iTunes 128Kbps AAC - is it really that bad?

iTunes music store finally became available in Portugal a couple of weeks ago so on the weekend I signed up and checked it out. Just to try it out I downloaded the new Caetano Veloso album "A Foreign Sound" (see here).

Cost me €9.99 for a 23 track album. Not bad considering that to download each track individually would cost €0.99 (which, at around 66p is considerably cheaper than the 79p per track that iTunes charges in the UK ).

So, how bad is this 128Kbps AAC rip then? Well, it's not bad at all. In fact it's pretty bloody amazing . I used iTunes to burn the album to a CD - it does the conversion to 16/44.1 WAV and the burning, really, the more I use iTunes the more I love it. It's just SOOO easy to use and everything just works so easily and intuitively.

Anyway, for comparison, I bought a copy of the original CD yesterday and this evening I've been comparing the two listening to them on my system.

The first thing that amazed me is that the CD that was made from the 128Kbps AAC files (I'll call it the rip CD from now on) sounds very, very good. If I'd bought that as a CD I wouldn't have any complaints. There's nothing obviously "wrong" with the sound and absolutely no audible compression artefacts such as the digital "phasing" sounds you can sometimes get on poor quality internet radio.

Playing the the rip CD and the original CD back to back I really do struggle to tell the difference if the truth be told. The original CD is perhaps marginally better in terms of focus, soundstage depth and "ambience" but it really is the tiniest of differences. Tonally there's nothing to tell them apart.

So, the next thing I did was open the same track from each CD up in SoundForge to compare the WAV files and, as you can see, they are as good as dammit identical.



The top one is the rip CD, the bottom one is the original CD and the portion of the WAV shown is the first sounds from Track 1 (Carioca).

I'm not that familiar with using SoundForge (just downloaded the 30 day trial version) so if anyone knows a better way to compare two .WAV files let me know. Still, from what I can see it's clear the differences between the rip CD and the original are miniscule.

So, the 128Kbps files on iTunes Music Store that we've all been slagging off as unlistenable rubbish just about tolerable for portable listening, are actually pretty damn good. So, before you knock compressed formats, give them an objective listen. You might be surprised

Michael.

 

Can't really add any input on the quality issues etc however you can get practically every album you can imagine on allofmp3.com for around $0.60 128 AAC. Thats a whole album for around 35p! When you think that on itunes you pay double that for just one track its worth a look. Its apparantly legal too.

 

Just checked out allofmp3.com and it looks decidedely dodgy. It's based in Russia for a start. It's run by a company called "Media Services Inc." in Russia and if you go to their homepage (http://www.mediaservices.ru/index.shtml) they boldly claim a THAWTE SSL certificate but if you click on the link, it's an invalid cert. There's also no physical (address, phone no.) contact information.

Lastly, there's simply no way it can be legal with entire albums being sold for $0.60. With iTunes struggling to get the record companies on board I just don't see how they could possibly have agreed to those absurdly low prices.

Michael.

 

You might as well just use P2P, it may be illegal, but its free. I wouldnt pay anyone for music compressed using a lossy format to be honest.

And Mike, you like a 128K file? You call yourself an audiophile?

 

Try spectral analysis. You can use EAC for this task.

Here's what you should do, to get the frequency and spectral analysis:
1. Open EAC
2. Click on "Tools->Edit WAV" ( a file-open dialog will open)
3. Choose you wav-file (A new dialog is opening)
4. Click on "View> Spectral analysis" (the presentation of the wav
file will change)
5. Make a screenshot and save it
6. Click on "View->Frequency analysis". (the presentation of the wav
file will change).

You might also want to look at some of these links

Source analysis utilities

http://www.ff123.net/peacefulspecview.html
http://sd.lecastel.org/mp3.html
http://www.lincomatic.com/mp3/mp3quality.html
http://www.personal.psu.edu/users/b/x/bxg178/MP3freq/mp3wavcomp.html
http://www.raze.host.sk/mp3facmp.htm
http://www.audiohub.org/get/fa/sa.htm
http://www.audiohub.org/get/fa/fa.htm

AuCDtect http://www.true-audio.com/index.php?menu=aucd

EAC Used for ripping CD-DAs to HD, it can also do spectral and frequency analysis, see "Tools". http://www.exactaudiocopy.de

Instructions and examples on detecting MP3 sources in your .wav files.
http://www.soppybagrecords.net/pages/mp3_tutorial.htm

 

Michael,

The differences you described could just be down to the fact that you actually knew which CD was the original and which was the downloaded copy. I don't want to get into DBT issues, but if someone else was doing the swapping for you, I doubt if you would perceive differences in focus, soundstage depth and "ambience" as the graphs that you have posted are even much closer than your descriptions of the two discs.

However, things like reduced focus, soundstage depth and "ambience" (if true) are enough reasons to make many audiophiles avoid such downloads.



Enjoy the music,

Lawrie.

 

I'd challenge anyone here to be able to tell which CD was which, played through a system of their choice. I reckon most people wouldn't be able to reliably identify the 128Kpbs rip CD.

EDIT: You see Lawrie, I pre-empted you

Cheers Lee - I'll play around with EAC a bit.

Michael.

 

Mike make a screen shot of each wav in SF import to photoshop and make each one a layer, reduce transparency of top layer and superimpose one over other, any discrepancies should be obvious.

 

OK, spectral analysis reveals all (cheers Lee). I don't claim to understand exactly what I'm looking at in a spectral analysis (someone who knows, please chip in) but it would appear that pretty much everything above about 15kHz has been unceremoniously dumped .

I still stand by what I said before though. I doubt that many people would be able to reliably tell which track was from the rip, but the differences I heard when I knew which was which (reduced focus, soundstage depth and "ambience") are consistent with reduced high frequency information.



Michael.

 

The comment as I'm sure you realised was made in jest. In effect you are preaching to the converted with me at least, since I happily have listened to 192K MP3's for years and never felt unsatisfied.

 

I did realise Paul

Michael.

 

remastered edition of fleetwood mac rumours arrived today. figured i'd try itunes to rip it as a test. and it completed quite quickly. so now i have it in 128kbps aac formats. i then ripped it again with cdex in 128kpbs mp3's, so i could play both back to back.

can i tell the difference ? no.

soundcard is a videologic sonicfury, which apparently plays mp3 files in hardware, so maybe that helps ?

kept the cd in the drive, and used itunes to play the cd itself.

can i tell the difference ? no, not really.

now let's try the cd in the arcam ....... i think i must be going deaf

dont seem to notice going from cd to pc, but pc first then playing the cd afterwards, and i can.

conclusion ? gonna stick with mp3's rather than ripping them again in aac's, as i've done nearly 200 cd's already. (for my pc library, which is just for background "party" music).

 

Just a quick word about that, the Sonicfury will only decode MP3 files in hardware, not anything else like AAC, and even then IIRC it will only do it with certain drivers and player software.

It doesnt matter anyway, because decoding MP3's only takes up a very small fraction of what a modern CPU can process anyway.

 

The difference is quite marked, as much as I like apple, AAC is not the be all and end all.

I purchased Millions now living will never die (Tortoise) and played it about three times before putting it away. A month ago I bought the actual CD and its totally different and a stunningly good album.

Basically all the elements are there, but there is no soul, or decent top or bottom end LOL.

 

Interesting the difference between the two spectra...

The AAC is clearly strongly band-limited to 16kHz, with a few higher-frequency artefacts appearing when it was converted to CD format. At the scale, it's difficult to see other differences for sure although there appear to be some.