Hi all, I think this may interest you...

Today my syster showed my a notice about a guy from
Israel that says that he can extract audio from Vinyl discs
scanning them with an "normal image scanner"and analizing
the resulting image!!!


Maybe fake?.
Theorically its possible (I think) but I don't know if technically is
possible in real life.

For those of you interested in more info, the explanations directly
from the developer of this method.

http://www.cs.huji.ac.il/~springer/

Best regards from Spain
b:.

ZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZzzz
Sorry! Sorry!
I use to visit the forunm frequently, but I didn't saw any post
about this.

Next time I'll do a search before post.
Best regards.

This was posted about just yesterday , see HERE

and also on hydrogen audio on the 6 september 2002!!!

Try searching before you post.

Cheers,
Kristian

OMG! People won't give up reposting these news.

Maybe I should put it in the FAQ.

Along with portable players that support AAC and requests for a Seekbar in Foobar2k.

Too true, man. Merely reading the term "seekbar" has recently started to cause me skin rashes.

Just for grins, have you thought about compiling the thing, Roberto? The author offers the source code, but no binary... although to get enough visual resolution to match a true aural representation of the soundstage, I suspect you'd need something on the order of an image twice the size of what the pure, uncompressed WAVs would be. Scanners simply aren't up to the task, yet. Maybe someday; there's no fault with the theory behind this software, as long as people are realistic about the limitations, as well.

To deal with the issue of reflection from within the groove, this fellow scanned his image in four separate regions. Not enough, by far! A minimum of 32 passes would have been recommended (to get the arc of the groove down to a possibly manageable segment), but again, more is better - as long as you are dealing with a fixed axis of rotation. How about 128 segments, just to be safe? (360 or more would be best, but I'm trying to be realistic.) Let's see... you'd need to mount a post on the glass of your flatbed scanner, and rig an indexing mechanism to ensure the rotation was no more and no less than 2.8125 degrees. Then you'd need a software solution for accurately rotating and compositing the image segments. Don't forget, your scanner's optical resolution (no interpolation, since that wouldn't give a very accurate representation of the original sound!) needs to be somewhere around 40,000+ dpi. And you can't save in JPEG, unless you are willing to endure some slight artifacting.

Am I being sarcastic? Only slightly; this will all undoubtedly come to pass, and probably sooner than any of us would expect. (Just look at the past five years of digital audio evolution, if you need to put things in perspective!) But as for a real-world application, it isn't there yet.

- M.

Don't forget there's this laser turntable which kind of "scans" the record So the whole idea isn't totally insane - it's just a matter of resolution as M mentioned...

Resolution and proper, uniform lighting, I would say..

I tried. But Borland is a bitch, so I gave up.

I suggest someone with the necessary BCB skills to build it. I can host the binaries then, if needed.

The source is incomplete File with IDE-generated objects' properities (*.dfm) is missing.