Live: Pro Colorist AMA

[00:00:08]Oh, hello everybody. I don't believe I'm live until I can see my own face on my own live stream, which means that there's 5 seconds of awkward silence at the beginning of literally every single grade school in >> Oh, there you are. time.

[00:00:21]>> You're up.

[00:00:22]>> There we go. Now we're live. What's up everybody? Pumped to see you guys this morning. It's a good Friday to talk about color grading and all kinds of fun stuff in Resolve. Uh what's new with you guys? What's exciting? We're going to dive right in in a second here. Uh I've got some cool stuff to share with you guys first. So, I'm going to share something. This is kind of just for you guys here in grade school today. I'm going to share something that the rest of YouTube isn't going to find out about until Monday.

[00:00:51]But if you're here, how many of you guys are here for your first time in grade school today? let me know in the chat.

[00:00:57]Hopefully, we've got a good combo of OGs and uh newcomers to the grade school scene. But if you are here, then I'm going to assume that you are worthy of uh a little bit of an inside track. So, some of you guys might know if you were here for my last grade school session, I've got something really cool that we're doing right now that we are in pre-launch for right now that we are going full launch with on Monday. I am taking all of the different stuff that I do and offer and talk about and create and make and teach and bundling it into one place where you can have regular ongoing access to it and get 1% better at color grading and at the business of color grading every single day. Uh a quick stat something you guys might find interesting. If you just do a quick assessment of like all right most of us what we do if we're like here to improve our color grading craft what are we going to do three or four times a year we are going to invest in something whether that's a plug-in whether that's a course some kind of resource to better ourselves as colorists right and if we're lucky we'll invest in something really good that makes like big gains in our practice in a short period of time.

[00:02:01]So like let's go crazy and say like hey you're able to get one and a half times better. Let's go nuts. Say you're able to get twice as good at what you do because you signed up for a really good course or plugin or whatever it is, either from me or from whomever else, and it just really delivers and you're able to level up significantly. And let's say you do that four times a year.

[00:02:22]Do you want to know what wins between that and getting 1% better every single day?

[00:02:28]It's 1% better every single day. And it's like not even close. 1% better every single day. that is a 30x increase or excuse me 38x increase ver versus a 16x increase of making these big spiky uh investments in yourself and in your career a couple times a year. So that's really the idea behind why I have chosen to bundle everything that I do into one place where you can have ongoing access to it and get a little bit better every single day. Whether that's through courses, every single plugin that I make and offer, every single plugin that I will make, uh like I said, courses, AMAs, live streams, like everything that I do to serve my community and try to empower filmmakers to be the most powerful professionals on the planet is going to be available inside of this community. In fact, it already is. It's live. It's happening. It's in pre-launch right now. I'm announcing it to the rest of YouTube on Monday, but I'm sharing it with you guys right now. It's called Colorist Inc. I'm going to drop a link to it in the chat. Actually, my buddy Chad is here with me. chat, if you wouldn't mind, drop a link to that in the chat for me. And what's up, by the way, buddy?

[00:03:26]>> Copy. One sec. Hey, what's good, man?

[00:03:29]Let's go.

[00:03:30]>> Yes, sir. All right. So, Chad's going to drop that in the chat for for us. If you guys want to get in on it is the cheapest price right now that it's ever going to be again. Right now, you can get in way cheaper than ever before. And here's the thing, y'all. I I want to be clear, like, I am pumped to be here with you guys regardless. I recognize and I want you to recognize what I'm talking about is not for everybody. Like if you're not trying to see a 38x increase in your color grading craft and enjoy all the benefits that come with that, then it may not be worth the investment that is going to be required for you to be a member of this community. This is for working professional colorists or for people who are all in on that level of color grading and getting to that level. So, it's not a fit for everybody.

[00:04:09]It doesn't have to be a fit for you and we can still be friends. We can still hang out in grade school. But, I did want to give you guys that inside line on it because I'm super pumped about it.

[00:04:16]Chad's actually in there with me. We've been having a blast. We just wrapped up our private AMA that we uh do before grade schools uh just a couple minutes ago for Colorist Inc. this new private community. But uh we had a blast with it and I'm now pumped to be here with you guys regardless of uh anything else. So let's get into it. Chad, what do my friends in the color grading world want to know about today?

[00:04:40]>> Um I let's see. I think, you know, compass placement has been a a common topic uh in terms of like where to put that in a like the context of a larger tree, you know, like um you're a little laggy. I don't know if it's you or me, but um >> it might be me, y'all.

[00:04:59]>> My my internet is not great. Let me see if I can >> Oh, yeah, people are commenting lag. But yeah, that's been a common uh topic lately has been uh like you know the best uh practices as far as compass placement in terms of the context of a larger tree. You know, should it be earlier in in the in the in the in the tree or later? My philosophy is earlier, but you know, I would love to hear I personally would love to hear your take.

[00:05:26]>> Yeah. Cool. All right. So, let's do a let's talk quickly about that. And I'm going to try also to get this lag a little bit happier for us. Y'all, everyone hang tight for just a sec.

[00:05:34]Let's see if we can figure out what's going on cuz this ain't right.

[00:05:40]>> Yeah, you kind of look like Max Headroom.

[00:05:45]>> We don't need that.

[00:05:48]Okay, everything looks good, y'all. This might just be good old-fashioned bad internet service. Uh, let me do this. Hang tight one minute.

[00:06:01]For any of you guys who don't know, I just moved and it's entirely possible the internet is not as bountiful.

[00:06:09]>> You got fires out by you.

[00:06:11]>> Yeah, we got we got I don't know. I don't know what's I don't know what the story is. Okay, I'm going to I'm going to try one other thing, you guys, and then we'll we'll dive in here. Let's just try going good old Wi-Fi.

[00:06:23]Oh, it looks like is it maybe getting better, you guys? There's like this lag.

[00:06:26]So, let me know in the chat.

[00:06:29]Oh, I feel like it looks maybe better.

[00:06:33]All right, I'm going to dive in. You guys, please keep us posted in the chat how it's looking, how it's feeling out there. I don't want to do a live stream for one. I don't want to do a live stream between me and Chad. That's kind of boring. I want to make sure it's awesome for all you guys. All right, so let's talk about some >> I would I would enjoy it.

[00:06:47]>> Yeah, we we would have some fun. Yeah.

