ISO reading

[Miguel "Macgregor" De Olaso]

Hi all, while shooting last week i found myself incapable of measuring the light with my cine lightmeter. THe readings i was getting did not match what i was seeing in the RED LCD. At least i could say that what i see in the RED LCD is what i get. That´s good.
But the light meter set at 320iso was producing images at least 2 stops underxposed in the red raw footage.
So finally i had to throw away my lightmeter and make exposure by looking at the LCD and checking for the highlights not to be clipped. After each shot, i quickly gave the card to Blair or Brooke who told me if the shot was ok or not by cheking the histogram in red alert.
I think we did not have any shot under or overexposed by more than half stop.

SO my question is: has anyone checked the ISO by himself?

[Stephen Williams]

Hi all, while shooting last week i found myself incapable of measuring the light with my cine lightmeter. THe readings i was getting did not match what i was seeing in the RED LCD. At least i could say that what i see in the RED LCD is what i get. That´s good.
But the light meter set at 320iso was producing images at least 2 stops underxposed in the red raw footage.
So finally i had to throw away my lightmeter and make exposure by looking at the LCD and checking for the highlights not to be clipped. After each shot, i quickly gave the card to Blair or Brooke who told me if the shot was ok or not by cheking the histogram in red alert.
I think we did not have any shot under or overexposed by more than half stop.

SO my question is: has anyone checked the ISO by himself?

Hi Mac,

It's exactly the same with a Viper. When I shoot with a Viper I know that I can safely work with a spotmeter set to 320 asa, Highlights can be +5 stops over. On a low contrast studio shoot then the asa may well be nearer 80, pushing up the lighting budget somewhat.

However you don't have to fully expose all the time.

Stephen

[Miguel "Macgregor" De Olaso]

Thanks Stephen, you know a lot more than me, so that is why i ask you.
If i expose as 320 asa, i get an image aprox 2 stops underxposed. Sure this helps to prevent clipping the highlights BUT i also get a lot more noise when later i have to recover the image from the left part of the histogram (dark tones, shadows).
You mean this is they way of working?

I prefer to have as much information on the right part of the histogram as long as the highlights are not clipped.
For this red1 shots i did last week, the image however just stayed in the center of the histogram by overexposing 2 stops what the lightmeter read at 320.

Or maybe i have to throw that lightmeter away... XD

[Miguel "Macgregor" De Olaso]

Uhmm i just read this from another user, so i might be not the only one:

Meter scene using incident meter at 320 ISO as I would with Film
Meter says to shoot at T4 at 320 ISO. Shoot test with subject
camera set to 320 ISO lens set to T4. Image is almost black on camera LCD (my only live viewing option) I understand this is fixed now.
Open footage in Red Alert using settings for REC709. Looks way too dark. Change setting to LOG looks better but still too dark.
So I just rated the red ISO at 100 instead of 320 and the images looked great. I had no clipping in the whites of the skin tone at all and images are grain free.

[Stephen Williams]

Thanks Stephen, you know a lot more than me, so that is why i ask you.
If i expose as 320 asa, i get an image aprox 2 stops underxposed. Sure this helps to prevent clipping the highlights BUT i also get a lot more noise when later i have to recover the image from the left part of the histogram (dark tones, shadows).
You mean this is they way of working?

I prefer to have as much information on the right part of the histogram as long as the highlights are not clipped.
For this red1 shots i did last week, the image however just stayed in the center of the histogram by overexposing 2 stops what the lightmeter read at 320.

Or maybe i have to throw that lightmeter away... XD

Hi Mac,

Digital cameras don't really have an asa rating, just a total dynamic range. Thats why people need to test the cameras. themselves.

Obviously you cant light a huge night scene for 80 asa on an indie budget! So you have to make decisions how to make the exposure call. A spot meter will be the more useful IMO.

You also have to see how skin tones look exposed to the right, I have a theory that sometimes exposing them nearer the middle can have advantages.

Basically there is not right & wrong it's a gut feeling at the time.

Stephen

[Scott Webster]

Mac,

Loving your work. Latest firmware released this week has a number of tools to aid in exposure. Histograms, waveform and false colour. Will be very useful when the EVF comes online with assignable button control.

