Cinematography Mailing List - CML
    advanced

Reflectance vs Volts & RGB Values

That one will creep up forever; but I'm curious to poll you guys as to what you believe is the correspondence between subject reflectance (as in 18% gray card) and the voltages on a waveform monitor. To make things more interesting, how then this all relates to RGB values.

The lack of consensus here is astonishing. Why do you believe Kodak's TEC system with the Gray Card Plus is engineered for specific values, and how do you believe this all relates to shooting in HD, when scanning film, etc...

Enjoy.

Philippe Panzini


>That one will creep up forever; but I'm curious to poll you guys as to >what you believe is the correspondence between subject reflectance >(as in 18% gray card) and the voltages on a waveform monitor.

The relationship is non-linear, and you can think of it as basically being the same as the characteristic curve of a film (ie. as a log brightness/log density or log voltage plot).

In the old days, most of the non-linearity was due to the camera tube, and the data sheet in the RCA handbook gave you curves for various operating levels. I probably have curves for some plumbicons here in my office.

On modern CCD systems, I don't know what the curves look like, but they MUST be documented somewhere in the camera documentation.

Scott Kludge

>That one will creep up forever; but I'm curious to poll you guys as to >what you believe is the correspondence between subject reflectance >(as in 18% gray card) and the voltages on a waveform monitor.

In SD I think of 18% gray being around 52-55 ire. Someone taught me that years ago and it's always worked out fine, although I rarely use it. I'm always looking at the top and bottom of the range instead of the middle.

Not sure what you mean by RGB values.

Art Adams, DP
Mountain View, California - "Silicon Valley"


>In SD I think of 18% gray being around 52-55 ire. Someone taught me >that years ago and it's always worked out fine, although I rarely use it.

A related question : I'm trying to make some custom video charts.

Anybody know what 18% is in Photoshop terms? Also, what would 71 IRE (same as Zebra's) in Photoshop terms (HSB or whatever).

Here's what I'm trying to do. I want to make a grey card that shows me 70 or 71 IRE (what would be typical skin tone) so that I can compare it with typical reflectance in the scene. Also to set exposure before the talent arrives, at least for testing.

Once I print it out, I'll be able to test with a spot meter, but I would rather avoid an endless series of tests, if I can.

Blain Brown
DP
LA

If you look in a high end video camera which puts the zebras in the bars you will find that 70% ire is in the yellow bar give or take a couple of points.

What you could do is print a series of yellow bars and compare that with the 10% grey and a white card. This might be a good place to start.

Here's what I'm trying to do. I want to make a grey card that shows me 70 or 71 IRE

Paul M. Zenk
DP
Los Angeles, Reno, Sacramento
www.paulzenk.com

>Once I print it out, I'll be able to test with a spot meter, but I would rather >avoid an endless series of tests, if I can.

Hmm. I used to have a Russian spot meter which was calibrated to (caucasian) skin tones. It was bloody useful when in a hurry.

Not that I am answering your posting properly but hey, neither does anyone else.

Regards

Chris Maris
UKDP (soon to be based in Sweden too)

Is there any correlation between an f-stop's worth of exposure change
compared to an IRE level change?

David Mullen ASC
Cinematographer / L.A.

>Is there any correlation between an f-stop's worth of exposure change >compared to an IRE level change?

Not in my experience. Part of the reason is that IRE is a linear scale, while f stops are essentially logarithmic. Another part is that video color correctors have separate control over blacks, whites, and midranges – which means that contrast can be readjusted. Yet a third factor is that different film stocks react in different ways, so the densitometric (is that a word?) result of a stop change is different on say, a stock of ASA250 than it is on, say an ASA500.

If what you're asking is more related to video origination, say HD, I think you still run into the same issues in that camera sensitivity (i.e., the camera's "D log E curve") causes the relationship to be different depending upon what camera and lens combination you're using, just as film stocks do.

Now, having said all that, I do recall that when Kodak first tried to bring out calibration charts for telecine (the original Telecine Toolkit) they had a display that represented one stop as, I think, 15 IRE. And Yuri Neyman's chart (the Gamma & Density chart) has specific IRE values for each gray chip. However, I still don't believe you can come up with a specific translation, because there are just too many variables. Less so on film scanners, though.

Mike Most
VFX Supervisor
IATSE Local 600
Los Angeles

>Is there any correlation between an f-stop's worth of exposure change >compared to an IRE level change?

