MP3 or AAC?

January 24, 2006 1:42 AM

MP3 or AAC?

I've noticed that AAC (.m4a) files tend to be smaller than MP3 files with the same bitrate. Is the sound quality the same?

Is there any reason to prefer MP3 to MP4 files? I would guess that the former are more widely supported?

posted by Grinder to computers & internet (21 comments total)

Yes, MP3s are much more widely supported. When I started using a non-Apple portable music device, I sure wished that I had been ripping my CDs in MP3 instead of AAC. Just a thought if you want to be a little less dependent on Apple (although some other companies make AAC-playing devices and software).
posted by grouse at 1:45 AM on January 24, 2006

From the wikipedia entry on AAC

AAC was designed as an improved-performance codec relative to MP3... Advanced audio coding (AAC) was promoted as the successor to MP3 (ISO/MPEG Audio Layer-3) for audio coding at medium to high bit rates.

Indeed AAC encoded files are higher quality as similar bitrates. The same is also true of Microsoft's Windows Media Format and the patent-free and fully open Ogg. But none are supported as well as good old mp3, even though they are all much better.

I personally prefer Ogg, because it's open, free and works real good.
posted by MetaMonkey at 1:48 AM on January 24, 2006

(*at* similar bitrates)
posted by MetaMonkey at 1:50 AM on January 24, 2006

I used to rip to Ogg. Then I bought an MP3 car stereo and I wished I'd just used MP3. MP3 is the lingua franca of music. Unless you know for sure that you'll never need to play your music on a different device, it's safest to stick to it.
posted by salmacis at 1:59 AM on January 24, 2006

Any two formats will be the same size at a given bitrate. The bitrate merely defines how many bits of file it plays a second!

AAC does sound better at a given bitrate than MP3, or alternatively, is smaller at a given quality level. Pretty much any lossy compression algo is better than MP3, since MP3 was the first. Anything that's worse basically doesn't get past the drawing board stage.

I still use MP3 though since it's more versatile. Everything supports MP3, discmen, dvd players, you name it.

Also note that all these formats have a lot of wiggle room: the encoder has a lot of choices to make about what information to throw out, so two MP3 encoders will probably come up with different results.
posted by aubilenon at 1:59 AM on January 24, 2006

That's what I thought - thanks for confirming.

If I decide to convert from MP3 to AAC later, will I lose quality?
posted by Grinder at 2:23 AM on January 24, 2006

The mere act of converting an MP3 to an AAC file may cause some minor dissolution of quality, but unless you're listening far too closely, I doubt it would severely impact your experience that much.

It depends a lot on the bitrate of the source MP3, though. Certainly, converting existing files to AAC wouldn't serve much of a point, unless you have a device that has a particular liking towards AACs, since the quality will not improve—the compression on the source is completed, the tops and bottoms of the frequencies that it cuts have been lopped off, and there's no getting them back. Unless you believe Creative's X-Fi soundcard-based hype. But that's a different story.

The way things are moving nowadays, versatility makes more sense to me than file size. Hard drives are getting dramatically larger, especially in portable devices. If you absolutely require the entirety of your library at your hip throughout the day, and have a massive quantity of songs, you might want to see how much you can squeeze out of a single KB, but MP3s will work in any device that dares to play music. Unless it's made by Sony. *cough*
posted by disillusioned at 2:32 AM on January 24, 2006

The most recent versions of LAME have shown a pretty dramatic improvement in consistency. If I understood the test I read correctly (I just skimmed it), LAME is now as good, with MP3, as iTunes is with AAC.

Apparently, earlier versions of LAME were a bit inconsistent... most files sounded awesome, but it had trouble with some pieces. It now apparently does an excellent job with everything they've tested it with.

LAME at --preset-standard is good enough for practically any use. If you're on good equipment, you might like --preset-extreme. These give you Variable Bit Rate(VBR) files... they use more bits in the tough spots, and less in the easy ones. I believe --preset standard averages about 145Kb or so, and --preset-extreme is about 185Kb.

Personally, I've just gone lossless with everything, as drive space is so cheap these days. I use all lossless in my home stereo, and ripped from the CD images to MP3 for the iPod. (I use --preset-extreme.)

Most people can't hear lossless versus MP3 (including me in most cases), but ripping music that way means you'll never have to do it again. You can easily and simply convert between lossless formats without losing anything... since they are, after all, lossless. :) And if you make an CUE/WAV image file with a program like EAC, you'll be able to recreate your original CD if you lose or damage it.
posted by Malor at 3:34 AM on January 24, 2006

Oh, upshot... you should probably use LAME and MP3, since it sounds just as good, and works on everything.
posted by Malor at 3:35 AM on January 24, 2006

You may want to experiment, but I'm pretty sure that if you compress your files from mp3 to ACC they will sound like utter crap (double compression = utter crap). You should re-rip from original .wav files or download acc files directly ripped from .wav.
posted by sic at 5:21 AM on January 24, 2006

The mere act of converting an MP3 to an AAC file may cause some minor dissolution of quality, but unless you're listening far too closely, I doubt it would severely impact your experience that much.

This is called transcoding, and it's a bad idea. It's like taking a Xerox copy of a Xerox copy. Things go from OK to bad quickly.
posted by deadfather at 6:17 AM on January 24, 2006

If you talk to the guys at Hydrogren Audio, they'll tell you that AAC offers an advantage at low bitrates (<128) but that advantage starts becoming less pronounced as you get higher and is pretty much gone around 192.
posted by deadfather at 6:23 AM on January 24, 2006

I rip everything into AAC. Then if I need mp3s for any reason, I use iTunes to convert the file to mp3. It leaves the original AAC behind and makes a duplicate in mp3.

