MP3 vs WAV: The Blind Test - Can YOU tell the difference?

[Mr. Darcy]

iTunes Music Store gives you 128kbps files by default (in AAC, I know ... we'll come to that test in time Wink).

Not sure if y'all were aware that much of the EMI catalogue on iTunes is now (optionally) available in DRM-free 256 kbps AAC as part of so-called "iTunes Plus:"

http://www.apple.com/pr/library/2007/05/30itunesplus.html

[Mark]

Not sure if y'all were aware that much of the EMI catalogue on iTunes is now (optionally) available in DRM-free 256 kbps AAC as part of so-called "iTunes Plus:"

http://www.apple.com/pr/library/2007/05/30itunesplus.html

Yep, knew that, thanks.

[George]

Not sure if y'all were aware that much of the EMI catalogue on iTunes is now (optionally) available in DRM-free 256 kbps AAC as part of so-called "iTunes Plus:"

http://www.apple.com/pr/library/2007/05/30itunesplus.html

Yes, though I don't understand why it should cost more. 

[Mark]

Yes, though I don't understand why it should cost more. 

You'll never make a ruthless businessman, George.

[Mark]

Okay, ladies and gents. I've fixed the download issue.

If you go now to the Blind Test Site, you'll find you can play each sample right there in your browser window.

Hope this helps. And keep those guesses coming.

[Mark]

Quick update: There have been 70 visitors to the Blind Test Site (thank you to all), and a complete set of correct guesses which I received via PM. Someone else also posted correctly in this thread ... beyond which I can say no more. If you think it was you, PM me and I'll tell you. But please, don't then spoil it for others.

[Mark]

Okay, as some of you have complained about slow downloads from my site, here are the four clips via RapidShare:

Clip A

Clip B

Clip C

Clip D


If you click on the grey arrow download links on the Blind Test Site, you'll now be pointed to the above RapidShare links. And of course, you can listen from your browser at the site itself, too.

[Mark]

Another update: I've received a second complete set of successful guesses by PM ... and there have now been 81 visitors to the Blind Test Site. I can't say too much about the two sets of perfect guesses, except that they were arrived at under VERY different listening conditions. So, it currently looks like you don't need top-notch kit to take part in this experiment.

[George]

Another update: I've received a second complete set of successful guesses by PM ... and there have now been 81 visitors to the Blind Test Site. I can't say too much about the two sets of perfect guesses, except that they were arrived at under VERY different listening conditions. So, it currently looks like you don't need top-notch kit to take part in this experiment.

Does this include perfect guesses on this thread too?

[orbital]

No idea. C sounded better than D the first time I've heard, but then on second listen I can't really differentiate 

[M forever]

If you think it was you, PM me and I'll tell you. But please, don't then spoil it for others.

Don't tell *anyone*. They will communicate it to other people so they can all be "winners", provided they have any online "friends" here.

[Bunny]

And yet, there are so many people who confidently claim that they can hear the difference between compressed and uncompressed sound when comparing these common MP3 file bitrates to source WAV files.

I can't tell the difference on my computer.  This was not something I didn't know before.  The sound card and my desktop speakers just aren't good enough.  I would have to download the samples and burn them to cd and then play them on the big speaker system.  I know from experience that I can tell the difference between 128 mp3 and mp4, 128 and 192 mp4 and also 192 and 256 mp3.  I can't hear the difference between ALAC and wav files at all, but there is less "air" around files that are 256 than 320 kbps MP3.  I don't bother with mp3 because you really can hear the differences in that format.  Mp4 compresses more information into the smaller files so for anyone who's going to rip a lot of music, I would recommend that they not use the mp3 format at all.

Edit: For some reason, the crappy speakers in my husband's car really point up the differences in quality.  In the car, we can tell the difference between larger compressed files and ALAC pretty easily, which is scary.

[71 dB]

I can't hear the difference between ALAC and wav files at all.

No wonder, since ALAC is a lossless format. There is no difference.

Edit: For some reason, the crappy speakers in my husband's car really point up the differences in quality.  In the car, we can tell the difference between larger compressed files and ALAC pretty easily, which is scary.

This is not weird at all. Audio compression is based on masking effect, and poor frequency/time response can twist masking so that unwanted sonic things are revealed. The coding scheme assumes you to have a neutral sounding system. In that sense, a very good system may reveal less differences than a coloured and distorted system.

[The Mad Hatter]

No wonder, since ALAC is a lossless format. There is no difference.

How can a format be lossless? I would have thought that any compression ultimately involves taking something out?

[Tancata]

How can a format be lossless? I would have thought that any compression ultimately involves taking something out?

