Red MX reading 3/4-stop under

[Andreas Mendritzki]

Ok, so I'm trying to get to the bottom of an issue that has come to my attention:


Our camera, it seems, is consistently giving me about 2/3-1-stop under what it 'should'.

Here's how I'm coming to this conclusion:

If I shoot a gray-card, with the camera at ISO 800 with VM=Red Color, and turn on false color on the cam, I get 18% gray (Green, in False Color) at, lets say, T 5.6.
However, doing a reflective reading with a light-meter off the gray card @ ISO 800, it suggests a T-stop of 5.6 and 2/3.

Same thing at ISO 320 with VM=RAW. The light-meter suggests a T-stop that's 2/3rds of a under what I actually need to open up to to get 18% on the False Color meter.



At first I chalked this up to just a difference between the way the RED and the Sekonic where interpreting their ISO. However, we recently had a client with two REDs side by side and the DP said that he felt like our camera was consistently 1-stop under the other Red.

Unfortunately I don't have another Red here to do a quick and easy A/B set up to check this.


Does this seem like its within the realm of normalcy? Because its not looking that way to me. . .

thanks

 

[Alexander Ibrahim]

Check to see if you have a custom look loaded into the camera.

That sort of thing. I think you can do a force upgrade and wipe out all your settings in the camera. Doublecheck that.

If you are clear of any customizations that might affect exposure ... then open a ticket with Red.

 

[Deanan]

Many reflected meters are calibrated for 12% and not 18%.

 

[Andreas Mendritzki]

Right... I mean, its pretty clear that the only way to be sure is to test it side by side with another camera.

(for what its worth, yes, I re- installed firmware, black-shaded, etc)

 

[Alexander Ibrahim]

Many reflected meters are calibrated for 12% and not 18%.

I think you chose to respond to the less interesting part of the problem description.

The card could well be off, the meter could be off, the camera could be off. With that set up, who knows. I have issues with the described rating test ... I'll address them separately.

As I understand it, the problem is less with the specific reading ... but rather that this Red's response differs from the response of another Red. They should be pretty close to identical.

Let me say that differently: It doesn't matter what you are metering or what the specific value is. If this camera is consistently seeing things 2/3 of a stop differently from another identical model - that's a problem.

Now, it could be that the other DP doesn't know what they are talking about. It could be that the other camera had a preset loaded on, or just that it was set up differently.

 

[Alexander Ibrahim]

Right... I mean, its pretty clear that the only way to be sure is to test it side by side with another camera.

(for what its worth, yes, I re- installed firmware, black-shaded, etc)

You may have done what I am about to write about ... but I think its worth writing anyway.

Well, as Deanan pointed out we don't know how the gray card was calibrated.

We can eliminate that using the light meter as a reference point - provided you have confidence in that device.

How? Candles of course!

First off, I'd set your meter to give readings in 1/10 of a stop for this exercise. (I use a Sekonics L-758C, several other meters allow the same operation with approximately the same precision.)

Get out your ASC manual and pick the stop you want to test at. I'll suggest t/5.6. At ISO 800 this is 50 candles. At ISO 320 this is 125 candles.

Light the card using the incident meter to those light levels.

Now use the reflected meter and determine what ISO the meter gives you a reading on stop for.

This effectively calibrates the card. Because you are using the same meter, you should get a spot on result if the card is exactly 18% reflectance. Your card might be 12% reflectance, 18% reflectance ... or it may have originally been one of those reflectance levels and faded/darkened over time.

Now, if you have separate incident and reflective meters ... there is some ambiguity. The meters may not perfectly coincide. In this case, you can't tell if the meters are set up to different standards, or if the card differs. Unless you are 100% confident in your card (i.e. its a new DSC card or other high quality card. The el cheapo Kodak cards don't count!), then I recommend you assume the error is in the gray card for this exercise. (Meters tend to be more precisely calibrated in the factory, and their settings are more stable than a card's reflectance.)

Anyway ... you have an actual stop reading for your card/meter system at this point.

So, you know what you have lit the card to, and what it reflects. If you've lit to 50 candles and set your meter to read at ISO 800, then you should get within 2/10 of a stop of t/5.6. If you get close to t/4 instead ... then you have a 12% card.

Fine ... whatever. Now you know where you should set your camera's stop to get the current set up to read at 50%.

I suggest, finagling the lighting so that the reflected meter gives exactly t/5.6 at ISO 320. Now RAW false color should give 50% for the card at t/5.6. I've actually never tested this against RedLog ... but I assume that RedLog will vary from the RAW exposure, and its worth understanding how that varies.

Notice I've been using t stops. If you use f-stop lenses, you will have some variance, and that variance may be non-linear.

Now you are comparing the camera to a known standard - your card/light meter system.

That card/light meter system could still be screwed up though. Using multiple light meters and finding them in accordance can increase your confidence in the card/meter system. If you have two meters giving the same results ... you can really trust the system. (That's why I recommend having two meters if you are a DP, and calibrating them against each other!) Your AC or gaffer or another friendly DP may come to the rescue.

Additionally testing against multiple cameras can help identify whether or not one particular camera is out of calibration.

Its important to note that some older Red One M cameras have different results at a given ISO than Red One MX cameras. Particularly if the M has older firmware on it, but even with current firmware. They are different sensors, so this is expected. You also can't tell if a Red One MX is calibrated properly by testing it against a Scarlet, Epic or Alexa. These are all different systems ... you need to test against an identical camera with the same set up/firmware revision.

If these tests are inconclusive, then you need better tools than a typical studio has. Your rental house might be able to help refine your testing ... but I'd look to Red unless you are in LA or NYC and deal with a better house.

It should be noted that even with substantial care, these tools will leave a little variation in your data. Don't expect "theoretically perfect" results from these tests. If you need very high precision, you may as well start at a rental house and work your way up to vendor labs.

In my opinion some care and my normal working tools gives me the results I need.

 

[Andreas Mendritzki]

Hi Alexander

Thanks sir, for the very thorough response. I really appreciate it.

We've been using the same Sekonics L-758C - so no problems there!


The method you describe makes sense.

It still doesn't rule out the possibility that ALL Reds are simply a bit 'optimistic' in their read-out. But frankly, that's not my experience.

Tomorrow I'm going down to our local rental facility and going to do an A/B test with another MX. That should answer a number of questions.

 

[Andreas Mendritzki]

For what its worth, I did my AB test with another MX, and they're coming out the same.

I guess its just a systematic difference in calibration between Red and the Sekonic (as Deanan pointed out above) . I'm surprised more people havn't commented on this before.

 

[Alexander Ibrahim]

For what its worth, I did my AB test with another MX, and they're coming out the same.

I guess its just a systematic difference in calibration between Red and the Sekonic (as Deanan pointed out above) . I'm surprised more people havn't commented on this before.

I don't see that with my Sekonic.

Now ... I could have created a profile somewhere in the mists of time for the Red ... I don't recall anymore. Mostly these days I read the meter in candles and not exposures.