naturerecordists
[Top] [All Lists]

Re: 24 bit vs. 16 bit

Subject: Re: 24 bit vs. 16 bit
From: "Raimund Specht" animalsounds
Date: Tue May 15, 2007 12:25 am ((PDT))
Walter Knapp <> wrote:
>
> Posted by: "Raimund Specht"
> 
> > Sure, even 8 bits could be an adequate bit depth under certain 
> > circumstances... 
> 
> Works pretty good for recording at conferences...
> 
> > The fact that ATRAC and MP3 very often do not much affect the
> > subjective sound quality of nature recordings is an indication for the
> > low dynamic range that is actually in such recordings. Both ATRAC and
> > MP3 internally reduce the bith depth for certain frequency intervals
> > even below 8 bits.
> 
> What they do is make more intelligent use of the bits available. 
> Uncompresed sound uses the same bit rate even if it's not needed, say 
> like a silent part between the sounds as a example. In compression like 
> ATRAC bits are allotted more according to need. So you need less bit 
> rate for the same quality of sound. Not sure I'd say they were 
> necessarily low dynamic range.

Yes, it's also the limited bandwidth and low entropy of the recorded
sounds that allow the compression algorithm to allocate the available
bits more efficiently. The relatively high background noise floors (or
limited dynamic range) in nature recordings then helps to mask the
potential artifacts caused by the compression algorithm.

It is certainly interesting to know that MP3 and ATEAC work so well on
most kinds of sound recordings because there is usually very little
energy at the higher frequencies above about 10...14 kHz. 

Also, the highly regarded Sennheiser MKH microphone series has been
optimized for best noise performance at the lower frequencies below
about 10 kHz only. The noise level of those microphones at the higher
frequencies is even worse than in other less expensive microphones
(e.g. Sennheiser K6/ME6x series). That might also rise some questions
regarding the justification of 96 or 192 kHz sampling for normal
(human) listening purposes. But that would probably open another can
of worms, which could further damage a few more illusions on the
latest developments in audio technology. So, I should better shut up
for now ;-)

Raimund






<Prev in Thread] Current Thread [Next in Thread>
Admin

The University of NSW School of Computer and Engineering takes no responsibility for the contents of this archive. It is purely a compilation of material sent by many people to the naturerecordists mailing list. It has not been checked for accuracy nor its content verified in any way. If you wish to get material removed from the archive or have other queries about the archive e-mail Andrew Taylor at this address: andrewt@cse.unsw.EDU.AU