Debunking the higher the bit rate the better the sound quality myth

[Nostromo]

Can you really tell the difference between a song encoded at 128 Kbps and the same song encoded at 320 Kbps? Audiophiles will tell you that the one encoded at 320 Kbps will sound a lot better. Here's a way to find out, objectively.

Step 1. Download Foobar 2000. Don't click on the normal installer, click on the special installer instead. What you need for the test comes with the special installer.

http://www.foobar2000.org/

Step 2. Encode songs you know well at 128 Kbps and at 320 Kbps. I used different types of music, just to make sure:

- Siegfried's Death & Funeral March by Wagner

- Miles Davis' Blue in Green

- Nick Cave and the Bad Seeds' Where the Wild Roses Grow

- Boards of Canada's Music is Math

Step 3. Import these files in Foobar. Click on "Playlist" and then select "Add files..."

Step 4. Select the 128 Kbps and the 320 Kbps version of the same song, right-click and select "ABX Two tracks..." The ABX comparator will open. Its a program for double blind listening tests. The program assigns randomly one version of the song to A, and the other to B. It does the the same with X and Y. You have to find out if A is X or Y, and if B is X or Y. After you made your choice, the program tells you in the status zone if you were right. It also tells you the probability that you were guessing.

To make sure it wasn't a lucky guess, you have try the test a number of times. That's why you have to click on "next trial" after you give your answer.

The results? I admit I can't tell the difference, even though I have decent headphones (Sennheiser PX-100). Otoh, I have crappy integrated sound card. Anyway, if there's difference, its very subtle.

[Verto]

too much effort

[Urban Ranger]

Double blind tests are the w00t

[SlowwHand]

Per M.I.T , 224 is cd quality MP3, however a superior sound card is needed, 192 is just as good. 160 acceptable. 128 is crap.
So anything over 192 is overkill. All my mp3's are 192, for this reason.

[SlowwHand]

On top of that, the most accurate is actually vbr, set to a mean of 192.
If the ripping program or bitrate conversion program uses Lame technology as actual compression agent, it's possible to see the fluctuation in quality, not only between 2 entirely different cd's , but between tracks of any given cd.
This is more apparent in a ripping program such as Exact Audio Copy.
dBpowerAMP, a bitrate conversion program, uses Lame also.

[Nostromo]

Originally posted by SlowwHand
Per M.I.T , 224 is cd quality MP3, however a superior sound card is needed, 192 is just as good. 160 acceptable. 128 is crap.
So anything over 192 is overkill. All my mp3's are 192, for this reason.

FYI, I used AAC. Have you tried the test? Can you actually hear the difference between a song encoded in MP3 @ "cd quality" 224Kbps and the same song @ "crappy" 128 Kbps. You could be surprised...

[SlowwHand]

I guess my point is I try to have it reproduced as artist intended, giving myself the chance to hear it, whether I can or not; rather than degrading it and giving away all possibility.

[Asher]

AAC is much, much better than MP3. Try your test with MP3.

[Nostromo]

Some people claim they can hear the difference. From the iPod review at 6moons, a website for audiophiles.

http://www.6moons.com/audioreviews/ipod/ipod.html

FYI, he's comparing the iPod to a 4K$ CD player. So its not exactly a fair comparison But he plugged the iPod in a high-fi system.

Now I'll switch to my next question. How does the sound hold up when you reduce the bit rate? On most tracks, I found only small differences when moving from lossless compression to AAC 320kbps (the highest of the lossy AAC compressions). That's good news indeed since the file size is less than half that of the lossless compression. This means you can get more than twice as much music onto your iPod and not really be much the worse off for it. In fact here's a table showing you the file sizes of one track under a variety of sampling rates:

Uncompressed - 70.0MB
Lossless - 35.6MB
MP3 320 -17.0MB
AAC 320 -17.0MB
AAC 192 - 0.3MB
AAC 128 - 6.9MB
AAC 64 - 3.5MB
AAC 16 - 0.9MB


