Feature for lightness, contrast, etc, compensation between print sizes.

Ernst Dinkla

Sr. Member

Posts: 410

Feature for lightness, contrast, etc, compensation between print sizes.

« on: May 19, 2011, 01:51:54 PM »

There have been rule of thumb solutions for compensating color/tone/contrast differences between large and small prints of the same image. For example Caponigro has written something on that subject. But today I came across an interesting scientific article on the subject:
http://art-si.org/PDFs/Processing/MahdiNezamabadi_PhD2008.pdf

I do not grasp all of that article it but I think there could be a feature squeezed from that article for Qimage. Maybe not all the refinements but the most important ones at least, Suppose one has the perfect edited image/print on A4/A3 size on a specific paper. Image data stored for that format. The color/tone/contrast corrections could then be applied to the data on the fly for smaller and larger prints, based on the image content as described in the article.

met vriendelijke groeten, Ernst

Try: http://groups.yahoo.com/group/Wide_Inkjet_Printers/


	Logged

admin

Administrator
Forum Superhero

Posts: 4218

Re: Feature for lightness, contrast, etc, compensation between print sizes.

« Reply #1 on: May 20, 2011, 01:05:37 PM »

Holy Crapoli! That thing is a mammoth! Anyone have the Cliff's Notes version? Grin

Having read enough to think I know the basics... my current thoughts (before looking into it further) are that the other factors of viewing probably have a much greater effect than size-to-contrast. Things like:

(1) They talk a lot about LCD displays in that paper. LCD size, viewer position, and viewing angle are probably a bigger factor since LCD's differ in contrast/color based on viewer location and angle since LCD screens are known to change contrast and color from left to right/top to bottom anyway.

(2) Intended viewing conditions of a print: if a print is displayed on a wall under top/ceiling lighting for example, the fact that the top of the print is closer to the light source than the bottom and the associated higher brightness at the top of the print is probably a much larger factor than viewing distance.

(3) For prints that are not mounted, what do you do? Do you assume everyone will hold a 5x7 print a certain distance from the eyes? An 8x10? 4x6?

(4) For mounted prints, you'd have to even account for the height of the person doing the viewing and force them to stand at a certain distance.

So I think the environment would have to be completely controlled (you'd have to "force" the audience to stand a certain distance from the print) and the lighting would have to be perfect (print lit from top, bottom, and sides evenly) to get this to work properly. If your viewing audience is allows to go up and examine the print or walk back to a longer distance, you may have done more harm than good trying to fool with this.

Mike


	Logged

Ernst Dinkla

Sr. Member

Posts: 410

Re: Feature for lightness, contrast, etc, compensation between print sizes.

« Reply #2 on: May 20, 2011, 04:05:52 PM »

Quote from: Mike Chaney on May 20, 2011, 01:05:37 PM

So I think the environment would have to be completely controlled (you'd have to "force" the audience to stand a certain distance from the print) and the lighting would have to be perfect (print lit from top, bottom, and sides evenly) to get this to work properly. If your viewing audience is allows to go up and examine the print or walk back to a longer distance, you may have done more harm than good trying to fool with this.

Mike

Mike,

It is possible to criticise the tests and I considered the LCD use in that test and I thought of possible aliasing effects in the reduced size images. Nevertheless I think it has been done quite well though. Fact is something happens when prints are enlarged and I am not the only one who observed that. Proof prints at reduced sizes do not predict the effects larger prints show. At least there is some serious research done on that subject now and the outcome might be used for Qimage or another application for printing.

met vriendelijke groeten, Ernst

New: Spectral plots of +250 inkjet papers:

http://www.pigment-print.com/spectralplots/spectrumviz_1.htm


	Logged

admin

Administrator
Forum Superhero

Posts: 4218

Re: Feature for lightness, contrast, etc, compensation between print sizes.

« Reply #3 on: May 21, 2011, 04:48:36 AM »

Quote from: Ernst Dinkla on May 20, 2011, 04:05:52 PM

But how to you "explain away" the fact that holding a 12x8 print at arms length gives you the same viewing angle as a much larger 24x36 when viewed from 6 feet away? For larger prints, you typically stand farther away. For smaller ones, you move forward. I still want to know what is the market for software that equalizes contrast and color for print size, when to get it to work properly, you have to strictly control your photo-to-viewer distance?

Mike


	Logged

Ernst Dinkla

Sr. Member

Posts: 410

Re: Feature for lightness, contrast, etc, compensation between print sizes.

« Reply #4 on: May 22, 2011, 11:07:08 AM »

Quote from: Mike Chaney on May 21, 2011, 04:48:36 AM

When aiming for large prints you will create smaller proof prints. It is then wiser to have them on the same viewing board illuminated with the same lamps and judged at the same distance. The more if the large print will be a copy of a painting etc. standing there too. In practice now I will print strips at a 1:1 size to the original or intended reproduction size as a final step because the smaller print is not having the same effect. Using a smaller proof print at arm's length is not the way to do it in my experience.

The optimal print at a certain size should dictate what the print will be at another size and when displayed on the wall viewing distances will not be so strictly related to the print size, not for the large and not for the small one. If that rule was so strict then something could be said for not creating good upsampling routines (like available in Qimage) as people will see the large or small print at related viewing distances. In that case I have seen that argument used quite often in other camps, yet we appreciate the better extrapolation routines of Qimage.

We probably do not share the same opinion on this subject but at least the article didn't go unnoticed.

met vriendelijke groeten, Ernst

Try: http://groups.yahoo.com/group/Wide_Inkjet_Printers/


	Logged

Pages: [1]

« previous next »