[APML]: Second HP v. Kodak Test

[Benjamin Gomes-Casseres <ben@alliancerevolution.com>]

Many thanks, Dave Kodama, for your suggestion on how to make my rough 
scanner test a fairer race. I followed your advice and rescanned the HP 
image, this time adjusting in prescan so that the histograms of the 
resulting raw image closely matched the Kodak's raw image (that one I don't 
have the option to adjust, as it is done by the camera shop). Using tracing 
paper and lots of iterations I got a very, very close match in histograms. 
After further processing to produce a final picture, I again checked to 
confirm that the histograms are very close. (I know it's still not a fully 
fair race, because I am not optimizing the Kodak in prescan. But at least 
it does not prejudice the HP.)

The new HP images are on my site (I deleted the old ones). I think they are 
slightly better than the old images, but not much; they are still "coarser" 
than the Kodak scans. No change in chromatic aberration. So, I think my 
conclusions stated on the site still hold. They are under "Scanner 
Comparison" at www.alliancerevolution.com/astro

I am not sure what is causing the difference, because I am totally new at 
this stuff. The Kodak scanner actually has lower resolution (2000 dpi) than 
the HP (2400 dpi), ...but there appears to be a difference in bit depth? 
The Kodak has "12 bits per RGB channel", while the HP has "30 bits A-D". I 
don't know what that means...could someone explain this apparent difference 
to me and give an opinion whether this causes the difference in "texture" 
in my comparison?

By the way, these rough-and-ready tests have convinced me of the point that 
Mike Regish has stated repeatedly: The HP PhotoSmart scanner appears to be 
a tremendous price/performance value, even if it is not top-of-the-line in 
performance. (I'm not sure what the Kodak 2035 Plus costs, but the number 
$8000 comes to mind from earlier discussion, compared to $499 for the 
PhotoSmart.) Furthermore, it is a breeze to use. I decided to keep mine.

Even so, I hope to gain some insight from the Olympic judges as to how they 
score the performance of these two scanner images on my comparison page. 
Or, please tell me how my test is skewed (as Dave did) so that I can go 
back to doing more homework. (I've learned a lot from that already!) Thanks 
in advance!

I do look forward to the real Scanner Olympics, with standardized test 
negatives, experts running the scans, etc. It will be interesting to see 
whether resolution or some other machine features matters more, aside from 
operator experience, and whether you get what you pay for or not. (It could 
well be that HP is pulling the rug from under this market, as they did in 
printers and as Compaq and others did in PCs.)

Ben Gomes-Casseres
www.alliancerevolution.com/astro