[Mick van Rossum, NSC]

Hi all, while shooting last week i found myself incapable of measuring the light with my cine lightmeter. THe readings i was getting did not match what i was seeing in the RED LCD. At least i could say that what i see in the RED LCD is what i get. That´s good.
But the light meter set at 320iso was producing images at least 2 stops underxposed in the red raw footage.
So finally i had to throw away my lightmeter and make exposure by looking at the LCD and checking for the highlights not to be clipped. After each shot, i quickly gave the card to Blair or Brooke who told me if the shot was ok or not by cheking the histogram in red alert.
I think we did not have any shot under or overexposed by more than half stop.

SO my question is: has anyone checked the ISO by himself?


Did some serious testing today with two Reds. I also thought that 320 was too optimistic, when exposed to 160 it looked much better. I took a Macbeth chart, changed the IRE setting in the zebra setting of the LCD to 90 % and opened the aperture till the Zebra hit the 90 % mark on the macbeth chart; I got an exposure of 160 ISO. So the real question is ; is the RED as sensitive as acclaimed ? How do the Red people get their 320 ISO out of the box ?

Cheers

Mick van Rossum NSC

[David Battistella]

Hey guys,

The way I have understood it to now is that the camera is always rated at 320 and any ISO changes are really only affecting the monitoring of the image. The sensor itself is recording at 320 ASA all the time. But in reading this thread it seems that the tendency might be to overexpose the RED for a better overall histogram and a more pleasing recorded image with less noise.

This has been a very useful thread.

Thanks. Nice Work.

David

[Greg M]

Hi all, while shooting last week i found myself incapable of measuring the light with my cine lightmeter. THe readings i was getting did not match what i was seeing in the RED LCD. At least i could say that what i see in the RED LCD is what i get. That´s good.
But the light meter set at 320iso was producing images at least 2 stops underxposed in the red raw footage.
So finally i had to throw away my lightmeter and make exposure by looking at the LCD and checking for the highlights not to be clipped. After each shot, i quickly gave the card to Blair or Brooke who told me if the shot was ok or not by cheking the histogram in red alert.
I think we did not have any shot under or overexposed by more than half stop.

SO my question is: has anyone checked the ISO by himself?



When you were metering was the camera shooting at 1/48th (180 degree shutter) sec? If you were shooting at 1/500 did you compensate? If so how much?

[Miguel "Macgregor" De Olaso]

Yes.
How do i compensate? Changing the shutter speed in the meter. ;D

[Greg M]

Yes.
How do i compensate? Changing the shutter speed in the meter. ;D

what meter are you using?

[Emery Wells]

Hey macgregor, in my chip chart tests we found exactly the same thing as you. With the light meter set to 320, we we're consistently about 2 stops under exposed.

Graeme shed some light on the subject in an email:


...That chart has probably a 5 stop exposure range across the greyscale? so, rating at ISO 320 is protecting hightlights that are not in that image, so you could probably rate ISO 160 and still not clip. ISO is a funny beast - what you rate it at depends on the dynamic range you're capturing and whereabout you want middle grey to land in relation.

...Basically rate 500 for high dynamic range, 320 for medium 160 for low. It's basically just deciding where mid grey goes and how much highlight room you have.

[Greg M]

are you metering with a cine meter setting your fps to 24 and compensating for shutter speed via shutter angle or are you using a still meter and setting a shutter speed?

[Miguel "Macgregor" De Olaso]

Hi digitalfx.
My lightmeter has cine style and still style.
Either way is the same for this.
When i was metering i was at 24fps 180º 320 iso.
Later when i switched to 1/500 sec i had the meter in the pocket since it was not working for the shooting.

THanks Emery, what Graeme says might make sense.

[Greg M]

Thanks...I dont have my Red yet (next week)...just wanted to be sure I was on the same page.

[Miguel "Macgregor" De Olaso]

Thanks...I dont have my Red yet (next week)...just wanted to be sure I was on the same page.

Lucky you! XD

[Fergus Meiklejohn]

I don't See Objectively..

So as interesting a question for all you very experienced people: does the camera strike you as faster or slower than filming with standard film stocks or common video cameras?

"If I was filming this scene on 35mm 500ASA would it be easier than RED?"
"If I had the Viper...?"
"If I had a little Sony HDV Z1..?"

[Sebastian Cramer]

I would like to come back to the issue of the ASA rating. Has anyone had additional information here?
I you do a propper setting for the red with histogramm etc. and then compare the reading on the iris with an exposure meter (or spotmeter against grey) is this in the area of 320 or less?
Thanks.

