[Tig] Monitor calibration

Ian Richardson ian.richardson
Wed Jun 15 08:41:32 BST 2005

Hi all,
I thought I would throw in a bit from 'down under'
At this current facility we use a Minolta 2150 [not my choice but is OK]
Luminance we measure in cd/m2 not usually the good old Ft-L however......
We do not use 'illuminated surrounds' but rely on D6500 'spill light' with
neutral [grey] background walls behind monitors.
All 3 x T/C rooms are similar [ish], including Inferno's Flames, DS-HD's, Smokes etc.
We set on clean full frame Pluge not a 'combo signal' [ie with bars]......
Lum to 75 cd/m2 this equates to 21.89 Ft-L, say 22Ft-L  [1FL = 3.426 cd/m2 I recall]
This is in the bottom of the spread Martin mentions below.
It is good for all our rooms.
Black is set on Pluge at the 'nominal artiste viewing distance' 3% bar just visible
We set x=315 y=309. the x&y points differ from
D65 spec but these were arrived at 'years ago' by consensus between
all the [then] existing facilities and their respective instruments.
In the interests of interchange between facilities. [probably time to do it again
if there is anybody left to test with in our current business climate]
I/we find that the best results for tracking near black are done by 'eye'
[bottom Pluge step down to black] One gets quite good at this, and my opinion
is with all the D65 in your face as you do it it seems to result in correct
black bal. where it counts. Our meter cant resolve at this low level.
We then track up the brightness to apx 15 % and see if we get any hue shifts
or tracking errors. Then restore to correct brightness etc.

Ian Richardson
Chief Techo and apprentice cat herder
Digital Pictures

Martin Euredjian wrote:

> Thanks to James Braid of Oktobor for supporting the TIG
> --
> Thanks Rob,
> So, you are doing 35fl at 100% and 0.79fl at 20%.  Does this include
> BVM-D32's? 
> Now we have 35, 30, 28, 21, 20 and 15 footlamberts as quoted white point
> levels being used by facilities world-wide.
> While no numbers exist for target black levels, the grey-level method ranges
> from 0.1fl at 10% video to 0.79fl at 20%.  Mathematically speaking, if you
> take the 0.1/10 measurement and 30fl for white and calculate what the number
> would have to be at 20% it would be closer to 0.6fl rather than 0.79.  A
> similar error (can we really call it that?) is had if we take 0.79 as "true"
> and calculate the number for 10% video.
>>helps in big facility where multiple people will be calibrating displays
> since you can give them  
> absolute numbers to shoot for.
> This is the one argument I heard for using absolute number rather than using
> pluge to ultimately set blacks.  Of course, this only works if your viewing
> environment is reasonably consistent.
> Having said that, we ran some tests to look at variation in black level
> adjustment (via pluge) with various pluge configurations and room
> illumination conditions.  No conclusive data yet.  Just lots of numbers.
> I had a good conversation with Bill Hogan about this very topic last week.
> He is of the opinion that we need to fix pluge to have it be one digital
> value above and below digital black AND use full frame pluge rather than
> pluge off in a corner in SMPTE bars.  I don't think there is much of an
> argument against the idea of the +/- 1 pluge (not % but digital step).
> However, some I've talked to are of the opinion that setting pluge without
> at least a white chip somewhere in the frame is a bad idea.  I haven't put
> much thought into this at all yet, but we did run some tests with and
> without a white chip to get numbers.  
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> Martin Euredjian
> eCinema Systems, Inc.
> voice: 661-305-9320
> fax: 661-775-4876
> martin at ecinemasys.com
> www.ecinemasys.com
> participate in the tig wiki pages at http://tig.colorist.org/wiki
> _______________________________________________
> Tig mailing list
> Tig at tig.colorist.org
> http://tig.oktobor.com/mailman/listinfo/tig
> _____________________________________________________________________ 
> This e-mail has been scanned by MessageLabs.

More information about the Tig mailing list