[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Artisan Dreams

> Not that it's perfect just yet,  ITS STILL A COMPUTER!! But the folks from DV
> have been on the case...so, so far so good.  The more we make sense of it as
> colorists the more powerful the Artisan is...so, ask again in a month.
> Whew,
> Got a light?
> Andy Lichtstein
> Complete Post, LA

Well, as a Software Developer I have the freedom to create the control
environment and userinterface the way I want, but that may not be the
way the endusers wants it. What I need is more input from the endusers
about how they would like their control environment. I could implement
100's of different ways to control a colorcorrection unit. Anything from
commandline based to trackballs, mouse movement etc etc. But not many of
those would be usefull. When I sit down and play with our system and the
control environment, I find weak parts which need improvement. Like for
example, yesterday I was running the system and I found that our tape
transport control was actually very silly. It is a mirror of an ordinary
Betacam VTR panel with Rewind, Rec, Stop, Play and Ffwd. 

Question: How many times do you actually press Rewind? or Rec. 

I mean these are the buttons closest to your hand when operating the
tape transport. So I went back to my desk and started to implement an
improved tape transport. I decided to replace the Rec button with
Reverse playspeed (Which does not look too good on a Betacam without
framestore, but is usefull when searching the material). I replaced the
stop button with a still command. So when pressing stop, the tape does
not issue "STOP" just stops the tape motion. I replaced Rewind and Ffwd
with frame advance and frame stepback. Major improvement? Well better
transport control. But what I have noticed when talking to people is
that everybody wants thier own configuration, so I will probably have to
make the tape transport fully configurable.

Ok, that was just one example, and it was only something I did notice
when running the system. I do not operate the system 8-12 hours a day
5-6 days a week, so I cant see all the bottlenecks in the control

Ok, lets make the controlenvironment 100% configurable. It might be
good, some people like to be able to reconfigure everything to suit
there needs. But, it also have a drawback. All configuration options,
differences etc make the environment harder to understand for a new

What I need as a Software developer is guidelines about what you
operators and colorists want, which features that is usefull and which
are not. Do you want to colorcorrect a scene by adjusting the trackballs
until you like the result, or do you allready know what you want the
scene to look like and apply a colorcorrection based on what it is and
what you want it to be. 

I have not worked as a colorist, but I once worked as a copyist at a
1 hour photolab. We had old printers which basically had to be
controlled by simply looking at the negative and decide an exposure.
What we knew was that the machine always tried to pull the exposure
close to medium gray in brightness and color balance. What we had to do
was to overide that decision by adjusting the exposure and color balance
in different steps of RGB and exposure time. If there was a white statue
on a green lawn, the machine would make the statue magenta by its
internal measurements (Lots of green, increase green exposure to
increase magenta) so we had to manually reduce the green exposure
ourself by just looking at the negative and figure out what the machine
would do with it.

However, a colorist always sees the direct result of his or her
adjustments, and maybe film colorcorrection is not comparable to
negative printing, but there are similarities. What I liked with the
colorprinter I worked with, was that the printer always tried to make a
good exposure of a negative, the drawback was that the chosen exposure
was not always appropriate for just that paricular negative. Modern
printers perform some image analysis on the negative and make a more
qualified determination of the exposure. But the thing is, we would
never been able to print the images correctly if the printer didnt
perform its mesurement, beacuse the time between the exposure decision
and the final result was approximly 10-15 minutes, and to keep the 
1 hour promise and the waste low, you could not spend too much time
and copies on one every negative, at least not at the consumer level
which is the quicklab market.

So is automated support in colorcorrection of any interest? I am not
talking about removing the colorist, just give he or she a tool so that
he or she can work faster. I mean, the machine can measure colors,
calculate corrections but it is the human that must provide the
aestethic opinion.

Best Regards:
Mattias Ernelli
Software Developer
Digital Vision, DV Sweden

The telecine mailing list is run automatically by SmartList v.3.10.
It is available as a digest... questions to rob at xyzoom.alegria.com.