I bet it would vary depending on the CCD, DSP, and menu setups - I would expect that changing the gamma settings, for instance, could affect the IRE level change considerably.

George Hupka
Director/DP
Downstream Pictures
Saskatoon, Canada

David Mullen ASC wrote:

>Is there any correlation between an f-stop's worth of exposure change >compared to an IRE level change?

In a word Yes, but I forgot what it was. It was *years* ago that I checked. Too Bad Walter isn't lurking. Hey Lou C, you out there?? Lou Comenetz will know. I think its about 7 lines of video .

Mark Smith

>Is there any correlation between an f-stop's worth of exposure >changecompared to an IRE level change?

No, because gamma correction and knee/DCC tend to confuse the calculations.

The theory I've heard is that in a perfect world with linear camera response, a one stop increase at the lens will result in a doubling of the ire readout, but in practice I've never seen this correlation except at very specific starting ire levels.

Bob Kertesz
BlueScreen LLC

>No, because gamma correction and knee/DCC tend to confuse the >calculations

I agree and say yes but.... The way video cameras are these days with tons of menus and the ability to slice and dice the curve it gets to be a little hard to make one answer fit all. Like what is the difference if you make your gamma crossover at 55 or 60 instead of 45??

Still think if you look at the big pic, turn the variable knee off, and see the results you will find that that there is a pattern that emerges from all the clutter.

Now darn it I suppose I have to find a WFM and jog my memory as to what it is.

Mark Smith

>Is there any correlation between an f-stop's worth of exposure change >compared to an IRE level change?

I read 15-20 ire per stop somewhere. It holds roughly true for SDTV.

Art Adams, DP
Mountain View, California - "Silicon Valley"

70 ire= 467mv in RGB Colorspace

The equivalent RGB values in Photoshop are R=171 G=171 B=171

Regarding the waveform monitor portion of the thread and how it relates to television cameras : with gamma correction on, and a gamma crossover of 45-50 IRE, every 10 IRE change in level above gamma is 1 f stop. In other words, if you have a shot of a white wall, and the white level is 90 IRE, closing the lens 1 full stop will result in a white level of 80 IRE.

This is true in both SD and HD cameras.

Dave Satin
Video Engineer

>with gamma correction on, and a gamma crossover of 45-50 IRE, every >10 IRE change in level above gamma is 1 f stop.

I hate to be ornery (no, wait, that's not true!) but I've never worked with a video camera, SD or HD, that had five stops of latitude between 50 ire and 100 ire. 2.5 stops maybe, at a stretch, but never 5 stops.

Art Adams, DP
Mountain View, California - "Silicon Valley"

>Here's what I'm trying to do. I want to make a grey card that shows me >70 or 71 IRE

Many lab-provided gray scales, like FotoKem's, have an enlarged light gray square opposite the darker 18% gray square. I guess it's meant to represent caucasian skin tone reflectance and probably is one of the zone's in the
Zone System.

But anyway, I've found it to be about 70 IRE. If you set your zebras to appear at both 70 IRE and 100 IRE and set the exposure so that the white strip zebras in the viewfinder, you'll find that the light grey area will zebra as well.

David Mullen ASC
Cinematographer / L.A.

Art Adams writes :

>I read 15-20 ire per stop somewhere. It holds roughly true for SDTV...

I've been following this thread and asking myself: "If an f-stop's worth of change represents a doubling or halving of exposure, the translation formula to IRE has to be logarithmic, modified by the factor of the gamma curve at a particular point on the curve.

So how can it be anything as simple or linear as this -- roughly or otherwise? (Unless, of course, you mean VERY Roughly, INDESCRIBABLY Roughly, or OFF-THE-CHARTS Roughly!)

Dan Drasin
Producer/DP
Marin County, CA

>(Unless, of course, you mean VERY Roughly, INDESCRIBABLY >Roughly, or OFF-THE-CHARTS Roughly!)

That about sums it up. If I've got a light that's too bright and I look at the waveform and it says ten units hot, I tell my gaffer to drop a single in the offending lamp. That usually does it.

That's the most I've ever used that information for and it works fine.

Art Adams, DP
Mountain View, California - "Silicon Valley"

David Mullen ASC wrote:

>Many lab-provided gray scales, like FotoKem's, have an enlarged light >graysquare opposite the darker 18% gray square.

Thanks, I have a Fotokem chart, I'll give that a try.

Blain Brown
DP
LA


Copyright © CML. All rights reserved.