skintone

[David Battistella]

Graeme,

This is very obvious with the constant improvement at each level, hardware, firmware, software, and decoding. It's been a great pleasure to see the RED image aesthetic improve with every build. The color reproduction is what is most important to me and I think that RED reproduces colors very accurately.

I don't get much into the "this over that thing" or what is best out of the box, but more understanding the tool and what is required for each situation.

We can al probable agree that the camera produces stunning sharp images at with light sources 5000K and up and preferes more light than less (just like motion picture film). and this all of course depends ont eh creative choices of teh production and the individuals. We have all benefited from and appreciate you hard word and great passion.

David

Thanks Scott. I know that R1, R1 M-X, Epic and Scarlet can produce great skin tones (as many people in this thread demonstrate) and when you watch a lot of movies (I saw a trailer for a movie the other day, and I'll leave it nameless, but it's skin-tones were awful. It was obviously digitally shot, but I was happy that IMDB didn't list it as shot on RED) there's some stuff out there that's oh-so-much worse than any decent digital cinema camera is capable of producing it's not funny. That said, I work on a principle of constant improvement and feedback, and I'm sure everyone will be happy with how we progress in the future. With that regard, the tone of this thread and supportive attitude is refreshing and helps, rather than hinders the process.

Graeme

 

[Björn Benckert]

Wish we can load lut's into the dragon to use for previews.

 

[Dave Blackham]

I agree. If we could switch in and out a monitoring LUT this would be super helpful. Probably generate it in either Resolve or RC-X, Id assign it to a soft button on camera.

 

[Mike P.]

I went over this in another thread.....

1) There's no "native" ISO, because we don't employ destructive analogue gain, and hence there is no "least gain / damage" setting, as all ISOs have the very same analogue gain and are all non-destructive

2) Each ISO setting over a wide range is equally valid

3) what we suggest as an ISO starting point is determined by highlight protection and noise

Graeme

Ha! Thanks Graeme. I'm pretty sure you've gone over it in more than just one other thread

I understand and agree with why shooting at 800ISO protects highlights, and how there is little difference between 320-800 in terms of noise. The problem for me (conceptually?) is that if there were no native sensor ISO, than wouldn't/shouldn't ISO adjustments affect the RAW sensor capture? As far as I'm aware, they do not; they only affect the histogram. If I were "changing" the sensor's native sensitivity via ISO adjustments (okay, asa adjustments), it should affect my raw goalposts and stops lights, and not just my histogram (in the same way that changing from 100asa to 500asa film stocks - the "sensor" of film cameras - would.) In other words, 500asa film isn't just 100asa film with gain applied; it actually "absorbs" light faster to make the image brighter (or at least that's what I thought).

With digital, sure no gain is applied, but the raw capture is pretty much exactly the same regardless of what ISO you shoot at (in terms of where clip and noise floor are.) In film, you'd get an extra stop of detail in the shadows, while giving up an extra stop of highlights, right? That's why perceptually, the MX sensor seems to have a native ISO.

Of course I could be throwing fud and be completely and absolutely wrong (in which case, I'd redact this post to avoid confusion of future readers and to prevent people from reading my ignorance.)

 

[Graeme Nattress]

Well, film ISO works by changing chemistry. We don't do that. ISO setting for us is just electronic gain in development, so it doesn't impact the RAW R3D file at all. When there's a physical change, like chemistry, or analogue gain (or in some cases burned in digital gain) there is a sense of a certain setting being native, or that particular stock is native to a particular ISO value. So I think your argument above more applies to us not having a native "do least harm" setting, as compared to a DSLR say where some settings of ISO are done with pure analogue gain, (and hence the one with least analogue gain could most likely be called "native") and some with burned in digital gain which are often compared to the ones which don't burn in digital gain by saying they're "not native".

A confusing topic for sure.....

Graeme

 

[Mike P.]

I guess what I'm after is at what ISO is 18% grey at ~18% (when physical exposure is set and a chart illuminated accordingly)? That, regardless of how it's dressed up, is the native sensitivity of the sensor. The noise floor and the clip point are fixed regardless of the ISO setting (~13stops apart, between which 18% grey fluctuates based on ISO) and all raising ISO does to maintain highlights is force you to physically (f-stop/shutter/ND) "under" expose (which won't look underexposed when viewed at ISO800). Because of that, fill becomes more important, since everything below 18% is getting crammed in to fewer bits. HENCE, if you light for what is perceived as the "native" ISO of the sensor (where 18% = 18%), and then expose for ISO800 (where it looks brighter, so you expose down to compensate), you're essentially ensuring that everything below 18% is getting ample light... On the flip-side, if you light for 18% at ISO800, its actually already "under" exposed, and you'll invariably gives it the less than ample light (because it'll look okay at ISO800.)