The track in question is the second movement of Haydn
String Quartet Opus 20 No 2 which lasts 7m23s. On the Meridian, the period instruments exhibit subtlety of sound, strong dynamic contrasts and a sweet string tone. The treble is refined and extended and the image is suspended across the room. The harmonics are subtly captured and there is a most musical feel to the presentation. This is a superlative achievement since this is one of the hardest discs to get right. Many a system sounds edgy and thin on this material. Switching to the iPod Mini and using lossless compression, the first thing you notice is that the pace of the music has slackened and the image has receded towards the speakers.


The instruments maintain their color and the distortion remains low but there is a considerable loss of body in the cello and to a lesser extent, the violins. But there is no edginess to the sound and that is quite unexpected. Rather more disappointing is that when there is a strong sound from one instrument, it seems to drown out the quieter instruments, indicating a power supply that can't quite keep up. Switching quickly back to the Meridian, the image expands and fills in the gaps between the speakers. The dynamic contrasts are more fully realized and the string tone is sweeter, the treble more extended and there are harmonics I didn't know I was missing through the iPod


Stepping down a notch to AAC 320, the sound is not quite as clear as the Lossless Compression, the string tone slightly homogenized but the differences are small. The string tone remains sweet most of the time, with occasional patches of roughness, and there is a slight but discernable loss of body and purity of tone. MP3 is not as efficient as AAC and this was clear when I tried MP3 at its highest setting, 320kbps. While the file size is the same as the AAC 320, the sound quality is seriously impaired. The openness at the top diminishes and the viola loses presence, often being masked behind the violins in unison passages. There is less detail in the tonality of the instruments and some of the fine delicacy achieved in shaded passages is lost. It's harder to make out that these are period instruments. But -- and this is a big but -- the sound is still very listenable, coherent and colorful.


Having established the superiority of AAC over MP3 at least on the iPod, I concentrated on the AAC performance at successively lower bit rates. 192 kbps gives a further loss of body and an almost complete collapse of the image size. You could drive a truck between the speakers and not hit a musician. At this level, leading edges are lost and the music fails to convince. Even the color of the instruments pales.


128kbps continues the story, the sound diminishing to just suggestive of what might have been. At 64kbps, we are listening behind a thick curtain, all sense of pace is missing as are bite, detail, resonance and presence. I stopped at 16 kbps. It sounds like a wind-up gramophone without the hiss.

This guy either has magic ears or he's full of ****. I wonder if would be saying all this in a double-blind test? Maybe I could tell the difference on a high-fi system. Or at least with high-fi headphones...

[Nostromo]

Originally posted by Asher
AAC is much, much better than MP3. Try your test with MP3.

I can't tell the difference between a song encoded in MP3 @128Kbps and the same song encoded in AAC @128Kbps.

[Asher]

Then you must be listening to this on a crappy system (bad sound card or iPod).

[Nostromo]

That may well be it. Obviously, I did the double blind test on my PC, who has an integrated sound card. My headphones are decent, though.

[Mr. Harley]

Guys - you also need Good speakers, preferably the $500 plus type. If you are using the vast majority of headphones, or even the typical $70-$200 set of computer speakers, you may well be unable to hear the difference (which is why having all your songs for Ipod and auto play at 128 bit rate makes sense). However, run your output through a good stereo with great speakers, play the CD on your stereo while simultaneously switching back and forth to your computer. You will hear the difference in that case. So it's all context.

[Nostromo]

Originally posted by shawnmmcc
However, run your output through a good stereo with great speakers, play the CD on your stereo while simultaneously switching back and forth to your computer. You will hear the difference in that case. So it's all context.

Unfortunately, that wouldn't be a double blind listening test. You want to make sur its not the placebo effect you're hearing.

[Agathon]

Play it through a real hi fi system. You'll hear the difference immediately.