Grade School: Pro Colorist Film Look Q&A

are we live do I sound normal I really hope so that would that we had like a hi everyone good morning by the way um we had like a kind of weird grade school last week where like we had such good conversation at docking points that were completely uh fouled up by technical issues on this side so I'm really hopeful that all of our technical issues are laid to rest and that I sound beautiful and pristine and that my buddy Rafa sounds beautiful and pristine what's going on Rafa hello everyone I really tried to fix the audio for for the last week and it didn't work so hopefully at least people on each other say that the audio is perfect this week so it should be a no problem yeah no problem I'm very excited it was almost it's kind of like when you get sick it's almost worth it when you like get well again she's like oh my God I feel amazing that's how I feel right now having a show with clear audio and video um hi everybody I hope you're doing great I'm doing I'm doing really well uh it's been a fun week here at Colin Kelly color I'm getting my grade on I'm building Cool Tools I'm uh working on some uh changes in my personal life and in my time management uh that are uh really feeling like they're positive and helping me to be more efficient more effective more calm and spend more time with my family so I'm just feeling good on all cylinders and I'm excited to chop up today's topic about film emulation within resolve this is something that I get asked about all the time it gets lots of comments and questions here on the channel so glad we got to make a video on it and then we're going to get the chance to dive a little bit deeper into it today uh by the way I see some of my ogs some of my some of my Rags here on the channel anybody brand new anybody never been here for a grade school before put your hand up let me know if you come in later let yourself be known let us let us know that you're here for the first time we'd love to know we love uh getting uh new members of of the band uh here in the group so I don't know I feel like sometimes with grade school I have initial comments or remarks to share with you guys but I I don't uh I don't have any such comments today so if we got questions let's kick it off let's get this party started we have a few questions already uh but before that it seems that people cannot hear me so time to repeat questions okay that's let's see uh just say something else for me real quick let me check my levels over on this side let's try with this maybe yeah okay yeah I sure do see your levels and even when we Rafa and I did a quick test right before we went live and you sounded good so I'll do my best to repeat uh Rafa's question and and uh well well Soldier on all right let's get started with this one um from your last video you mentioned that once you mastered the look development part in result uh it's going to lead you out of result into foreign stranger territory could you briefly touch touch on uh where that leads can I touch on it yes can I briefly touch on it let's see I sure would like to um so here's the reality guys look Dev is like as most of y'all know look Dev is like a huge part of my professional practice these days and I'm so like happy about that like it's something I fell in love with very early in my career and it has indeed let me down some strange paths and uh like it's been a a slow and very non-linear learning process for me where it leads where the ideal place to go when you're sort of hitting the upper limits of what you can accomplish inside of resolve that's something without a clear answer unfortunately so just to give you some context I have in the past done courses on building show let's and Building looks within resolve where we basically take the type of material and techniques that we explored in this week's video and dive as deep as possible into that Corridor of techniques meaning look Dev that we can do with M resolve I've done entire courses on that I've done two entire courses like 16 hours a piece where we dive into using the fusion page of resolve and get into doing more complex customized things and we use dctls that I've built or provided we use the Custom Tool to apply our own math so that's a whole other Avenue of exploration if you're like my friend Steve Yedlin then you can do a really really Advanced look development in something like Nuke another compositing environment where you can really control and structure exactly what you want to see happen of course there's the aspect of coding that comes into play too and coding is highly related But ultimately a separate skill as well from look development so if that all sounds complicated none linear and not super clear as to where you would want to go when you hit the upper limits of what you can do within resolve you're right it's not very clear and it's pretty difficult and it's something that I'm actually really excited to be building a path for you guys to accomplishing in a more linear fashion right now so many of you guys know about the colorless career accelerator course that we began doing about a year ago that we're going to be doing again in the coming months that is really the introduction to building a professional color grading practice we're also working on a look development course that is going to be the follow-up to that course basically for alums for those of you who've gone through the colorist career accelerator there's going to be this second course that is available for more advanced look development but is going to try to keep it really simple and basically allow you to stay within resolve stay within the color page but use a custom set of tools that I've developed for doing really powerful but really intuitive look development without having to go and learn code or math or Fusion or nuke or any of the stuff that that I just rattled off so the path is there or it's being built anyway to getting to a really sophisticated look development practice without having to spend like the decade that I did uh doing it we've got our best men working on it and ladies working on it around the clock but that would be kind of the ideal answer is that there would be a framework in place that you could move through because otherwise for me it's been like hopelessly non-linear and exploratory and I'm not complaining I've had an absolute blast doing it at the same time I don't necessarily want to suggest to anybody here in grade school today go spend five six seven years and you'll start to get the hang of it just by beating your head against the brick wall I'd love to be able to offer you a more efficient and straightforward solution than the one that I uh the path that I took when I was figuring out how to do look development let's go to the female basically when building a palette a emulating print film um it's better to use infusion the color Matrix or uh just a hero versus Hue and if you choose the to use the color Matrix how can you use the power Matrix in a more optimal way oh great question so question is is it better when you're doing film emulation to use your Hue versus Hue versus an RGB Matrix so I don't think there's a single like correct answer to that I can tell you in my look Dev practice I use both tools I don't necessarily use the Hue versus Hue and resolve although sometimes I will but I will definitely use a hue versus Hue curve type of adjustment or an RGB mixer to accomplish the sort of rotation of Hues whether that's warming or cooling or getting my yellow greens more green or my uh you know like yellows more gold or whatever I'm trying to accomplish I'll try both methods depending on what I'm trying to accomplish here's what I would say if we kind of made a really brief survey of the strengths and weaknesses of each approach let's start with the RGB mixer the RGB mixer and let's like go to good old Isabella here the RGB mixer is not terribly intuitive so we talked about this even in the video this week we're like okay I want to warm my yellows that takes a little bit of thinking in the RGB mixer doesn't it say like well warming is going to come primarily from my red Channel and if I want to change my Hues I'm going to want to add some green into my red Channel because that is going to cause green stimuli to move further up in that direction right so that kind of takes some thinking it's not that like you can't think your way to a good solution there but it does take some thinking it's not the most intuitive adjustment in the world on the plus side the RGB mixer doesn't break it's like you I'm not going to say you you you could drive this hard enough that you could produce like some problematic results but in general because all you're really doing is applying gain to your red green and blue it's really really hard to get it to break or to get it to tear now if we contrast that with the Hue versus Hue it's kind of an opposite set of pros and and cons like if I look at Hue versus Hue like here's something I do this in session all the time when I'm color grading if I feel like oh what would we think if we warmed up the yellows I can tap my curves go to my Hue versus Hue grab my yellow curve and just rotate it a little bit and I'm done that's fast clean and very intuitive because I'm just saying yeah take my yellows and rotate them stand them up a little bit more then when they're aiming kind of slightly Northwest here at the yellow Target right that's very simple very intuitive but I can 100 percent rotate this too far and produce unnatural results or even get breakage if I go too far with the adjustment which is again something that I can't necessarily easily do in the RGB metrics like check check this out obviously this is beyond the realm of where my taste would take me anyway but like you can see there's this sort of like splotching beginning to form in her cheeks and really where the danger lies with a tool that can do this a tool can do 10 of that and sneak by you and you might not notice it and that's no fun to notice for the first time when you're like at the premiere at the film festival or whatever and you're like oh geez that's kind of noisy or chattery or it's not smooth that's no fun so that's kind of the trade-off there is a curves adjustment can be more intuitive but it can also be a bit more dangerous whereas the RGB mixer is a bit less intuitive but it has far more tolerance it's it's far more stable in terms of how quickly it will break or not so in terms of which is best for film emulation you could accomplish any explicit goal of film emulation whether that's warming your yellows or like making your Blues more cyan or whatever you're trying to do and that's really just going to come down to observation and saying like here's what I want to mimic or emulate from something that I see from a film or a film emulation that I like you can accomplish pretty much any Hue rotation goal with either of these tools it simply comes down to like all right what's one that feels reasonably intuitive to me that I can make work and that I can get clean results out of where there's no artifacting or tearing whatsoever so either can work I definitely encourage learning both and I can tell you you know like spoiler alert for uh any future look development uh you know like courses or videos that we do I definitely encourage spending the time on getting comfortable with a matrix with the three by three Matrix as a foundation just because it is so clean and if you spend that time to initially get comfortable with how to produce hero rotations then you have a very very clean tool that you can always rely on and always feel good about even if you're doing more aggressive adjustments all right also in your last video you mentioned um how to basically do a split toning uh but who could you execute a different kind of split tuning basically that is something that is not tilling the shadows and kind of orange in the highlights it's possible to create a strong look with for example blue slash yellow or the green magenta or some other kind of color scheme yeah this is a great question so we're asking about split toning and whether it's possible to do something different than the classic cool shadows warm highlights uh cocktail that I talk about so much here on the channel short answer yeah yeah it absolutely is and it's been done I've seen it done I'll tell you guys something this is something that I think about all the time where I'm like oh am I going to do that again like I do it in I won't say most oh actually yes most the majority of the projects that I grade are going to have some sort of cool shadows and or warm highlights not necessarily both of those things but almost always at least one of them and I think every time I go for my split toning I try to be in the discipline as a professional colorist as a professional creative of going like am I just reaching for something out of habit and is there actually something more inspired or creative that I could do here than what I always do and I often will try and for me it's even as simple as like oh what if instead of like a cyan what if I did like a pure blue push in the bottom generally I don't like it and same thing if I'm like oh what if I did a red push in the bottom I don't think I've ever actually been able to pull that off in the wild and produce an image that I liked that I don't know what that says about me whether that means I have narrow taste or little imagination but I can tell you I keep trying it and really the further I get away from cool shadows warm highlights the less psyched I am about the image in general now in some cases I will I don't always use split toning so I will let Shadows uh you know by comparison to a classic split tone in in the sense of comparing to that I'll have Shadow that to me feel very warm because there's not any cooling going on down there but that's probably the closest I come to doing warm Shadows is not doing cool shadows if that makes sense now in the top end cool highlights there's a little more play there like that's something that I will do from time to time although it's something that I'm probably more likely to do is like a shot level balancing thing than like a part of an overall look so my yeah the short answer can you do non-traditional you know like warm Shadows uh cool highlights or like pink Shadows green highlights you can of course do anything at all in my experience I haven't been able to find find much gold in those alternate approaches and I've in fact found more and more nuance and detail to drill into the more that I spend time just within the constraint of like well you should have a neutral segment the bottom should be some kind of cool the top should be some kind of warm there's so much variability just within those constraints that I've continue to find interesting aspects to explore just within that sandbox so that's my personal answer but I would never discourage you from exploring all possible approaches and I would love to hear and see if you come up with something that you really like that is outside the realm of that sort of classic traditional cocktail of cool shadows warm highlights let's go to a different topic for now if you are cover managing Rock 7 online material bringing in that shot into the ministry gamut do you still set the middle grade point two three four one or does that change since there is no real mid grade in display space oh this is a really good question so the question is you know we've we've uh we talked in this week's video and we talk all the time about color managing in um or rather about look development and some of the constraints that uh I think are really helpful to impose in look development one of those is uh constraining middle gray I talk about that I feel like in almost every video that I do mid gray is like one of my favorite topics in all of motion Imaging maybe my favorite topic but the question is when we are considering mid gray and considering anchoring mid gray in our look so again you know if we just go into our DaVinci wide gamut gray power grade here and create a new custom curves node and drop an eyedropper so that we've got that Anchor Point like so the question is do we need to bother with this if we've got a Rec 709 piece of material like this I would say yes absolutely and my reasoning for that would be to invite you to think about where in the process this is happening the reason that this question is being asked it's a good motivation for asking the question they're like well I don't know where the heck mid gray is in some random piece of Rec 709 footage as you guys may have heard me talk about before there is no explicit definitive mid-gray position in a display color space it just goes wherever you want it to go so the idea that oh my mid gray should reasonably be sitting at a 0.4 or wherever when I input map from Rec 709 that is a generous assumption because there's no guarantee that your mid-gray even in a well-exposed well-mastered image is going to be in any particular position maybe generally you could say it's going to be between a 0.3 and a 0.4 but there's no guarantee it's going to be anchored to any more specific point within that however if we think about our look and we think about you know like here's our contrast curve similar to like what we built this week right I'm going to do my my toe on my shoulder all that good stuff Let's uh do something like that so here's my contrast curve which is right now pretty mellow by the time this Rec 709 image receives this curve it will have already received my exposure adjustment and whether or not I am doing it consciously or by the numbers I when I reach for my exposure for my offset wheel or for my Global exposure whatever tool I'm using to adjust exposure when I reach for that my eye is going to gravitate toward positioning my mid gray my exposure my middle exposure of this image in alignment with all of my other images in my timeline my you know like camera log images or whatever else the case may be so so assuming that mid gray is at a particular position coming from a Rec 709 Source probably a bad idea assuming mid gray is going to be within a pretty tight tolerance of a Target by the time a graded image hits your look I think that's a pretty good bet to lay because your eye is going to find your way to that without you ever checking the number at all I hope that distinction makes sense but a great question um what is the is there a better way to create a single colored look other than just pushing the whole image towards a single color oh cool question so is there a way a better way to create a a whole colored look what I'm going to call I'm going to what I typically refer to these as like a wash look or a monochrome look where you're like hey I want this whole thing to have like a gold Vibe for example I think there is a way to improve on that base of like well if you want it all to be warm you're going to have to push it all warm sooner or later somehow some way right but let's talk about some principles in look Dev that we can use to maybe accomplish things a bit better in that regard so let's say I don't know let's go to like this image here we'll we'll audition this look on a few different ones but let's say we want to do like a really pushed warm look with very low color separation and really push it into that upper left hand quadrant okay two uh two things that come to mind uh kind of right off the the bat when we talk about this the first one well let's say let's leave our creative contrast curve I think just as a a base because that's generally there's going to be some kind of creative contrast curve in virtually any look that I'm going to build that one looks kind of nice but the two ideas that occur to me from here I'm going to label this and I'm actually going to make this the last thing in my chain this is something else you guys can play around with I encourage you to play around with and look Dev the order of stuff matters there's no color grading police who are going to arrest you for doing it in the wrong order but you can try out different orders of things and get different results from the exact same components that are incorporated into your look so here's what I would look at for doing a push or a wash number one get there by subtracting rather than adding so if we want to do a warm push you're going to get better results generally speaking more filmic I know that's a loaded term but what I would generally consider more filmic results instead of by going for your offset and pushing things like Northwest like so and you know kind of finessing something in like that like let's just grab a still of that instead of that you could go to your blue in your curves here and start pulling blue away and you're going to see that's giving me more yellow than like that gold that I was going for before so I'm going to go over to my green and I'm also going to pull some green away like so and kind of keep going there and then put that back and you can see I'm getting a different result I'm also using gain as opposed to offset which I was using before I probably could have done that better a moment ago but if I wanted to effectively get offset then I could also do something like this and work the bottom end as well but I would uh use my curves for sure I would think subtractively rather than additively if that makes sense so I'm just going to start like kind of pulling these things away so it's more of an offset esque adjustment and then something you can do is like if it's feeling really strong really fast then you know get your shoe angle looking somewhat proper like that's moving in the right Hue Direction it's just way too much of it and now you can just go to your good old key output gain and kind of pull that back into taste so I think that can be a really nice approach and if we wipe there like that's subtle but I hope you guys can see the difference here so uh right now I'm on my subtractive method using the curves versus the offset Hues aren't identical but also like the image feels like it's got too much color energy in it compared to where I'm at now so that would be a good principle there and then really what we want to do here the thing that we need to somehow deal with is when you do a warm push there's colors that are already very warm in the image like her hair is already very warm compared to the sky which is relatively cool or her jacket which is relatively neutral kind of olivey green so you kind of need to like level things out a little bit so that the stuff that's already warm isn't racing out to the edge of what's even like decent looking while the other pieces are struggling just to catch up and even take on a bit of a warm cast because we're trying to do a wash thing so if we're going for a monochrome thing there's lower color separation there so actually what we want to do here Downstream or what you can do is create what I would call like a saturation wrapper and you could do this in one of two ways or you could use both of these tools I'm going to suggest for starters that you use your sat versus SAT and you say well I'm going to tame in my Peak saturations and let the lower saturations be that's going to kind of like level the playing field out if that makes sense so that those things that are already quite colorful are going to be slowed down and prevented from getting too saturated while those things that are already uh I I should have said not not we're not slowing down the things that are already colorful really we're slowing down things are already warm and we're letting the things that are less warm get caught up you could it's it it's like I don't love the way these graphs work and resolve but this is really what we're doing is creating a shoulder for our saturation curve here and you can see when I combine those two now we're getting this warm push but without anything spilling and feeling really ugly you know so this is a good way to kind of compress those things into the pocket another way you could drive at this idea in a slightly different way either as an alternative or as an additional piece would be to do your