[Kevin Halverson]

SO my question is: has anyone checked the ISO by himself?

Hello macgregor,

I have just recently concluded a test to define a number of photometric properties of the camera. The problem with an ISO rating is that it is subjective in terms of what you define as the balance between highlight and shadow latitude. If you use the mean distribution, then the number will be considerably lower than has been published and some might jump to the conclusion that the camera is somehow inadequate (which is certainly not the case at all!).

As this test was conducted with the aid of a RED team member and not intended for public dissemination (at least at this point in time) I would prefer to discuss this off list if you have any interest in the measurements that I have made. Please feel free to PM me if you would like details on the testing regiment and the results.

Kevin

[Evin Grant]

From the time I've had shooting with the camera I'd say the ISO is in direct proportion to the amount of shadow noise you can tolerate. Most of us are not used to this amount of control of our images but rating the ISO of the camera is tantamount to choosing a particular film stock for the look of your piece. I chose to shoot the music video at ISO 640 becasue I liked the small amount of shadow noise, it looked more organic to me. If I was doing beauty I probably would have shot at ISO 320. You can expose flat scenes as low as ISO 160 and still hold your highlights but you really aren't gaining (No pun intended) much noise reduction by doing this, so unless you want to forgo some NDs I wouldn't rate it any slower than 320. A spot meter is going to be a very useful thing with Red, those of you so inclined might want to read up on the Zone system a bit. But it's not like any other camera that's gone before. It's like having Kodak's whole film library at the touch of a button.

[Deanan]

I've used the zone system analogy a few times to describe how one can think of exposure with raw. But people just look at me funny....

It basically comes down to what you want to protect, what you
want to push to the edge relative to your scene dynamic range.

[Chris Gearhart]

At first I was confused when I saw heavy grain popping up in relatively low ISO images. But I'm beginning to get my head around the nature of the beast.

This seems to boil down again to the flexibility and control inherent with RED as an acquisition system. I have used the zone system often enough, but I never thought of how that really impacts how you view the rating of the camera (image sensor) itself. Evin, I like your analogy to choosing film stock with the touch of a button--at least insofar as it describes grain ratios in a given exposure scale.

Kinda similar but different to the idea of "gain" up and down, no? (barring electronic voodoo from the A/D converter, on which I am somewhat fuzzy.) It just depends on where you locate "zero gain". Previous cameras hard-coded that in with some sort of switch that shifts the "ISO" up or down the overall range. I had always assumed there was some absolute, ontological ISO of digital cameras, and adding gain was compromising the "purity" of the image. Red combines a gaggle of stops with a "new" exposure paradigm that takes a bit to digest. Feel free to nuance me, here. Basically, we are in control. I'm finally getting what Graeme has been saying for over a year now. (Turns out Graeme is smart. Who knew?)

Now the question becomes, as has been said, how does Red compare with other cameras under similar exposures (let's say of relatively comparable digital cameras, like, say, Viper?) Is it's snr footprint-range less than or more than? By how many stops? Can't wait to test it myself.

(And btw, I must say I am very pleased with the grain patterns on what I have seen on low-exposure Red images.)

[Fergus Meiklejohn]

Now the question becomes, as has been said, how does Red compare with other cameras under similar exposures (let's say of relatively comparable digital cameras, like, say, Viper?) Is it's snr footprint-range less than or more than? By how many stops?

Yeah that is the question!

I'd also love to hear an opinion.

[Stephen Williams]

Yeah that is the question!

I'd also love to hear an opinion.

Hi,

Most digital sensors seem to be about the same rating @ 320 asa.

I have been in situations with a Viper, in a low contrast scene when I was at the equivalent of 80 asa, thats not fun in a small studio.

Stephen

[Esmaile Neissi]

gays all your opinions are great ,but there's a question comes up:as you all gays sad 320 asa seems to be the native ASA for digital cameras,if so,then changing the asa on set seems to be rung because in this way we just change the previewing of the image
in the monitor and no thing ales.
So what's the best solution to prevent this phenomena and get good images in your idea?
according to this discussion i conclude that we shoud
equip our selves with fast lenses and just work with
asa 320 $ dealing with f-stops,is it right or i misunderstand the case?

[Esmaile Neissi]

BTW,theres a good free tutorial for calibrating your digital camera with your light meter (sekonic),i think it's a worthy link to see.
Here is the link to download:
http://www.freshdv.com/2007/07/calibrating-a-light-meter-for-digital-video.html

[BASSAM MSSALATIE]

gays all your opinions are great


:as you all gays ?