[00:06:49]All right, so let's talk about Compass real fast. So, for those who don't know, I'm going to give you like a really brief rundown here. This is a uh plugin that I just released recently. By the way, it's included in that thing that I was just talking about, colorist inc.

[00:06:58]When you sign up for colorist ink, you get licenses to literally everything that I make. Kind of cool. Uh, very simply like placement of compass in Noderee. Let's talk about grading tree in general here. So, I've got my normal color management setup that you guys are familiar with here.

[00:07:14]And boy, oh boy, we're moving slow. I'm going to guess that lag is not fixing itself because even my resolve power.

[00:07:21]>> You put video lag. And let's see.

[00:07:26]Okay, let's see.

[00:07:31]All right, hang on, y'all.

[00:07:33]>> We're looking to stop motion.

[00:07:35]>> Is the Let me like question for the chat. Can you guys hear me? Okay. And most importantly, if I share my screen.

[00:07:44]>> Yeah, everybody can hear you. Okay.

[00:07:47]>> Okay.

[00:07:49]Let's see here.

[00:07:51]And then we're also gonna test things with.

[00:07:58]Let's see if we can see if Resolve is kind of happy. Is Resolve behaving itself? It'll take a second for y'all to let me know, but I'm I'm not going to dive into demoing stuff if we're not even seeing Resolve. As I'm flipping through things, are you guys seeing this in somewhat reasonable real time?

[00:08:16]>> It just popped up.

[00:08:18]>> Okay. Yeah. Let me know how >> that's reason. It's relative. Yeah, it's it's reasonable.

[00:08:24]Okay, >> it's just like you.

[00:08:27]>> Okay, well then that's good news. You know what? Me at five frames a second, that's probably all any of us really need. That's that's that's sufficient.

[00:08:34]All right, so let's talk about some grading tree, order of ops type of stuff. So question is, if we're taking a look at my base tree, by the way, I'm I'm in Resolve 21. I haven't done anything in resolve 21, so I'm going to have to go to someone else's grade school to have much insight to offer you guys on 21, but I am in it if we want to talk about anything in there for the rest of this session. I've got my base node tree in place. And so the question for like order of ops for compass, which is a tool for mapping in and out of film negative space essentially. Oh, and Chad, do me a favor in between uh prompts. Um, mute your mic for me. I'm I'm hearing just a little bit of like ambience.

[00:09:09]>> Yeah, no worries. One sec.

[00:09:10]>> Thanks, buddy. Um, so if we take a look at compass, if that power grid ever loads up for us, compass allows us to move in and out of film negative space. There's a number of different things we can use that for, but one of them is to move into film negative space to get true photochemical printer points and like film grading type of adjustments and then to move back out of it. So the fundamental question that is being brought up is where should we place that in our node tree? I'll show you the way that I like to do it. I would do it right here in this area. So there's just going to be this sandwich where I can do my photochemical adjustments, but I'm still going to use linear operations and I'm still going to want to do those earlier in the node tree. Not going to make too much of a meal out of that beyond that just because this is a tool that not all of us are using or have. But that would be kind of my guidance on it and the philosophy behind it for those of you guys who are interested in that question or interested in grabbing the tool.

[00:10:02]That's kind of how I would do it. Like I would do it medium I would do it like mid to later in the note stack as opposed to first thing just because there are linear operations that you really can't do once you've gone into or through compass that I want to do uh kind of right off the jump. So that's how I would think about that.

[00:10:26]What else we got? Don't forget to unmute if you're talking right now.

[00:10:36]No, the the zoom my zoom window like hit itself as I could find it. Um, let's see here. What do we got, guys?

[00:10:46]Uh, let's see.

[00:10:53]If you got a question, please uh tag me in the questions so I can actually uh find you because there's a lot of chat here.

[00:11:00]>> A lot of chitchat.

[00:11:01]>> Um a lot of chitchat.

[00:11:05]Do you watch movies in the grading environment?

[00:11:10]>> Yeah, guys. I mean, like short version of it there. Do you want to know that the the my two favorite things to do in my grading suite? taking naps on the client's couch. Don't tell my clients.

[00:11:21]And watching movies in there. It's awesome. Yeah, 100%. Oh, and in the uh in in our DI theater, like that one takes a little more coordination and I usually have to like convince the people my my partners. I'm like, "Hey, don't do anything else. Can I book the room for two hours so I can watch, you know, Die Hard or whatever it is." That one takes a little bit more coordination, but it it's definitely been done and it's definitely awesome. So, >> that's a Christmas movie.

[00:11:44]>> The ultimate Christmas movie. My wife fights me on that every year, but it does take place at Christmas.

[00:11:49]>> Okay, so was reference reference created using Genesis?

[00:11:56]>> Good question. No, reference wasn't created using Genesis, although it's very heavily inspired by uh when we released Genesis in August of last year, I went from I was just talking about this uh with my colorist ink crew earlier uh this morning. I went from being like, "Oh, this is awesome for like as an end user that is being like, "Oh, I love using Genesis." Obviously for any film emulation needs, that's what I will use it for going forward.

[00:12:22]But more and more even after we release Genesis, I realize like this is just great. Like there's so much stuff that just works in here. Like here's the way that I would think about this for you guys that this is less to do with Genesis and more about philosophy of like color grading and image rendering.

[00:12:37]when the way of reproducing motion images for movies was film, nobody was trying to do film emulation, right? They were like, "No, this is just what we use to get a good image on the screen." So, to me, that's what I realized Genesis was is like, "Oh, this isn't like a way of like quoting or riffing or like making a cover band for a thing. This is just a way of getting an image on a screen that looks really, really good and is incredibly versatile." So, that's how I think about Genesis. And at the same time, there are aspects of film rendering that are just kind of baked in and are what they are that are not consistent with some of the modern demands or needs of particular color grading pipelines or particular projects. So reference was really all about like, hey, how could I take the same concept like down to the color science that drives Genesis, how could I take like all the goodness that I love about Genesis and cook it into something that is a little more neutral and has like ticks a few more of the boxes that are sometimes important to us as like modern uh color graders. namely having a truly neutral tone scale, meaning like there's no tint in the shadows or the mid tones of the highlights and uh having a black that touches all the way to zero and a white that touches all the way to the peak at all times. So the idea was like, hey, can we, you know, like create something that has the essence of the film rendering that I love from Genesis that also checks those boxes and leaves us with a very beautiful rendering of image, but something that is uh got a little bit more room to build a look underneath or to just do something a bit more transparent if that's truly the intent.

