Exposing + Native ISO

[Gunleik Groven]

Not to practical three months later in a DI.

battistella

Could be...

Bt it is a good point to marks point that we want the "metering" from the RAW view in the exposure tools, not the image...

 

[Mark L. Pederson]

You know what is weird.

Since the very first day of owning a red (early adopter 758) and right up until today, the only place to see RAW view is in the camera.

You can not see raw view (or the specific RAW view lut) on anything other than the camera or camera monitoring devices. That means a DP uses that image as a reference image and will never have the chance to his reference image it again.

so maybe that needs to change, maybe it doesn't .

discuss...


battistella

Just eliminate RAW view - once the EXPOSURE tool are corrected to work the way they should - (RED is looking into this and testing now) - there is ZIP, ZERO, ZILCH reason to have RAW view other than to confuse people.

Please prove me wrong if you have any idea WHY this should stay -

 

[David Battistella]

Sorry for the long post but please take the time to read it.

There should be a view (LOG, RAW, PRO VIEW, B&W whatever) separate from the look placed on a monitoring path that allows a person who understands the tools, the DP, gaffers, operators, etc. which allows you to see "into the negative" a bit better on set.

The problems with applying LUTS and looks on set , etc, like Gunleik explained in his example of a large budget set that lit to the monitor is that this attitude moves us away from measuring light and understanding how the sensors photo buckets are absorbing the light hitting it (like light on silver at the film plane) (as an aside, getting rid of ISO is like saying let's get rid of the T-stop, f stop is fine).

otherwise working with. RED is like working with a 5D and that picture on the screen and the monitor is all you are going to get.

I'm anticipating your response will be that if the underlying tools, Exposure, ire, zebra, are all working on RAW, then the most critical factor is that a base sensitivity has to be established for the sensor.

If the dragon sensor indeed claims to replace all film stocks from 250-2000ISO , then they give users maximum flexibility to rate the sensor at whatever ISO they like then I would say that opens it up to a lot of subjective interpretation (and also creative choice). But I do think it is the manufacturers responsibility (like kodak or Fuji would do) to say, at this ISO, with daylight (or tungsten) sources, metered and lit according to industry standards, will yield the sharpest images with the most accurate color.

Like in film the creative choices can come in, push pull, 1,2stops, use filtration, bleach bypass (digital version) becomes a creative choice that deviates from the baseline set by the sticker on the film can.

i feel it is important that RED stay on their course to be the digital replacement for film, therefore a continuation of that language is necessary to continue that tradition (like a nikon D700 does not change the language from an F3).

In using film, the DP is the person responsible for how the light hits the film plane, the best ones actually work in blindness with their knowledge of how the film and processing would react. They had a shit BW video tap but made great images. Why should that knowledge or method deviate too much because of an HD monitor onset! which represents about 1/6 of the data being captured.

RED is in more peoples hands, but few people have really taken the time to know and understand it enough to be able to shoot it blind folded (in the sense of knowing how strictly a metered scene will interact with the sensor). Is that RED's fault, no. That is the fault of an industry which has become lazy toward craft (at the middle and lower end).

So give pro DP's the a professional tool and let them have a sensor view, they need it to help people understand that there is more going on under the hood than what the client sees on the monitor.

Battistella

 

[Mark L. Pederson]

There should be a view (LOG, RAW, PRO VIEW, B&W whatever) separate from the look placed on a monitoring path that allows a person who understands the tools, the DP, gaffers, operators, etc. which allows you to see "into the negative" a bit better on set.
Battistella

Well ... please define what looking "into a negative" means to you?

I'm not saying I disagree - but - since you MUST perform RGB processing to view it - you have to make a choice.

So ... what exactly do you want to see in this "LOG, RAW, PRO VIEW, B&W whatever" - and tell me EXACTLY how you want use it and why.

 

[David Battistella]

Could be...

Bt it is a good point to marks point that we want the "metering" from the RAW view in the exposure tools, not the image...

i really have not used any of the exposure tools outside of RAW view mode. Why would I care what is going on in the meta data when all I care about is going on at the sensor level? Give me a crappy image with all the tools available and set up and send whatever you want downstream to the pipe that feeds the non non critical monitors (commercial client). I don't care about the meta data views. I leave my camera set to 320, 5000k and work from there, but I've taken the time to understand how the machine works too.