For example, a friend of mine wanted to hear some music I had (by a band that allows recording and trading) but his portable player only would allow him to use mp3 (or WM). I converted the files to mp3, burned them to disk and then was able to delete the mp3s after I burned them to disk, leaving my AACs behind

It works for me. Just a thought. Good luck.
posted by terrapin at 6:33 AM on January 24, 2006

I rip everything into AAC.

Rip first to lossless (Apple/FLAC/SHN/etc). Then do your compression into lossy. Some programs will do this compression automatically during synch based on device profiles.
posted by meehawl at 6:48 AM on January 24, 2006

As most of my stuff is already ripped to MP3 (by iTunes), I think I'll stick with it for now. Re-ripping to lossless would take too much time and chew up too much disk space - neither of which are plentiful at the moment :-( - and I'd rather retain the flexibility of MP3.

I figured converting would be a bad idea. I must set up LAME though.
posted by Grinder at 6:55 AM on January 24, 2006

Oh, and thanks again for the tips.
posted by Grinder at 6:55 AM on January 24, 2006

I don't buy the WMA or AAC is better than MP3 because of age. I'd like to see an ABX test to confirm this, especially on typical consumer grade equipment.

MP3 is only a format and a slightly older one, but that doesn't mean all improvements on the encoding end have stopped. Considering most people can't even hear the difference between an MP3 and a CD, the difference between other lossy formats is trivial or at least just splitting hairs.

The nice part of mp3 isn't that its supported by everything, its that the format does not support DRM. So your mp3 is yours for all time. You can copy it here and there. You can share it on networks.

Using non-mp3 formats is just begging for vendor lock-in and once you go that route you really can't go back. Sure you can re-rip all yoru CDs again when you realize that your new player doesnt support that apple or ms stuff properly.

Also, when it comes to the format wars, your friends over at hydrogenaudio.com have something to say:

All modern encoders are not equals. Not at this bitrate obviously. Vorbis ends the test with a clearly superiority, whereas WMA doesn’t really show big difference compared to an old format as MP3. AAC (whatever the profile) is disappointing; the High Efficiency profile doesn’t help the AAC core to perform better at 80 kbps, and seems rather to handicap the format. Also interesting to note: MP3 at 128 (when encoded with LAME) is currently untouchable, except maybe by vorbis on various music. People shouldn’t seriously expect to put increase the musical content of their portable player by 100% and keep the same quality as MP3 128. Most people on HA.org probably knew that

The above is from a test testing 80kbs aac and wma to the standard 128 mp3. HA is full of other tests, but the general consensus is that there's nothing wrong with the mp3 format and that a 128 or 192 kbs mp3 is very close to the source as far as humans can tell. Thus the advertising that you can get away with 80kbs with aac or wma and get "128 mp3 quality" sound is bunk.
posted by skallas at 10:09 AM on January 24, 2006

I have a somewhat large quibble with skallas' post above. Nowhere does anyone state that 128kbps LAME-encoded mp3 is "close to the source." Far from it.

In fact, I'd venture to say that if you gave most people a nice, quiet room and good cans and asked them to ABX between the source and a 128kbps mp3, they'd do a decent job at it.

192kbps mp3 is a different story. Good ears can ABX it, but most probably couldn't. Variable bit rate mp3 at around 192kbps is hard for even good ears to distinguish, except on problem samples.

Furthermore, that test shouldn't be used as a basis for blanket opinions. It is a function of specific encoders performing at specific bitrates, and nothing else.
posted by deadfather at 12:21 PM on January 24, 2006

deadfather, in fairness I did write "128 or 192" kbps. Its way too easy to get into an audiophile wankfest over this but I was addressing the following.

1. The myth that AAC and WMA is so advanced you can get away with a lower bitrate and save disk space. At least one poster had claimed that in this thread Indeed AAC encoded files are higher quality as similar bitrates. The same is also true of Microsoft's Windows Media Format. This might be technically correct but real world testing shows that these lower-bitrate files do not match up.

2. The myth the because MP3 is slightly older means its obselete. MP3 is a good format to begin with and current encoders like LAME are excellent.

3. The myth that the MP3 format as used cannot produce excellent audio especially considering the hardware people are using.

4. The very real problem of vendor lock-in and using formats that support DRM.
posted by skallas at 12:57 PM on January 24, 2006

Be aware that none of the positive things people say about MP3's acceptability at 128kbps and up apply to the iTunes MP3 encoder, which has a reputation for being crummy. If you do MP3, LAME is going to be the way to go, ideally in conjunction with Exact Audio Copy (if you are ripping on a PC).

I think the DRM argument is a bit of a red herring. It is possible to DRM AAC, but iTunes doesn't do it when you rip your own CDs. I'm sure people could wrap DRM around MP3, if they wanted. In the process, they'd break playback on any device that didnt' support the DRM, just as FairPlay protected AAC is broken on any device that doesn't support FairPlay.
posted by Good Brain at 4:22 PM on January 24, 2006

I think I know what people will say, but what about the MP3 encoder built into Windows Media Player 10?
posted by flameproof at 3:52 AM on January 26, 2006