Welcome to our community

Be a part of something great, join today!

Log in
  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

question to graeme or deanan? TED said:rate red 250 in low con and 400+ in high con

David Mullen ASC said:
That doesn't sound right to me -- I'd like someone to explain why higher contrast scenes benefit from higher ASA ratings, because it seems to me that the dynamic range of the sensor is a constant and all you are doing is deciding whether to capture more highlight or more shadow information by playing around with the ASA rating. But you can't extend the dynamic range being captured with higher ASA ratings.

So what does low contrast versus high contrast have to do with ASA ratings? Are you basically saying that with a low contrast situation, you have more freedom to push the image "to the right" exposure-wise because you don't have hot highlights to worry about? But that really doesn't mean you are extending dynamic range in high contrast situation by using a higher ASA rating and thus exposing more "to the left" to protect the hot spots.
Click to expand...

thank you david.
you did say what i wanted to say.
 
I'm not Graeme or Deanan but if you want MHO...

I'm not Graeme or Deanan but if you want MHO...

I'd wager the perceived DR, not the actual DR, is improved in high contrast material by staying out of the clip. It is my theory that this impression is created mostly by what is revealed in the image by the second step in the process - the grading stage. Remember - with the RedOne we don't paint in camera, we tweak the RAW data set in RA, RC, Scratch, etc.

In low contrast scenes, where a hard clip is much less of a concern and more content of interest is in the bottom (left) half of the histogram, you want to expose to the right. This is even more critical due to the sampling math inherent in a linear device. Resetting the working ISO is simply a means to an end, it makes the exposure tools on the camera read lower so you open up the iris.

The native properties of the sensor do not change - it is what it is, therefore:

My own working style is to leave the ISO at 320 at all times. When shooting day exteriors or other high contrast scenes I take the clip indicators (stoplight primarily) more seriously and when in doubt I close down another half stop. When shooting under well controlled lighting and soft sources I make sure at least one of the stoplights is lit, usually all three.

Despite the many warnings about false color I use it all the time as a reference point. I want to see where 18% grey falls, I want to see what portions of the scene are reading down in the blues, I want to see if wardrobe or surfaces I care about are yellow, orange or red.

I do toggle to view RAW to check the clip point. I work in view RAW exclusively when there are no clients/producers camped in video village. In a perfect world I would assign different view options to various monitor feeds around set. This can be done with LUT boxes on bigger shows but it would be easier if it could be done by camera menu - I'm pretty sure the RedTeam all over this for Epic, they're sharp cookies.

In any case, I find that by leaving the ISO alone I don't have to make a second mental adjustment, YMMV.

Bottom line - I am a place and fall Cinematographer - which means I am focused on getting specific elements of the scene placed in the zones (Ansel Adams index) I want them in. BTW - I carry an Digi-Con when I need more working DR in high contrast settings where greater light control is impractical (where's that 80 by silk and 17 grips) and restore contrast in CC.

Cheers - Blair #19
 
David Mullen ASC said:
So what does low contrast versus high contrast have to do with ASA ratings? Are you basically saying that with a low contrast situation, you have more freedom to push the image "to the right" exposure-wise because you don't have hot highlights to worry about? But that really doesn't mean you are extending dynamic range in high contrast situation by using a higher ASA rating and thus exposing more "to the left" to protect the hot spots.
Click to expand...

Sure. You're just shifting things over, and protecting highlights at the cost of introducing shadow noise. The key is, though, that shadow noise is more acceptable in high-contrast shots. Not because of anything technical happening in the camera, but simply because of human visual perception. In a shot with very bright areas, especially one dominated by very bright areas, you can't see shadow detail as well because of glare.

Thus, in a high-contrast shot, you can expose to the left more to protect highlights, without worrying about overly noisy shadows.
 
David Mullen ASC said:
Are you basically saying that with a low contrast situation, you have more freedom to push the image "to the right" exposure-wise because you don't have hot highlights to worry about?
Click to expand...

Exactly. It means in Red Alert, you can bring the image back down, and the image will be less noisy than if you'd shot higher ISO.
Try it for yourself!
I'm telling you, on that foggy lake, the difference in results between setting the camera at 400ISO and at 100ISO were chalk and cheese. On my little 5.6" monitor, the two versions looked the same. Viewed on a decent size screen, it was another story.
The more light you can give the sensor, without clipping being an issue, the creamier will be your results, especially if you're shooting 3K or 2K.
I guess it's mostly because your pushing the sensor noise in RedAlert or whatever, down into the shadows, but also possibly partly due to what Redcode is doing (being less "lossy" to the mids and highlights than the shadows). I could stand corrected on that part.
 
