Converting Red to how Alexa handles saturation

[Christoffer Glans]

So, this is an old article about how Alexa treats colors and saturation.
http://www.dvinfo.net/article/production/camgear/what-alexa-and-watercolors-have-in-common.html

I was thinking about how to match or create some kind of Red to Alexa LUT or a node structure in Resolve to replicate how Alexa handles saturation and colors. I was wondering if there's someone who's been trying this out? What could be done to Red material in order to bring it closer to how Alexa looks as a starting point before grading.

 

[Hrvoje Simic]

I was thinking about how to match or create some kind of Red to Alexa LUT or a node structure in Resolve to replicate how Alexa handles saturation and colors.

Waste of time.

What could be done to Red material in order to bring it closer to how Alexa looks as a starting point before grading.

If by "closer to how Alexa looks" you are referring to certain image properties and not the "look", one could apply a very specific image optimization.

 

[Christoffer Glans]

Waste of time.



If by "closer to how Alexa looks" you are referring to certain image properties and not the "look", one could apply a very specific image optimization.

The baseline is replicating the initial starting point you have with Alexa. It does handle colors differently, so it ease ups the grading and create an initially better looking skin tone primarily. Whenever I start out grading Red footage, the saturation is all over the place, most skin colors feel flat and unnatural. If I could get closer to Alexa as a first node structure that would help a lot. I don't think it's a waste of time since most clients like the look of the Alexa over Dragon. I know most people don't or won't care, but there are some who does and there's no reason not to be able to deliver.

 

[jussi rovanpera]

Notice that Art Adams is specifically writing about 709 gamma with Alexa, even thou he doesn't mention it in the article. I haven't seen this kind of "Lum vs Sat" behavior with log c, and it would not make sense with log footage - if you'd have exposures couple of stops apart, you'd have differently saturated colors.

But I use a Sat vs Sat curve all the time in Resolve to tame saturated colors, and also the Lum vs Sat curve to get more "filmic" saturation.

 

[Sabovic Adis]

The baseline is replicating the initial starting point you have with Alexa. It does handle colors differently, so it ease ups the grading and create an initially better looking skin tone primarily. Whenever I start out grading Red footage, the saturation is all over the place, most skin colors feel flat and unnatural. If I could get closer to Alexa as a first node structure that would help a lot. I don't think it's a waste of time since most clients like the look of the Alexa over Dragon. I know most people don't or won't care, but there are some who does and there's no reason not to be able to deliver.

Are You getting extra paid for matching Red to Arri?

 

[Hrvoje Simic]

The baseline is replicating the initial starting point you have with Alexa.

That is not a realistic baseline.

I don't think it's a waste of time since most clients like the look of the Alexa over Dragon. I know most people don't or won't care, but there are some who does and there's no reason not to be able to deliver.

The look is related to certain image properties. That is understandable and much of those properties can be achieved. Waste of time would be the assumed route to get there.

 

[Chuks Ude]

Hi Christoffer, have you tried this:

https://vimeo.com/182168023

An explanation here:

http://liftgammagain.com/forum/index.php?threads/color-space-transform.7431/

 

[Ugo Del]

The baseline is replicating the initial starting point you have with Alexa. It does handle colors differently, so it ease ups the grading and create an initially better looking skin tone primarily. Whenever I start out grading Red footage, the saturation is all over the place, most skin colors feel flat and unnatural. If I could get closer to Alexa as a first node structure that would help a lot. I don't think it's a waste of time since most clients like the look of the Alexa over Dragon. I know most people don't or won't care, but there are some who does and there's no reason not to be able to deliver.

Two different looks for two different cameras. The only way would be to modify the hue/saturation for each color, and work on the saturation curves, which will lead to something a bit unnatural and even noisy if pushed too far. The solution is to buy an Alexa if you want that look, or a Varicam like I did. I always prefered the look of the Alexa, I tried to mimic it for years on my Reds, and since I work on the Varicam I'm now certain that I wasted my time trying to make Red footages look like Alexa or Varicam, there's more than the saturation involved, there's the color separation, the hues, the DR and roll off, the initial softness. Red cameras are great, they are small with crazy HFR and many other cool things, but you have to accept its look. If you don't, just buy something else.

 

[Christoffer Glans]

"Solutions" like buying an Alexa instead or similar responses aren't helping, this isn't about which system is better and I couldn't care less about that kind of dicsussion. The idea here is that Alexa skin tones and color handling is better than Red, that's just a fact (fact in that the majority of filmmakers and cinematographers, me included, like how Alexa handles the colors, i.e making it more filmlike organic than standard RGB sats. If we were able to come close to a neutral image that's close to how the Alexa handles colors that would give a very pleasant starting point.

I'm trying to combine the best of both worlds; Reds powerful R3D, HDRx, resolution and so on, together with Alexas nicer colors. It would be far more interesting to hear what actual tweaks in Resolve you could do in order to bring them closer, if only as to increase our knowledge about the differences between the two systems.

