naturerecordists
[Top] [All Lists]

Re: [gear] one bit technology for field recording [long msg]

Subject: Re: [gear] one bit technology for field recording [long msg]
From: "Jeremiah Moore" jeremiahmoore99
Date: Wed Apr 4, 2007 2:18 pm ((PDT))
>From a demo of DSD technology I had at Airshow Mastering in Boulder in the
late '90s when the technology was sony proprietary, I learned that the DSD
process is actually extremely noisy.  But with the noise shifted into an
extremely high frequency range.  So yes, there are boatloads of distortion
at frequencies above 1xFS, but it doesn't matter.

It took me a while to understand delta-sigma - I was assuming that a "bit"
was a tiny amount which was nudged up and down by the incoming voltage.
This is not the case.  Each sample is full-scale, a "one" representing the
extent of the scale in a positive direction, "zero" representing the extent
of the scale in the negative direction.  A recording of digital silence in
DSD is a 1.4MHZ square wave.  A recording of all ones would be maximum
positive DC offset.  A higher incoming voltage will be represented by more
ones; a lower by more zeroes.  The "average" of the ones and zeroes over
time is the audio signal, somewhere between one and zero.

One interesting thing about the format:  a DSD D-to-A converter is a lowpas=
s
filter.  That's it.  You could plug a cable containing a digital DSD signal
into an analog lowpass filter, then into your amp.  (Sharp minidiscs with
"1-bit" headphone amps do just that.)

DSD sounds phenomenal.  But it will only be as good as the analog circuitry
ahead and behind it.  I've not listened to the Korg units so can't comment.


-jeremiah


On 4/3/07, oryoki2000 <> wrote:
>
>   Sigma delta encoding gets its accuracy simply through the enormous
> number of observations per second.
>
> A continuous voltage input would be represented as an oscillation of
> zero, one, zero, one, etc.
>
> Anyway, the proof is in the pudding. We need to get one of the
> MR-1000 models into Rob D's hands for some real world (and lead-lined
> narration booth) testing.
>
> --oryoki
>
>
>



--
-----------------------------------------------------------
jeremiah moore | SOUND | 
http://www.jeremiahmoore.com/  professional site and reel
http://babyjane.com/timeweb/  personal site
http://northstation.net/  music project







<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