[00:14:12]So that's kind of the idea of the way we built reference. reference, by the way, for anyone who doesn't know, is my uh viewing transform to get you from Da Vinci wide gamut out to Rec 709? It's kind of like a color space transform, except it's beautiful.

[00:14:26]>> All right, we got some good questions coming in. Do you always color uh balance correct with the look L already applied?

[00:14:36]>> Oh, I love this. I'm so glad this came up. I feel like this came up in a couple comments this week and I did not get to all of them. So, short answer, yes, 100% all of the time. If you'd like to know why, stick around. We're gonna dive into it right now. So, here's the thing, you guys. This is one of the benefits for me of like my color science journey and really trying to understand specifically the way film colors were achieved and can be achieved. and just in general wanting to be a student and an expert in the way narrative like motion images are rendered pleasingly if that makes sense.

[00:15:11]Like how was that done 10 years ago? How was that done today? How was that done 50 years ago? One of the big benefits to my practice and to my sort of like philosophy of grading that came from that is realizing there is absolutely no neutral technically accurate like transparent way of putting the data from the original scene onto a screen. You follow what I'm saying? There is no way to one to one just put what the camera saw onto the display. And I'm actually going to give you a example of what I mean here. So, I'm going to use a different DRT than reference. I'm going to use a DRT called open DRT for a moment here. Again, if my Power Grades will ever load up, I think that's directly to do with the terrible internet service that I'm getting in the IATE old man call that I'm going to be making when we wrap this live stream. Do you know who I am? I'm just kidding.

[00:16:07]That's not that's that's not my vibe, you guys. I'm I'm I'm not that guy. I'm not that guy demanding service at the restaurant and demanding to be sat down.

[00:16:16]This is ridiculous. I refuse to wait. I promise. That's more like my my wife is usually the bad cop in those situations.

[00:16:23]Anyway, uh if we take a look at OpenDRT, I'm going to go to a particular mode of OpenDRT. Open DRT, by the way, this is a really, really good DRT. It's also free.

[00:16:33]You can go and download this from my friend Jed's GitHub. It's awesome. Uh it's right up there with uh my friend Juan's $24.99. These are both free >> Smith kills it.

[00:16:42]>> Yeah, absolutely kills it. These are both free DRTs and you can get them uh uh from the respective GitHubs for these creators and they're like they like crush 10 times out of 10 are better than uh built-in resolve CSTs. So if you don't do anything else, go grab these or again go grab reference. Uh that's a free download as well. That is what I was starting with today. But I'm using uh OpenDRT specifically because it's got these different modes within it. And there's a mode that is highly I'm going to use a fancy word for you guys, colorometric. Okay. Uh that's actually this base mode right here. So this right here is pretty darn close to a true colorometric as close as possible to a colorometric rendering of this image. Meaning colorometric meaning measurement of color. Right? So, if we were to actually go through the academic exercise of like, hey, let's measure what those colors were on set and reproduce them as accurately as possible so that when I measure the calibrated monitor and the image coming off of it, I'm seeing the same colors. That would be like the intent and the goal of a colometric rendering. As I said before, that's actually very hard to do and it really can't be done except in like certain core areas of the image. But the bigger point is, even if you do it, this is what you get, y'all. Like, let me do a little layer mixer here to give you a comparison. I'm just going to give you open DRT in this mode. Now, again, I want to emphasize this is a mode of Open DRT that actually is useful for certain things. So, this is in no way knocking on Open DRT. I really like Open DRT in part because it helps me to explore and demonstrate concepts like the one we're talking about today. So this is not a matter of me saying hey look how much better my thing is than than this thing this thing is awesome. Uh it's also helping us to illustrate the point that I want to make. So if I compare reference which is again designed to be opinionated but also designed to give you a sensible sound base image on the screen. Here's reference. Here's the open DRT rendering. Like it's not just like a contrast thing even though that's kind of what grabs our eyes. Like the skin tone and like the way that colors render is totally different. like skin tone is going to tend to go very kind of like yellowy greeny and just have all these kind of problems. So I'm making a big meal out of it but essentially there is no way to put an image on the screen without imposing an opinion is my point that I want to make here. Once we accept that and believe me that is a hard pill to swallow. If you get a little bit closer to accepting hey there is no technically neutral just sound baseline way of putting the image on the screen.

[00:19:16]If you just get a little bit closer to that in the course of this session today, that was worth your time in my opinion. But once you do accept it, the natural thing that comes after that is like, okay, well then all that matters is the preferential rendering that I choose and that the other stakeholders for this project choose with me. Once we've chosen that, how do we know whether the shot is balanced? Whether it looks balanced through that preferred rendering. Does that make sense? like the fact that it looks balanced through some other rendering or through an a objective or neutral rendering. What difference does that make? That's not the way that we're going to see it. So, we want to have that full context and we want to embrace the creative lat latitude and creative responsibility that we have to put a pleasing image on the screen with our pipeline and then to do our balancing and shaping of the individual shots underneath that. Does that make sense you guys? This is like a core part of my philosophy. It's slippery at first. It can even be counterintuitive, but it's super important to grasp. And it's actually the secret to like if you go look at go pick any movie that came out like longer ago than around 20 years ago, that would that means it was shot on film and printed on film. That's the exact system, the exact practice that was used to put those images on the screen. No color timer in the history of color timer has has ever gone. Well, yeah, it looks good to add a couple points of red to it when it's going onto this print.

[00:20:43]But what about the neutral rendering of the image? That would be nonsense, right? Because there's no such thing as a neutral rendering of the image.

[00:20:48]There's only what the film print was going to do, which is all they care about and how their work is going to interact with the behavior of that film print to get a pleasing image on the screen. So yes, always 100% you want to build your look, your macrolevel pipeline and creative transforms before you do any color grading and balancing and things like that.

[00:21:09]>> Okay, we got a lot of questions coming in. So um question Colin before told that he doesn't like to adjust luma channels separately. What is the reason for that?

[00:21:20]How would you access bleachy looks in that case?

[00:21:25]>> Yeah, great question. So the the observation is that I've said in the past, I typically don't like to adjust luminance channels when I'm grading. So there's lots of different forms that that can take. But if we were to look at, you know, go into uh our primaries here, let's go to the timeline level, the clip level rather. And uh you know, if we look at, let's just go to our custom curves. We can do it over here.