If it needs to be dumbed down for some DPs or clients, fine, do it on another feed.

Battistella

 

[D Fuller]

Yes it is my understanding that the analogy applies. Each line of chip (M, MX, Dragon) has it's own response curve that defines it just as each film stock has it's own defining response curve.

You do not actually change the sensitivity of a film stock either. When I set the ISO on my old film Pentax, what I am doing is changing the way the internal light meter responds. My understanding is that this is the same with Red. The difference is that the ISO setting can also directly effect the way the captured values are displayed, which does not happen with film.

Am I still on track?

Yes you are on track. But I'd modify a couple of these observations. The ISO setting on your Pentax changes the way your light meter responds, that's true. But it's important because it changes that way you will expose the film. And it very much changes that way the captured values are displayed (though not until after developing and printing).

To quote myself from another thread: After some detours, I've come to think about exposure on Red in much the same way that I used to think about it on film. A film has a fixed sensitivity (ISO), but knowing that, I can choose to expose at a different ISO and push or pull in post (either chemically or in transfer or printing). I know the image will have different characteristics if I do that, and that's part of my exposure calculation. I can over expose and under-develop, and get more highlight compression and more shadow detail. I can under expose and overdevelop and get more contrast, more grain, but with the risk of blocked highlights. All from the same film stock.

I expose Red with very similar thinking, and that's where ISO has become important again -- not the camera's ISO setting, but ISO as a reference for light levels in front of the lens. I know that the sensor is about 320 ISO native, but I don't love the completely-clean image I get at 320 for most things. At 800, I get a bit of very pleasant noise, which I like, and the added benefit of some headroom that lets me roll the highlights off nicely. But for things like tabletop and package shots, I usually want a "fatter negative" with more color and a bit less noise, so I expose for 640 or 500, knowing I have to watch the highlights more carefully, but knowing that it will do things to the image that I want. Or as with Gunleik's image above, I can underexpose and get an image where darks cluster together and mix with noise a bit, and the hilights occupy most of the long slope of the tonal curve. All with a sensor that has a fixed response.

What I can't do is expect that a scene I "shoot at ISO 400" will look like a scene that I "shoot at ISO 3200". That won't happen. (And it won't happen with any other camera on the market either, by the way.) Light matters.

 

[Mark L. Pederson]

In using film, the DP is the person responsible for how the light hits the film plane, the best ones actually work in blindness with their knowledge of how the film and processing would react. They had a shit BW video tap but made great images. Why should that knowledge or method deviate too much because of an HD monitor onset! which represents about 1/6 of the data being captured.

Battistella

That knowledge and method does not deviate at all.

If a DP can expose 35mm film without a monitor and nail the exposures - than that DP already has the exact understanding of how to work with a RED camera - sensor = film stock - and like every DP does when a new film stock is released that they are interested to use - they shoot a few simple tests using their specific exposure method and subjectivity - and they understand exactly what to do.

And those folks do not need a histogram, monitor, false color/exposure meters or stop lights to nail their exposure perfectly every time. I've seen it done. Lots of the big dogs hate monitors on set.

Having more tools for exposure is GREAT - as long as everyone understands them - and they work as advertised. We are getting there

 

[David Battistella]

So ... what exactly do you want to see in this "LOG, RAW, PRO VIEW, B&W whatever" - and tell me EXACTLY how you want use it and why.

I want a flat logish style image shows me something more complete than linear light. RAW as it is or RC4/REDlog film would be perfect.

I want to see this so I have a sense of the total frame and I can later decide what I have in the blacks and the highlights so that I can adjust later in the grade, know what is blown out and what can not be recovered.

I want the brightness of that image to be as close to the best possible image the camera can produce (sensor native 5000K, baseline ISO accurate to metered light, noiseless image) that yields the sharpest image with the best color fidelity and the widest latitude, without coloring it up for me. (If I want to deviate from that I will do my own testing and cook my own look, but as a start I want to know what the manufacture thinks will make the camera/film stock shine in terms of optimum technical precision, by adding a simple s-curve to make one lights easy.

I want two monitoring paths, the professional one as above with menus, guides etc and a clean cooked look for clients.

