Sorry, another Dragon optimal ISO thread...

[David Battistella]

Thanks David, that's very helpful. Makes a lot of sense.

But with your ISO 1000 example, for a properly exposed image less light is hitting the sensor than at a properly exposed ISO 320 right? You'd have to open up the aperture more to bring the exposure up for the ISO 320.
So I would think that in your example above shooting at ISO 1000 would preserve the sky better but give you less shadow control than at ISO 320. As you said ISO 320 would lose the highlights faster.
But you mentioned that you rated at ISO 1000 to see into the shadows better? Am I missing something?

Yes, less light is hitting the sensor so here I rated at 1000 and stopped down. This gave me a flatter histogram. If I opened up and rated at 320 I probably would have a harder time pulling out blue skies and I'd be trying the flatten the his tram in post, which is not the same thing. It's an easy test to do in a high contrast scene. The key here was the midpoint for that scene and holding enough above and below avoid more of a shillohette image which a 320 rating might give you.

 

[Ulf Krentz]

I still think where and how you map middle grey at the time of exposure has a pretty major influence. It depends if you are looking for flatter histograms or ones that are bunched to the right or left, but Bjorn makes a good point about finding the midpoint for that. The more you experiment the more you find the sweet spots.

Yes sure, I think I have a clear view of what is happing here, just don´t understand Michaels argumentation at all.
As long as the dynamic range of the scene is smaller than the sensors DR there is room for exposure variation. My guts feeling especially with the STH OLPF is the right side of the histogram is kind of compressed by the tone/data mapping to protect the highlights. That´s where you loose highlight details when you strictly expose ETTR. I definitely need to get and test the Standard OLPF...

 

[Hrvoje Simic]

ETTR is a trend born in times of lower sampling, smaller DR and hard clip. Now it just saves the data no one in the audience cares about and can lead to underexposed humans in unlit interiors with nice fluffy clouds in the sky.

ETSP
Expose-To-Shot-Priority

 

[Jake Wilganowski]

ETTR is a trend born in times of lower sampling, smaller DR and hard clip. Now it just saves the data no one in the audience cares about and can lead to underexposed humans in unlit interiors with nice fluffy clouds in the sky.

ETSP
Expose-To-Shot-Priority

Mmm, in my opinion ETTR is strictly to combat noise and nothing else. With weapon noise isn't an issue with the standard. In your example it sounds like you are protecting highlights at the expense of the human subject, which would be the opposite of ETTR

Unless I'm misunderstanding your reference

 

[Mike Nagel]

At capture, ISO is nothing but a brightness setting for the LCD.

ETTR only in order to get a fat negative. Observe the histogram and goal posts.

lot of peeps focus on highlight retention, it is even more important to focus on mid tones this is where the most important meat is.

there is so much data in the raw that u can retrieve it is a beauty (if u know what ure doing in post), just don't clip and don't crush (if possible).

 

[Brett Harrison]

In the world of linear light, ISO is arbitrary, and not as useful as it was in the world of film. The Wikipedia film speed entry is a good read, here's the link to the digital camera section

https://en.wikipedia.org/wiki/Film_speed#Digital

The Cineon system was built so that you could 'unbuild' from the characteristic curve of a film stock, i.e. work backwards to linear light based on densitometer measurements of the given stock.

https://en.wikipedia.org/wiki/Cineon

That's something that the old system of metering with an ISO/characteristic curve in mind couldn't do, which is why with film you have to know ahead of time and expose accordingly, but with digital you can simply maximise the signal to noise ratio and decide after the fact, how you will characterise your information.

 

[Mark Toia]

Everyone has their different ways of doing things... but sharing is caring.

I'm a simple shooter and I keep my exposure methods the same. I call it the K.I.S.S method. KEEP IT SIMPLE STUPID

RED have graced us all with a very good histogram level at the bottom left of screen. People that are getting blown out highlights, over exposure or even under exposure and noise in anyway are obviously not looking at this histogram. Simple as that.