loom versus set and say like hey in the upper end just take it easy like I don't need to see much of this wash up there just let it go and fall a little bit more neutral that could be a good ingredient as well and so if you start to look at the net of what we're doing here that's starting to come together to me in a much more elegant way than just grabbing that you know offset adjustment like we did initially and getting something like this so here's like just pushing offset around and here's our kind of more refined approach and if we wanted to you know like this is where look Dev gets sort of fun is you can now go back and increase your key output gain for example or you could also see if you get a different result by working Downstream of these pieces and now doing something doing like a little bit of gamma or gain pushing in that positive direction actually better would be to go back again to your curves and do more of the same thing we already did subtract some blue subtract some green I would hope that this is not going to produce much of a different result than just increasing the key output gain over here but again something to play around with but these like kind of Main Ingredients subtracting the things you want less of to get the thing you want more of and then controlling your saturation curves like we see over here I think that's a pretty good recipe for like a more of a washed look when we're grading or look look deving I should say another question from Jim uh Luke Development question as well um he's taking that he finds the worker um he finds that the worker is the easiest tool to what he used to neutrals by locking pins on the warm side and grabbing the center and pulling it to the cool side why don't you use the worker for that oh that's a good question to be honest I would be open to using that I think for me if we look at like you know we did our comparison a couple minutes ago of like okay RGB Matrix is like pure as like the driven snow the Hue versus Hue is a little bit more prone to making a mess but uh can get you great clean effects as long as you're mindful of it for me the color warper just in my experience of using it not only in result but I use the color warper a lot for look Dev like years ago when it was still just 3D luck Creator and it didn't exist in resolve and I just found it to be artifact prone and that's uh maybe a assessment that I need to revise or update but that's generally why I steer away from it is because even in situations like you know with the Hue versus Hue I showed you the situations where I'm like yep I know that the type of adjustments where I'm going to get problems are more aggressive rotations you know with the color warper my assessment in past usage has been even in unexpected situations even when I'm not necessarily doing anything that big or that would be obviously a bad idea even in those situations I can sometimes get what feel to me to be like kind of arbitrary artifacts noise breakage because of the models that are being used down here or because of the way that we are moving in or out of them um so that's my assessment I will fully cop to that maybe being overly superstitious and maybe I need to put a fresh look on there but that's generally why I don't use the color warper is I just it hasn't uh I I just don't have enough trust for it to use it and be confident that it's going to always work on every shot and never introduce problems even when it really shouldn't introduce any problems so oops I was muted you were muted what is loop development give an example of that for sure so you know like a custom gamut in look development is something that sort of borrows from more technical color science so if we think about like we know when we're mastering for rec 709 while Rec 709 is a gamut right there's a reddest red a greenish green and a bluest blue the rec 709 can produce and uh if you have something redder Greener or Bluer Than those points it's outside of the rec 709 gamut in general right that's a technical gamut that gamut exists to describe the performance of a display right Rec 709 is important because I need to know that if I send a red value that is redder than the red of Rec 709 to a Rec 709 display what difference will I see it's actually a good slightly geeky question for you if I have pure Rec 709 red 100 Rec 709 red and I send it to my Rec 709 display I'm going to see the reddest red that that display can reproduce right what if I send a red that's just a little bit redder than Rec 709 red to the display what will I see then think about that for a second I wish we were in Zoom you guys should all join me for uh the next Workshop or course that I do so that we can do my my favorite pop quizzes live and actually give me the chance to hear your answers hopefully you've had a chance to answer to yourself the answer is nothing you won't see any difference at all right because you've already hit the edge you've already hit the limit you've already hit the max of what that display can do in terms of the reddest red that it can reproduce so that's the reason that the gamut of Rec 709 exists that's the reason we have technical gamuts is to describe the capabilities and the limitations of a device whether that's a camera or a display a creative gamut has a different context creative gamut is not about what's the limitations of what the device can do whether that's the camera or the display the creative gamut is about what are the limitations that I want to impose based on my taste and this is a slippery thing within look development that I I also think simplifies look development once we begin to understand it there for example even in Rec 709 which as we all know is not not the largest color gamut in the world and there are important colors that would be great to see that are outside of rec709 right even in Rec 709 there are colors that I literally never want to see as a colorist they just don't look good to me if that sounds prejudicial it kind of is but that's what a creative gamut is about for