But everyone is assuming that in a high contrast situation, you're going to want to bias the exposure to protect the highlights. But I can imagine a high-contrast scene where all the important detail was in the shadows and the areas of overexposure were limited to small areas that did not have important scene information. Imagine a room full of black bowling balls piled high with small areas hit by bright spotlights but the bulk of the room was rather dim and underexposed. You'd set the exposure so that the main area of visual interest was going to reproduce with good detail and low noise -- which in this case may be the dark bowling balls sitting in the dim shadows.
 
David, yes. I think the "high-contrast shot" most people are probably picturing in their heads in this thread is more of a daylight exterior sort of situation.
 
Absolutely, David, your second paragraph puts it right.

Noise is not setting in suddenly, like saturation, but the ratio of signal vs. noise is gradually getting worse the more you expose to the left. So, a low contrast scene profits from staying away from those areas while still avoiding saturation.

Changing ISO just helps you judge it on the screen, but looking at a histogram in RAW mode and exposing to the right is always correct – as long as you have a grading session on your schedule.

The only exception to this rule: you want to deliver useable footage right from your camera. i. e. use the RED like a video camera. Then I'd switch to Rec709 and expose visually while adapting ISO to the scene.
 
I did some exterior daylight tests lately and had the impression that there is not such a big difference in the histograms between RAW and 709. It seemed to me the 709 offered a more realistic image that RedSpace, where the colors were highly exaggerated and highlights clipped too early. So exposing to the right in 709, even a bit "over the right" produced a good exposure in RAW. Can anyone confirm this ?

Frank
 
If you take a step back and think how the exposure process is then maybe it is time for some new ideas. The aperture of the lens is supposed to control and regulate the light in to the camera. But usually the light is adjusted before it get there with filters. The aperture is used as a creative tool to give the scene the proper feel with depth of field. If the lens had a active filter (like a transparent LCD that can go from clear to black) light regulation built in the camera or the lens. Then you get new levels of control without carry many cases with filters. ..or if you think more about the "digital filter" thing, you could probably build stuff with same tech that is in normal flat-TV that have multi properties like a grad filter, dynamic limiting of high light leves areas etc..
 
Exposure "rules" are only for a mythical "general" scene exposed at the mythical "correct" level. As David points out, it's the art and skill of the cinematographer to either use the "rules" or not to achieve an exposure that meets the needs of the shot in terms of the story it has to tell.

It's relatively ok to give technical advice, but your common sense and aesthetic have to take over when it actually comes to setting that f-stop on a real lens pointing at a real scene.

Graeme
 
Jones said:
I did some exterior daylight tests lately and had the impression that there is not such a big difference in the histograms between RAW and 709. It seemed to me the 709 offered a more realistic image that RedSpace, where the colors were highly exaggerated and highlights clipped too early. So exposing to the right in 709, even a bit "over the right" produced a good exposure in RAW. Can anyone confirm this ?

Frank
Click to expand...

Well, I believe RAW is (of course) the widest exposure range being captured, so Rec 709 gamma will be a bit truncated in terms of what gets displayed.

Actually that's more or less how I like to work, which is have a display -- whether thru a LUT or some other method -- which is higher in contrast than what is being recorded, so I know I'm giving myself some extra detail to work with in post if I can make it look right for the more limited dynamic range that the display is showing. Then in emergencies when the contrast is over the top, I can check the RAW image to see what I will actually get to work with in post.

The trouble with monitoring in RAW is that you have to train yourself to remember that some visible shadow detail is buried at the noise level and will probably be less visible once you make the blacks nice and rich in post. Plus you might not give yourself any leeway on the highlights in post if you lit the scene so that you need every bit of overexposure detail visible in RAW.

On the other hand, I think, in order to get a film look, you're going to end up compressing (with a knee control or something) every bit of over-exposure detail to be visible in the final display gamma of the project, so one LUT I've sometimes used (for the Genesis) is where I crush the shadows down a little to force myself to add more light there and get detail recorded above the noise floor, but used all of the highlight information in RAW. But then, I have no wiggle room in post because now I've incorporated every bit of overexposure detail into the planned image.

I monitored my RED projects using Rec 709 and generally that worked out in terms of how I like to expose. I only wish I had crushed the blacks on the set monitor a bit more to force myself to add more fill. Trouble with a lot of LCD monitors being used on sets is the black level and backlight when viewed off-axis often lifts the image and shadows, giving you a false impression that you aren't adding enough contrast to the lighting.
 