[00:21:49]And I'm going to adjust my Y only of my curve. And let's actually go to a preset.

[00:21:56]By the way, as I'm waiting for this to load, I'll point out to you guys any of these power grades and lots and things that you see me always messing around with. A lot of these I've made available here on YouTube. Any of them that I haven't, they are available inside of Colorist Stink, which I mentioned at the beginning of this call, it's the place where I am uh gathering up all of these like part of the reason that they're that they're not always available everywhere is because I'm always making more of them and I'm always iterating on them. And I've now got from my selfish perspective a place where I can just drop that stuff and share it immediately without having to like strategize or like figure out how to deploy it. It just is live for everybody's in my community. Anyway, if we take a look at my contrast curve that I'm going to start shaping here, I'm doing a Y only adjustment on this. Oh, and I actually need to turn my loom mix, which is defaulted to zero, back to 100 for now.

[00:22:41]But if we look at making this adjustment, this is an example of adjustments that I typically don't like to make. And also an example of the bleachy type of look that is being asked about like oh well how would you do like a bleachy type of look if you are not doing Y only adjustments. So here's what I would say. It's not that I never do Y only adjustments but I recognize them for what they are which is a I'm going to put it to you guys this way. We're going to get kind of deep for a second here. We're going to get like lightly color sciency here in grade school.

[00:23:13]Everybody buckle in. So the ground truth of images today captured and reproduced images is RGB. What do I mean when I say that? What I mean is that camera sensors see certain amounts of red, green and blue. That is what gets captured by the original sensor and stored as an image that we then manipulate. Right? What does our display ultimately put out?

[00:23:34]Red, green, and blue. Right? So the end toend journey is three pipes, a red, a green, and a blue. And along the way, we manipulate those channels to get what we're looking for. You with me so far?

[00:23:46]It's pretty basic stuff. Hopefully, there's a point in the color grading process where to achieve certain effects, we have to do what's called derive channels like we have to infer channels that don't exist. So, the easiest one for me to grasp there, do you guys know that cameras don't see saturation?

[00:24:06]They don't. Neither do displays.

[00:24:08]Saturation is a derived or an inferred element. They're not it is not native to the RGB of the camera or the RGB of the display. We have to use some arbitrary mathematical formula that says hey there's this thing called saturation and the way that you define it based on the red green and blue the value of a red green and blue pixel is as follows. And you use some math to get there. Okay.

[00:24:32]The other thing that is in that exact same category is to say, "Hey, there's this thing called luminance." And the way that you define luminance is by applying the following math to the red, the green, and the blue of the input pixel. Okay? So, luminance is a derived or an inferred element. And so, there's there's two things that come about here.

[00:24:51]Number one, when we are using those derived or inferred elements, we are a little bit further away from the ground truth of the image. And there's the potential for weird stuff to happen because we're making things up, right?

[00:25:02]And then we're doing stuff to that madeup thing and then we're trying to turn that madeup thing back into what it originally was. It is a can be a bumpy ride for the image, right? It doesn't have to be like we spoiler alert, I use saturation in my color grading all the time. Like I'm not saying you can never do this. Like you it's a big part of color grading. But my preference whether it's luminance channels or saturation, those are the two easiest examples. is if I don't have to do that, I'd rather not, right? And I don't see it as as fundamental to the color grading process as RGB because RGB is like definitionally more fundamental to the process. And we already talked about color timers earlier today. The color timers, the people who have mastered some of my favorite motion images of all time, most of my favorite motion images of all time, the parameters that they had to work with were red, green, and blue. That was it. They couldn't work with contrast or luminance or saturation or any other derived element. I'll leave contrast off that list. That's a little more complicated, but certainly luminance and saturation as we've been talking about. They didn't have control over those things and they were able to achieve incredible images. So for me, I go to that second rung of like, all right, we just need to trim saturation here or like I need to achieve a bleach bypass look as the original question was talking about. That's a point where I'm like, all right, I need that. There's no way for me to get it with my ground RGB.

[00:26:24]So, I'm going to make the conscious choice to climb the ladder to that next wrong of an inferred element of luminance or saturation. But if I don't have to do that, then put it this way, no one has ever broken an image by just sticking with red, green, and blue. It's just red, green, and blue. That's why there's never been an artifact in a film print. It just is red, green, and blue channels. Well, that's not the only reason, but it's a good reason for it.

[00:26:45]So that's my philosophy there and why I try to avoid luminance and saturation when I can and then consciously use them when I've decided, hey, I need this and simpler tools and solutions won't do.

[00:26:58]Hope that helps.

[00:27:02]Okay, this one might be a little bit deep for today, but um someone's asking uh regarding insight on export settings, regarding shift and you know preferences and settings.

[00:27:16]>> Yeah. Okay. I mean there's a lot to unpack there. So, I'm going to give you guys the principles, and this is something that uh we actually do have like dedicated videos coming out on very soon, and we can talk more about in grade school in like specific aspects.

[00:27:30]Cuz here's the thing, what you're talking about, I'm going to assume, is some variation on what is often called the gamma shift problem. Here's the problem with the gamma shift problem.

[00:27:39]It's not a problem. It's like 20 problems. It's kind of like saying, you know, like I I I sort of laugh every time I see these big forum posts about stuff like this because I'm like, "Oh, that this is the equivalent of saying someone," when when are we going to fix that stupid car problem? Cars run out of gas and cars need their oil changed and tires go flat and sometimes the windshields break and sometimes the alignment is off and sometimes the electrical system doesn't work. When is someone going to fix this darn car problem? Well, that's not a problem.

[00:28:11]That's a whole network of problems that may be related or may be completely independent and have all kinds of different variables that go into them.

[00:28:18]Right? So the issue of the fundamental issue, the thing we can agree on that is what we are all talking about when we say that is what I saw when I said this is the grade is not what I'm now seeing when others look at it. Right? That's like the only sort of like loose definition that we can agree on when we talk about export shift or gamma shift.

[00:28:39]Right? But even within that, there's a million reasons for that, right? Like there's that and you really have to tackle it systematically and one by one.

[00:28:49]So here's here would be like the hard prescription that I would uh give you guys or the tough love that I would give. Most of the time when this comes up, the so-called gamma shift issue and people are having problems with it, they could solve it very easily, at least like they could they could put the biggest dent in it by far by simply having an accurate grading setup. And like I'm not an elitist. I recognize not all of us are professional colorists.

