Art, thanks for the very thoughtful post. I think you've hit
most of the high points.
At 5:01 PM -0800 11/8/00, Arthur Entlich wrote:
>As much as I like science and the scientific method, it is, when it
>comes down to it, just another religion, and I don't like science
>zealots (I'm not suggesting Roger is one, BTW).
>
>The problem with scientific method is that in its own special arrogance,
>it attempts to control all variables. Then it justifies the result based
>upon the purity of the empirical data. Only one problem, we cannot
>measure that of which we are not aware. That is the flaw, and why
>science has a "faith" component, and why I distrust the "absoluteness"
>of science zealots.
I've been called a lot of things, but "zealot" isn't one of
them! After spending over a year trying to get consistent results
with my various scanner/printer/profiler/pigment ink/special paper
combinations, I'm willing to agree that "eye of lizard and toe of
newt" is probably the most important factor in my rate of success
(except I think it's "eye of newt").
>I'd agree that the
>367.9 dpi (not downsampled in Photoshop) version was slightly superior
>to the 360 dpi version. which had been both mildly downsampled in
Photoshop, and then manipulated further by the Epson print driver.
>So, what does this possibly mean? It might mean that not downsampling
>in Photoshop and just printing the image at the right dimensions and
>having the Epson print driver do the downsampling might give a superior
>result, which is what Austin reports and suggests.
>It might mean that with this particular sample, using that paper, and
>driver settings, with that printer, those inks, and that particular
>Epson driver, the results improve by leaving the image at full
>resolution.
I think the difference, small though it is, is consistent. I
did the same experiment with the PDI printer test target that has
some tiny print on some of the gray scales, and again the 360 ppi
print was noticeably superior to the 240, and very slightly inferior
to the non-downsampled print. (This was determined by examining with
a loupe, not by scanning the print and looking on the monitor). As
Austin said, it was worth it to find evidence that printing with very
odd resolutions (e.g. 367.9 ppi) has no deleterious effect on the
quality of the print, and the "magic number" concept doesn't seem to
be valid. That doesn't explain why some people insist they get better
results when they use 240 or 360. Must be the "eye of newt" factor
again.
>The interesting part, is that I was able to open both files and I
>compared them in Photoshop. They were not the same. The JPEGed
>versions were higher contrast, and actually looked very slightly
>sharper. I still would come to the same basic conclusions, but it shows
>the problems with variables being confounded by all sorts of factors.
>This might explain why many of the disagreements we experience here
>occur in terms of issues of "sharpness" and "detail" and of course,
>"color rendition".
Another good point. I used Photoshops's "Save for Web"
function, set at 100, which I presume is as good as it gets with
JPEGs. Even so, there were differences. I'll bear that in mind.
Now to end this before Lawrence Welk gets involved and we get
off the topic of Epson printers and upset Tony again <g>.
Roger Smith