"I only wish I had crushed the blacks on the set monitor a bit more to force myself to add more fill. Trouble with a lot of LCD monitors being used on sets is the black level and backlight when viewed off-axis often lifts the image and shadows, giving you a false impression that you aren't adding enough contrast to the lighting."

Thanks David. Is this true for all digital cameras? Do they all need a little more fill than you would intuitively expect? And can contrast be increased in post with no loss in image quality?
 
Within limits -- and it's not particularly a digital camera issue, if you want to slightly crush the blacks in a D.I. with something shot on film, it helps to have a little extra fill to compensate. I'm not talking about using flat lighting either because you can't really fix that in post, it looks too artificial.

There is a classic connection between deep blacks and shadow detail in that the richer the blacks in the print or in a digital color-correction, the more low-end detail is buried. This isn't necessarily a bad thing either, it's why many cinematographers used to overexpose negative film when doing a photochemical post and contact print -- you end up with deeper blacks once you print down and detail seems to fall-off nicely into pure blacks rather than into mud.

With film, the low-end detail tends to be grainy and in digital, it tends to be noisy, so the problem is similar -- you want the flexibility to bring down the blacks yet retain some shadow detail, so you want to use fill light carefully and thoughtfully with this in mind while avoiding the look of obviously filled-in shots or flattened lighting without good contrast.

But it's partly a taste thing -- I like really deep blacks. I like ENR prints for example, or prints using Vision Premier stock. Solid blacks in a frame add depth and dimensionality to the image.

I once talked to Roger Deakins about use of fill light and he said that he really didn't think of lighting in those terms.

If you watch his movies, he likes things to roll-off into black -- what he does to reduce the need for fill is to wrap his key around, sometimes by putting a soft light that is on the same side as the key light, just adding to it more frontally but less intense -- so you have the effect of a side-lit face but both eyes are lit because the key feathers off as it becomes 3/4 frontal, yet there is no light on the shadow side, so 1/4 of the face stays nice and black. Of course, other times he will have scenes filled in more naturally just from the nature of the room lighting and the ambience. But you rarely see what could be perceived as a fill light, just there to fill in the shadows. Instead you see varying degrees of softness from the key light. It's a good way to think about lighting.

I'm a bit more technical about it, I tend to always have some fill from behind camera, but knocked so far down as that it barely registers. It is sort of my safety net because if I see in the first take that the actors are doing something where they end up facing away from their key, I have the option of slightly brightening the fill to compensate. But sometimes the electricians look at me funny because I set up this fill but knock so far down as to make you wonder if it is doing anything. I find that at the least, it tends to create a nice glint in the eyes and clean up some noise and color cast in the shadows.

But there are other times where I just put some white cards or cloth down off-camera to increase the ambient fill in the room. And if you are shooting in a white-walled space, sometimes there is plenty of fill without adding anything - in fact, you sometimes need to add negative fill to get some contrast back.
 
incredible, all this. I used to be fine until I started reading all the threads sanjin mentioned. I've decided to go back to simplicity again. rgb parade in rec 709, and if I appear to have been slightly off, exposure adjustment in redcine. Work needs to be done!
 
David: "I tend to always have some fill from behind camera, but knocked so far down as that it barely registers. It is sort of my safety net because if I see in the first take that the actors are doing something where they end up facing away from their key, I have the option of slightly brightening the fill to compensate."

That's why you get the big bucks! Great advice.
 
Graeme?
"I'd not ETTR by adjusting ISO though, but use the raw clip meter or histogram in RAW to achieve it precisely."

A question about exposing with histogram and stop lights.

Do the "stop lights" go off as soon as any one pixel in that color peaks at 100?

I have a problem using histograms, for the following reason: The place you want to see the most critical info (at the right) has the smallest peaking. It's so small, it's hard to see if it's there or not. Is there a way to highlight what has peaked so that it is more obvious?

Like -- could this turn into a FEATURE REQUEST to ZOOM IN to the right most part of the histogram, so that it is easier to see exactly what's there in the most critical area? You could call it the ETTR ZOOM MODE or something.
 
I think Stuart knows the numbers, but a certain small percent of the pixels have to clip before the light activates. A single pixel clipping will not turn the light on.

I tend to get very good results using the raw vertical bar graph indicator myself.

Graeme
 
You must log in or register to reply here.
Back
Top