[00:29:13]Not all of us have the budget to go out and buy the latest and greatest and the best uh equipment. But it's also like it's sort it's sort of like I'm again going to belabor my car analogy because I think it's good in this case. It's like, "Oh man, I just can't seem to solve this 200 mph problem. I just cannot get my car to go 200 mph." Like, but you are not driving the right car.

[00:29:31]Like, there are cars that go at 200 m an hour without problem. Like, you can go and buy them today and you will be driving away from that car dealership going 200 m an hour. Hopefully not.

[00:29:41]Nobody go drive a car 200 m hour. This is a very abstract example. You can go buy that car, right? But your car is not ever going to go 200 miles an hour.

[00:29:47]Doesn't mean you're supposed to buy a faster car. Doesn't mean you're a bad person because you have a slower car. It just is the reality of the equipment that you are using. Right? So if you are not seeing an accurate image when you are making the decision that this is your image, then your hopes of ever seeing that exact same kind of inaccurate everywhere else are basically zero. Does that make sense? So if you just start with that, now the thing is there is indeed a whole other galaxy of problems and issues and practices that can be put into place to help ensure that the image you saw on the calibrated display in the suite is indeed what gets seen by the greatest amount of people uh out in the wild. But it's got to start with that foundation. Otherwise, you're just chasing your tail. you are like there's nothing harder in imaging than saying hey I want to get the same type of inaccurate in a wide variety of inaccurate scenarios does that make sense like it would be easy to say hey I want to get the I want I I I want this to be wrong everywhere but if you're saying I want it to be the same kind of wrong which is really the argument there you're just going to run into problems so the solution here is to tackle it systematically start with getting a good reliable image when you are actually making the decision that this is the image and then you can work the problems from there and say okay why is this output deviating from the correct output that I saw when I graded the image why is this one doing it why is that one doing it and tackle them one by one and tackle them with priority because like we don't need every single possible viewing experience to be accurate we need the ones that most of the people who watch our thing are going to watch it on to be accurate hope that helps >> yeah totally I mean my philosophy is you calibrate calibrate for your target and then hit that target and you know and be as close as you can be. That's your best your best bet. Um >> 100%.

[00:31:38]>> All right. So the next one is um I think this might be a data levels issue, but he's like why is it when I watch films and analyze their waveforms or false color with omniscope the contrast is richer but way lower on the waveform.

[00:31:56]But my own work is stretched out higher.

[00:31:59]This sounds like a data problem.

[00:32:02]>> That could be. I love that you're doing that. Whomever is asking that question, that is one of the best things that you can do. I had a teacher in college who gave me some writing advice that I've transposed into color grading that has just served me really well for a long time. His advice is if for anyone who wanted to be a professional screenwriter at the time, which I was considering at that point, he's like, you only got to do three things. Write, rewrite, and watch movies like a writer. You can transpose that one to one to color grading and have a very good career just like repeating that same mantra. Grade grade grade some more and watch movies like a grader. So that's what you're doing by analyzing waveforms and scopes.

[00:32:36]And here's the thing, scopes are and like get getting insight into why you're seeing what you're looking at in that movie scopes versus in yours. It can drive you nuts, right? Like I know what you're talking about for sure. like maybe not the exact thing you're describing, but I I used to do this like before it was easy. I used to have like a whole like Blu-ray capture rig so that I could capture accurate stills from Blu-rays of movies and I would look at the scopes and be like, how are they getting that shape of scope and how does that correlate to the image that I'm seeing on the screen? And a lot of the times it was very non-intuitive and counterintuitive and a lot of the time it has to do with the processes that are being used more so than the actual color grading techniques that are being used on an individual shot. But I so I would say you can confirm that you are on the right track with stuff. It's pretty difficult to say, "Oh, because I'm seeing this in the scopes, I'm going to try to move my scopes to be more in alignment with that and hope to get closer visually to it."

[00:33:36]That can work, but it's not guaranteed to work by any stretch because this is a very simple like this is a single dimension. And it's a very dumbed down simplified rendering of what the image is actually doing. There's probably another car analogy in there, but I'll spare you guys since I feel like I've probably hit my aotment for the moment.

[00:33:57]But I would say like, you know, start with seeing if you can match that in.

[00:34:01]And then if you want the bonus, ask the question of like, okay, my scopes are sitting high, their scopes are sitting low. I like how their scopes are sitting. What is the ideal tool? what is the ideal lever for me to pull on to make my scope my scopes look more like theirs instead of just hey what's any old thing that I can do to make my scopes look like theirs because that you can do with you know like dumb example you can do that with offset you know but that's in most cases not going to work especially there's a lot of reasons why that won't work so like I would ask the question what tools will be to my best advantage to make my scope shape look a little bit more like the shape of uh the images that I'm liking seeing in these references.

[00:34:43]>> Yeah. And I think it's really important to analyze the scene because a lot of people look at like, you know, the waveforms and try to push things a certain way, but it's not always necessarily applicable to the scene in question, you know? So, don't like I I always preach trust your scopes, but like you got to also be careful to be beware, you know, aware of what the scene in front of you actually is because if you overstretch your highlights and it doesn't make sense, then you're actually doing more damage than good. So, that's my philosophy. Um, how to be organized with Genesis if you have series of 10 episodes, for example, like grabbing stills. H you know what? I want to kick the can on that one just because that's only going to apply to my Genesis users and I want everybody here to benefit from the conversation. So, let's find another one.

[00:35:36]>> Okay, cool. Um, let's see here. If a good DRT can change radically the perceptual quality of a final image, do you think the input transformed from a camera's color space into DWG can have a similar impact?

[00:35:52]>> Oh man, we're getting to get juicy today. Let's chop that up for sure.

