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

Telecine No- Drift !



In response to Robert Lovejoy's statements about telecine repeatability
over time.

> On a Rank, the CRT ages significantly over its lifescan.
>
> There are so many variables on a telecine - ---- whose values drift over
the lifetime > of the device.
>
>CCDs do not drift over time as much as tubes and it is likely that their
corrections  > will hold much longer.  Tube ageing is the most significant
reason for corrections  > having errors over time.

Robert implies that telecines are not stable in terms of their
representation of transferring film image densities to video levels. 

He advises this to be a problem machine to machine and for a same machine
over a period of time and for this to be more true of CRT based systems
than CCD.

Well Robert if we were talking 1970's & 1980's CRT or CCD telecines  I
would have to agree.

However modern CRT based digital telecines have built in absolute
references and drift correction. The earliest version of this was Ursa's
automatic alignment.

This procedure could be used as often as required, usually at a gate change
or at least on a daily basis.

This correction took into account CRT level, and color balance (I have yet
to see a Cintel/Brimar CRT phosphor that's color balance changes with age
or temperature ). The correction also took into account PEC and any analog
front end discrepancies. This correction would always pull the telecine to
within about 1% of any previous default correction regardless of any device
ageing, or even the use of a different gate with a different lens.

BTW this type of correction is identical to CCD telecines which use a
system known as FPN (Fixed Pattern Noise) correction, and I can assure you
light bulbs age a lot faster and over a much greater brightness range than
CRTs.

Modern telecines, both CRT & CCD based utilise digital color channels, no
drift here.

The latest CRT telecine designs also include means to stabilise second by
second between auto aligns. Ursa Diamond in particular includes systems
that remove even the minutest second by second variations, Diamond Glow or
Twigi remove ALL CRT variation, as well as getting rid of the last vestige
of CRT grain, while Diamond set removes even the smallest variation in
detector level regardless of how the detector is abused in terms of light
hitting it or it's gain being changed.


Bear in mind that CCDs are analog detectors that do drift quite
considerably, so while the latest CCD telecines have feedback to help avoid
the relatively large variations in their light sources, what is done second
to second or even hour to hour to remove CCD drift ?  Maybe something or
maybe not.

Therefore assuming the telecine you use has had an auto alignment performed
and the color channel has not had some engineering trims changed since your
last session then you should get the same result each and every time. 

> I would really love to be able to see/have some kind of readout/printout,

> whatever, of the session.

Well I guess this is up to the Da Vinci's and Pandora's of this world.
In theory a whole list of data could be obtained but it would not look like
a simple lab RGB printer point readout.

A minimum meaningful colorimetric readout, excluding secondaries would be
:-

Detector level : Red, Green, Blue
Gain level : Red, Green, Blue
Gamma setting : Red, Green Blue
Lift level : Red, Green Blue

Bear in mind that one facility may choose to adjust detector level while
another may choose gain level to achieve the same result. And if
secondaries are used then it becomes a whole new ball game.

> Modern color correctors also offer global correction ability
>
> This could quickly return a corrupted-by-time set of corrections to
viable ones.
> It is unlikely that using this method will result in a 100 per cent
recovery of the 
> original scene.

Robert also refers to global corrections to an existing list

Global trims, I understand, usually cause colorist to perspire to the
extent that maybe we should make telecine controllers rust proof. Why are
global trims such a problem.

When it comes to color correction we need to consider whether a trim should
be added/subtracted mathematically or multiplied/divided. This is an almost
impossible decision as often a mixture of both are required.

Take the examples of two scenes, one with 100% luminance content one with
30%.




We want to make the 100% luminance 80%, 20% less than originally set. Easy,
just subtract 20% from the luminance. But if it is a global trim do we
really want to subtract 20% of peak luminance from the second scene and
make it 10% or do we actually want to subtract 20% of it's 30% luminance
and make it 24%.  The answer is probably somewhere in between, or more
likely the voices from behind will say no, no , we don't want to change the
luminance at all of the second scene, only those scenes at about 100%
luminance.

A difficult call I think you will agree.

So rest assured modern telecines are repeatable and becoming more so, not
only second to second hour to hour and day to day, but month to month and
why not decade to decade.

We have progressed since TOPSY, and you ain't seen nothing yet.

Peter Swinson.


 

 
++
thanks to Dave Corbitt for support of the TIG in 1997
---
mailinglist digest available......posting guidelines on the webpage
the Telecine Internet Group  <http://www.alegria.com/telecinehome.html>


Follow-Ups: