ARRI Christmas LUT

[Anton Esteban]

I have a Dragon and Helium Weapon that I've been testing Johns LUT extensively. I think it does work a lot better, out of the box, on Helium footage. But I found that my Dragon footage just needs a little more saturation in the node before the LUT and a little Hue vs Hue for the skin tones--it renders my Dragon footage skintones a little red/magenta for my taste so you just have to give it a nudge yellow. I've also found that I needed to boost the saturation of Blue and Cyan a bit. Overall John's LUT is amazing. I just don't understand how he was able to map the luminance values per color without breaking the image. When I tried making my ARRI like LUT I noticed that the Alexa renders blue and cyan darker than RED's. Its nearly impossible to translate that change into a LUT in Da Vinci Resolve without all kinds of bad translation artifacts. John obviously used the highest quality tools to make his LUT.

 

[Anton Shavlik]

I never found a way general purpose way to use the hue vs. luminance curve in resolve, nor to key a color and darken it. For me it's worked to use the color channels themselves as a matte to darken saturated blues and cyans. Using blending modes like lighter color, subtract, multiply etc you can create some nice logic to clean up the image based on the color channels themselves.

 

[Anton Esteban]

Very interesting. Do you use a layer mixer or parralel mixer to do that? I got to try that out. I always thought doing something like that would have unintended consequences with the green and red channels. But maybe that's the whole point of separating them.

 

[Anton Shavlik]

You have to think about how you want to approach it. Like in my mind you take the red and green channels and use 'lighter color', and then subtract that from the blue channel. What's left is anywhere the blue channel was greater than either red or green. If you invert that and multiply it by the luminance of your original images you've made anywhere that feels blu-ish become darker. You can use the log controls on the layer you inverted to change the cutoff of what feels 'blue-ish' as well as how strong to effect the blue-ish parts of the image.

I hope that makes sense. I'm by no means a colorist but I got frustrated by some of the controls in resolve and I find doing things like this sometimes gets me what I want in a cleaner quicker way.

 

[Anton Esteban]

I always loved playing with the soft light and multiply mode in the layer mixer because of the way it gave more film like contrast. I would then key the layer to taste. I never thought of using it with specific channels of color. I got to see how this works on John's test image. He did open my eyes to unintended consequences of corrections. Of course if it looks good then it doesn't matter but in making a general fix like a LUT then it matter a lot. I think your strategy offers a much cleaner way to control color luminance. I've been using very soft qualifiers of colors from the preset menu--yellow, red, green, magenta, cyan, blue--and then playing with their luma response with curves. It works great in a scene to scene basis since you can refine them to get it right but is not great as a general fix.

 

[Marc Wielage]

I never found a way general purpose way to use the hue vs. luminance curve in resolve, nor to key a color and darken it.

For me, hue vs. lum is a pretty dangerous mode. However, you can use it as an overall brightening tool if you make an adjustment to the whole thing (not one specific hue). It is a very, very dodgy mode.

 

[Anton Esteban]

I don't mess with Hue vs Luma at all but your ideas of using it as a brightening tool makes total sense. I find qualifying each color as a specific node along with a 'skintone' qualified nod grouped together through a parralel mixer and then adjusting the colors there works best for me. You can then select the whole mess of nodes and then turn it into a compound node that you can use as a preset for future corrections.

Changing the luminance of color was a revelation for me. I used to just adjust Hue and Saturation based on a color chart like the DSC labs oneshot. After matching the colors of an ALEXA chart to a RED chart I realized how important it is to adjust the luminance values of color for better color accuracy. This is why I'm about to buy this guy:

http://dsclabs.com/chromamatch-pro/

DSC LABS supply you with a testing image that you can compare hue, saturation, and luminance of all colors.

 

[Marc Wielage]

All the DSC charts are useful, but there is a point where you have to embrace the practicality of spending too much time on one shot. If you have a 10 shot :30 second commercial and can spend 6 hours on it, that's great. But if it's an indie feature or a documentary with 2000 cuts that has to be done in some reasonable amount of time -- like 7-10 days (which is a luxury on some projects) -- you are simply not going to have 5 minutes per shot to infinitely adjust the luminance of specific colors. You're going to have to do an overall, fix certain lighting problems, give the shot a look, and move on. Not every frame can be a Rembrandt, every day, or an entire project.

There is also the risk of over-processing material to the point where it gets pretty stomped on and has kind of a non-organic look. Sometimes, simplicity is a better direction in which to go.