[00:35:58]Okay. So, I'm going to give you guys the overview and then I'll give you the why. So, the short version of it, I would say in theory, yes, in theory, the way you map from any one color space into another color space can, you know, there's like different ways you can do that supposedly. And certainly when it comes to like, well, I'll leave it at that for now. In practice, no. We have a very well- definfined and very simple way that we move in between color spaces specifically between scene referred color spaces meaning camera color spaces and working color spaces. Color spaces where we have not yet accommodated for the display. That's what I mean when I say scene referred. That's scene referred as opposed to display referred where you're typically going to have tone mapping and gamut compression and essentially an image that has been shaped to fit into the boundaries of what the display is physically capable of doing. If we're in a scene referred space like when I map this image from area wide gamut 4 into da Vinci wide gamut intermediate these are both abstractions. Neither of these need to be represented on a screen quite yet. So I don't need to manipulate or distort or bend the signal in order to accommodate physical limitations of a display. I'm just in some abstract land where numbers mean a certain thing. Okay. So the way that we do that, which is not perfect, but is uh probably the best way, the most principled way to do it, the way that it's done in literally any color management framework is you strip off the log encoding, that would be log C4 to get back to linear. You apply a 3x3 matrix, which is essentially what we have in our RGB mixer, and then you apply a log encoding to your destination space. Okay, that's it. That's all that we do. So in that sense, there's really no fancier option to avail ourselves of that is going to maintain the linearity of the original image. So remember, as you guys know, we already talked about this today. I like to balance my image in linear and I like to set my exposure in linear, meaning that I want the original red, green, blue ratios of my source image to be preserved.

[00:38:01]If I was using any more complicated of a vehicle to move from area log C4 in this case into Da Vinci intermediate, I couldn't do what I'm doing right now, which is manipulating the linear ratios of red, green, and blue as the camera saw them. What I'm doing right now, my mental model is essentially reensitizing the camera to the light used in the scene. Either for balance purposes or for explicit creative purposes where I'm like, hey, I want this thing to be much more sensitive to cool light than to warm light. So that we end up with a rendering like this. It can go either way, right? So in that sense, if I want to maintain that linearity that I'm talking about, there's really nothing else that we can or should do for input mapping stuff other than stripping off log encodings, applying a linear 3x3 matrix, and uh re-encoding back to log.

[00:38:51]Although in this case, it's that that log encoding is being stripped back off again somewhat confusingly. So I'm adjusting original linear ratios of red to green to blue just with a matrix applied. So the intuition is a good one and it's a good question to ask like oh if there's all this money to be made with the way that we put the image on the screen are there similar opportunities and pitfalls with the way that we are mapping our camera into our working color space. And here's the thing, y'all. Like, I'll put a bow on this by saying like, if you go like Google around, you'll find people who are selling LUTs that are like, "Hey, this is a lot that's going to map you from like whatever, like your Blackmagic camera color space into your working color space better than a color space transform." Like, there are products out there that claim to do that. The problem is definitionally, those products are not maintaining the linearity of the original signal, which creates its own problems. Even if there are visual benefits, you're better off getting those benefits later. and by other means and instead prioritizing maintaining the linearity of your signal so that when I do stuff like adjust exposure or when I adjust red to green to blue it's clean and I literally am uh working with like the as close to the physical properties of the scene as possible at least at moments when I need to like there's going to be other times pretty quickly in the color gra I'm like all right I don't care about that so much anymore like I don't care about when I go in and I draw this curve I don't really care about scene linearity. I'm just trying to make something that looks good, you know, like that's kind of nice. Doesn't matter to me that it's distorting like this is explicitly distorting the image.

[00:40:23]That's fine with me because that's what I mean to do. But I do need to have that opportunity before this happens to manipulate the original linear ratios.

[00:40:31]At least that's important to me in my practice.

[00:40:37]Okay, so let's talk about highlights.

[00:40:39]How to control highlights? I always need to use one to two nodes to get back details when it's very sunny outside, but after adjusting the highlights, it doesn't look like a very sunny day.

[00:40:51]>> H re read the the characterization of the problem again. I always need to use one to two nodes. Read that again.

[00:41:00]>> Yeah, it sounds like um I always need to use one to two nodes to get back the details. So, he's it sounds like he's looking for like highlight compression and um recovery and uh when it's very sunny outside, but after adjusting it doesn't look like it's very sunny, like a very sunny day.

[00:41:18]>> Got it. Okay, cool. Yeah, I mean the guys this is awesome to talk about and these are some of the core things that wherever you are today in your color grading journey, you've either had the chance to explore and to find your feel and your taste for these things or that's a chapter that you've yet to get into in your color grading journey that you're going to have a lot of fun with.

[00:41:38]The first thing that I would recommend as like a broader solution to your problem, it's going to have more to do with your eye than with your hands. What I would do is do uh what one of our other uh questioners, one of our other grade schoolers today was already uh talking about, which is to watch reference movies and go look at the way highlights are shaped and ask yourself the question specifically like go watch a Roger Deakons movie. Go watch a movie that is really artfully shot that you love. Like I I gave a reference just now that I love that maybe isn't the one for you. Go find a movie that where you really love the way the visuals are achieved. And specifically look at highlights. look at things like, you know, what's out what's on the outside of a window when day interiors are happening and ask the question, how much am I seeing versus how much is just being sort of rolled and compressed into this nice kind of creamy matte thing that I know is bright but doesn't necessarily carry a ton of information.

[00:42:33]The reason I bring this up is I'm going to give you I'm going to inject some of my tastes that you can choose to take or leave. The reason I bring this up is I think highlight recovery as a concept has become kind of an odd fetish in modern color grading because sometimes it is possible. But like my preference generally, let me see if we can find a great example of this.

[00:42:55]I mean, this one's going to be kind of extreme, but it's a good one to look at.

[00:43:00]Like if we spin this negative, let's just straight up look at what's in the negative.

[00:43:05]Look how much detail is in there, you guys.

[00:43:09]So, it's blurred because this is it's like out of focus and they've got these like uh you know like glass bricks and stuff happening, but there's lots of detail that was captured there. This is something that drove me crazy in my color grading practice for like a long time like long after I started working professionally. I could never answer the question like wait there's that's there.