The problem is, you can't just blast the scene full of light (e.g. light for ISO100, expose at 800) because then you'll massively under expose when you switch to ISO800... plus it becomes more difficult to shape light and it would require even more than you need. With daylight balance sources this is less of an issue... With tungsten, your blue channel gets severely low (causing noise even at 800ISO, where people assume they're safe/clean because 'that's the MX starting point') and I find skin tones take the brunt of (blue) light deficiency.

By the way, I'm writing this out (needlessly elaborating) more so to try to show/explain how I'm thinking of things, and not to "prove my point."

 

[BRANDON JAMESON]

Whoa, whoa, it's not a technique and definitely not a rule (I amended my statement to make it less definitive)... It's basically just me saying, in my testing I generally find that, 'more light is better than just-enough light' and it goes under the assumption that you're still shaping the image (with light and ratios) just as much as you would at 800ISO. The problem I'm finding when lighting to ISO800 is that I often underestimate the amount of fill needed and just riding the contrast/curve to compensate doesn't give amazing results (usable, but not amazing). The other thing is that saturation in both highs and lows tends to disappear on RED footage (this may be a characteristic of all digital cameras though I've never noticed it as much as I do with RED footage), which means you need to fill them to maintain tonality/colour in those areas and then bring them down in post. Technically you could rely on your colourist, but that doesn't seem like an ideal solution to me (plus what they can do depends a lot on the source footage you give them).

[Off topic; this is also why I still think raw stills controls (like for Canon and Nikon stills... and Cinema DNG) with their "fill" and "recovery" sliders is still a better implementation than RED's controls. Both fill and recovery (I believe) increase/maintain saturation and (micro)contrast within the shadows and highlights dynamically as you slide those sliders... With RED/RCX you have to mask those areas and (re)apply saturation/contrast... Similar result, just a pain in the ass and much longer than just "increasing the fill slider."]

So again, nothing formal and it's not a technique, I'm just saying that when I light to 320 but then bump the ISO to 800 (and expose there), the results tend to be "thicker" when I manipulate them afterward (especially in the mids/skins). Conversely, I think that's what these film LUTs essentially do (they expand the mids, while trying to roll off lows and highs), but they do it after the fact.

Thank you for the clarity and the wisdom..much appreciated Mike.

 

[Graeme Nattress]

"I guess what I'm after is at what ISO is 18% grey at ~18%" - is that 18% in the RAW data code values, or 18% in the developed image? Because if you put it at 18% in the RAW, you are now ISO200 say, and you have 2.4 stops of headroom over mid-grey, and no headroom at all over 100%. Obviously that's a lower limit of usefulness, and I'd also suggest it's probably only useful in low-contrast scenarios. But it's not less or more valid than putting mid grey at 4.5% in the raw, and having two more stops above mid grey (and correspondingly two more stops above 100%) with an ISO of 800. Of course, ISO is not determined by raw code values at all, but by the code values in the developed image, mapping 18% light to ~44% in the image.

Graeme

 

[Vico Martin]

Actually, I find More DPs that ask me to rate MX to 320 ISO arguing that RAW view is rated at this ISO. Normally I advise to follow the manufacturer's instructions, but it's a conversation that I need to have increasingly often.

Congrats everyone for this thread

 

[Graeme Nattress]

The "RAW view" hasn't changed since the original R1, and of course, it's not really a RAW view because RAW cannot be reasonably said to be viewable. It's primary design use is to see exactly what is going on in the highlights, so has a very low gain development curve to allow you to do just that.

Graeme

 

[Bob Gundu]

Perhaps the RAW view name has to change? Maybe sensor highlight view?

 

[Filip Orlandic]

Bad lighting mostly HMI and flourescent, wrong white balance with print out kodak and fuji 3d luts produce ugly skin tons on any digital camera even on 35mm. I don't want to mention shots under mixed lighting conditions. If you prefer saturated images you need to take care of it a lot more than usually.