ALAC and FLAC compression is analogous to ZIP or RAR compression of other digital files - completely lossless. That is, on decompression a bit-perfect equivalent of the original WAV is returned.

Illustration: 11111111110000111 might be more efficiently stored as

"Ten 1s, 4 0s, 3 1s".

[71 dB]

How can a format be lossless? I would have thought that any compression ultimately involves taking something out?

As Tancata said, that's not the case. In lossless compression the data is stored more efficiently. Redundance is removed etc. Lossless music files are usually 50-70 % of the original size so the compression isn't huge but on the other hand all the information is there, 100 % quality. 

[rappy]

I think D sounded worst. Concerning the others... I don't know

[Bunny]

No wonder, since ALAC is a lossless format. There is no difference.

This is not weird at all. Audio compression is based on masking effect, and poor frequency/time response can twist masking so that unwanted sonic things are revealed. The coding scheme assumes you to have a neutral sounding system. In that sense, a very good system may reveal less differences than a coloured and distorted system.

Thank you for explaining this!  No one else has been able to tell us why this is so.  This is why I have ALAC files on my ipod whenever we do a long car trip, and am reluctant to buy any lossy compressed files.  The compressed files all have artifacts and sound awful; the ALAC files sound like a cd.  Btw, my computer automatically plays the files on the Windows Media Player which has DST (oops) SRS to improve the files, so it's really impossible for me to distinguish the files directly from the streaming audio.  In fact the files sound exactly the same and even audio clips on sites like Amazon sound decent on the computer.

I have a question about Lossless files: I know my audio players will play ALAC files when burned to cd or dvd, but is FLAC supported by most players?  Also what about WMA 9 (Windows Media Audio 9) LosslessWav, Pack and other lossless formats like MPEG 4 ALS?  How well supported are they by audio players?  And what is Ogg Vorbis -- lossy or lossless?  And what's MLP?  I saw a reference to in some literature about streaming video and satellite radio.  AFAIK, MLP is just a typo for MLB.

[The Mad Hatter]

ALAC and FLAC compression is analogous to ZIP or RAR compression of other digital files - completely lossless. That is, on decompression a bit-perfect equivalent of the original WAV is returned.

Illustration: 11111111110000111 might be more efficiently stored as

"Ten 1s, 4 0s, 3 1s".

As Tancata said, that's not the case. In lossless compression the data is stored more efficiently. Redundance is removed etc. Lossless music files are usually 50-70 % of the original size so the compression isn't huge but on the other hand all the information is there, 100 % quality. 

Ah, that makes a lot of sense. Thank you both

[71 dB]

Thank you for explaining this!  No one else has been able to tell us why this is so.

That's because I am an acoustics engineer. 

This is why I have ALAC files on my ipod whenever we do a long car trip, and am reluctant to buy any lossy compressed files.  The compressed files all have artifacts and sound awful; the ALAC files sound like a cd.

High bitrate lossy formats should not sound awful. Something is wrong if they do. In a car there's a lot of background noise etc. Who needs High End sound in those situations?  Frankly, if you find 320 kbps AAC or mp3 files awful in car the problem might be between the ears...

I have a question about Lossless files: I know my audio players will play ALAC files when burned to cd or dvd, but is FLAC supported by most players?  Also what about WMA 9 (Windows Media Audio 9) LosslessWav, Pack and other lossless formats like MPEG 4 ALS?  How well supported are they by audio players?  And what is Ogg Vorbis -- lossy or lossless?  And what's MLP?  I saw a reference to in some literature about streaming video and satellite radio.  AFAIK, MLP is just a typo for MLB.

I haven't used much lossless formats. I find them very weakly supported. My iPod does not support lossless sound.  On the other hand high bitrate AAC or mp3 gives good sound, at least for noisy outdoor listening.

Ogg Vorbis is a lossy format similar to mp3 and of similar sound quality.

MLP is Meridian Lossless Packing. A proprietary lossless compression technique for compressing PCM audio data developed by Meridian Audio, Ltd. MLP is the standard lossless compression method for DVD-Audio content (often advertised with the Advanced Resolution logo) and typically provides about 2:1 compression on most music material. All DVD-Audio players are equipped with MLP decoding, while its use on the discs themselves is at their producers' discretion.

MPEG-4 ALS is similar to FLAC in its operation. Simply put it is a quantized LPC predictor with a losslessly coded residual using Golomb Rice Coding or Bounded Gilbert Moore Coding (BGMC).

Lossless formats are similar to each other, some algorithms are just little more "intelligent" than other giving better compression ratio. In my opinion there is too many formats and they are too weakly supported. One well-supported format is what we need.