I want histogram, goal posts and Exposure tools. No ire, no zebra or any of that crap leftover from shitty video cameras.

battistella

 

[David Battistella]

That knowledge and method does not deviate at all.

this goes to my point of RED taking a leadership position as the replacement of digital film. I feel they should be on this like a pit bull on a bone. Needs to be established right now, with this sensor and this system. Put the eggs in that basket.

You hear me on that one RED.



battistella

 

[Elsie N]

Yes you are on track. But I'd modify a couple of these observations. The ISO setting on your Pentax changes the way your light meter responds, that's true. But it's important because it changes that way you will expose the film. And it very much changes that way the captured values are displayed (though not until after developing and printing).

To quote myself from another thread: After some detours, I've come to think about exposure on Red in much the same way that I used to think about it on film. A film has a fixed sensitivity (ISO), but knowing that, I can choose to expose at a different ISO and push or pull in post (either chemically or in transfer or printing). I know the image will have different characteristics if I do that, and that's part of my exposure calculation. I can over expose and under-develop, and get more highlight compression and more shadow detail. I can under expose and overdevelop and get more contrast, more grain, but with the risk of blocked highlights. All from the same film stock.

I expose Red with very similar thinking, and that's where ISO has become important again -- not the camera's ISO setting, but ISO as a reference for light levels in front of the lens. I know that the sensor is about 320 ISO native, but I don't love the completely-clean image I get at 320 for most things. At 800, I get a bit of very pleasant noise, which I like, and the added benefit of some headroom that lets me roll the highlights off nicely. But for things like tabletop and package shots, I usually want a "fatter negative" with more color and a bit less noise, so I expose for 640 or 500, knowing I have to watch the highlights more carefully, but knowing that it will do things to the image that I want. Or as with Gunleik's image above, I can underexpose and get an image where darks cluster together and mix with noise a bit, and the hilights occupy most of the long slope of the tonal curve. All with a sensor that has a fixed response.

What I can't do is expect that a scene I "shoot at ISO 400" will look like a scene that I "shoot at ISO 3200". That won't happen. (And it won't happen with any other camera on the market either, by the way.) Light matters.

I'm going to copy and paste this into my reference notes because it spells out a plan of action that I might use... assuming these same numbers correlate to Dragon

 

[Mark L. Pederson]

I want a flat logish style image shows me something more complete than linear light. RAW as it is or RC4/REDlog film would be perfect.

What is wrong with REDlogfilm right now? What do you want to see that you can't see in REDlogfilm ?

I want to see this so I have a sense of the total frame and I can later decide what I have in the blacks and the highlights so that I can adjust later in the grade, know what is blown out and what can not be recovered.

Only way to do this is false color meters - just like being discussed in Truth ... or False Color thread. This can not be indicted accurately enough by a single RGB representation alone.

I want the brightness of that image to be as close to the best possible image the camera can produce (sensor native 5000K, baseline ISO accurate to metered light, noiseless image) that yields the sharpest image with the best color fidelity and the widest latitude, without coloring it up for me. (If I want to deviate from that I will do my own testing and cook my own look, but as a start I want to know what the manufacture thinks will make the camera/film stock shine in terms of optimum technical precision, by adding a simple s-curve to make one lights easy.

How much for the goose Wonka

This is a pretty flawed request IMO because " best possible image" and " best color fidelity", etc. are SUBJECTIVE. You would need to articulate technically what you want to see - in technical terms - not subjective.

I want two monitoring paths, the professional one as above with menus, guides etc and a clean cooked look for clients.

You can do this now - just uncheck "tools" in the monitor setup. You gotta go one HDMI and one HDSDI - but you can convert HDMI to HDSDI with those bad ass small ATMOS converters.

I want histogram, goal posts and Exposure tools. No ire, no zebra or any of that crap leftover from shitty video cameras.

battistella

You have a histogram now. You don't need to use any of the IRE tools - but they can stay there for folks that like to use them.

 

[Mark L. Pederson]

this goes to my point of RED taking a leadership position as the replacement of digital film. I feel they should be on this like a pit bull on a bone. Needs to be established right now, with this sensor and this system. Put the eggs in that basket.

