href="index.php?act=findpost&pid=343389"][{POST_SNAPBACK}][/a]



I'd be careful around here about using adjectives such as "grainy," "dark," "smooth," et al to describe encodings without providing evidence from a blind test. Encodings don't typically have sounds signatures like you describe, especially at high bitrates: they either sound transparent to the source, or artifact during particularly difficult passages (and at these bitrates, those artifacts would be rather subtle). Seeing as how -V2 --vbr-new was often rated a 5/5 by Guru in his most recent listening test (averaging around a 4.6 throughout), it should be quite difficult to ABX from 192kbps VBR AAC.

Also, MP3 is technically worse at handling transients than AAC, which theoretically contradicts the statement "the transients are quicker." MP3's short blocks are 192 samples long, compared to 128 for AAC, which could lead to pre-echo and "thicker" sounding transients compared to AAC, if anything.

Itunes AAC

Reply #15 –
about itunes with eac, is there a better readme file than the one included with the encoder? I want more details on changing bitrates. also can i just use eacs id3v1.1 tagging options? does that work with aac? or do I have to type the artist name everytime in the command line options?

Itunes AAC

Reply #16 –
Quote
Quote
I did comparisons between 192 vbr and 160 vbr then threw in lame 3.97b1 --preset fast standard. 192 vbr wins. lame is second then 160 vbr aac third. The 2 aac files sound very close with the same characteristics but the 192 has a brighter sound and attacks are more defined. The mp3 file has as much highs as the 192 vbr file but with less body. The 160 vbr has the same rich sound as the 192 file just with less definition. For all types of music I tested, 192vbr AAC wins in all tests by a slight margin. 160vbr AAC and the mp3 file is a tossup. music that needs richness benefit from the ever so slightly darker sound of the 160 aac file and ones that need detail will benefit from the mp3( the transients are quicker). The mp3 is second to the 192 vbr aac though cause it sounds closer to the original cd than the 160 vbr aac.

Whats nice about the 160vbr aac is that if it is not compared to the original cd, It sounds nice and rounded and better for a lot of music especially electronic. 192vbr has it all compared to the others here and is therefore my choice so far.

Just to addon, I feel that the overall distortion of aac errs to the smooth and sort of chalky side while mp3 errs to be a little grainy.

On my quick tests of 128 aac vs 128 mp3, the mp3s noticeably roll off the treble more than the aac and when complex passages come, the aac sounds like it clips the sound a bit(sounding like an amp clipping) while the mp3 gives its more digital sounding mess of an artifact. With the aac, it fooled me into thinking that my headphone amp clipped while the mp3s distortion was unmistakeably mp3s own thing. stepping aac up to 160 got rid of the clipping for the specific passage.

I'll keep posting my findings...
[a href="index.php?act=findpost&pid=343389"][{POST_SNAPBACK}][/a]



I'd be careful around here about using adjectives such as "grainy," "dark," "smooth," et al to describe encodings without providing evidence from a blind test. Encodings don't typically have sounds signatures like you describe, especially at high bitrates: they either sound transparent to the source, or artifact during particularly difficult passages (and at these bitrates, those artifacts would be rather subtle). Seeing as how -V2 --vbr-new was often rated a 5/5 by Guru in his most recent listening test (averaging around a 4.6 throughout), it should be quite difficult to ABX from 192kbps VBR AAC.

Also, MP3 is technically worse at handling transients than AAC, which theoretically contradicts the statement "the transients are quicker." MP3's short blocks are 192 samples long, compared to 128 for AAC, which could lead to pre-echo and "thicker" sounding transients compared to AAC, if anything.
[a href="index.php?act=findpost&pid=343464"][{POST_SNAPBACK}][/a]


Well, Thats what I hear and have proven to myself with foobar abx. 160 vbr aac really does have a softer quality to it than 192 vbr or the preset standard mp3 file. mp3 when it reveals its limitations softens transients more but it does it in a different way. mp3 has that underwater or talking in the fan type of coloration while aac sounds like an analog coloration and sounds as if the 160 file is recorded with a slightly lesser quality cassette tape than the 192 file. Try listening for yourself, it is subtle but apparent in all the music i abxed.

I didn't say aac was worse at handling transients. remember the 160vbr aac file is considerably smaller than the preset standard mp3 (which is slightly larger or about equal to the 192 vbr aac). What I said was that the softening sounded more like it wasn't digital in nature like mp3s artifacts are which makes it better in my opinion.

I know most people here don't believe in different sounding cables but the 160 vbr and the 192vbr aac files sound as if the same music was playing on the same stereo system with the same levels but with one difference in the signal chain.... The interconnects being used is one step lower in price of the same series of a companys cable with the 160  Hehehehe I'm gonna get flamed but I'm serious.

Well, anyways I haven't detected the differences between 192vbr and 224 yet. It looks like I'll be using 192 vbr aac when I get my ipod

Itunes AAC

Reply #17 –
Quote
about itunes with eac, is there a better readme file than the one included with the encoder? I want more details on changing bitrates. also can i just use eacs id3v1.1 tagging options? does that work with aac? or do I have to type the artist name everytime in the command line options?
[a href="index.php?act=findpost&pid=343465"][{POST_SNAPBACK}][/a]

When using iTunesEncode, the taggind is done through iTunes, I don't know why you are typing the Artist name in the command line, those letters on the command line are to pass all the disc info to itunes so it can tag the files. To change the bitrate, you have to set up iTunes with the encoding settings you want and then select the AAC or whatever encoder with iTunesEncode.

