PDA

View Full Version : Just in case you were wondering how accurate ISO 800 is on Dragon...



Evin Grant
02-01-2014, 03:33 PM
I decided to do a quick and dirty ISO verification of my Dragon last night, simple setup using a 500W Lowel Omni and a Macbeth chart.
This is obviously tungsten only and I'll try to do a sunlight and HMI test as well but it may take a week or so.

Notes:

Lens used was a 50mm f1.8(T1.9) AFS-G Nikkor so I could show the displayed stop on the 5" touch. (any discrepancy at f8 between the f & T stop is under 1/10 stop and within the margin of error of this test.)

The 18% chip exposure was verified by a Sekonic incident and Minolta spot meters (see pic) as well as the "video" exposure mode and a Zebra set to 41-43 IRE.

The frame grab was made in RC3/RG3 but was the same result in RLF.

This is only a reference to 18% grey placement as it relates to the indicated ISO, all other tone placements above and below will be determined by the curve in use.

http://www.evingrant.com/pics/Dragon800.jpg

Brian Merlen
02-01-2014, 04:09 PM
thanks for the test! obviously much appreciated!

Martin Stevens
02-01-2014, 04:50 PM
Great info Evin!

Andreas Mendritzki
02-01-2014, 04:55 PM
Thanks. That's good to see

Kwan Khan
02-01-2014, 04:56 PM
Thanks for sharing

Marc Wielage
02-01-2014, 05:49 PM
I decided to do a quick and dirty ISO verification of my Dragon last night, simple setup using a 500W Lowel Omni and a Macbeth chart.
I've often said that I think the standard Macbeth chart is worthless in an all-digital world. It had a meaning in a photochemical film lab, but not for digital.

If you're trying to measure exposure, get the DSC Labs grayscale chart (http://www.dsclabs.com/grayscales.htm):

http://www.dvxuser.com/articles/rescharts/Toni_050511_E2W_3ins_sm.jpg

I much, much prefer a greyscale chart to a chart with random chips of color paints that have no bearing on modern-day camera chips. The DSC people are pretty sharp.

Evin Grant
02-01-2014, 06:17 PM
I much, much prefer a greyscale chart to a chart with random chips of color paints that have no bearing on modern-day camera chips.

I wasn't using the color chips, just the 18% grey chip. With the backup from both the incident and spot measurements I can say with certainty that chip is accurate.
The rest of the chart is of limited use I agree, but it is a known quantity and that gives it some bearing as a reference outside the particular electro-optical system you may be testing.

The truth is I couldn't find my 15 year old Kodak grey card and I'm not sure I'd trust it at this point anyway.

Martin Stevens
02-01-2014, 07:11 PM
I wasn't using the color chips, just the 18% grey chip. With the backup from both the incident and spot measurements I can say with certainty that chip is accurate.
The rest of the chart is of limited use I agree, but it is a known quantity and that gives it some bearing as a reference outside the particular electro-optical system you may be testing.

The truth is I couldn't find my 15 year old Kodak grey card and I'm not sure I'd trust it at this point anyway.

Whoa. I have a gray card from 1982. Perhaps I should toss it. :)

Gunleik Groven
02-02-2014, 02:48 AM
Thanks Evin!

KETCH ROSSi
02-02-2014, 02:52 AM
Nicely done Evin!

Sanjin Jukic
02-02-2014, 03:12 AM
Nice!!!

Michael Tiemann
02-02-2014, 05:24 AM
Whoa. I have a gray card from 1982. Perhaps I should toss it. :)

Definitely! After all, black is the new gray ;-)

Jacobo Martinez
02-02-2014, 09:47 AM
Guys, is 18% grey in the exposure meter ( GREEN HIGHLIGHT ) of red cameras 43 IRE? or is it 50 IRE? When you rate at 800iso in the camera, am I suppose to see 43 IRE in RCX or should it be 50 IRE?

Evin Grant
02-02-2014, 01:02 PM
Guys, is 18% grey in the exposure meter ( GREEN HIGHLIGHT ) of red cameras 43 IRE? or is it 50 IRE? When you rate at 800iso in the camera, am I suppose to see 43 IRE in RCX or should it be 50 IRE?

42-43 IRE is what Red's meters considered middle grey, I believe Arri uses 45 on the Alexa but it's widely considered anywhere between 42-50 but I've not seen a camera that uses anything higher than 47.

Remember the legal IRE range is 7.5-100 and it was originally an analog video brightness scale which is arbitrary when applied to the digital space.

Jacobo Martinez
02-02-2014, 02:03 PM
Thanks Evin. Just to understand what is going on the raw side, when i expose to iso 800, and middle
gray is 42 IRE, i am actualy exposing to about 30 IRE in the raw right?

