Acc or MP3 , which is better sound quality?

GO TO ADMIN PANEL > ADD-ONS AND INSTALL VERTIFORO SIDEBAR TO SEE FORUMS AND SIDEBAR

CraigNB

New member
Joined
Oct 14, 2005
Messages
487
Points
0
Age
37
Location
Boston, MA
Website
www.nbrock.net
AAC provides audio encoding that compresses much more efficiently than older formats, such as MP3, yet delivers quality rivaling that of uncompressed CD audio.

a 160 kbps AAC file is similar in quality to a 192 kbps VBR MP3 file (with the AAC file weighing in at less than 90% of the equivalent MP3 file size)
 

lkalliance

Maker of Grilled Cheese
Joined
Oct 22, 2004
Messages
382
Points
16
Location
Minneapolis, MN
As Craig mentioned, it would depend on how you're compressing. According to some, ACC sounds better than MP3 at the same bitrate. But as it is subjective, there's no universally agreed-upon answer. Some people will also be quick to point out that MP3 is a more commonly-used standard, which would outweigh any miniscule sound quality improvement.
 

kornchild2002

New member
Joined
Feb 21, 2004
Messages
12,193
Points
0
Location
Cincinnati
Please, please, please, please, please search the audio forums for many common questions/debates such as this one as this topic has been discussed 2398475982347968032098 times.

However, since I like helping people pick a digital audio format for their iPod, I will answer.

First off, it all depends on what mp3 encoder and mpeg-4 AAC encoder you are talking about. There are many mpeg-4 AAC encoders out there: WinAmp, FAAC, Nero, iTunes/QuickTime, etc. There are also a lot of mp3 encoders out there: Lame, Fraunhofer (used in Windows Media Player, free version of Winamp), iTunes (based off of Fraunhofer), etc.

As far as mpeg-4 AAC encoders go, the iTunes and Nero AAC encoders are statistically tied but iTunes is free and provides slightly (not statistically) better sounding files. As far as mp3 encoders go, the iTunes and Fraunhofer versions absolutely suck. These two mp3 encoders belong in the group of 10 worst mp3 encoders out there, and this has been proven with numerous tests. The Lame mp3 encoder is the best mp3 encoder out there as it is free and provides high quality files.

Now, lets compare iTunes AAC with iTunes mp3. Craig would be 100% correct on this. A 128kbps VBR AAC will have the same, if not better, sound quality as a 160kbps mp3. A 160kbps VBR AAC will have the same sound quality as a 192kbps VBR mp3. This is saying a lot that a new format can compete like this. However, the iTunes mp3 encoder hasn't seen a major update since 2000 (had a bug fix in 2001) making it pretty damned old.

Now, lets compare the Lame mp3 encoder and the iTunes mp3 encoder. The same thing pretty much goes. A 128kbps VBR Lame mp3 would have the same sound quality, if not better, than a 160kbps VBR iTunes mp3. Why would two different encoders who use the same format (mp3) have different sound? Well, the Lame mp3 encoder is constantly being developed. The Lame mp3 encoder is more efficient, has been tuned over the years, and has much better VBR encoding capabilities.

Last part, lets comapre the Lame mp3 encoder and the iTunes AAC encoders. Here, Craig would be 100% wrong. Statistically, the Lame mp3 encoder and iTunes AAC encoder will both provide the SAME quality at the SAME settings. A 128kbps VBR iTunes AAC will sound the same as a 128kbps VBR Lame mp3. The iTunes AAC encoder will provide slightly higher quality than the Lame mp3 encoder but this is not much: on a scale of up to 5, Lame 128VBR received a score of 4.55 while iTunes AAC 128VBR received a score fo 4.68. This small (0.13) difference is not enough to be statistically different.

Who conducted all these tests? Two groups/people. One person over at hydrogenaudio.org (THE definative source for audio comments/questions/news) conducted one set of tests and another group over at hydrogenaudio.org conducted the other set. The group of people were comprised of average listeners and hard core audio enthusiasts (I hate the term audiophile as true audiophiles won't admit that a lossy encoder [mp3, aac, etc] can perform well and they like to use fluid terms). So, just for you knowledge, these tests conducted are accurate and can be trusted.

