Newsgroups: comp.graphics,sci.image.processing
Path: cantaloupe.srv.cs.cmu.edu!europa.chnt.gtegsc.com!news.sprintlink.net!tpoint!pm2e-1-36.tpoint.net!user
From: lensnut@tpoint.net (Steve Strickland)
Subject: Re: Unhalftoning Halftones
Message-ID: <lensnut-1305950041510001@pm2e-1-36.tpoint.net>
Sender: news@tpoint.net
Organization: Cyberoids of Prepress
References: <3o756l$eot@mack.rt66.com> <1995May3.130026.1020@csginc.com> <D8CEDy.HJ@ucc.su.OZ.AU> <grunes.208.800121134@news.nrl.navy.mil>
Date: Sat, 13 May 1995 05:40:30 GMT
Lines: 19
Xref: glinda.oz.cs.cmu.edu comp.graphics:76519 sci.image.processing:14602


> A good solution would identify the grid, and set the digitization pixels
> to match them.  Its a little messier if the grid isn't rectangular.  
> For example, a checkerboard grid can be transformed into rectangular form
> by rotating the picture 45 degrees before the scan.

In practise, this won't work because printed halftones undergo anamorphic
distortion due to mechanical stresses encountered on the printing press,
rendering precise pixel location irregular and unsymetrical. Moire
results.
 
> In summary: I think the problem is mostly solvable, but it would involve a 
> fair amount of work to do it right.  Is it possible that commercial
> packages like Adobe Photoshop or Illustrator [sp?] have already solved 
> it?

Photoshop is not there, yet. Photomultiplier scanners can certainly record
the individual halftone dots accurately, but at a huge cost in memory and
processing time.