At the moment I'm trying out RWG / Log3G10 with a 709 LUT using only the different curves methods (SAT vs LUMA and so on) to try and match it.

 

[Sabovic Adis]

But why make apple look like orange, whyyyy? :rofl:

:thumbsup:

 

[Ugo Del]

"Solutions" like buying an Alexa instead or similar responses aren't helping, this isn't about which system is better and I couldn't care less about that kind of dicsussion.

Well, 2 years ago you were already asking the same question. Did you make any progress in 2 years? Apparently not. Maybe it's a clue. That trying to match them "half way" by making the Alexa images closer to the Red is possible, but having Red footage look like Alexa is not. And that Red is not interested in working on it. The new color science will improve things a bit but it will have the same Red look that seems to please Red team and most Red owners. It's not about which system is better, it's about the fact that they are different. It's been discussed for years.

 

[Hrvoje Simic]

Glad to see Chris has faith.

a) It doesn't have to look like Alexa to posses the image parameters which seem pleasant with Alexa image and analytical part of mind recollects with Alexa
b) It goes from RWG/Log3G10 but needs special treatment, not a generic treatment

 

[Christoffer Glans]

Well, 2 years ago you were already asking the same question. Did you make any progress in 2 years? Apparently not. Maybe it's a clue. That trying to match them "half way" by making the Alexa images closer to the Red is possible, but having Red footage look like Alexa is not. And that Red is not interested in working on it. The new color science will improve things a bit but it will have the same Red look that seems to please Red team and most Red owners. It's not about which system is better, it's about the fact that they are different. It's been discussed for years.

That's because I'm not primarily a colorist. I shoot, direct and edit, occasionally doing grading jobs, which mean I have a lot on my table. So me asking is hoping there are some magicians out there far better than me who has been able to do it. The demand for it isn't minimal. The only thing that Alexa has that makes it such a huge part of the industry is how it handles the colors, no one can deny this fact. It doesn't matter what individuals think about it, the Alexa "look" has been the gold standard of colors for digital cinema for a while now. What I mean by gold standard is the neutral basic look you get from it. Most of the time it's rather easy to spot the difference between Alexa and any other brand. It's harder to spot the difference between Red and other brands like Sony in some cases. Just look at "The Crown", which felt a lot like Red. In lack of better terms; there's a luminance to the skin tones on Alexa that's not present in any other camera tech out there. Probably due to the color science mentioned in the article. But if it's only about different levels of saturation, luma values etc. it's quite possible to change it if you have all the information available as you do with RWG/Log3G10. But the "what" is the question. You can do a "fit all" pre-LUT for it. If a conversion is tweaking the colors in just the right way, then the input will always follow that tweak. It's not a change in luma values but in color hue and saturation on different luma values, which means exposure shouldn't be affected and therefor will wok on any exposure. But as I said I'm no expert in this field, I just know that Alexa "colors" are very sought after and should in my opinion be important to study. Resolution and other stuff aside, they've managed to keep their sensor tech for a few years now without the need for an upgrade, while Red and other brands upgrade the hell out of their cameras. I haven't seen the new color science for Helium, but using RWG/Log3G10, I don't feel the jump from Dragon is that big and it's not beating Alexa in this regard. Helium is cleaner and sharper but colorwise, I think there's still work to be done.

That's one reason I would love to try out and see the Panavision color science they developed with Light Iron.

Glad to see Chris has faith.

a) It doesn't have to look like Alexa to posses the image parameters which seem pleasant with Alexa image and analytical part of mind recollects with Alexa
b) It goes from RWG/Log3G10 but needs special treatment, not a generic treatment

Haha
Yeah, I know that it can't be Alexa, but the colors on Dragon aren't handling close to that at all. So I think we could come much closer than we are now.

Just did a try to see if I could manage this with the curves alone (just did a minimal change on the luma, but the rest is just hue and sat balance on different luma values so the change would carry over on any shot.

RWG / Log3G10 - Output LUT to 709

RWG / Log3G10 - Tweak - Output LUT to 709

Seeing them together the "tweak" one feels a bit desaturated, but her skin tones in the original can sometimes be borderline extreme in the red, yellow and magenta areas. On a person with much more uneven skin with no makeup, Dragon can look really awful.

 

[Anton Esteban]

Have you tried to use a LAB color node to experiment with saturation? There is an interesting tutorial on it on Mixinglight.com. Coupled with a saturation vs saturation node after the LAB saturation node you get some very interesting more film like colors. Also, always shooting with some kind color chart helps to bring back the skinniness shot on RED from its natural 'yellowish' tone. However, don't use the auto color chart function in resolve, the best way to correct the hues in RED footage is to do it manually. You can also find a tutorial on it on how to do that in mixinglight.com. If you follow the above steps with an initial node that bumps up the exposure of the shadows you get color that's a little bit closer to Alexa footage. I've gotten amazing results with this strategy, especially in daylight. The problem I could never get around is in darker lighting situations especially with a mixed lighting scenario. This is why I'm hoping that this new Helium sensor will solve that problem. From what I've seen so far, it is much better in lowlight and is definitely more sensitive to color. Skintones seem less 'splotchy' and have much nicer transitions between colors and shades even in non daylight scenarios.

