Scanning a print VS. Interpolation?

Discussion in 'Digital Darkroom' started by hique, Jan 10, 2005.

  1. Hi there.

    I would like to ask you about a question that is becoming a myth to
    me.

    It's the second time I hear about it from two different working
    people of the graphic industry.

    What they tell me is that if you want to make a big print it's better
    to scan a smaller print with a good scanner than to interpolate the
    file.

    For instance, if you need to make a 20x30cm print at 300dpi, but you
    only have a file with 10x15cm at 300dpi, you should print it with
    10x15cm and then scan the print with a good flatbed scanner setting
    the size you need to print (200% in this case, at 300dpi).

    They suggest that this will create a file with better quality than
    interpolating with photoshop.

    Actually I don't think it's possible since we would only lose quality
    when we are transfering the data from media to media. Another fact is
    that it would be necessary to GAIN resolution in the scanning
    process, what I believe, is impossible.

    Did anyone ever heard about this theory. Is there any logic behind it?

    Cheers.
     
  2. Please send me a couple ounces of whatever they're smoking. It sounds like they've got
    enough supply to go around.

    Goes to show that even people who do something for a living will fall for the stupidest
    urban legends sometimes.
     
  3. Me thinks you should stop talking to these people. You can only lose information every time you perform an output and/or acquisition procedure. You are correct. The only way printing and scanning could be better is if you use a bad interpolation method.

    By printing and scanning, you are effectivly using the printer's interpolation method (if any) and the physical properties of the print medium to interpolate. The only logic I can see behind this is that this could provide more "visually pleasing" interpolation. But, again, that just means you should be using a better interpolation method in Photoshop. I would recommend using Qimage for print interpolation.
     
  4. I'm with you Marcio. The fewer steps you can take to the final print, the better off you are, in general. I think interpolating in Photoshop will give you a better final print.
     
  5. Multiple passes of 10% and tweaking tend to produce better results than one large pass.
     
  6. You have nothing to gain from printing then rescanning the print! It's a matter of Entropy. Every change you make, particularly such a dramatic physical change as printing, degrades the image quality.

    A print has a very limited dynamic range, perhaps 2.5 to 3 stops, and relatively poor resolution, perhaps 40 dpi considering dithering and dot gain. An optical print is no better. In all, that's considerably less than you could expect from a digital image, scanned or otherwise, at the same size.
     
  7. Actually, I think Marcio is right. I went to Art Center and that's what all the graphics and advertising kids are doing to achieve large prints larger than 3 x 5 feet in size. They said it's their own little secret. I've never done it myself because I never needed a image that size but from what I've seen, it looks great. I also know a editorial photographer by the name of Chris Buck, www.chrisbuck.com, that uses that process. He would make a 11x14 C-Print and then scan the print instead of scanning the film. He said in his opinion it gives a smoother transition between colors because instead of going from a small 6x6cm film source, he's now working from a 11x14 inch sourcs. I don't know if this works the same using a digital print instead of a C-Print and then scanning it.


    For those accusing them of "smoking some stuff," sometimes lots of good ideas comes from "smoking" that stuff. I don't know if this one is good cause I never tried it, but it sounds like if done right that it might work.
     
  8. I just scan my "stuff" and then smoke that. ;-)
     
  9. "He said in his opinion it gives a smoother transition between colors because instead of going from a small 6x6cm film source"

    He could have done a much better job starting with a good film scan, and waste way less time and effort. He may be a good photography, but what he said is against science.
     
  10. To answer my own question, I did an experiment in my work today.

    I printed a 6x4cm print in the Frontier and then scanned with a Microtek 6800 flatbed (not the best scanner, just the one that was available).

    The results show that there is no benefit from scanning the print. The transitions are indeed smoother as someone said, but that comes together with a huge loss of sharpness. Interpolation caused the picture to pixelate a little but provoked less loss of sharpness.

    The friend of mine who defended the theory pointed in his defense that the scanner is not very good, but I hardly believe that it is the limitating factor (probably the print is. Maybe 40dpi as someone pointed).

    He also said that the digital printing (Fuji Frontier) renders some artifacts, and that the original print should be better using an analog minilab or enlarger. I can even agree with him, but anyway the print would still hold little detail.

    Thanks for all the responses.

    Cheers.
     
  11. I think it only works if the smaller print is made with a traditional optical enlarger and a good one at that.
     
  12. Scanning prints often is what we do; our only inputs. A digital file that produced that print is a better input. It has no printer artifacts; no printer calibration errors; no lost shadows or highlights. The rage of scanning prints is stronger today; and is very common; instead of a raw digital file. Some folks eat vomit too. It doesnt mean we need to.
     
  13. I think it only works if the smaller print is made with a traditional optical enlarger and a good one at that.

    Well sure, if you've got a good enlarger and a cheap scanner, yeah, this can be effective--I remember back 6 or 7 years ago I was able to clearly outresolve a 2800 dpi Nikon with an Epson scan of a good (not machine) 8x10. I think there's very limited benefit vs. a 4000 dpi scanner--and huge drawbacks--and still absolutely zero possible benefit scanning a digital print, the recommendation that started this discussion.
     

Share This Page