[00:43:29]It's right there. I can see that the camera captured it. I just want to see it represented. I just want to find a way to bring it into range so that I can see it along with a healthy exposure of my subject and it just like why can't I seem to accomplish that? I'll give you guys a good example. Like you want to see a good counter example, the thing that like the the version of uh doing this that I hate, go look at real estate photography. Like go find any real estate website. go on like I don't know like compass.com or you know whatever like like Trullia or whatever and go look at the home interiors specifically anytime you see a home interior you're going to see somehow the inside of the house is really well exposed but also it's a perfect bluebird day outside and the exposure level doesn't seem to be that much different than the inside of the house and if you feel as I do it looks really really weird even though mechanically the box is being checked it looks very odd. So, the reason I'm bringing this up to you right now is because it it kind of starts with like you need to know what you're trying to accomplish and know that it's possible and know that it's desirable cuz like there is a way for me to slice into this image and pull this so that it's alongside this. But what it's going to do is feel exactly like what you talked about in your question. It no longer feels like I forget the exact expression. No longer feels like daylight essentially because our eye knows well that's sort of odd. I know that ambient daylight and you know like whatever a a you know like exterior landscape out this window. I know visually that that is many many many many stops hotter than the exposure on a face inside. And if you try to smash those things together in your color grade it just feels weird. People tend to think it's like oh this is what HDR will solve. like no like HDR can solve it by reproducing that stuff uh without it needing to get dimmer but then it just hurts your eyes you know so it's it's kind of another problem but like point being I would start by really refining your eye and figuring out what you want to look at and then let me put a little bit of a practical kind of button on that and let's use this as our example you want to know like the the the practical maneuvers of it you guys know my expression god is a curve fundamentally what you are talking about doing once you've sort of like accounted for taste and target and what you actually want to see, what you expect to see and and sort of what your aesthetic is. The mechanics of implementing it are very very simple and resolve actually and like the most of the discussion around this makes it way more complicated than it needs to be. God is a curve. What do I mean when I say that?

[00:46:05]What I mean, let me go back to my I don't even need my anchor right now.

[00:46:09]Really, I'm going to pick an arbitrary anchor point here on this custom curves.

[00:46:14]By the way, this isn't the actual way that I would do this. I have custom tools for it, but these custom tools are fundamentally doing the same thing, just a little bit better and faster. I'm going to look at bringing in my highlights while not having unwanted influence on the darker areas of the image. So, if I let me get rid of this for now, I'm holding option to drag this control point and I'm basically looking at like, okay, so now you can see I'm having tons of effect on my subject. That's not really my intent. I want her to stay more or less the same. And I'm going to do another control point here to ensure that this is kind of locking and nothing off here to the left is actually changing.

[00:46:51]And so like this is how I would manage those highlights.

[00:46:59]And what's fun about this, if you try this, if this is your first time trying it, more than likely you're going to go like, gosh, that's hard. That feels weird. It's really easy for me for me to make adjustments that don't look right.

[00:47:11]like, yeah, but at least you're getting the true interface of what you're doing, which is shaping the shoulder of the shot. This is the only thing that gives you that feedback as opposed to a highlight knob or a qualifier, which are more abstract and kind of harder to understand what you're doing. This is what we're doing. And the degrees of control are how long do you want this segment to be before you force things to behave linearly?

[00:47:35]And how much are you dropping this? And how much are you compressing this here?

[00:47:41]Does that make sense? Just as a geometry problem to you guys. I'm throwing a lot at you right there. But this is what it is to shape a highlight. Like that's why I love curves because they are both a tool and a uh visual aid. This is a way of achieving that. But it also is what you are. It is it is the result itself if that makes sense. You are shaping this in this way. So that's how I would think about it. and it will self-limit you and and let you know if you're like, "Oh, yeah, but I would like to bring that in even more, you know, to use my example, like I'd like to get to somewhere like this." It will self-limit you and say, "Yeah, but that looks weird. I can't do that without giving you an impossibly tight curve or whatever it is." All right, that's plenty on that. Hope that helps.

[00:48:24]>> Yeah, 100%. I was also in the same boat.

[00:48:27]I used to always like get really wrapped up on the highlights and there's a point where you just have to look at the scene and be like just let it go. Like it it is what it is. Like I've watched movies like Blackhawk Down or The Long Walk recently and the sky is just white, you know, and there's just it is what it is.

[00:48:45]Um so sometimes you just got to not get too wrapped up in it in my opinion. But uh next is uh why do you use uh Da Vinci wide gamut versus aces?

[00:48:57]Oh, good question. So, here's the real reason. It's not a perfect reason, by the way. Uh, but the actual reason for this, let me give you a a practical example right now. Remember how we were just talking a moment ago about wanting to um about how we handle our input mapping, put it that way. Let's switch to an ASUS pipeline for a second here.

[00:49:20]I'm going to grab a still of this and then I'm going to turn off everything for a moment.

[00:49:26]And we're just going to try something here.

[00:49:30]Let me turn whatever's going on here off. We'll turn this off. And we're going to go to this looks wacky on the screen because it's a linear image. I'm going to do an ASUS transform. And we're going to say no input transform because that means AC is linear. That's what this image is. And we're going to go to ASUS CCT. And now I'm going to go over here to the clip level and we'll do another ASUS transform from ASUS CCT to rec 709.

[00:49:58]Now they've gotten better, but I've still got artifacts here in this cyan area. See what we've got going on here compared to what we had before. Those kind of harsh biting edges. The only way for me to solve that in this pipeline would be to do some nonlinear stuff right here at my input mapping because ASUS CCT is a physically bounded working space. Like not an improbable one. It's basically RC 2020 for any of the geeks among us, but it's physically bounded in a way that most camera color spaces blow way past. which means that if you're using it right out the gate, you're going to have to handle those edges, handle those negative values from mapping when you map from camera color space into working color space. And I don't like to do that. I don't want to disturb the linearity of the image before I've even had the chance to touch it. I want to wait and have the opportunity to manipulate the linear state of image and then touch it uh after that point. Another way to put it would be a working space that requires these kinds of compensations and nonlinear operations. It's not a very good working space because that's literally its one job is to be a big enough container for whatever needs to go inside of it. The fundamental like easiest uh sort of like diagnosis to put on why this is happening is because ASUS CCT is not a big enough container for in this case as linear or whatever the original source camera data was. So that's why I prefer not to work in aces as a working space uh because it is uh too small to perform well under all circumstances. I mean look up here as well. You can see other kind of icky splatty stuff going on with this image.

[00:51:37]Again, you could solve that, but you'd have to do it by uh making a compromise that I don't want to.

[00:51:50]Okay. I'm not sure. Are you a a tablet like a pen tablet user?

[00:51:54]>> I am. Yeah.

[00:51:56]>> So, somebody's asking like what's your recommendation for first time uh pen tablet user? Uh buy buy like buying solution.

[00:52:05]>> Oh yeah. I mean like honestly go grab whatever like let me see if you can see my can you see my screen there? I'll I'll it doesn't matter. It's a tiny tablet. It's like the size of it's like like smaller than a piece of paper. uh this tablet like it's maybe half a size of piece of paper. I think it's the smallest tablet.