You hear me on that one RED.



battistella

Well ... in all honesty David - most experienced cinematographers understand ISO on the RED, Alexa, F65, etc. and have no real issue with exposing on the sensor(s) and system.

I think it IS established with the sensor and system.

I just think there is a lot of confusion over ISO with folks who don't understand the basics - which is okay - it can and will be confusing if you never really understood the basics - but this forum exists, right? Reducation exists, right?

 

[Scott Crawley]

Yes David, thank you. Those things all being about equal are understood.

For me the crux of it (Which Mark kindly clarified for me.) was this idea of adding gain to the signal coming off of the sensor... but we are not adding gain to the signal coming off of the sensor... at least not directly to the recorded image. We are recording the straight signal produced by the sensor, but adding gain in the monitoring path both during production and post. It may seem a fine distinction, but it is a key distinction. Granted it isn't something I think about every day, since I routinely expose to RAW view with an eye on RG3/RC3 only as a window into post processing. That way I don't get caught up in exposure errors due to poor ISO selection. Generally speaking I plan for a basic exposure that is correct and use that as a basis for any look that I am trying to create. The more creative or extreme the look the further I deviate from that plan but that is the basis of my process.

Incidentally, I frequently grade my own work and my procedure for that is generally the same.

Yes you are on track. But I'd modify a couple of these observations. The ISO setting on your Pentax changes the way your light meter responds, that's true. But it's important because it changes that way you will expose the film. And it very much changes that way the captured values are displayed (though not until after developing and printing).

To quote myself from another thread: After some detours, I've come to think about exposure on Red in much the same way that I used to think about it on film. A film has a fixed sensitivity (ISO), but knowing that, I can choose to expose at a different ISO and push or pull in post (either chemically or in transfer or printing). I know the image will have different characteristics if I do that, and that's part of my exposure calculation. I can over expose and under-develop, and get more highlight compression and more shadow detail. I can under expose and overdevelop and get more contrast, more grain, but with the risk of blocked highlights. All from the same film stock.

I expose Red with very similar thinking, and that's where ISO has become important again -- not the camera's ISO setting, but ISO as a reference for light levels in front of the lens. I know that the sensor is about 320 ISO native, but I don't love the completely-clean image I get at 320 for most things. At 800, I get a bit of very pleasant noise, which I like, and the added benefit of some headroom that lets me roll the highlights off nicely. But for things like tabletop and package shots, I usually want a "fatter negative" with more color and a bit less noise, so I expose for 640 or 500, knowing I have to watch the highlights more carefully, but knowing that it will do things to the image that I want. Or as with Gunleik's image above, I can underexpose and get an image where darks cluster together and mix with noise a bit, and the hilights occupy most of the long slope of the tonal curve. All with a sensor that has a fixed response.

What I can't do is expect that a scene I "shoot at ISO 400" will look like a scene that I "shoot at ISO 3200". That won't happen. (And it won't happen with any other camera on the market either, by the way.) Light matters.

 

[D Fuller]

I'm going to copy and paste this into my reference notes because it spells out a plan of action that I might use... assuming these same numbers correlate to Dragon

The numbers might not correlate exactly, but the methodology will. For example, if the Dragon native ISO is 800, it would just shift the whole thing up the scale a bit. Very much like switching to a higher-speed film.

 

[David Battistella]

How much for the goose Wonka

This is a pretty flawed request IMO because " best possible image" and " best color fidelity", etc. are SUBJECTIVE. You would need to articulate technically what you want to see - in technical terms - not subjective.
.

Mark

the technical responsibility is not mine, it's REDS responsibility.

Look ok at it like this. Red makes the camera and they make the film stock.

Actually, we hear a lot about the body form factor, etc. the way arri or Aaton would but we also hear about image, the way kodak or Fuji would. So let's say the red sensor is a film stock.

As as I stated, the same way kodak or Fuji provides an Asa value, for their film with a sticker on it that says 500asa daylight, they are telling us that with daylight sources at meters set to 500asa and exposing this film in that way and processing it with the correct chemistry will yield a correctly balanced image and in this way the stock will perform at or near its optimum.

Thats for for the beginer, it's a baseline.

The he expert says ok. What if I make this 2 stops over, cross process push, pull and find their own style and method.

But it a baseline exists.