Evin Grant
02-02-2014, 04:48 PM
The "RAW" view is a bit of a misnomer, it's not linear light and still has a curve applied plus it's not tuned to the Dragon sensor. I don't use the RAW view at all, I use the histogram, a meter and a calibrated monitor for exposure. If your big concern is noise then you need to test each EI and then bring it through your specific post chain and evaluate them in your specific delivery format on a suitable display for that format. Your tolerance for noise has as much to do with how the footage is being developed, delivered and displayed as it does what EI you shot at. So far I've projected (2K) up to EI 2000 Dragon footage and been very happy with the results as long as it's properly exposed.

Robert Ruffo New
02-03-2014, 07:03 PM
Very smart thing to test! Cool! Thank you!

Marc Wielage
02-03-2014, 07:53 PM
I wasn't using the color chips, just the 18% grey chip. With the backup from both the incident and spot measurements I can say with certainty that chip is accurate.
A full-frame grayscale chart like the one I suggested would be infinitely more usable.

Some interesting tips from DP Ryan Walters on how he uses Kodak grayscale charts to measure different cameras:

http://www.ryanewalters.com/Blog/blog.php?id=3551583675371023276

Evin Grant
02-04-2014, 02:26 AM
Marc, I suppose you missed the "quick and dirty" part of my opening sentence :toetap05:

I'm well aware of the limitations of the the grey card, but I think this is the most salient point in Ryan's article...

"1. If it ain't broke, don't fix it.
If your current metering methodology is yielding consistent, predictable results, more power to you! Don't sweat the fact that 18% really isn't 18%. In the end it is about the results you produce, not the technicalities."

The point being that Red placing their "middle gray" exposure at 42 IRE appropriately compensates for any reflectivity difference, so call it 12% or late for dinner, with all three measurements in alignment you can feel confident using your incident, spot or in-camera meter for proper exposure without compensation, assuming you have tested your 12-18% grey card.

And yes, I need to buy a new grey card, probably this one from X-Rite though, those DSC charts are bit overpriced these days...
http://xritephoto.com/images/products/M50103_M1.jpg

Morning Glory
02-04-2014, 08:39 AM
Thanks, Evin. Very Quick. and Dirty. informative, too!

Peter Lyons Collister, ASC
02-04-2014, 09:12 AM
Thanks Evan,

I used 800 ISO on my meters throughout my little RKO film in Atlanta and was very pleased with how it processes out. I have yet to run some shots all the way through 2K DI to see how the noise is but you are correct that one's own tolerance for noise from a creative/Story standpoint and how the post work and delivery system (vimeo all the way to 2K DCP for theaters) is the true test.

Thanks again for taking the time to test and share.

Peter

Marc Wielage
02-04-2014, 09:11 PM
Marc, I suppose you missed the "quick and dirty" part of my opening sentence...
Well, is it "quick and dirty," or is it "accurate"? I'm not sure you can have both. Here's a grayscale chart that used to be very, very standard for HD cameras on TV sets, and it's $39:

http://www.markertek.com/productImage/HI-RES/GSHD.JPG

http://www.markertek.com/Tools-Test-Equipment/Video-Test-Charts/GSHD.xhtml

Talk to any experienced colorist in the world and they'll most likely tell you that this is far more useful than just a standard Kodak 18% grayscale chart or a Macbeth chart.


The point being that Red placing their "middle gray" exposure at 42 IRE appropriately compensates for any reflectivity difference, so call it 12% or late for dinner, with all three measurements in alignment you can feel confident using your incident, spot or in-camera meter for proper exposure without compensation, assuming you have tested your 12-18% grey card.
I don't think it matters as long as you have a 100% white field and a 100% black field on the chart, with various steps of gray all the way up. Seeing that on an RGB parade scope and a histogram will give you a much better understanding of how the camera is exposing and how it's going to fare in color correction, as well as where the real-world limitations are going to be.

Exposure is a wide-open subject (no pun intended), and there are so many ways to interpret virtual ISOs in digital cameras, there's no one method that's 100% foolproof. I wish that Red and Arri would provide more information as to how they rate their cameras and where these numbers come from. I don't think either camera can really do 800 reliably, not without noise issues. In a lot of cases, the noise doesn't matter, so it's not really that much of an issue.

Evin Grant
02-04-2014, 11:02 PM
Well, is it "quick and dirty," or is it "accurate"? I'm not sure you can have both.

I can, thanks.

Keith_D
02-05-2014, 12:09 AM
Thank you Evan, this was very informative and much appreciated.