Re: [ARSCLIST] Optical Groove Digitization

[Jon Noring <jon@xxxxxxxxxxx>]

[Project Gramophone cc'd]

Kurt Nauck wrote:

> Have you seen this?
>
> http://www.eif.ch/visualaudio/
>
> Perhaps now the tinfoil playback problem has been solved!
>
> There is a description of this system in the July 2003 IASA Journal.
> Fax: (425) 930-6862

I've been intrigued for a long time with "optical" methods to preserve
and to recover the "sound in the grooves." This work is definitely the
next step in the right direction.

My current thinking, however, for the most advanced playback of older
recordings is to get a true, high resolution, 3-D characterization of
the entire record. In essence, to create a very precise topographic
"map" of the record. This way, the full information of the recording
is preserved, and it may be possible, using a super-computer and the
right algorithms, to do the highest possible quality "transfer" of the
recording (interestingly, one may be able to precisely correct for
complex warping of the records.)

The problem with the current playback methods, such as a stylus
running in the groove or the optical method noted in the article
described above, is that they still "linearize" and mix information
together, and once one does that, one loses a lot of information that
otherwise will be useful for the best possible restoration.

With the 3-D topographic approach, in principle one should be able to
best ascertain what the original signal was based on full analysis of
both walls of the groove, from the top of the record to the bottom of
the groove. It may be possible to recover the original signal from some
types of groove damage, and to minimize the effects of nicks and
impurities in the record substrate (e.g., the abrasive particles records
used to use to polish the steel playback needle -- these stick out like
boulders in the groove path.)

Of course, some challenges to this 3-D topographic approach are:

1) Huge amount of data: To get adequate 3-D resolution will require
   sampling a *lot* of data -- I haven't figured it out yet, but we
   are talking about many gigs at the minimum for a typical 10" 78rpm
   disc. (But disk space nowadays is readily available.)

2) Coming up with the program and requisite algorithms to analyze the
   data. It would not surprise me that it would take a super-computer
   (such as a cluster of PC's) several hours to extract the optimized
   signal from the grooves (if this is the case, then this project
   could tap the huge reservoir of volunteer pc's out there, such as
   what Project SETI uses.) Maybe I'm being overly pessimistic, and
   that in the distant future this can be done quite fast and by the
   "ordinary" restoration engineer.

3) And of course, how to actually scan for this data. There is no
   doubt equipment/techniques that will do this, but it is unknown if
   there is any "commercial" equipment that does this (there might be);
   rather, it may still be in the province of advanced research at
   places such as the National Labs, for example (I used to work as an
   engineer at Lawrence Livermore Laboratory, and making high resolution
   3-D "topographic" maps of surfaces is something that someone there
   may have expertise in -- one of these days I plan to check if this
   is so.)


Just my $0.02 worth.

Jon Noring