 

[Gunleik Groven]

Actually, I find More DPs that ask me to rate MX to 320 ISO arguing that RAW view is rated at this ISO. Normally I advise to follow the manufacturer's instructions, but it's a conversation that I need to have increasingly often.

Congrats everyone for this thread

If you're off 5600 WB, that can be a risky business, if you expose all the way "up".

To me the 800 advice is good, because it gives headroom for WB without clamping any channels too much in the process...

Now... Let's talk WB.... Hahahahaha!

 

[jake blackstone]

Taking to DOP's who've used both Alexa and Epic, there does appear to be a common thread in that they think Alexa really nails skin tones. Feedback from colourists we've worked with also appear to back this up - they tell us RED footage is that little bit more difficult to work with than Alexa's.

That said, I would still pick our Epic over Alexa any day )


Scott

That's funny, that first you solicit advice of DPs and colorists and then promptly dismiss all of them, when you don't get the response you like without offering a reason for the Epic choice
For the record, I concur on the Alexa advice. Said that, there is no reason why you can't get great skin tones from Epic without much effort.

 

[Kevin Marshall]

That's funny, that first you solicit advice of DPs and colorists and then promptly dismiss all of them, when you don't get the response you like without offering a reason for the Epic choice
For the record, I concur on the Alexa advice. Said that, there is no reason why you can't get great skin tones from Epic without much effort.

He didn't necessarily dismiss their opinion out of hand - he didn't say that he's choosing Epic's skin tones over Alexa's. I interpreted it as him preferring the Epic camera system over the Alexa, which is made up of more things than out-of-the-box skin tones.

 

[jake blackstone]

He didn't necessarily dismiss their opinion out of hand - he didn't say that he's choosing Epic's skin tones over Alexa's. I interpreted it as him preferring the Epic camera system over the Alexa, which is made up of more things than out-of-the-box skin tones.

I know, that the name of the thread is the "skintone". But to me image is much more, than just that. Skintone is just a small, but important part of the big picture, that determiners the choice of the acquisition platform. Blindly championing one camera over another, while dismissing opinions of other professionals, for no apparent reason, strikes me as puzzling and insincere. He just said "I would still pick our Epic over Alexa any day " without giving his reasoning for that choice. If he's made up his mind regardless of any other opinion, then just say so. But if he cares more about "Epic camera system", whatever that means, then why bother asking what other people think about skintone from Alexa vs Epic?

 

[Graeme Nattress]

Please don't send a good thread off into pointless arguments.

Graeme

 

[Scott Crawley]

As RedUsers we are committed to a system and this thread is attempting to educate us all in ways to maximize our facility with that system by focusing on one of its perceived weaknesses. It is for the betterment of all Red shooters, not for Alexa comparisons. Frankly that bit is played out and not particularly useful.

 

[Adrian Jebef]

The "RAW view" hasn't changed since the original R1, and of course, it's not really a RAW view because RAW cannot be reasonably said to be viewable. It's primary design use is to see exactly what is going on in the highlights, so has a very low gain development curve to allow you to do just that.

Graeme

I'm working on a job right now with the Sony F65. In the camera's newest firmware Sony added a "High Key / Low Key" toggle. Basically it's the same as the EPIC's RAW View with the added benefit of being able to see the top and bottom (highlights and shadows) that are being captured before clipping. You hit the button once and the video out feed goes into "High Key" mode which pulls the gamma curve way down and gives you a glimpse of all the fine highlight tonalities that are being recorded in the raw files. Hit the button once more and the feed switches to "Low Key" which bends the gamma curve way up and shows you all the shadow detail. A third punch puts you back to normal Rec709 view.

After working with this function for a couple of days now I have to say it's awesome. Not so much for myself on set when confirming proper / expected exposure but more so for the DP, Director, and Producers when they ask "Are the highlights clipping?" or "Is that too dark?". Now a quick push of a button gives a very accurate high and low range of our recorded image. I would love to see RED incorporate a High / Low raw view mode in an upcoming firmware release. Considering the expected expanded DR of the Dragon sensor this functionality would be really helpful.

 

[jake blackstone]

Please don't send a good thread off into pointless arguments.

Graeme

You're correct about pointless argument. It's pointless trying to have an unbiased conversation on this site...