Please Correct your spelling mistakes ..:ohmy:

[David Mullen ASC]

It may not be that the images were two stops "underexposed" at 320 ASA but simply that that's how a RAW file converted is supposed to look to protect highlight information.

I was color-correcting a 35mm project that was transferred to HDCAM-SR in 10-bit LOG mode, 4:4:4, and I was amazed at how murky and dark all the scans were before we color-corrected them, for a movie where the negative was not underexposed in general. They reminded me of the dark & murky dailies I got for "Big Love" where the 35mm negative was being transferred to HD to hold all the highlight information without any clipping.

So you may simply be seeing how 4K RAW 12-bit to 4K RGB 10-bit LOG looks in terms of luminence values until you start to color-correct it for whatever display format you are intending. It also suggests that if you are planning a film-out, you want to make sure that your final color-corrected master hasn't been permanently corrected to look good on a video monitor instead of for output to a film recorder.

[RFK]

David,

So we should use the reading out of the light meter at 320ASA and ignore the dark look on the monitor to achieve a "correct" exposure??? Let's say I get a T4 at the meter, I should set the iris at 4 despite the ISO rating at the camera and the look on the monitor.

[John Waterman]

My understanding is to determine the ASA of a video camera you place a grey card filling the field of view. Then you light that grey card and measure that light with an incident meter in ft candles. Then you adjust the iris on your camera's lens until the grey card reads 50 ire on a waveform monitor. (or for red, 50% value in the raw data) Once you have done that middle grey is now in the middle of your camera's response (latitude). Knowing the shutter speed, the lens T stop, and the ft candles you can mathematically determine the native ASA number for the camera with a formula (in the asc manual).

That will give you the native sensitivity of the camera, but not necessarily the best asa to RATE the camera at. that depends on the latitude of the subject. If you are shooting a low contrast subject, lets say the highlights are only two stops over middle grey, then you could choose to move middle grey further "to the right", toward the highlight side of your camera's latitude scale. By moving middle grey you are effectively subtracting highlight latitude and adding shadow latitude. But in this case, it is a good move, because you only need two stops over grey. By moving grey up the scale you are essentially boosting the whole signal of the image, which will increase the signal to noise ratio, and provide an image with less noise in the shadows. However, because you moved middle grey up the latitude scale, you need more light to create middle grey, thus a lower "slower" ASA number. You are essentially purposefully over exposing the image, because you know that you do not need all of the latitude above 50%. Then in post you are compensating and bringing the image back down.

Does that make sense?

[David Mullen ASC]

David,

So we should use the reading out of the light meter at 320ASA and ignore the dark look on the monitor to achieve a "correct" exposure??? Let's say I get a T4 at the meter, I should set the iris at 4 despite the ISO rating at the camera and the look on the monitor.

The on-set monitor is showing a conversion to 4:2:2 / 720P / REC709 through HD-SDI I believe and shouldn't look dark if you expose at the ASA rating that the camera is set for (whatever ASA setting you choose). I was referring to how it looked converted later from 12-bit 4K RAW to 10-bit LOG 4K RGB in RED CINE / RED ALERT.

[RFK]

The on-set monitor is showing a conversion to 4:2:2 / 720P / REC709 through HD-SDI I believe and shouldn't look dark if you expose at the ASA rating that the camera is set for (whatever ASA setting you choose). I was referring to how it looked converted later from 12-bit 4K RAW to 10-bit LOG 4K RGB in RED CINE / RED ALERT.


I understand the conversion from 4k RAW to 4k RGB, but I'm still trying to understand the ASAxISO Issue. On film I always have had just my lightmeter as a reference of the amount of light that would hit the negative. So now I'm thinking more and more that the 4K RAW is my new negative, it seems that the ISO actually is not the same as the one on film. I'll have no problem having a dark or murky picture on the monitor if I can trust that the RAW material will have the right exposure for a future film transfer. But I'll have to understand exactly what is being recorded at the 4K RAW.

[David Mullen ASC]

I understand the conversion from 4k RAW to 4k RGB, but I'm still trying to understand the ASAxISO Issue. On film I always have had just my lightmeter as a reference of the amount of light that would hit the negative. So now I'm thinking more and more that the 4K RAW is my new negative, it seems that the ISO actually is not the same as the one on film. I'll have no problem having a dark or murky picture on the monitor if I can trust that the RAW material will have the right exposure for a future film transfer. But I'll have to understand exactly what is being recorded at the 4K RAW.