On the other hand: I'm in the middle of a project where just yesterday, I was forced to lighten the luminance value of the fleshtone in a character because he's deeply-tanned, has bad makeup, and looks terrible. The extra effort on some of his shots avoided letting him have leather-like skin, which I suspect is how he may look in real life.

 

[Anton Esteban]

One thing I discovered about being a RED owner and using the DSC charts on jobs is I started to notice that beyond some dramatic lighting changes--i.e. going from daytime to tungsten or using HMI, Fluorescent, or LED (which I always found you need to check the charts for)--I noticed general trends in the way my camera handles the charts and that if I made certain camera bias adjustment for a given lighting setup I usually found it worked exactly the same for the rest of the shots. The only thing I would check from time to time, especially if it was outdoors, was the white balance. I've also found this to be true as a 'general' rule for other cameras that I've had to correct for. If I see a problem that jumps out at me that's when I go back to a shot of the charts to see why it looks odd.

What I'm excited about the DSC Labs Chroma match chart is I never really tried to correct for the luminance bias per color that my RED's have. I'm hoping that once I get a general bias correction going, I'll just have to tweak that preset slightly for more dramatic light source changes.

That being said, you are absolutely right when saying simplicity is key. I remember color correcting one hour reality TV shows many many yrs ago in one day. I had a rule to make it 3 nodes max--white balance, exposure, simple final tweaks--otherwise I knew I'd be screwed. I guess it just depends on the project. I work namely in fashion were the spots are so short but color and image is all my clients, sadly, really care about.

 

[Marc Wielage]

That being said, you are absolutely right when saying simplicity is key. I remember color correcting one hour reality TV shows many many yrs ago in one day. I had a rule to make it 3 nodes max--white balance, exposure, simple final tweaks--otherwise I knew I'd be screwed. I guess it just depends on the project. I work namely in fashion were the spots are so short but color and image is all my clients, sadly, really care about.

I think it depends on the nature of the reality show and the length of the day. If it was 1200 shots in one show (which is possible), then you'll be lucky to do :30 seconds per shot, which is 10 hours (without a render). If it's just 500 shots, and they're well-lit, then you you could spend 2 minutes per shot and still wind up at 10 hours.

Number of nodes is a personal choice, and I know of colorists who use a "fixed node" setup of a dozen nodes or more on every shot. The trick is some of the time, the nodes are bypassed and they only turn on the nodes when they really need it. The key is that a LUT will not solve these problems -- it has to be active color-correction with dynamics and keyframes and windows, all over the place. Sometimes the color-correction is very simple; sometimes it's very complex. The nature of the shot changes, just as the nature of the initial photography changes.

 

[Anton Esteban]

I usually got 3 days to do the show but their were a couple of occasions when their post ran late and they would pay a much larger rush fee to make the deadline in one day. It did take me about 10 hours to do one pass and then I simply rendered it to Final Cut Pro and then I'd watch and rerender scenes that needed big fixes. it was an Alaskan survivalist show which was well shot and 90% daylight. The main thing I did was apply a cool/desaturated node that they wanted as a look. I'd been doing the show for a minute and it was so formulaic that I already had a basic look/correction dialed in for each of their given setups. Beyond the look, most shots just needed some quick tweaks and that's the only way I was able to do it in such a short time. It definitely wasn't my best work but they were always happy.

These days I can spend 1-2 days on a :30-:60 video depending on the complexity of the project. The image and color is everything for my fashion clients.

Now as far as LUTs go, who says that you still don't do active color correction even with the LUT. In the case of the True Color LUT i see what John did as a first step transform for people who aren't happy with REDs starting point look. I use it because of the way I like how he has mapped the luminance, hue,and saturation of each color. This doesn't mean that after applying his LUT I call it a day. It's just the beginning. I just find it a much better starting point, especially in regards to skintones, overall tonality, and anything blue.

 

[Anton Esteban]

On another note, I noticed quite a few people selling their Helium sensor cameras. I hope this is not buyer's remorse. I think this sensor is a big improvement from the DSMC 1 Dragon (DSMC 2 Dragon is less of a gap in my opinion). I guess this is one of the reasons I tried to make an Alexa like LUT. I wanted to show how much more dynamic range (from digging in the shadows) and color flexibility this sensor has over previous sensors. The Dragon is still a beast and that's why I kept one of them, but the Helium Weapon definitely gives you plenty of new options especially because of its low light sensitivity. I shoot a range of higher frame rate footage and the fact that I can push the ISO higher is a lifesaver when you simply don't have enough light or need to get deeper focus. Yes the grain is bigger once you crop in but its still a lot better than the 4k 120 from my Dragon DSMC 1. I mean the fact that you can finally use LUT's on camera almost made the upgrade worth it just for that. I was sick of telling my clients that I'd make it look nice in post. Now I can show them something in the ballpark of what I'm actually going to deliver.