me is about shaving away sculpting away Those portions of my gamut so that they never make it out to the display at all so it's about taking yeah I understand here's like the full box of crayons that are available to me in Rec 709 or P3 or whatever it may be but a creative gamut you could think of is essentially looking at that box of crayons and taking out all the Crayons you don't like and throwing them away so it's the same principles it's the same way of describing what's the reddest red the green is green the blue is blue of my look that I'm going to reproduce but it's not about technical capability it's about creative limit rotation or rather limiting the application of an aesthetic in order to get a particular result so creative gamut is really just about sculpting away starting with everything that's possible and sculpting that down to everything that you want to see and removing everything that you don't want to see because then you don't have to deal with that problem at the shot level you're not going to have if like you're not going to have a garish red stop sign if you have removed garish red from your creative gamut slippery subject but I hope that helps illuminate a little bit um like Alexa for lowering DSLR cameras and what kind of tools could you use in resolved that is a good question how would you what was the the how would you emulate color points basically just the how for example an Alexa looks like 709 compared to uh regular pc DSLR yeah yeah so this is a good question with a a a a deep answer worthy of exploration you know like basically the the simpler question that we have to start with if we're asking how would I model the complex behavior of an Alexa and make the behavior of a Canon DSLR for example align with the behavior of the Alexa well that's two things that I need to deal with there right first I need to essentially strip away and neutralize or level the behavior of the Canon DSLR that's number one right then I need to apply the complex behavior of the Alexa that's something that's not terribly easy to do you know like it's funny because on its surface it would be as simple as a color space transform saying like oh wow we have Canon and we have Arie are we done is that it do I take my Canon footage and move it into Arie there are even colorists and practitioners who do this because they are mistakenly under the impression that moving into the area wide gamut color metric is going to give them better colors or more filmic colors or get them closer to an image that looks like it would if it had been shot on an Alexa but the reality is it's not that simple the truth is that the color gamuts that we have for uh you know like different cameras as described by camera manufacturers they're simplified idealized versions of a much more complex animal and in fact there's some interesting emerging research about how we can more robustly account for and uh describe the non-linear behavior of a camera because the thing that you will never get with something like this is this is a purely linear transform this is the equivalent of basically the RGB mixer plus a tone curve okay so you're never going to get the non-linear behaviors that are present in both of these cameras whichever cameras we happen to be talking about but in the hypothetical case of our canon in our Alexa there are going to be non-linearities there in terms of the behavior of the the sensor its response to light so accounting for those things those things there's a lot of the essence of what makes a camera look like it does are in the non-linearities that are present there at the same time cameras are trying to be linear to split your camera manufacturers are trying to make the sensors be more linear it's just that it's impossible to do that perfectly or they the you know like tolerance of a given sensor coming off the production line doesn't actually do like perfectly embody the design of the camera there's all kinds of variability in there but to go back to the question to how would you fit that one method would be basically brute forcing and just looking at like a bunch of color points on an Alexa versus on a Canon and then making the Canon fit into what the Alexa did instead that's essentially I mentioned uh Steve Yedlin a couple minutes ago when he did his initial round of emulations like generating emulations of uh Kodak film that's essentially what he did except the two camera systems were a film system and an Alexa and he basically did a Brute Force point-by-point color analysis that said hey go analyze you know whatever these couple thousand color patches that I've measured under equivalent conditions for each camera system and make the way that the film system responded make the Alexa produce that same result take the Alexa starting point and map it to the result that the film system arrived at so that would be one way of doing it but that's not something you can really do within resolve the closest you could come in resolve would be a color space transform or if you wanted to do a color chart but that's really going to be a similar result to doing the color space transform so a long answer that I I hope is better than saying you can't and not giving you any reason why but uh the reality is there's you're better off whether you're talking about emulating the behavior of an Alexa or bet or emulating the behavior of film you're better off identifying on an aesthetic basis what is it doing that I like and then manually trying to incorporate that aesthetic into your look development stack and figuring out what tools you can accomplish that particular component or that particular component of the quote unquote Alexa look or the film look or whatever the uh camera in question might be all right let's go to Loop develop in the game if you create a look for the little car mode or races and the home timeline is for example Alexa 35 but then you want to use it um the whole Loop there for let's say about Magic 4K how does it transfer between cameras is dynamic range an important factor