[Openexr-devel] Careful description of your concepts for us to consider

[gary demos]

Dear Florian,

I enjoyed the well-attended birds-of-a-feather this morning.

I think it would be very helpful if you could write up your thoughts
about how you concieve the meaning of the OpenExr data when
used in the manner that you were proposing.  I'll not comment on
the proposal for a language to interpret the OpenExr data, because
I found that idea thought provoking, and I want to think about
it some more.  However, I think you did not really mean "scene referenced"
for OpenExr, but on the other hand, you didn't mean any other common
meaning for the state of the data either.  I view your concept of the 
OpenExr
data, as defined by the output process for film recording to density (and
possibly the process for display as well), as being a new idea, which
is not "scene referenced" but neither is it "output referenced".  I believe
that you described OpenExr data as having some fundamental linear
properties, such as the ability to use proportions for correct "pixel 
coverage"
anti aliasing.  You also described it as not intended to be device 
independent,
but rather as being specific to the film-out chain.

These are all good thoughts, worthy of further careful consideration.  
As such,
I would request that you write up, in your own words, as precisely as 
possible,
what you have in mind for OpenExr and the CTL to define it for film-out.

Then, based upon your concept, I believe we could also begin discussing
what this means for actual scene-referenced data, such as Viper 
quasi-log-to-linear
data, or other digital pixel sources where the data is strictly linear 
in proportion to
the light.  We can also extend the output discussion to include a TI 
DLP output
path, in addition to film-out, and possibly display-out.  Thus, while 
not device
independent,  this is "several-device".  There is also the concept of a 
digital intermediate
composite, where two disparate sources must be brought into a common 
space.  Are
both sources and the output of the composite all in OpenExr in your 
concept?

I look forward to your further thoughts, which you can post on your 
website, or
perhaps just send to us over the reflector.  I think you will stimulate 
interesting
email discussion of these fundamental issues.

I believe that the large attendance was a clear indication of the 
success of your
group's work on OpenExr, and its obvious growing popularity (me included).

-Gary Demos