 

[Marc Wielage]

Now as far as LUTs go, who says that you still don't do active color correction even with the LUT. In the case of the True Color LUT i see what John did as a first step transform for people who aren't happy with REDs starting point look.

I haven't seen anything done in a so-called "look LUT" that you can't just do by hand from scratch with the controls available in any decent color-correction program. Any experienced colorist can whip out two dozen prefabs looks in a couple of minutes to give the client the range of looks available. I generally prefer that the client give me a day to do a pass by myself, then I can match all the shots then work on a look as a second pass the next day -- if we have that luxury. But I've done thousands of projects without LUTs that came out just fine.

I concede that for neophytes and for people who have to work extremely quickly (particularly editor/colorists on a deadline) a LUT is a quick way to get some kind of look that can get through the show. And no question, a custom-made LUT specifically designed for a particular shoot, specific exposure, and a specific DP can work to establish the general tone. I generally try to talk the client out of it provided I can exactly match it with a PowerGrade, and so far, I've been able to do that. The advantage is using a look that's not destructive, as some LUTs often are.

In LA, I'd say 3 days is a luxury for reality shows. Two days is pretty typical for an average network scripted 1-hour drama. There are always exceptions, like Game of Thrones (which takes a week) and other very high-caliber programs. Once the show has been on a few weeks and the looks have been established, it's not hard to keep them on file and dial them up when the client says, "let's go for that super-blue desaturated look we had on show #2" or "can you show me that brown, vignette flashback look we did in the pilot?" Once they have a shortcut like that, the sessions can go very quickly. No LUTs per say, but pre-fab looks are definitely used and can save time.

 

[Stacey Spears]

Changing the luminance of color was a revelation for me. I used to just adjust Hue and Saturation based on a color chart like the DSC labs oneshot. This is why I'm about to buy this guy:

http://dsclabs.com/chromamatch-pro/

DSC LABS supply you with a testing image that you can compare hue, saturation, and luminance of all colors.

You should also check out the ChromaCineX chart from DSC Labs. (Shameless self promotion, though I make nothing from it.) It does not have the reference you lay over it like the ChromaMatch or ChromaMatch Pro, but it does allow more in-depth analysis. I use it in front of every doc interview. As Marc points out though, not everyone project has the luxury of time.

 

[John Schick]

Now as far as LUTs go, who says that you still don't do active color correction even with the LUT. In the case of the True Color LUT i see what John did as a first step transform for people who aren't happy with REDs starting point look. I use it because of the way I like how he has mapped the luminance, hue,and saturation of each color. This doesn't mean that after applying his LUT I call it a day. It's just the beginning. I just find it a much better starting point, especially in regards to skintones, overall tonality, and anything blue.

Thanks for all the feedback, and that was my intent exactly - to create an alternative starting point for grading.

I haven't seen anything done in a so-called "look LUT" that you can't just do by hand from scratch with the controls available in any decent color-correction program. Any experienced colorist can whip out two dozen prefabs looks in a couple of minutes to give the client the range of looks available. I generally prefer that the client give me a day to do a pass by myself, then I can match all the shots then work on a look as a second pass the next day -- if we have that luxury. But I've done thousands of projects without LUTs that came out just fine.

I would classify these LUTs more as technical transforms than mere “look LUTs”. They have accurate gamma conversions (REDlogFilm / LOG3G10 to AlexaV3LogC), as well as colorspace transforms (REDWideGamutRGB to ARRI’s Wide Gamut / FilmMatrix). The exposure is increased by +1 stop (to better match the Alexa at equivalent ISOs) and the LUTs implement tone-mapping to retain out-of-gamut colors.

You could try rebuilding all that manually on a per shot basis, though the benefit here is this applies a known, standardized transform to the image every time - achieving more consistent results between shots.

This is akin to a colorist applying ARRI’s Film Matrix LUT on some Alexa Wide Gamut footage or selecting DRAGONcolor instead of REDcolor4; Log3G10 instead of REDlogFilm - it’s just a different input option for the footage, prior to any creative look development.

