[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Date Index][Thread Index][Author Index]

Re: OT That 'sizzle sound' of Mp3s

Andy,  thanks for putting some fish in the glasses....

The result of the process is not as bandwidth limiting as people 
assume (except at very low bit rates the HF response is filtered to 
prevent aliasing).  I am surprised by how frequently people say that 
they think MP3s have no bass.   It simply isn't true.

The white noise test will show up  artifacts because the process is 
based on a theory of perceptual encoding which states that certain 
loud sounds will mask certain quiet sounds and this is determined on 
a frequency by frequency basis. (in FFT 'bins').  A swept sine wave 
allows the process to deal with a single frequency at a time.  Given 
equal lengths, I wonder which test signal makes a smaller mp3 file.


>Charles Zwicky wrote:
>>You should record a 20-20k sweep tone and convert it to MP3.  You 
>>would be surprised.
>You're right, I was :-)
>Didn't expect it to reproduce a smooth sweep.
>It's a good test, but doesn't show up the limitations at all.
>(dunnow, maybe in a better listening environment, there certainly were a
>  few artefacts revealed when mixing out of phase with the original)
>Next I tried with white noise,
>and the metallic sizzle appeared like magic.
>andy butler
>( tests done with Audition, and mid-fi speakers, sitting
>  by noisey pc.  mp3 at 128bps
>  well, really it was 64bps mono,)