In a sense red is handing us the dragon and saying (quoting Phil) the Asa rating of this film is 250-2000 rate it however you like. I just think the manufacturer of the sensor (like film) should determine through their own internal testing the baseline for the novice user. Like, here, 5600k sources, 640 Asa and set aperture and shutter accordingly and you'll get decent relatively noise free images.

Shoukd i and 5000 other users be telling them that? no. If they want people to get successful images out of their sensor , the way lkodak and Fuji would wanttheircustomers to experience, then you test and you determine what is a. Best starting point.

Otherwise se you'll get, I started at. 2000 and pushed it to 6400 and now it's really noisy all day here.

Its up to us to discuss and push the expert stuff once they lay the groundwork.

Battistella

 

[D Fuller]

Well ... in all honesty David - most experienced cinematographers understand ISO on the RED, Alexa, F65, etc. and have no real issue with exposing on the sensor(s) and system.

I think it IS established with the sensor and system.

I just think there is a lot of confusion over ISO with folks who don't understand the basics - which is okay - it can and will be confusing if you never really understood the basics - but this forum exists, right? Reducation exists, right?

A lot of confusion comes from the fact that Red has these three different customer sets: people who learned on film, with its fixed ISO per film stock and workflow that includes post-processing; people who come from video, where ISO has no meaning at all, there is just gain; and the DSLR group, for whom ISO is a baked-in gain-with-noise-reduction setting. They all understand ISO differently, and I think we often talk past each other because we don't understand the terms in the same way.

 

[Gunleik Groven]

A lot of confusion comes from the fact that Red has these three different customer sets: people who learned on film, with its fixed ISO per film stock and workflow that includes post-processing; people who come from video, where ISO has no meaning at all, there is just gain; and the DSLR group, for whom ISO is a baked-in gain-with-noise-reduction setting. They all understand ISO differently, and I think we often talk past each other because we don't understand the terms in the same way.

+ 112

And that is REDs strength and weakness.

Problem is when the film guys get "taught" by a videoguy to unlearn all that lovely knowledge.

Or when experienced TV guys start shooting RED

(both cases will work, but there is room for confusion)

 

[Mark L. Pederson]

As as I stated, the same way kodak or Fuji provides an Asa value, for their film with a sticker on it that says 500asa daylight, they are telling us that with daylight sources at meters set to 500asa and exposing this film in that way and processing it with the correct chemistry will yield a correctly balanced image and in this way the stock will perform at or near its optimum.

Thats for for the beginer, it's a baseline.

The he expert says ok. What if I make this 2 stops over, cross process push, pull and find their own style and method.

But it a baseline exists.

In a sense red is handing us the dragon and saying (quoting Phil) the Asa rating of this film is 250-2000 rate it however you like. I just think the manufacturer of the sensor (like film) should determine through their own internal testing the baseline for the novice user. Like, here, 5600k sources, 640 Asa and set aperture and shutter accordingly and you'll get decent relatively noise free images.
Battistella

Well ... like I have said over and over - there is not really a NATIVE ISO. There is RANGE. If you want to pick ONE baseline - you really need to pick that based on your exposure method and subjective preference. If you are a novice - set the Dragon at 640 ISO and have a great time. You will shoot great images. Done.

But these cameras are not "point and shoot" cameras - they are very high end tools.

I'm just unclear based on your post - and it could be a language thing - exactly what you want RED to do that they are not doing right now?

 

[Mark L. Pederson]

A lot of confusion comes from the fact that Red has these three different customer sets: people who learned on film, with its fixed ISO per film stock and workflow that includes post-processing; people who come from video, where ISO has no meaning at all, there is just gain; and the DSLR group, for whom ISO is a baked-in gain-with-noise-reduction setting. They all understand ISO differently, and I think we often talk past each other because we don't understand the terms in the same way.

Agreed.

The answers, IMO, are ALL in the roots of shooting film.

Still or Motion - but FILM.

RAW is digital FILM.

If you understand the fundamentals of shooting and exposing FILM you are all good with RAW.

RED cameras are digital motion picture cameras. They do not actually shoot "video". And the EPIC is not DSLR camera.

 

[David Battistella]

Mark,

Are you saying ISO 640 is a base rating or is Red saying that ?


Battistella