[LAU] How bad is mp3/ogg

[S. Massy]

On Wed, Oct 12, 2011 at 12:19:55PM +0200, Philipp ??berbacher wrote:
> Excerpts from S. Massy's message of 2011-10-12 00:23:42 +0200:
> > On Tue, Oct 11, 2011 at 05:35:09PM -0400, S. Massy wrote:
> > > On Tue, Oct 11, 2011 at 09:07:33PM +0000, Fons Adriaensen wrote:
> > > > On Tue, Oct 11, 2011 at 10:57:38PM +0200, Atte Andr? Jensen wrote:
> > > > 
> > > > > My question is: is this really a fair way to judge the artifacts  
> > > > > introduced by encoding?
> > > > 
> > > > No, it's completely invalid.
> > > > 
> > > > The correct way would be a double blind A/B/X test between the
> > > > original and the encoded versions.
> > > With suitable hardware. What I mean is, I think a great way of
> > > demonstrating the difference between lossy compression and uncompressed
> > > audio is to do an A/B test through a consumer device and then do it in the
> > > studio. The difference can be striking.
> > > 
> > > I like to think I have decent ears, and I can only very rarely tell the
> > > difference once over 192kbps. Though I've also found that bitrate isn't
> > > always everything (i.e some audio seems to respond better to a given
> > > compression algorithm than other). I wonder what other people's
> > > experience has been in that respect.
> > Responding to myself here. Out of curiosity, I just did some AB'ing
> > (hadn't in a very long time) and you can hear a difference even above
> > 192 kbps. A lot of it seems to have to do with emphasised frequencies. I
> > think Jostein had a very good point about these formats potentially
> > actually being pleasing to people because they make things sound
> > "bigger/punchier".
> > 
> > Very interesting...
> > 
> > Cheers,
> > S.M.
> 
> Was this a double-blind test? Otherwise you don't really need to test at
> all. The idea of a double-blind test is that you can't fool yourself.
> With a simple A/B test you might not want to fool yourself but still do
> it.
No, I'm afraid it was nothing of the kind and extremely unscientific. I
certainly cannot discount the possibility that my brain bypassed my ears
and fooled itself. What is interesting, however, is that I actually did
*not* expect to see a difference between uncompressed audio and high
bitrate ogg/mp3, based on experiments I had conducted several years ago,
but did. I would really be interested in taking part in some kind of
double blind testing some time and see what the outcome is.

Cheers,
S.M.