Granted, these particular LUTs aren’t for everyone - but I just wanted to clarify that there’s a lot more going on here than just another arbitrary “look”.

 

[Anton Esteban]

I haven't seen anything done in a so-called "look LUT" that you can't just do by hand from scratch with the controls available in any decent color-correction program. Any experienced colorist can whip out two dozen prefabs looks in a couple of minutes to give the client the range of looks available. I generally prefer that the client give me a day to do a pass by myself, then I can match all the shots then work on a look as a second pass the next day -- if we have that luxury. But I've done thousands of projects without LUTs that came out just fine.

To be honest I never use LUTS except for the ARRI Rec 709 Transform and now John's RED to Alexa transform. Like you, I do everything with corrections, which I have a vast array of pre-fabricated corrections that I use depending on the project and scene.

I concede that for neophytes and for people who have to work extremely quickly (particularly editor/colorists on a deadline) a LUT is a quick way to get some kind of look that can get through the show. And no question, a custom-made LUT specifically designed for a particular shoot, specific exposure, and a specific DP can work to establish the general tone. I generally try to talk the client out of it provided I can exactly match it with a PowerGrade, and so far, I've been able to do that. The advantage is using a look that's not destructive, as some LUTs often are..

Mark have you even tried John's LUT? I'm really curious to see what you think of it after giving it a try.

In LA, I'd say 3 days is a luxury for reality shows. Two days is pretty typical for an average network scripted 1-hour drama. There are always exceptions, like Game of Thrones (which takes a week) and other very high-caliber programs. Once the show has been on a few weeks and the looks have been established, it's not hard to keep them on file and dial them up when the client says, "let's go for that super-blue desaturated look we had on show #2" or "can you show me that brown, vignette flashback look we did in the pilot?" Once they have a shortcut like that, the sessions can go very quickly. No LUTs per say, but pre-fab looks are definitely used and can save time.

The usual schedule I had was 2 days to make the corrections and then a few hours on the 3rd day to sit with a producer and watch the whole thing so he could call anything out. The 1 day turnaround I was talking about literally meant I received a hard drive that early AM with the said show, I spent a good 10-12 hours going through it all and then the producer would come that night to go through it before a final export. When you say 2 days in LA does that mean the next day is when you go through the show with the network? Just curious.

 

[M Most]

In LA, I'd say 3 days is a luxury for reality shows.

I worked on two reality shows. On one of them I was given 8 hours, but that included playout to tape (this was a few years ago) and addressing client notes. So it was more like 6-7 hours. On the other I was given 6 hours.

Most colorists who have never worked in the major L.A. television post facilities are not aware of the time pressure that's applied. If you add up all of the seats in LA facilities for final colorists on broadcast shows, there are probably only about 30-35 colorists in those seats. That's a pretty exclusive group. To occupy one of those seats, you need to be able to display a combination of quality and speed, and both are equally important. Every one of the colorists in those chairs has both, and if they don't, they don't stay in the chair. As a colorist on any one of those shows, you need to develop a sense of what's important to address and what isn't, you need to be able to address those things very quickly and decisively, and you need to do it regardless of the quality of the original photography (although that is usually pretty high on studio level work, not necessarily high on reality shows). The bottom line is that anyone working on studio level shows is pretty darned good. Some are truly outstanding, some are less so, but they are ALL very good. And fast.

 

[Anton Esteban]

I can only imagine how good you have to be to work at that level.

 

[Kenneth C Merrill]

The exposure is increased by +1 stop (to better match the Alexa at equivalent ISOs)

Just to clarify, you're saying that at equivalent ISO settings, the Red exposes a whole stop darker than the Alexa? In other words all other variables being equal, to get the same luma values as an Alexa at 800 ISO you would have to rate the Red at 400 ISO?

That may have been true for the Dragon (which I have heard elsewhere exposes dark compared to other cameras), but I would be surprised if the Helium sensor rated that slow.

I have a feeling I may be misunderstanding things a bit.

 

[Anton Shavlik]

Kenneth, I think it's the opposite. Red at 800 iso 'looks like' arri at 400 iso as far as the brightness of the image, but that has nothing to do with noise or clipping i.e. how to rate the sensor. Arri looks like it has a brighter midtone as well as lifted shadows compared to red. You could put a curve on the red to get the same effect, so that 400 iso on red looked like 400 iso on arri.