Digital BW, The Print

Paul, 

As for sharpening, my scanned images normally look plenty sharp globally so I don't sharpen during scanning. They also look plenty sharp globally when being edited in PhotoShop so I normally apply only minor local sharpening, and often not even that. I sharpen moderately during printing, however, using QImage. I did not modify the file to obtain the data posted below other than changing the file size because I thought that might confuse data interpretation. I started with one original file and successively reduced its size incrementally using Bicubic Sharper to obtain the data. 
 
Randy. 

--- In DigitalBlackandWhiteThePrint@yahoogroups.com, Paul Grant <studiopbg@...> wrote:
>
> Randall,
> 
> Good discussion.    Although I don't have files that large I have seen the problem before.    I think alot depends on the methods of Size reduction that you use.   What methods are you using for file size reduction?   Do you do any processing afterward ie: sharpening etc.    What method of redcution are you using if in photoshop.   I think the default for photoshop is "Bicubic" and I think that tends to soften.  You might try Bicubic Sharpener?
> 
> Paul
> On Aug 24, 2010, at 12:15 PM, lgrrrb@... wrote:
> 
> > Several people commented that my website images lack contrast. At first, I thought this was the result of poorly calibrated monitors (their monitors, not mine!). After more people made the same comment, however, I looked more carefully at image contrast on other photo websites and saw that images on the majority of them were more contrasty than mine. My original files displayed good contrast on their histograms and printed with plenty of contrast so I suspect that something else may be afoot.
> > 
> > I believe that the following discussion applies to nearly everyone but think that large format film people and drum scan people who generate huge image files may have a greater problem than others. In my case, most of my original images approach 1GB in size so it is a great problem for me. 
> > 
> > To prepare an image file for web display, I have to reduce an enormous flattened, full-resolution grayscale file a tremendous amount. I believe that averaging such a large number of pixels during the image size reduction causes the bright pixels to be averaged down and the dark pixels to be averaged up enough to cause image contrast to be reduced substantially. 
> > 
> > To test this hypothesis, I used PhotoShop to open an original 714MB file which had lots of high values. The image, "Bright Cloud Along David Thompson Highway," was one that several people said had low contrast. The image can still be seen in the Western Canada Gallery at www.RandallRBreseePhoto.com 
> > I found that the original 714MB file had 45,735 pixels with a gray level of 255 (white). I reduced the image size to 135MB and found that this file had only 3,655 pixels with GL=255. I reduced the image size to 34MB and found only 106 pixels with GL=255. A reduction to 1.4MB showed only 1 pixel with GL=255. Further reductions showed few or no pixels with GL=255. 
> > 
> > When the number of GL=255 pixels is expressed in terms of the total number of image pixels (which gets smaller as the image size is reduced), the percentage of image area which is white drops steadily as image size is reduced. The percentage of white area in each file dropped from 0.012% for the original 714MB file to 0.0000014% for the 1.4MB file. So, people who said that some of my website images lacked contrast were right. 
> > 
> > I suspect that techniques used to downsize images is a whole topic in itself and there may be better ways to do it that mine. I'll leave that discussion to others. 
> > 
> > Most of you may already know that downsizing reduces image contrast but I seldom downsize images much unless I absolutely have to so I never thought about it before. The experience of downsizing images severely for display on the web has told me several things. 
> > 
> > First, think about what happens when a nice image is reduced to a 200 KB file for display on the web. It is safe to say that what you see on the web is going to be greatly different in MOOD than a print made to the same physical dimensions as the displayed web image except made from a larger file. For me, mood is crucial so this is a problem for my images. 
> > 
> > Second, I think that such a drastic size change for web display necessitates the application of a new curve and other adjustments to try to match the "mood" of a print made from the full resolution file. Of course, it is a loosing battle because a web image just won't have the pixels to do images justice but you have to do the best you can. 
> > 
> > Third, I would rather make a print using the largest file size that the printer can handle so I routinely print at 600ppi (the maximum my printer will allow). I know others will argue that this is a waste of resources but the data shown above tells me that, for lack of a better way of saying it, mood depends on file size. 
> > 
> > Fourth, I think I'd be careful about purchasing a print online based on a tiny downsized image which is displayed on a computer monitor. I think it is reasonable for people selling prints to routinely allow a buyer to return a print for a full refund if he is unsatisfied because he really doesn't know in advance what he will get. I also suspect that people who go to the trouble of capturing large sized images with good quality pixels are extremely limited in the ways they can communicate that quality on a website. Clearly, they can not display much TECHNICAL quality in tiny images on a website. As one group member told me, technical quality must be described in words and, hopefully, the words will be read. 
> > 
> > I want to thank all of the people who said many kind things about my new website and offered helpful comments. I hope the ideas in this post build upon some of the previous comments. 
> > 
> > Randall R Bresee
> > www.RandallRBreseePhoto.com
> > 
> > 
> 
> 
> 
> [Non-text portions of this message have been removed]
>