[00:52:25]It's the Inos Pro. Yeah, it's Waycom Inos Pro. I love it. It's a great tablet. I'm not particularly particular about the tablet. Like I like Waycom tablets. They're the only ones that I've ever used. I think the main thing is you just have to spend a week being uncomfortable and then after that you'll never go back. Or at least that how it was for me. But yeah, Waycom Inos Pro works great.

[00:52:48]>> Yeah, they definitely takes some getting used to. I use XP pen on PC and I quite like it but that's uh yeah >> again it's neither here nor there per se. So what are the challenges working on an animated show?

[00:53:04]>> Oh that's a fun one. So the I've done one animated show and we had a lot of fun with it. It was a stop motion animated show that we were doing an HDR finish for and they were shooting bracketed exposures of it. And there's lots of stuff that we had to do to get that workflow working. But to I'll I'll put it to you guys simply. This kind of connects up to, you know, we we've talked a lot about linearity and when we break it and when we uh, you know, really prioritize keeping it today. That was a scenario where really fundamentally the issue was like it was dressed up in all kinds of different ways and there was different language and like uh things we had to talk through as a team to figure it out, but fundamentally the challenge was that you couldn't really get linear data off of these cameras and you had to merge two nonlinear bracketed exposures together for every shot. And then you had to have that as your starting point for your color grade. So even a basic thing of like, okay, I've got two stacked exposures, neither of which are linear and which have been merged together in a kind of handwavy way. Take that as your starting point. And even a simple thing like, okay, let's let's like shut down by half a stop. That actually becomes non-trivial because the source image doesn't really have a we don't know what its tonal distribution is, if that makes sense. Meaning like is that thing that right now is twice as dark as the subject. Was it actually twice as dark as the subject or was it 1.7 times as dark as the subject or half as dark as the subject? It's really hard to uh know. It's really really hard to color grade from that vantage. Now I know that's actually a pretty common thing at least with stop motion because the cameras that are used for that don't tend to be modern cinema cameras with uh where you can easily access the linear data. Um, but that to me was like the beginning and end of it from there.

[00:55:00]Everything else from a creative point of view felt pretty easy. I was like, "Cool. This is like a completely different pallet and world and feeling than I'm used to and creatively responding to it." That didn't feel hard. And there didn't really seem to be any other technical challenges other than how do I get this darn linear data or get somewhat of a linear representation of this data so that I can work with it in a principled way in my first couple of move which moves which are like the most critical moves in a color grade. So that was that that's literally the only animation project that I can speak to firsthand but that was my experience with it.

[00:55:39]Okay, this will probably have to be our last one for the day. Um, how are you dealing with Blackmagic RAW footage in the RAW tab? Do you use highlight recovery or gamut compression in the RAW tab? And how do you input map the footage?

[00:55:57]>> Yeah. Yeah, great question to to chop up. So, I'm going to uh put it to you guys to at least answer to yourselves, even though I might not uh be able to wait for your answer in the chat while the the latency catches up. But if we look at our Blackmagic RAW here, highlight recovery and gamut compression. Let me ask you this question. Do you think highlight recovery and gamut compression, what do you think the relationship of these adjustments is to a true linear state of original image? As we've talked about today, remember like literally in that last example, I'm like all I want is the linear sensor data from both of those bracketed exposures. I can blend them together in a principled way or in general, all I want is the linear data off of the camera so that I can make my first few critical exposure and balance moves. What do you think is going to be done to that data that I otherwise could access in a true linear fashion with these?

[00:56:48]It's going to distort it in some way, right? It's going to distort the color boundaries or it's going to distur distort the tonal boundaries when I adjust these. Now, that doesn't mean I would never or have never used these. I have 100% had Blackmagic footage from like, hey, something's like broken up there and it's like in the neg like before it gets touched in the node tree, something's acting weird. And I've been able to fix it by turning on one or both of these. So, I'll use it as uh you know, like a tool when I need to. But in general, I like to turn it off because these things are presumably going to apply nonlinear distortions to the image in an effort to control very bright stuff or very saturated stuff or very bright saturated stuff. So I typically am going to leave those off. And you can see my settings here. I just debearer directly into Da Vinci wide gamma intermediate because and you can test this out if you want to like go do the math debeayering into blackmagic's like native format of like you know blackmagic design and you know whatever like I haven't even done this in a while yeah like blackmagic design film you can do it this way and then do a color space transform from this into Da Vinci wide gamut but it's going to be forensically identical not close like forensically identical and not by coincidence because in both cases the exact same process is being applied So, I figure, hey, I'll save myself the CST and just have it land directly where I want. So, that's how I work with Blackmagic footage. And then other than that, um, I will try to honor the white balance as shot. L source is not going to matter. Uh, and camera metadata, I will use color temp and usually turn uh tint off, I think. Although, it looks like I don't even have that that option right now. Maybe because I'm not on a Blackmagic shot. No, that doesn't make sense. Yeah, looks like those are going to be left on no matter what I want. Uh, so that's the story there. But yeah, that's how I debayer that stuff and that way it just lands right in Da Vinci wide gamut and I can start working on it. Uh, okay you guys. Was that fun? Anybody learn anything today? Did my feed ever get better, Chad? My my video feed?

[00:58:56]>> Not really. No.

[00:58:57]>> No. But the but but the Resolve feed seemed to hold up. Okay.

[00:59:02]Yeah, it was still a little laggy, but I mean, you know, we'll get it sorted.

[00:59:07]>> Okay, sweet. Well, I hope regardless of uh technical issues, you guys were able to pick up something, learn something new, find out something cool. If you want to learn something new, pick up something cool, use premium tools, have an elite community around you every single day and get that 1% better that I was talking about at the beginning of this session. Go check out Colorist Ink.

[00:59:24]We'll drop one more chat and or one more link in the chat to it for you right now. Go check it out this weekend.

[00:59:29]pricing goes up forever on Sunday night.

[00:59:32]The price that it's at right now is the best price that it will ever be at uh ever again. Um so go check it out if you are interested in making those kinds of daily consistent gains that net out to huge gains over time and that allow you to operate at a truly pro level. It's a really cool group and we've got already a bunch of cool people in there. So I hope to see you in there if that's a fit for you. And if not, I will see you next week here on the channel and for our next grade school. Thanks y'all. Take care. It's It's awesome.