Now, what the hell does all this mean? Personally, I wouldn't use anything other than Exact Audio Copy (EAC, exactaudiocopy.de) and Lame 3.97b2 (rarewares.org) to rip a audio CD. The Lame mp3 encoder will provide you with high quality files that can directly compete with the iTunes AAC encoder. If you want high quality, universal compatibility, and the fealing of safety when ripping your audio CD's, then use EAC and Lame. However, if you find that you will be living in a iPod world, plan on playing back your music with your iPod and computer, and don't really care about universal compatibility, then the iTunes AAC encoder will still give you high quality files.

Hey, but what do I know? The best way for you to pick your format and bitrate is to conduct a BLIND ABX test. That way, when you are listening to a ripped track, you can personally judge if it sounds like the source CD or not.

You can listen to all the people you want, I just hope that I gave you enough information to start. Out of all the people on these boards, only one can be trusted; yourself. Only your ears will tell you if something sound "CD quality" or not. I personally find that 128kbps VBR AAC sounds pretty good but 160kbps VBR AAC is more "CD transparent" and that 192kbps AAC is "CD Transparent." -V 2 --vbr-new Lame mp3's are my personal choice for backing up my music. This setting will provide mp3's that are around 190kbps VBR.

Good luck.
 

RGertner1

New member
Joined
May 3, 2005
Messages
25
Points
0
Location
Poughquag, NY
Website
homepage.mac.com
Also, one thing to add (just from personal expirience) I have found encoders work better for different types of music, and/or what you prefer to hear in your music.

I performed the following, and these were MY results:
I took 2 songs that I had the original CD and purchased from iTunes Music Store. These were Michael Bublé - You and I and Rob Thomas - Lonely No More. I made the following Imports: iTunes/AAC 192 CBR, Apple Lossless, EAC/LAME 3.96.1 192 VBR, and AAC 128 CBR purchased from iTunes Music Store.

I then created 2 playlists and put all 4 copies in each playlist. I then played each on my iPod Mini using Sony Studio Headphones
Rob Thomas:
MP3 (LAME) - 4 Stars
AAC (iTunes) - 5 Stars
AAC (ITMS) - 2 Stars
AAC (Lossless) - 4 Stars

I actually really couldn't tell the difference between any of the non-iTMS songs, I think it was just subjective on maybe some imaging (echo's in voice or base, etc) But I did chose the iTunes AAC as the best sounding.

Michael Bublé:
MP3 (LAME) - 3 Stars
AAC (iTunes) - 3 Stars
AAC (ITMS) - 4 Stars
AAC (Lossless) - 5 Stars

This one kinda suprised me when I sync'ed my iPod back to iTunes. The Apple Lossless sounded great, but the file I purchased from iTMS I thought sounded the most like the Lossless or at least the 2nd best to me. Where the 192 AAC and 192 MP3 (VBS) sounded fine but seemed to be missing something that the Lossless and iTMS version had.

Personally, until I performed this test, I never really noticed a difference in my iTMS vs music I have imported at AAC 192 CBR. I would/am perfectly happy with any of the above formats and won't be re-importing any of my CD's anytime soon.

Just my 2 cents on the whole matter.
 
Last edited:

mnhnhyouh

New member
Joined
Feb 27, 2006
Messages
1,759
Points
0
Location
Chengdu, China
Website
www.geocities.com
You need to have another program run the LAME program, or supply it with a command line from a dos prompt.

You can use Exact Audio Copy (see button in my sig) or many other programs. Unfortunatly you cannot get iTunes to update its encoder :(

h
 

DrewT

Member
Joined
Jul 21, 2004
Messages
230
Points
16
Age
36
Location
NC, US of A
Website
cntp.rh.ncsu.edu
CDex provides a decent interface to LAME (at least used to)

Even if iTunes AAC is slightly better than LAME MP3's I'd stick to the MP3's (hey, if you don't have a lot of songs, you can bump up the bitrate!), simply because if you buy another brand music player basically ever, the files will work and won't have to be reencoded. Plus - I doubt most people will listen to their iPod on studio headphonse in a quite room.