EDIT: I don't think ID3v1.1 will work with AAC files.
we was young an' full of beans

Itunes AAC

Reply #18 –
Donunus, if you still have them, would you mind posting a log of your ABX session with foobar? There's been very little (if any?) ABX data posted on the board regarding the new VBR settings above 128kbps, so any information is appreciated.

Itunes AAC

Reply #19 –
Quote
Donunus, if you still have them, would you mind posting a log of your ABX session with foobar? There's been very little (if any?) ABX data posted on the board regarding the new VBR settings above 128kbps, so any information is appreciated.
[a href="index.php?act=findpost&pid=343472"][{POST_SNAPBACK}][/a]


I will do another abx test and post the results because I passed the abx test even with my ears still clogged from coming from the swimming pool  and I didn't bother saving the logs.  I will do a more formal test with various songs. The tests I did for these so far were only three trials per sample and i rejected the differences if i didn't get it perfect 3/3. The songs I tested so far were from different sounding music but the sound characteristics of the files were consistent with all of the music. Music included were from bjorks greatest hits, Travis the invisible band, michael buble, ronnie laws tribute to eddie harris and luis giraldos daydream.

Itunes AAC

Reply #20 –
Quote
Quote
Donunus, if you still have them, would you mind posting a log of your ABX session with foobar? There's been very little (if any?) ABX data posted on the board regarding the new VBR settings above 128kbps, so any information is appreciated.
[a href="index.php?act=findpost&pid=343472"][{POST_SNAPBACK}][/a]


I will do another abx test and post the results because I passed the abx test even with my ears still clogged from coming from the swimming pool  and I didn't bother saving the logs.  I will do a more formal test with various songs. The tests I did for these so far were only three trials per sample and i rejected the differences if i didn't get it perfect 3/3. The songs I tested so far were from different sounding music but the sound characteristics of the files were consistent with all of the music. Music included were from bjorks greatest hits, Travis the invisible band, michael buble, ronnie laws tribute to eddie harris and luis giraldos daydream.
[a href="index.php?act=findpost&pid=343652"][{POST_SNAPBACK}][/a]

3/3 means nothing. When you get at least 8/8, then you can claim you ABX'd something.

Itunes AAC

Reply #21 –
Quote
It seems that the encoder is trying to hard to replicate the hiss, when it could really just be filtered to produce a better sounding file (if not a file that is perfectly accurate to the original CD).
[a href="index.php?act=findpost&pid=342846"][{POST_SNAPBACK}][/a]

Hint: Run a noise removal tool on your files before encoding them. It's not the encoders job to remove any hiss from your files. The encoded file should sound identical to the original in the ideal case.

Itunes AAC

Reply #22 –
Quote
Quote
Quote
Donunus, if you still have them, would you mind posting a log of your ABX session with foobar? There's been very little (if any?) ABX data posted on the board regarding the new VBR settings above 128kbps, so any information is appreciated.
[a href="index.php?act=findpost&pid=343472"][{POST_SNAPBACK}][/a]


I will do another abx test and post the results because I passed the abx test even with my ears still clogged from coming from the swimming pool  and I didn't bother saving the logs.  I will do a more formal test with various songs. The tests I did for these so far were only three trials per sample and i rejected the differences if i didn't get it perfect 3/3. The songs I tested so far were from different sounding music but the sound characteristics of the files were consistent with all of the music. Music included were from bjorks greatest hits, Travis the invisible band, michael buble, ronnie laws tribute to eddie harris and luis giraldos daydream.
[a href="index.php?act=findpost&pid=343652"][{POST_SNAPBACK}][/a]

3/3 means nothing. When you get at least 8/8, then you can claim you ABX'd something.
[a href="index.php?act=findpost&pid=344013"][{POST_SNAPBACK}][/a]


In that case, I can't do a 10/10 i dont think with these bitrates because after 7 repeats fatigue sets in and my hearing gets less acute to the differences. I can do it with lower bitrates better.

The thing is, I know its not placebo if i get 3/3 all the time with all the samples i tested. Going up to 320 vs 192vbr is a harder task. My abx results with those if i do get a 3/3 will be a very good guess. i think i can hear the difference but i find out it was actually just the music changing tone and not because of a different bitrate

So are you saying that since i only get a 3/3 but repeatable 3/3 after some times rest that the differences are not there with these files? They are subtle enough that a little fatigue can make the differences insignificant but I really do hear the difference. It is consistent with all samples i tested.

All I'm doing here is testing for myself which bitrate makes me content that I am not missing anything from the original sound and sharing my results for people that don't want to test for themselves.

Itunes AAC

Reply #23 –
Quote
So are you saying that since i only get a 3/3 but repeatable 3/3 after some times rest that the differences are not there with these files? They are subtle enough that a little fatigue can make the differences insignificant but I really do hear the difference. It is consistent with all samples i tested.[a href="index.php?act=findpost&pid=344078"][{POST_SNAPBACK}][/a]

What's so difficult in getting a 8/8 score if you can get consistent 3/3? When you get tired you take a break and then continue after you rested...
 

Itunes AAC

Reply #24 –
Quote
Quote
So are you saying that since i only get a 3/3 but repeatable 3/3 after some times rest that the differences are not there with these files? They are subtle enough that a little fatigue can make the differences insignificant but I really do hear the difference. It is consistent with all samples i tested.[a href="index.php?act=findpost&pid=344078"][{POST_SNAPBACK}][/a]

What's so difficult in getting a 8/8 score if you can get consistent 3/3? When you get tired you take a break and then continue after you rested...
[a href="index.php?act=findpost&pid=344082"][{POST_SNAPBACK}][/a]

In that case, I can do it. I'll make it 30 guesses. I'll get back to this thread in a week