Frank,
Thank you for your notes and perspectives on digital vs. film resolution.
I regret that my posts both here on on the www.vrphotography.com web
site may have drawn your ire. I assure you that such was not the
intention.
Such comparisons of technology are usually subjective to some extent, no
matter how much objective data might be gathered. My intent with this
particular tech note above was to address the choice of digital vs. film
from the level of the photographer who is still undecided about which
format might best suit his/her VR photo efforts -- not one who is already
sure of a choice made. It also is approached with the assumption that
most individuals who might refer to it (which is why it can be called a
"reference") are probably less interested in the technical minutiae of
programming code and visual engineering, but who simply want an idea of
what the most efficient, effective and straightforward tools might be to
serve their VR purposes.
My posts were not intended to start another great debate on digital vs.
film technology. And I won't continue such a debate here beyond simply
responding to some of your concerns.
Digital imaging today has reached the level where it equals, and
sometimes surpasses, that of film in some applications. However, the
choice is always up to the individual as to what best suits his or her
purposes. Choices of hardware... choices of software... they're all
exactly that -- choices. The one's I make are not necessarily the ones
you'll make. And that's OK.
On Sept. 21, Frank van der Pol wrote:
>Color balance:
>"Auto white balancing" is someting one would never do shooting a
>panorama. You don't auto white balance. You set a custom white
>balance for all shots. Auto white balancing suggests you use
>the 'auto white balance' setting.
Many photographers DO in fact use auto balance when shooting digitally
and get perfectly acceptable results (although I neither like nor
recommend shooting VR sequences this way). In fact, with many consumer
cameras, it is impossible to turn auto white balancing off. Custom or
preset white balancing are certainly preferred, in my opinion. Note that
there's a difference between these. Many higher end digital cameras come
with "preset" color balance selections, such as "daylight", "tungsten"
and "Cool White fluorescent." These are different from "custom" settings
which the photographer or camera create based on actual color temperature
of the scene.
>Resolution
>In practice, resoluton of a finished panorama is infinite and
>entirely independant of the film/digital question. Resolution
>in practice only depends on the internal memory of the computer
>used for the stitching process. This practical limit applies
>to both film and digital.
I disagree. True resolution (as represented by resolving power, rather
than mere pixel count) is entirely dependent upon the collective
resolving power of the various components used, and does not exceed the
resolving power of the lowest component. Again, using the example of a
lens with a resolving power of 200 lines/mm recording onto a film with 50
lines/mm yields a combined resolving power of 40 lines/mm.
The formula is:
1/R = 1/r1 + 1/r2 + 1/r3...
(where R is the resolving power of the system and r1, r2, r3, etc. are
the resolving powers of the individual components)
So while it is possible to shoot an image at a certain resolution and
then add & change pixels digitally via interpolation, sharpening
algorithms, etc., the true resolution of the image is still very finite,
and is very much limited to something less than the least capable
component.
>"film-based models not usually capable of fisheye or 3600x3600
>coverage". This should be 3600x1800 coverage, I think.
I agree. The more correct term should be 360x180 degree coverage, since
it's 360 degrees around and 180 degrees from zenith (vertex) to nadir.
The 360x360 degree description has been used by hardware and software
companies to make their field of view description sound more "complete",
but I've never liked it. That's the hard part about writing things like
this -- using terminology that is familiar to users (often because of
advertising and sales efforts by others) while trying to maintain the
best accuracy.
In this instance, it'd be easier to simply say that it captured a
"complete spherical view", but then that might not be fully accurate if
you were using a cubic projection (plus, you might run afoul of
proprietary claims to spherical technologies). "360x360 degrees" is not
*incorrect* either, because you DO get full 360 views around both the
horizontal and vertical axes. But I agree that "360x180" is more
descriptive (especially since that better describes the range of pan/tilt
in most viewing windows).
>But I am still wondering why you said that "it's virtually
>impossible to stitch a dozen 41 MB images together, such as
>those from the DCS 14n, using current VR software applications."
>...
>If it's virtually impossible to stitch these images, how would
>a person using film then be able to stitch high res scans of a
>series of 35mm negs or slides?
Perhaps it would be better to say... "it's a pain in the butt" <grin>.
It's *possible* to stitch almost anything, but often impractical.
Several years ago, I stitched 72MB scans of medium format images (shot
with Pentax 67 film camera) to create ultra high res panoramas for a
client. The effort repeatedly crashed QTVRAS (even with help from Apple
engineers), and after two days of attempted workarounds, including other
VR software and Photoshop, the client realized that this much data was
unnecessary anyway.
Today, computers are faster and memory/processing capacity far greater,
but it is still an extremely slow and time consuming process to stitch
ultra high res images. This is not something that most VR photographers
are going to be doing in any volume. Yes, it's possible, but it's likely
an exception for most of us.
Photographers who DO choose to attempt it will have to jump through quite
a few hoops to be successful (newest computers, memory & processing speed
requirements, matching of capable software applications with available
OS, high end capture or scanning system requirements, and of course, the
time to deal with it all).
Again, it comes down to choices. It is *possible* to mount an engine and
a propeller on a brick and get it to fly, but is it really practical (or
necessary)? For some, it always will be, but for most, there are
probably better means to an end. That's not to say that *ultra-high* res
panoramas aren't useful or practical -- only that the need for them is
limited, and may mainly be of immediate concern to specialists.
Thanks again for your comments, Frank.
Regards,
Scott
Scott Highton
Author, Virtual Reality Photography
E-mail: email@hidden
Web: http://www.vrphotography.comhttp://www.highton.com
_______________________________________________
quicktime-vr mailing list | email@hidden
Help/Unsubscribe/Archives: http://www.lists.apple.com/mailman/listinfo/quicktime-vr
Do not post admin requests to the list. They will be ignored.
