[ILUG] [OT] [flamebait] iPods

[fuzzbucket]

Colm MacCarthaigh wrote:
> 
> A month ago, Richard Bannister created such a blind test, and posted 
> it on the IIU mailing list [1]. It was a well-encoded 10-second clip,
>  *tough*.  I was the only person stupid enough to take it [2], and 
> Vorbis won [3].  It was definitely the best encoding. AAC came last. 
> But that's to my ears.
> 
> [1] http://iiu.taint.org/pipermail/iiu/2005q1/004226.html [2] 
> http://iiu.taint.org/pipermail/iiu/2005q1/004232.html [3] 
> http://iiu.taint.org/pipermail/iiu/2005q1/004242.html
> 

Yikes! I wouldn't even attempt to call that - plus it would be pointless
given I know which is which. That said, I agree with you - the first 2
seconds give it away... :-)

I wonder how these files were encoded. The thing is that once you go
above a certain bitrate the differences between the formats become more
blurred. Most tests I've read of use a low or medium bitrate (usually
~128kbps).

This presents its own problems in that a format like Vorbis doesn't lend
itself well to CBR testing. You can spend an age experimenting with
maximum and nominal bitrates or try playing with the quality parameter
until the average bitrate is exactly 128kbps. This could be deemed an
unfair test as more complex sections of the track will have more data
describing them where as CBR MP3 doesn't have that property. MP3 doesn't
have to be CBR so using VBR would be more fair, although you could argue
we're comparing common use and most folk go with the defaults.

Still, an interesting test. I'd like to see Musepack compared in more of
these test as in my experience it's a bloody fantastic format.That said,
we should all be using MAC and FLAC - lossless is the future! (Er...
anyone have a LOT of disk space I can borrow?) Can I get vinyl on
iTunes? Analogue is the future! ...*trails off, rambling about wow and
flutter*...

-- 
ICQ:139064407 Jabber:fuzzbucket at jabber.org Yahoo:rapidyokes