Re: arsclist Transfer of multiple copies, was: Full 3-D mapping of groove?

[Don Cox <doncox@xxxxxxxxxxxxxx>]

On 10/12/02, Doug Pomeroy wrote:

> In theory (best of all possible worlds), the audio combined from two
> pefectly syncronized transfers of two different copies of the same
> recording, will exhibit a 3 dB improvement in signal-to-noise ratio.
> The first problem, is syncronization, which must be exact. This has
> meant that all the source discs be must transferred *simultaneously*
> (getting several turntable platters to rotate in perfect sync is not a
> big problem). Even so, there would be sight timing differences, such
> as presented by off-center pressings, which Steve mentioned.

IMO the effects of off-centring would make trying to sync the turntables
a waste of time. Better to make the best possible 24/96 recordings and
then use correlation software to bring them into line.

That means a continually varying scaling/pitch factor is applied -
rather like applying a "wave" function to an image in Photoshop.



 
> It is gratifying to learn that CEDAR is thinking about this subject.
> It seems to me their existing Azimuth Corrector is actually part of
> the solution, as it can correct very small timing errors between two
> sources; what's needed is a much more powerful processor to deal with
> larger corrections over time.

Yes, but the Azimuth corrector is not changing pitch.

 
> Such a processor could sync transfers made at different times and on
> various turntables, greatly simplifying the process and increasing the
> potential usefulness of this technique. (Application of CEDAR
> noise-removal processing to the various source transfers, before
> syncronization, would no doubt contribute to optimum results.)

I suspect it would be better to do the correlation first - noise
reduction will remove low level sounds which might otherwise be
detected.

 
> The theory also states each doubling of the number of sources improves
> the s/n by another 3 dB. So, *four* copies of the same recording could
> produce a whopping 6 dB improvement! 

If it's a mono recording, you have two copies on each disc.



> Finding four different copies of
> the same record is not impossible to imagine, in some cases. (Pull out
> all your copies of those King Oliver Gennetts!) And, if they didn't
> all have to be transferred simultaneously, this technique could prove
> to be useful. Extending the theory, we could expect to see a s/n
> improvement in the range of 9 dB (!), if we could syncronize *eight*
> different copies of the same recording; but, obviously, this begins to
> get rather unrealistic.

I don't think it's unrealistic. If the maths can be worked out, the more
copies the better. Calculation time is not important - who cares if it
takes an hour to process a 3 minute track.

Specialist collectors like Ward Marston seem to keep several copies of
important discs.

I do think the software will be hard to write.

Regards
-- 
Don Cox
doncox@xxxxxxxxxxxxxx

-
For subscription instructions, see the ARSC home page
http://www.arsc-audio.org/arsclist.html
Copyright of individual posting is owned by the author of the posting and
permission to re-transmit or publish a post must be secured
from the author of the post.