My advice: stick with MP3

-Drew
 

vic6string

New member
Joined
Apr 1, 2006
Messages
2
Points
0
Actually, I think compatibility will be less of an issue as tiem goes by. Apple has such a huge market share that the other brands will have to offer AAC compatibility to draw people away from th ipods. Most newbies leave all the default settings at first, so there are millions of people out there who have ripped their CDs to AAC. If Sony, RIO, and the rest of em don't play AAC, they will never get these people to switch. Personally, I just stick to the default 128k AAC since 99% of the time i use the iPod for just casual listening and in that regard you really can't tell a difference. If you only use the ipod for jogging or listening in the car (remember that if you use an FM transmitter you will never get CD quality no matter how you rip it, in fact it will be less than high quality audiocassette) than the format really makes no difference. When I want to really listen to music (which I rarely have the time to do anymore) I bust out the actual CD and pop it in my yamaha with the MIT cables and Boston Acoustics reference speakers and the Energy Sub, or pull out my Senheiser headphones. Going with the 128k AAC gives me near CD quality and allows me the space to have ripped movies and TV shows on my 60 gigger.
 

Menko

Music Pimp
Joined
Jan 9, 2005
Messages
391
Points
0
Location
UK
I rip my music promarily for listening on my computer, so I need something a little better than 128Kbps AAC. After my tests I found that the next best thing to lossless (not an option with 10,000 songs!) was 192Kbps VBR AAC.

Advanced Audio Coding (which is apart of the MPEG-2 and MPEG-4 standard) was developed by the same people who developed MP3 (which is a part of the MPEG-1 standard). AAC is much more efficiant than MP3 and shouldn't take too long to become the new standard.

The only reason AAC isn't as common as MP3 at the moment is because "MP3" has become synonymous with "digital music", and is even often used to mean "digital audio player". MP3 is actually short for "Motion Picture Experts Group Layer-3"... just think about that before you use say something stupid like "Oh yeah, I've got it on my motion picture experts group layer-3".
 

kornchild2002

New member
Joined
Feb 21, 2004
Messages
12,193
Points
0
Location
Cincinnati
I really want to put an end to all these efficieny posts. Sure, the iTunes AAC encoder provides better efficiency when compared to SOME mp3 encoders (Frounhofers, Helix, Blade, etc.). This means that a 128kbps iTunes AAC has the same quality, if not more, than a 160kbps mp3 made with those useless encoders.

Now, THIS IS NOT THE CASE when comparing iTunes AAC with the Lame mp3 encoder. Not all mp3 encoders are the same, it doesn't matter if they all output to the same format (mp3), each encoder performs differently. The same is said for AAC encoders, not every AAC encoder performs the same.

Now, in recent tests, the iTunes AAC encoder and Lame mp3 encoder were tied at the 128kbps VBR bitrate. What does this tell us? It means that the Lame mp3 encoder will provide you with roughly the same efficiency as the iTunes AAC encoder. Sure, the iTunes AAC encoder did provide SLIGHTLY higher results (a difference of 0.2 is not big) but these results were NOT statistically significant. So the iTunes AAC encoder and Lame mp3 encoder will provide you with the same results at their respective bitrates (ie 128kbps vs 128kbps, 160kbps vs 160kbps, etc.).

I do agree that the iTunes AAC encoder is high quality, there is no denying that. There is also no denying the potential of the AAC encoder. The only problem is that all the "great features" that are in the AAC format aren't really seen yet. The AAC encoder may have 9374597 more features than the mp3 format but these features aren't heard yet when comparing it against the Lame mp3 encoder.

Now, I have to say this. If it wasn't for the Lame mp3 encoder then all these efficiency posts would be 100% correct. However, the Lame mp3 encoder is widely used and heavily developed and is free. If it weren't for the Lame mp3 encoder than all of my music would be in the AAC format as it does provide better quality than the iTunes mp3 encoder (Frounhofer), Windows Media Player mp3/wma encoder (Frounhofer/Microsoft respectively), MusicMatch mp3 encoder (Frounhofer), RealPlayer mp3 encoder (Helix), and the Blade mp3 encoder (there are 100's of mp3 encoders out there).

So, to re-define it all. The AAC encoder is more efficient than MOST mp3 encoders out there but it is not more efficient than all of them, mainly the Lame mp3 encoder.

I don't want the OP to get confused as to all the contradicting posts. iTunes AAC = Lame mp3 in terms of quality and efficiency.
 

mnhnhyouh

New member
Joined
Feb 27, 2006
Messages
1,759
Points
0
Location
Chengdu, China
Website
www.geocities.com
kornchild2002 said:
Sure, the iTunes AAC encoder did provide SLIGHTLY higher results (a difference of 0.2 is not big) but these results were NOT statistically significant.
Just to make clear what statistical significance means.