The little traffic lights you see next to the histogram are another tool to keep and eye on, If you see RED, your dead...(well you picture is now over exposed in other words)

Oh.. and remember, don't think every shot of yours has to be have perfect highlights, I quite often let the sun burn out or I also let specular highlights from light sources or sun bounce blow out.. why.. Well if It's blowing out with my own eyes, why can't it blow out in camera. Let the sun be the sun and be white hot ! You don't need to stop down to protect bright highlights from peaking. Just protect the actual parts of the image you want to see information in.. thats it.
If you over compensate that protection of highlights, you will end up with an even worse problem, And trying to lift your blacks back up which intern gives you ugly noise. YUK !

So... the balance of light is the key.

I personally never shoot past 320. Mostly I'm at 250 outside. I pretend I'm shooting film, A roll of Kodak 250D kodak. I made sure I always found the sweet spot on film and I stayed there. And Yes... I will lift my exposure in 1/2 stop ISO increments as the light dropped, But I never pushed anymore than 1 to 2 stops past that sweet spot. 250 + 1stop is 500, 500+1 stop is 1000... simple math for a simple problem.
Before I even looked like pushing my ISO, I looked at my lens F stop first. Open up...

Amazingly gifted people invented lenses that go down to at least F1.2.. Some even brighter. Use them ! Don't leave your 2.8 ZOOM on during a low light shot, because then you'll push past 800iso and start to get noise, your starting to compromise your image the longer you leave that on, next thing your at 1600 and its all over... Blaming the camera and not yourself for being clever enough to put brighter glass on. If you don't have brighter glass, add lights... Like we all had to do in the FILM days.

Because the next thing your doing is sitting in the grade or edit with loaded up blacks and shit loads of noise, kicking the cat because you know you could have avoided the problem by swapping that lens out. Or Adding lights to your scene.

But number 1 thing is, Reading that Histogram correctly.

Did you know that that histogram is showing you a good solid 3-4 stops... if you park your image curve (lump) in the middle of the histogram your are effectively 1.5 stops under exposed. So if you have your camera set at 800iso with the histogram bump placed in the middle, your technically now exposing light at 1600-2000 iso. Hence why people are always complaining about noise. This is a common problem I see when people send me their R3D files to check.

So to avoid this, if you are only wanting to shoot at 800ISO, push the curve to the right of the histogram until you start to see the traffic lights peak. When they do.. pull back just a touch. But know that shooting at 800iso.. you can push past the traffic lights a tiny bit, that's the safety in shooting at 800ISO.

But remember, if the sun is peaking your histogram, let it.. its the sun ! Just make sure the rest of the image is correctly exposed. You still wont get images as clean as people shooting at 250-320. But at least your not swaying over into NOISE LAND.

If you perform and maintain this same thinking while shooting at 250-320 ISO (watch those traffic lights) you will have super clean images. And better yet, if you are underexposed, you just pull the blacks up in post with little to no noise at all. Because your at 250iso, meaning you have at least 3 stops to drag back information in the blacks.

Simple as that...

Happy shooting.

 

[Nick Morrison]

Did you know that that histogram is showing you a good solid 3-4 stops... if you park your image curve (lump) in the middle of the histogram your are effectively 1.5 stops under exposed. So if you have your camera set at 800iso with the histogram bump placed in the middle, your technically now exposing light at 1600-2000 iso. Hence why people are always complaining about noise. This is a common problem I see when people send me their R3D files to check.

So to avoid this, if you are only wanting to shoot at 800ISO, push the curve to the right of the histogram until you start to see the traffic lights peak. When they do.. pull back just a touch. But know that shooting at 800iso.. you can push past the traffic lights a tiny bit, that's the safety in shooting at 800ISO.

This right here is amazing advice.

Many would be wise to heed it.

 

[David Battistella]

That's something that the old system of metering with an ISO/characteristic curve in mind couldn't do, which is why with film you have to know ahead of time and expose accordingly, but with digital you can simply maximise the signal to noise ratio and decide after the fact, how you will characterise your information.

True, but things like shallow or deep DOF are factors in creating an image as well, not just ISO, there are times you want to stay at a fixed stop. So using a meter and ISO as part of the overall exposure is still valid, the way it is on any DSLR. It's a little trickier with motion images and a 180 degree shutter, but there are times you want deep focus to take advantage of the resolution available with the sensor.

You summary does describe why in digital there is a bit more of a working range, but that was true with film and there are plusses and minuses to pushing and pulling in film and in digital.

It all comes back to the images you want to create and what is filling your frame.