 

[Hrvoje Simic]

Is that a bounce for fill or artificial light source on the lower right ?

The problem I could never get around is in darker lighting situations especially with a mixed lighting scenario. This is why I'm hoping that this new Helium sensor will solve that problem.

It won't. Mixed light pollution screws up color with every camera. That is not the sensor issue.

 

[Christoffer Glans]

Have you tried to use a LAB color node to experiment with saturation? There is an interesting tutorial on it on Mixinglight.com. Coupled with a saturation vs saturation node after the LAB saturation node you get some very interesting more film like colors. Also, always shooting with some kind color chart helps to bring back the skinniness shot on RED from its natural 'yellowish' tone. However, don't use the auto color chart function in resolve, the best way to correct the hues in RED footage is to do it manually. You can also find a tutorial on it on how to do that in mixinglight.com. If you follow the above steps with an initial node that bumps up the exposure of the shadows you get color that's a little bit closer to Alexa footage. I've gotten amazing results with this strategy, especially in daylight. The problem I could never get around is in darker lighting situations especially with a mixed lighting scenario. This is why I'm hoping that this new Helium sensor will solve that problem. From what I've seen so far, it is much better in lowlight and is definitely more sensitive to color. Skintones seem less 'splotchy' and have much nicer transitions between colors and shades even in non daylight scenarios.

Haven't tried it, thanks for the tip
The stuff I've seen so far from Helium doesn't really look that different from Dragon in my opinion. Maybe we have to wait for the new color science, but it's not how it was going from MX to Dragon.
I agree about the mixed lighting situations, that's where Alexa really shines and why I think this is important to look into.

Is that bounce for fill or artificial light source on the lower right ?

It won't. Mixed light screws up color with every camera. That is not the sensor problem but the problem of bad lighting.

The images? That's just the sun through foliage with a simple white bounce on her right side.
Mixed lighting do screw up things for any camera, but Alexa seem to manage it much better. Whenever I've graded Alexa footage with mixed lighting I've never really had that much of a problem, but it's always a struggle with Dragon.


Here's two others. The original look is how Dragon looks through the RWG / Log3G10 with 709 Lut (it's a work in progress from Graeme) together with a node that blocks saturation at the shadow/highlight extremes. The tweaked is the same but with changed balance values for reds and yellows together with some experimentation with luma vs sat.

RWG / Log3G10 - 709LUT

RWG / Log3G10 - Tweak - 709LUT

 

[Anton Esteban]

A good analogy for me to explain the difference between Alexa and RED cameras is similar to the difference between a good shotgun mic and an amazing shotgun mic. A good mic picks up all sound frequencies with clarity but an amazing mic will pick up all those sound frequencies but suppress the frequencies it thinks you don't want to hear, in this case it lowers frequencies that are not close to the sound of the human voice. RED cameras seem to be very sensitive cameras that show all the colors it picks up equally. Alexa's seem to use some kind of color matrix that supresses colors that it deems less pleasing. If you use a hue vs saturation curve in resolve and mess around with lowering and increasing certain hue's saturation you'll see what I mean. I've read in many places in REDUSER how the Alexa camera is closely mapped to look like modern Kodak film stock, hence its film look. Hopefully this new color science coming out will make RED footage more cinematic looking right out of the box.

 

[Justin McAleece]

The images of the dancer look very interesting. If you were to show the first one to most people they would bias it toward RED and the second toward Alexa. I think you are on the right track.

 

[Christoffer Glans]

A good analogy for me to explain the difference between Alexa and RED cameras is similar to the difference between a good shotgun mic and an amazing shotgun mic. A good mic picks up all sound frequencies with clarity but an amazing mic will pick up all those sound frequencies but suppress the frequencies it thinks you don't want to hear, in this case it lowers frequencies that are not close to the sound of the human voice. RED cameras seem to be very sensitive cameras that show all the colors it picks up equally. Alexa's seem to use some kind of color matrix that supresses colors that it deems less pleasing. If you use a hue vs saturation curve in resolve and mess around with lowering and increasing certain hue's saturation you'll see what I mean. I've read in many places in REDUSER how the Alexa camera is closely mapped to look like modern Kodak film stock, hence its film look. Hopefully this new color science coming out will make RED footage more cinematic looking right out of the box.

Good analogy and to the point. Having all the colors to the max might be good from a purely mathematical and scientific standpoint, but creating something from that which feels pleasant in how we feel film feels pleasant is a very complex process compared to working with Alexa, which handles colors not as pure math but as a replication of film colors.

 

[Anton Esteban]

This is why you need to check out LAB color. What it does is somehow magically give colors greater separation without necessarily over saturating them. In many way, this is one of the reasons I love film so much, you have color separation without increased color saturation. Skies are deep blue but skintones and earth tones never look garish.