Digital audio listening test

[Martin Czech]

We had a good laugh about that article.
AFAIK the mpeg algorithm was developed by the Fraunhofer
Institute. At least that's where we licence our code.

These guys know the weaknesses in and out.

It's like TV 2D picture filters. They look nice (better then
plain processing) if you feed the right test case into.
But there are also known cases where these algorithms
are not so nice.

Now, is it a fair test if you ommit the "critical" signals?
Then you get what the average program material will look
like. As an engineer I'm interested also in the worst case.

So -back to audio- this means recording and testing known critical
signals. Solo drums etc. If you do this you don't need a fancy audience
to tell the difference. And more bandwidth won't save you either.


m.c.



:::The German computer mag c't (issue 3/2000) did some extensive tests
:::comparing MP3 and non-reduced digital audio. They had many expert