Do digicams take "more realistic" pictures?

Do digicams take "more realistic" pictures?

Post by Anonymou » Wed, 28 Jul 1999 04:00:00



Hi, I'm Jim Mulch, head of Imagining Science at the WestBoy Colduck
Company, I will try to answer your questions. bs


Quote:>When I took my first pictures with my CP 950 and printed them out, I was
>a bit disappointed.  The prints looked sort of flat, lacked "snap."  But
>then I took the picture to the spot where it was taken and noticed that
>the world looked just the same, sort of flat (not particularly
>saturated).  I realized that I had become so conditioned to high
>saturation, high contrast prints from film that the digital pictures
>But I would rather have a camera
>that gave as accurate a rendition of the scene I'm looking as possible,
>rather than leave it to the film manufacturer to define "reality."  

Digital cameras have FAR more color accuracy potential than film. It is
simple to take a set of rgb values that have passed through a camera CCD
filter defined by video standards or something close to tristimulus values
(how the eye sees the world), and mathemtically convert them to actual
tristimulus value RGB, and then through color profiles
(http://www.color.org) to any desired color space. On the other hand it is
not only difficult to design a film to do this, considering capabilities
of chemicals chemical models, and chemical vartiabilities, it requires
higher order math and knowledge of the films characteristics, in order to
reverse model scene colorimetry from film dyes if you want to scan the
films back to the original scene color. This path has been established
with PhotoCd and PrintCD, but the models have far more noise than a
digital camera model ever will. Dynamic range is not as big of an issue
that film companies make it out to be, but exposure latititude is.

Quote:>Of course, it is a simple matter to make the digital
>prints look like film prints with a few computer operations.

let's use an example, the wedding/portrait markets in the United States
have grown very accustomed to how Vericolor III VPS film prints onto a
portrait paper. You could develop a color transformation that results in
cameraRGBvalues------->VericolorRGBvalues, when you acquire your camera
image into Adobe Photoshop, you can apply this transformation to the image
and your camera image then looks like a Vericolor film wedding shot.

There are three ways to accomplish this

1) trial and error use of photoshop to develop a set of balances, very
iterative, and not quanititative

using color profiles as defined at http://www.color.org

2) you can empirically model both systems, I have tried things like this
at the Westboy Colduck Company, you run into some target issues, and
issues related to the dimensionality and processing of targets through
both the digital and photographic system, but it is purely empirical, and
requires no film specific information from a film manufacturer to pursue.
Not a canned technology, but nothing is.

3) having spectral and color design information of the camera and the
film/paper combination, you can mathematically calculate this model.
Requires information from camera maufacurer and film/paper manufurer. The
Westbof Colduck Company would never make a digital camera that spits out a
Vericolor portrait/wedding look, so this appraoch is somewhat of a moot
point, unless one of the photographic universities got involved.

I know how to do all three of the above methods.

So what about mis-exposed digital pictures, or off-color digital pictures?
The reason that film prints so
easy, is that there are scene balance algorithms that correct for exposure
and off color. If these algorithms continue to be linked to metrics
specific to film, such as the density of which the film prints on a
specific paper, then truly digital images will never print without
muddling around in photoshop for hours. In order to enable digital
workflow, scene balance algorithms in a digital non-film color space are
paramount. You will not see a film company drive these algorithms.

It is clear that the strategy of film related companies with regards to
digital, is reactionary at best. You will not see a digital camera from a
film manufacturer, before someone else releases the technology, ala
Dicomed Bigshot a couple years back. Similiary, you cannot expect them to
be leaders in true digital imaging workflows, so mark your investments
appropriately, the sale of the Elmgrave equipment plant is a clear message
that
there is no committment to digital imaging at the WestBoy Colduck Company.

It is also worth noting that said film companies have had scan-only type
films, chrome-like, no need for masking couplers, get rid of DIR, DIAR
couplers and interimage, because they complicate by adding higher order
math in the algoriothms as described above. This technology has also been
around for 10 years, and has not been commercialized at the WestBoy
Colduck Company, because just like any other non-positive film, you need
color processing algorithms available as defined above to use the film
properly, and enablement of the algroithms, means enablement of digital
workflow. So if/when you see these films, be sure to ask for the
algorithms to go with them, otherwise, the lowest priced private label
film, is still all the quality you need.

If someone like a digital camera or electrophotographic or inkjet printer
company is interested in enabling digital workflows, I am available for
consultation. I am particularly committed to enablement of digital capture
and workflow in the mass portrait markets, school pictures, etc. So if you
have a digital portrait system, and want to fiddle around with how to tack
on a digital camera, drop me a note to rec.photo.film+labs, ATTN:
Imagining Science. I could also be easily convinced not to give such
advice.

 
 
 

1. Do digicams take "more realistic" pictures?

[...]

I'm reading this in comp.os.ms-windows.programmer.graphics.

Would you folks in:
    comp.sys.amiga.graphics
    comp.sys.mac.graphics

please remove comp.os.ms-windows.programmer.graphics
from the header for this and other similar discussions. It's
completely off-topic c.o.m.p.*

Thanks.

--
John A. Grant  * I speak only for myself *  (remove 'z' to reply)
Airborne Geophysics, Geological Survey of Canada, Ottawa
If you followup, please do NOT e-mail me a copy: I will read it here

2. absolute coordinates of points in 3D world

3. 3D & Anim. sig

4. Kai Power Tools 3.0 need input

5. panorama picture converter?

6. ****"""""BUG IN PROGRAM"""*****