Again, when you say "monitor" do you mean the on-set monitor showing you the live downconversion to 720P? Because that shouldn't look dark.

Basically you test different ASA settings and determine which level of noise vs. highlight clipping you prefer in different situations.

As for checking on highlight exposure, we had this discussion before -- the problem (I think) is that there is no way to monitor a RAW image live for exposure information, you have to use a converted video signal. So the waveforms, histograms, and false color info that the camera gives you are reading the 720P video signal being generated for the monitor on set. If you are shooting around the "true" sensitivity of the camera, more like 250 to 320 ASA, then the 720P signal will be similar enough to the RAW recording for those exposure indicators to be accurate. If you are selecting a high ASA rating, like 500 ASA let's say, then the RAW conversion to 720P is being boosted to look correct, exposure-wise, on the monitor but the exposure indicators are reading that boosted 720P signal, whereas the RAW signal is still be exposed and recorded at the native sensitivity -- different ASA ratings are simply metadata.

So for now the solution, some people are telling me, is to temporarily switch the camera to 250 or 320 ASA when you want to double-check the exposure using the histograms, waveforms, false color to watch for true clipping. Since you might be shooting the scene at 500 ASA, of course, monitoring the exposure at a lower ASA setting temporarily will make the monitor image look dark but you'll notice you have less highlight clipping because of the underexposure.

There is no "real" ASA setting per se, only the setting you choose that has the balance of highlight and shadow information that you want in conjunction with the amount of noise generated. For most digital cine cameras, it seems that 320 ASA is a good base to work around.

My point before was just that at 320 ASA, a conversion to 10-bit LOG may possibly look dark at first in post just because of the way the luminence values are being placed in the 10-bit LOG signal. But what you see on the set is not a conversion from RAW to 10-bit LOG, it is gamma encoded (sometimes mistakenly called "linear") video designed for viewing on a monitor.

[RFK]

Again, when you say "monitor" do you mean the on-set monitor showing you the live downconversion to 720P? Because that shouldn't look dark.

Basically you test different ASA settings and determine which level of noise vs. highlight clipping you prefer in different situations.

As for checking on highlight exposure, we had this discussion before -- the problem (I think) is that there is no way to monitor a RAW image live for exposure information, you have to use a converted video signal. So the waveforms, histograms, and false color info that the camera gives you are reading the 720P video signal being generated for the monitor on set. If you are shooting around the "true" sensitivity of the camera, more like 250 to 320 ASA, then the 720P signal will be similar enough to the RAW recording for those exposure indicators to be accurate. If you are selecting a high ASA rating, like 500 ASA let's say, then the RAW conversion to 720P is being boosted to look correct, exposure-wise, on the monitor but the exposure indicators are reading that boosted 720P signal, whereas the RAW signal is still be exposed and recorded at the native sensitivity -- different ASA ratings are simply metadata.

So for now the solution, some people are telling me, is to temporarily switch the camera to 250 or 320 ASA when you want to double-check the exposure using the histograms, waveforms, false color to watch for true clipping. Since you might be shooting the scene at 500 ASA, of course, monitoring the exposure at a lower ASA setting temporarily will make the monitor image look dark but you'll notice you have less highlight clipping because of the underexposure.

There is no "real" ASA setting per se, only the setting you choose that has the balance of highlight and shadow information that you want in conjunction with the amount of noise generated. For most digital cine cameras, it seems that 320 ASA is a good base to work around.

My point before was just that at 320 ASA, a conversion to 10-bit LOG may possibly look dark at first in post just because of the way the luminence values are being placed in the 10-bit LOG signal. But what you see on the set is not a conversion from RAW to 10-bit LOG, it is gamma encoded (sometimes mistakenly called "linear") video designed for viewing on a monitor.

Yes, I'm talking about on set monitor. So far this is just a theoretical conversation since I don't have the camera yet, but I'll do a film transfer on the my first RED job and I want to really understand the 4K RAW material as my "negative". I don't mind a less than perfect picture on the on set monitor and also in post just before the color grading session if I know that the RAW is properly exposed, the same as when you got best light dailies of low key scenes on film and you would know that there was a lot a latitude for the final print.