David

 

[Ulf Krentz]

One more thing to be aware of is that the histogram and goalposts show the interpreted values of ISO and curve settings. They are NOT showing your actual exposure unless you set it to RAW which is easy enough to check in between. The RAW histogram shows you the actual DR of your scene and where it is placed within the DR of the sensor. Actual exposure is a variable of the amount of light in your scene, your aperture setting and shutter speed, nothing else matters when shooting a RED. There are a lot of new shooters joining our group so I think it is really important to understand the basic of how RED works. Like Mark pointed out, just think about having one RED film stock.

Cheers
Ulf

 

[David Battistella]

There are a lot of new shooters joining our group so I think it is really important to understand the basic of how RED works. Like Mark pointed out, just think about having one RED film stock.

Cheers
Ulf

Ulf,

Could I just make the suggestion that most of the "new shooters" never shot a frame of film in their lifetime? They might not have film as a reference. It's not "bad" to describe the camera as a "variable range film stock" because that make it a more flexible tool.

A couple of months ago I use a 1000ISO base rating on a shoot because I wanted deeper DOF, that was not a disaster, that was using a modern tool in a way it was designed.

There are fewer hard and fast rules that with film, especially with Dragon and everyone will settle in to what they need, want and like on the day.

David

 

[Brett Harrison]

Definitely, histograms are often misunderstood. A histogram is not exclusive to images, it's actually a fundamental graph in mathematics - specifically probability. That's why you often see the bell curve when you look at a scene, because statistically you're more likely - by definition - to see 'shades of grey' than extremes of black and white.

https://en.wikipedia.org/wiki/Histogram

So each bar represents the probability - if you closed your eyes and pointed at the screen, you can predict what shade your finger is likely to land on.

Not all scenes adhere to the bell curve - sometimes there's no bell curve, particularly when you have a scene that is lit with mixed levels of intensity over a space.

I totally agree with what Mark's saying about putting the bell curve to the right, which is in keeping with the theme of ETTR - the bell or lump of the scene represents the shades of grey in your scene, including flesh tones, and the closer you can place that group of reflectance near the right, the richer that information will be.

That's presuming that the middle grey info in your scene is what you want to read as middle grey, that it's not a high key scene or mixed illumination etc and so forth. That's why as others have said, you need to analyse the scene for yourself to decide what percentage is what, so you can read the histogram like Cypher 'sees' The Matrix

 

[David Battistella]

mostly it depends on how you want things to look.

 

[Ulf Krentz]

Ulf,

Could I just make the suggestion that most of the "new shooters" never shot a frame of film in their lifetime? They might not have film as a reference. It's not "bad" to describe the camera as a "variable range film stock" because that make it a more flexible tool.

A couple of months ago I use a 1000ISO base rating on a shoot because I wanted deeper DOF, that was not a disaster, that was using a modern tool in a way it was designed.

There are fewer hard and fast rules that with film, especially with Dragon and everyone will settle in to what they need, want and like on the day.

David

David, but to be clear again: You´re in fact not shooting an higher ISO, just lighten up a too dark original. This is what folks lack to understand. In the analog world this would be to push a too dark slide in lab and/or scanning it brighter. I agree some of the new shooters won´t have a working background in classic techniques but I see a lot of young people very interested in it, too. I don´t think the RED system is hard to understand at all if you are willing to spend a bit of time at the very beginning. If you just start working without any research you probably will end up in having heavily underexposed data sooner or later without that being REDs fault.

 

[Michael Tiemann]

What you are missing is that, as with the Zone System, there are two systems in play: the "negative" and the "print". Exposing at ISO 320 (which means metering at ISO 320) gives you a fat negative that comes with a lot of forgiveness. Finishing at ISO 1000 (which means interpreting the RAW data at ISO 1000) opens up the shadows that would appear to have less detail if finished at ISO 320.

Not sure if I understand anything of what you try to explain. I don´t even see something like a negative with digital. The sensor pretty much behaves like a reversal film. Ansel Adams zone system was basically build on exposure and adapted development to create the perfect negative for print. This is what we do today when we tweak a gamma curve.

