Re: [ARSCLIST] looking for advice on cylinders

["Steven C. Barr" <stevenc@xxxxxxxxxxxxxx>]

----- Original Message ----- 
From: "Mike Richter" <mrichter@xxxxxxx>
To: <ARSCLIST@xxxxxxxxxxxx>
Sent: Tuesday, September 20, 2005 10:36 AM
Subject: Re: [ARSCLIST] looking for advice on cylinders


> Susan Stinson wrote:
> > Hello, Tom:
> > Actually, no, they don't.  As we are experiencing in the development of
> > our cylinder laser playback system, the laser beam picks up whatever the
> > groove modulation has to offer, including particles of dust, wax, and
> > malformations caused by mould activity.  It seems, reasonably, that once
> > the mould has arrived and done its _thing_, there is no fix.
>
> With the advantage of perfect ignorance, I suggest that you may be too
> pessimistic. The day may come when sophistication in the use of the
> laser will pay off. On the simplest level, scanning with two lasers at
> different frequencies will allow differentiation in a remote-sensing
> effect: the mold will have different transmission and reflectance
> properties in the two (or more) colors and that may be exploited. More
> sophisticated techniques, even tomography, may become possible in which
> the mold would become transparent to the pickup. Certainly, the cost
> would be high, but it might not remain so and it may be worthwhile.
>
> Some years ago, a friend was hoping to recover audio quality from an
> optical track on old film. (He was with Warner at the time, is now with
> Microsoft.) We investigated and had some success with gray-scale pickup
> and algorithms for inferring the shape of the waveform and correction to
> the two (symmetric) profiles - one on each side - which gave the
> highest-probability reconstruction of the original. Our tools were
> simplistic and the results were preliminary, but they certainly went far
> beyond the de-facto thresholding of conventional sensors.
>
The point here is that the mould actually eats (literally) the wax
compound that constitutes the cylinder itself, thus changing the shape
of the groove surfaces...in worst cases effectively eliminating the
groove itself and leaving no trace of what it might have lokked like!
Since the grooves on cylinders are vertical-cut, so that the signal
is carried in variations of groove depth rather than path, anything
that obliterates the depth variations effectively renders the
signal inaccessible.

Steven C. Barr