On 21/06/2012, Michael Biel wrote:


> But getting back to the non-groove pictures, I opined that fine grain
> photos of 700,000 78s exist in the Rigler-Deutsch Record Index films
> but that they all have light-wedges from point lighting instead of the
> ring lighting I had suggested. I wonder if Patrick and his
> technological wizards might try to see if there is an algorithm that
> might be able to even the lighting problem out. Since the same camera
> and light sources were used on all of the records, maybe that is
> possible. (And thanks to Chuck Howell for saying he would have
> listened to me!! I wish you had been a part of the AAA back then!)
>
I would be surprised if these photos had enough resolution to extract
good audio from them, regardless of the lighting.

Ballpark mental arithmetic suggests that you get up to 50 cycles of
audio per mm of groove. If the disc is 300mm wide, the camera needs a
resolution of 15000 line-pairs or cycles across the width of the photo.

In digital imaging that works out at about 225 Megapixels - not an
impossible figure for a technical camera, but more than an ordinary
film camera would give even on microfilm. A big copy camera such as was
formerly used by printers, taking large sheets of film, could manage it,
especially with monochromatic light.

A high resolution scan on a scanner big enough to hold a 12 inch disc
could easily have enough resolution. The highest resolution on current
commercial cameras is around 80 megapixels, but cameras in satellites go
higher than this.

Regards
--
Don Cox
[log in to unmask]