Hydrogenaudio Closes Doors For Now

verloren writes "The Admins at Hydrogenaudio, the community site discussing audio compression and related issues, have temporarily closed the site. They've posted a notice stating that they're rethinking the standards of the community, and how they're enforced. It seems to have been sparked most recently by a debate over what media players to use, but has been brewing for some time as the objective standards required at the site have been overlooked by many posters. The sister sites Foobar2000 and Rarewares are still available."

Comments

Re:Forced?

[Anonymous Coward]

There was always one big issue: (unintentional) bias

When testing audio codecs it's important to do a double blind test. (On HA they would call it "ABX", I think that was because that's the name of the program)
The easiest way to do it is to give a program the source file and the encoded version.
(Or two encoded files with diffrent codecs, depending on what you wanted to accomplish.
For transparency, ie can you tell the diffrence between a MP3 encoded with X options and the source file?, you would give source and encoded.
For comparison between 64 K Vorbis and WMA youi give two of those files)
The program would play them 8 times (or so), first one then the other, but you can't know which file is which since it would alternate.
You listen carefully and give it a score.

At the end the program tells you which file you preferred.

With psychoacoustic encoding that's the whole deal: you want to encode the file in such a way that for a human it seems as close to the original as possible.
Because of bias, a double blind test is THE ONLY way to accurately rate lossy encoders

But the newbies would come and say:
Hey this codec is better then that codec.
How do you know?
It sounds better
Have you done a ABX?
No, but my ears are golden

Or they would come:
I compared a graph of the source with the encoded file and this codec produces files that seem the most identical.

Both of these are totally wrong, and sometimes you would have dudes that would insist that one of these methods are good, no matter how much you try to talk some sense into them.

And another problem: misinformed audiophiles
"Oh no the stereo image is holy, joint-stereo is from the devil!"
Never mind that LAME has an excellent joint-stereo system.
And they would come with other crazy theories, for example challenging Shannon's theory that to encode a X Hz signal you only need 2X of bandiwdth.

So that's a part of it