Fun listening test - MP3 vs. CD

I have a portable CD/MP3 player which doesn't handle any other file format...

I have a portable CD/MP3 player which doesn't handle any other file format...

Well, yeah.  So do I.  I basically meant "the most widespread compressed lossless format".  I THINK that'd be WMA.

I basically meant "the most widespread compressed lossless format".  I THINK that'd be WMA.

Is it really that widespread?  For starters, I never even realised that there was a lossless WMA.  When talking about lossless formats, I'll immediately think of FLAC and Monkey's Audio (APE) for starters, and I know there are several others.  But WMA I associate with platform dependencies and DRM.  So maybe I'm just biased against it.

I've certainly read more about live recordings distributed in FLAC format than other lossless formats. And I believe that FLAC, by design, is patent unencumbered.

While Blurk makes a great point, we would do well to consider that the under-publicized WMA-lossless is playbable now on millions of Windows PCs with no additional software needed (though maybe a codec) -- and that makes it the most widely playable lossless format.  But that doesn't mean that it is the most widely used, or most popular, of course.  Funny how little things like semantics gets in the way, you know?

Bootleg recordings are something of a specialty of mine.

FLACs are, by far, the most popular lossless compression format for bootleg recordings.  SHN <Shorten> and APE <Monkey's Audio> are the two other commonly found and accepted formats for most bootleggers who demand lossless compression.

Blurk is also correct in his comments regarding FLAC's being patent unencumbered.

I couldn't detect the change, or ABX the two files.

Not surprising, given the high bitrate, unchallenging content, and near mono recording encoded in joint stereo!

Two (slightly constructive) comments...

1. In theory, you shouldn't use noise shaped dither when creating files which will be mp3 encoded. It looks like you have here. Use spectrally flat dither, or an encoder which accepts 24-bit input files.

(I say "in theory" because in practice, it sounds fine - though some people/bats might hear the problems!)

2. The artefacts from the audio restoration (identical in both the .wav and .mp3 versions) are far far greater than any problems you would ever hear from mp3 at that bitrate (which, apart from very extreme problem samples, is nothing). That's not to say they're really objectionable, but I can hear the de-noising. I think you'd have to get below 80kbps on that sample before mp3 introduced anything more annoying.


I don't know how you convince people who have heard "mp3 sounds bad" - I think your experiment page is a good start! Mind you, if they're the kind of people who believe that audio cables are directional, you're going to have a hard sell!

The reason for buying the CD would be if they wanted to re-encode the content themselves for their own use, either in a different codec or a lower bitrate. Here uncompressed .wavs are superior, because transcoding reduces quality. That said, I think they'd get away with it here!

Cheers,
David.

For the record, the final 16-bit files are dithered in preparation for CD burning - we then use these as the masters for MP3s, hence the noise-shaped dither...