It is based on the presence of variability in results.

If I have a coin, how do I tell that if is biased? I toss it three times. It comes up with three heads in a row. Does this mean it is biased?

Statistically heads will come up three times in a row 1 in every 8 lots of three tosses. So if I do 800 lots of 3 tosses, 100 of them will have 3 heads. That is with an unbiased coin!

Statistical significance is (mostly) based on there being a less than 1 in 20 chance that the result comes from no difference. So back to the coin analogy, 4 heads in a row is a 1 in 16 chance, so that would not be a significant result, but there is only a 1 in 32 chance of 5 heads in a row. So if we were testing coins for bias we would throw out all those that gave us 5 in a row, knowing that we would be getting it wrong less than 1 in 32 times.

Now back to the original 0.2 difference in the test above. Given that the results are not significant what we are saying that there is no difference. Calling it a slight difference is not true. When viewing non-siggnificant results we should realise that it is likely that if we repeated the test there is an equal chance that the results would be the other way around ...

h
 

kornchild2002

New member
Joined
Feb 21, 2004
Messages
12,193
Points
0
Location
Cincinnati
mnhnhyouh said:
Just to make clear what statistical significance means.

It is based on the presence of variability in results.

If I have a coin, how do I tell that if is biased? I toss it three times. It comes up with three heads in a row. Does this mean it is biased?

Statistically heads will come up three times in a row 1 in every 8 lots of three tosses. So if I do 800 lots of 3 tosses, 100 of them will have 3 heads. That is with an unbiased coin!

Statistical significance is (mostly) based on there being a less than 1 in 20 chance that the result comes from no difference. So back to the coin analogy, 4 heads in a row is a 1 in 16 chance, so that would not be a significant result, but there is only a 1 in 32 chance of 5 heads in a row. So if we were testing coins for bias we would throw out all those that gave us 5 in a row, knowing that we would be getting it wrong less than 1 in 32 times.

Now back to the original 0.2 difference in the test above. Given that the results are not significant what we are saying that there is no difference. Calling it a slight difference is not true. When viewing non-siggnificant results we should realise that it is likely that if we repeated the test there is an equal chance that the results would be the other way around ...

h
Thanks for explaining the statistical significance jargon. What I ment by slight difference was that, to the average person, when they see iTunes AAC getting ~4.5 with Lame getting ~4.3, they see it as being that iTunes AAC has a slight lead over Lame. However, as you pointed out and as the results explained, this is not true. The test could be conducted again and the results could be switched, they could be exactly the same, or they could be totally different.

I was just tired of all the "efficiency" posts as, it is true that iTunes AAC is more efficient than some mp3 encoders, but it is not more efficient than the mp3 format as the Lame mp3 encoder has proven this in recent tests.

Again, thanks for explaining the statics.
 

Ice Weasel

New member
Joined
Jul 23, 2007
Messages
1
Points
0
mnhnhyouh said:
Just to make clear what statistical significance means.
...
Now back to the original 0.2 difference in the test above. Given that the results are not significant what we are saying that there is no difference. Calling it a slight difference is not true. When viewing non-siggnificant results we should realise that it is likely that if we repeated the test there is an equal chance that the results would be the other way around ...

h
If the difference is not significant- it is a slight difference and there's no reason to say there's "no difference". The difference is slight - as revealed by the data - just not significant - as according to the stat test. There may or may not be a "real" difference - it just wasn't detectable by the statistical test. We don't know how things would come out if we did the test again (we don't know if this is like flipping a coin or not), and in fact, the more comparisons you make, the bigger your sample size, and the more power/ability you have to detect (statistically speaking) a small difference.

None of this changes the bottom line that the two encoders perform basically the same at those settings, but if we're talking Stats...
 

kornchild2002

New member
Joined
Feb 21, 2004
Messages
12,193
Points
0
Location
Cincinnati
I still hate it. I took Stat 391 last semester and that was the last math course I will have to take for my undergrad BS work. I am glad too. I am really tired of math and Statistics almost killed me (I am going for a Civil Engineering degree). The kind of math that we do in our engineering classes is totally different than the math they teach us in the separate math courses.
 
Top