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

Re: FWD>RE>a few issues from Da

Hi Dave, 
your comments welcome, and not un -anticipated (g).

Let me say that our perspectives are of course different. I will not try to
argue with you on technical grounds, my observations are purely aesthetic, plus
opinions based on feedback  and interaction with other colorists. Personally I
am a colorist and not an engineer, so I cannot rely on science or theory, only
what appears real to me. No what I mean ?

>>What Grading practices cause noise?  etc etc.

In simplistic terms, setting up the PECs is not just a question of taking them
out of clips, and color balancing is not the same when done in (RGB) Ursa as
when done in (RGBY) Da Vinci. I'm not pointing fingers here, just using what I
know as a case in point.
If you set the PECs for low contrast, they have a high noise content: If you
then use colorgrade, da vinci etc etc to add contrast, you are amplifying the
signal, and the noise as well. My point is only that now we have so much
processing power, and so many ways to do the same thing, it is easy to forget
that the output, even though we can work wonders, is dependent on the input. I
am all for you cleaning the signal to noise with greater light output, better
balanced PECs etc, but a colorist who pays little attention to PEC setup,
colorgrade balance, even maybe Digital primaries, then corrects everything with
half a dozen secondaries (I'm sure nobody here does but I've seen it happen) is
going to have  more noise,grain whatever you want to call it, than somebody who
gets the same result with just PECs and one set of primaries. (My guess is that
the Ursa needs a fine resolution filter much earlier in its processing path, it
is trying to deal with frequencies it cannot resolve. But that is unfounded in
scientific fact) It is not a comment on hardware but simple common sense. You
can wash a shirt 10 times and make it 10 times cleaner, but you lose
*irretrievably* some of the color. If you get my meaning.

>>The color separation filters and optics in the Cell Box for URSA are

That may be true. My guess here is that the Ursa Gold has a better internal
layout that reduces crosstalk, hum , or I do not know what.

My experience on each of the machines in question, is that when I use Rank
output without noise reduction, into a Da Vinci with Kilovectors, an Ursa Gold
separates colors better than an Ursa, and Ursa better than a Mk111. I was able
to do the Mk111 to Ursa comparison in the same facility. (Actually I had to
redo a job on the Ursa, when the client pointed out "its like comparing 1"C
with D1" his words not mine. I did my best on the Mk111, I had nothing to gain
by doing the job again, and lost a nights sleep - literally). My observation of
Ursa to Ursa Gold was over a number of facilities, but by now I have a
collection of difficult material, that I know well, and which I have been able
to put up on a number of machines. For me seeing how real film behaves is more
useful than resolution test films. I admit that there was probably room for
improvement on all machines, engineers can always do something to get that
little bit more, but my observations stand.

>>Noise shouldn't be dependent on focus of the CRT unless....

Come on Dave, this is the real world. Noise is any unwanted data in the final
image. Technically you are of course right. But since film has a higher
information resolution than video, some of that information if allowed to pass
through the system will end up as just noise. A clear way to see what I'm
referring to is to look at a detailed and saturated D1 picture, in component
digital, then compare it with a coded analogue output. There are artifacts
which are purely noise in the coded domain. In the coded situation a notch
filter can solve problems, in the film domain **some people** recommend a
reduction of electronic focus. And it works, but I do not agree with it myself.
It should not be necessary.

Incedentally, since this has become an essay :-}  (anybody still reading?) I
also perceive this as a significant difference between Rank and BTS machines.
BTS appear to have "cleaner" pictures, but on examination it is at the expense
of fine detail. To me it seems that they have the coring way up. Aperture
correction coring is as far as I know adjustable on Ranks, but not BTS
machines, and BTS inevitably show there machines with noise reduction where
Rank do not. I'm curious as to what a naked BTS Quadra really looks like!

>>Wouldn't it be preferral to see the scanned film image with no electronic
>>noise added by the telecine .....

Of course it would, I hope we can have that tomorrow. But right now, in the
real time world (as opposed to scanning or pinning at a few frames a second or
a minute) we colorists get the best we can from what we've got. If you can do
better, I am 100% behind you. Given freedom of choice, I prefer to work with
Ursa Gold, Da Vinci 888, and DVNR 4x4. It is not perfect, but it is the best I,
a colorist, with no engineering expertise, have been able to achieve. In truth
I am surprised that this combination has remained unchallenged for very nearly
2 years. A pretty healthy lifespan in this industry.

>>I just
>>feel it's time to clean up the URSA signal path. I feel noise shouldn't be a
>>part of the signal anymore and I view it as a challenge to tackle and resolve

What was the Mission Impossible theme. "Your mission should you chose to accept
it...this tape will self destruct in 10 seconds".

Your job/mission makes our jobs easier and better. I do not belive in a system
that is so noise free, the grading is irrelevant. The essence of film to tape
is not just the difference in resolution, but also the mediums themselves.
Current day tv has a very poor contrast and chroma response compared to
projected film, or for that matter computer graphics. 

I see the future in truly higher resolution viewing systems, and with out
losses due to noise or for that matter compression (Please somebody say
something about compression - the artifacts of compression make telecine noise
look insignificant, are we seriously unconcerned?)

OK monologue over. Dave your comments are taken as well meaning, and I would
like to restate I can not nor wish to dispute any of the facts you quote. 

My opinion is that a colorist is required to do more than operate a machine,
he/she is required to create an impression

Kevin                "The cruelest lies are often told in silence"
Zurich                -Robert Louis Stephenson (1881)