Of course the RED sensor acts as a reversal medium, not a negative, and of course its image, being presented on a screen or via a projector is more like a slide than a print. Which is why I put quotes around those terms. The point I was making is to highlight that exposure and finishing decisions have their respective implications, and the fuller description of the Zone System comprehends those differences, even though it introduces many other details as well. I was hoping that the larger insight of the difference between exposure decisions and finishing decisions would transcend anyone's tendency to be quite so literal in their interpretation of the context of the insight.

 

[Patrick Grossien]

I haven't read the whole thread, but I think it's wise to point out to everybody that there's some really good information "hidden" on red.com
Sean M made these great 101 lectures/articles which are worth a read and re-read... At least start here, when starting new with a RED camera. And concerning ISO and RED, read the article from the last link

red-101/exposure-with-red-cameras

red-101/red-camera-exposure-tools

red-101/exposure-false-color-zebra-tools

red-101/redlogfilm-redgamma

red-101/iso-speed-revisited

 

[Patrick Grossien]

...

The little traffic lights you see next to the histogram are another tool to keep and eye on, If you see RED, your dead...(well you picture is now over exposed in other words) ...

Great write up, Toia
Just to clear things up a bit, because one could misunderstand your info about the traffic lights.
The term traffic light is just being used because it looks like one, not because it works like one.

So, the exposure level is not "green" = safe / "red"=dead but it states which channel is clipping!
So it's : "Red"= red channel clips highlights, "green"= green channel clips highlights...

 

[David Battistella]

Look. I think I stated about 50 times RATING, not shooting, RATING, that means this. In some scenes I can RATE THE CAMERA at 250 ISO ( and I do, as Mark suggests in his posts) at other times I have RATED the camera at 1000ISO to create deeper focus.

What does that mean.

Sometimes, I set the camera to 250 ISO and I meter and expose much the way Mark describes, pushing the image to clipping Stoplights and then taking it back so that there are no "stoplights" activated, which is showing you which channel is clipping as Patrick points out.

Other times I have set my camera to 1000ISO and lit the scene around 1000ISO meter readings and stopped the lens down to get to an acceptable level of noise and the DOF I wanted. IN that case, in your theory I was underexposing the sensor, or not moving the RAW histogram to the RIGHT, but in that case I was doing what was acceptable and called for for the situation. Like instead of shooting at 1.4, maybe shooting at F4. All this after testing and running it through a post pipeline.

So.

Getting back to the original question of "what is the best ISO". It's the one you decide to rate your scene at. 250ISO is a pretty clean and polished look for commercials, product, and lots of forms of production, it's not suited to every budget or style of filmmaking. Just as 1000 ISO might not be suited to some styles.

The overal point is (and phil states this often and I happen to agree) is that there is a WORKING RANGE of ISO you can use to create you images.

David

David, but to be clear again: You´re in fact not shooting an higher ISO, just lighten up a too dark original. This is what folks lack to understand. In the analog world this would be to push a too dark slide in lab and/or scanning it brighter. I agree some of the new shooters won´t have a working background in classic techniques but I see a lot of young people very interested in it, too. I don´t think the RED system is hard to understand at all if you are willing to spend a bit of time at the very beginning. If you just start working without any research you probably will end up in having heavily underexposed data sooner or later without that being REDs fault.

 

[Mike Nagel]

U can create whatever images u want from the raw in post if u capture correctly in the first place (capture as much clean data as possible) ... But u can't do that the other way around (when u compromise the capture)...

ISO means nothing in RED capture... It's pure meta data

shoot by observing the raw histogram - u see what the sensor is getting. RED Sensors are light hungry across the stimuli range. Shoot fat.

 

[David Battistella]

You know, I know this stuff ten times over. But if you use a meter and the meter is set to 800 and the camera is set to 800 then the ISO does play a factor, even though it can be changed after the fact and is a metadata component, like White Balance.

When you say shoot by observing the RAW histogram then what you are saying is this

I RATE THE CAMERA AT 800 ISO with a Dragoncolor/REDLOG FILM curve applied at 5600K because that is what RAW view is.

and I am sorry, things like DOF can't really be done in post.

David

U can create whatever images u want from the raw in post if u capture correctly in the first place (capture as much clean data as possible) ... But u can't do that the other way around (when u compromise the capture)...

ISO means nothing in RED capture... It's pure meta data

shoot by observing the raw histogram - u see what the sensor is getting. RED Sensors are light hungry across the stimuli range. Shoot fat.