in the in how they look troubles between cameras oh this is a fun question so how do looks travel between cameras assuming that you know like we're in davinci-wide gamut or Aces so we have a you know let's just make a specific example of it let's say we're in DaVinci wide gamut and we've got a timeline of all uh do I have that Alexa 35 shot in here yeah so we've got an Alexa 35 shot let's say we've got a timeline of Alexa 35 shots for which we have developed a look in DaVinci wide gamut intermediate and the question essentially is if I get some black magic footage and I properly input map that into DaVinci wide gamut can I use the same look that I have been using for the Alexa 35 short answer yeah yeah absolutely remember that fundamentally the the we need to operate from the premise that cameras are effective accurate captures of scene data the reality is that's not always true and that's sometimes more true or less true depending on the camera but we have no choice as post-production creatives but to somewhat place our faith in the idea that a decent quality camera is a decent capture of scene data if we're not willing to accept that then we can't color manage at all and what and and we're literally just better off saying well no one really knows what this is or if it accurate if it accurately captured the scene so all I have are my hands and my eyes and my color correction tools to somehow make it look good on my display that's a far worse uh conclusion to arrive at than saying I'm going to assume the camera did its job camera doesn't always do its job perfectly in fact it virtually never does its job perfectly but the alternative of saying I'm just gonna feel my way to a good result shot after shot after shot and try to beat what the camera should have done is an even worse idea so if we think about our specific scenario of like oh does the dynamic range of the Blackmagic cameras that is it not enough for our look development here's the reality guys it kind of depends on the type of look development that you're doing and it depends somewhat on whether you're doing SDR or HDR but even in HDR and even if you like your highlights nice and Pinky and hot the reality is cameras like any prosumer up through high-end cinema camera they capture enough data and again if we kind of like stack on top of our belief system about cameras and what they are capable of and what we can trust them to do well if we're saying we trust them to capture scene data well we can stack on top of that and say we trust them to capture enough dynamic range for us to grade sculpt a look and get it out to display with enough dynamic range at the end point that we're seeing what we want to see right that's another belief that like can like have shaky Foundation sometimes but we should generally accept that too we should say you know a black Magic Camera a Canon camera Sony camera those are going to capture enough dynamic range for our needs even if we want lots of dynamic range even if we have an HDR display we need to trust that that dynamic range is there because what's the alternative the only real alternative would be to say hey filmmaker I'm not going to grade your stuff because I don't think you shot it on a camera with enough dynamic range we kind of got to work with what we got in the case like in my case it gets a little bit simpler because I'm not the biggest fan of one-to-one linear dynamic range over a long run I generally like to see compression in the bottom a fairly short like linear segment in the middle and then I like to see highlights roll over fairly quickly so I mentioned pingy highlights a moment ago I generally prefer kind of creamy or more rolled off highlights so for me it gets a little bit easier I don't even need as much dynamic range as I would if I wanted to get a really really linear reproduction of the original scene up to every inch of capability of my display because I actually want that shoulder I want it to roll over before the top end of my display but even if you want to get every inch of dynamic range available to you in the display in the specific example of an Alexa 35 versus a black magic I think you can trust that that dynamic range is sufficient there because even with HDR even trying to get one-to-one linear reproduction of your highlights an HDR display play has way way less dynamic range than the cameras that we're talking about that's just the reality and the dynamic range gets less and less important like we tend to think of like oh a 4000 nit display is that uh you know whatever that ends up being is that 40 times brighter than a 100 nit display well no it's not that much brighter and even if it is it's not like you are exposing up by that many stops to get a Brighter Image that's really just twinkly stuff in the top so the difference between four thousand and two thousand or one thousand it becomes less and less significant as you climb because there's less and less of the image that's going to be living up there anyway another long answer hopefully some helpful Concepts as you're thinking about look Dev for camera a versus B let's go to the black Topic in film brain emulation a couple of questions here the first one is our neutral flux a key element of emulating a film Loop and the second one is do you often obtain Shadows at the short level or you correct the Shadows at the shot level if so what tools do you use for that oh good questions so uh the the main one here is are neutral Shadows an important characteristic of a film system or of a film emulation no definitely not if you look at like you know I can show you like any data set that I've ever looked at for any measured film system uh like if we look at let's see our let's just look at our good old uh Kodak 2383 right here within resolve you're going to see these Shadows are nowhere near neutral there's a significant cyan push down there isn't there that's just innate of what are called the unstable primaries of a film system so no that's uh not only is that not important that's the opposite of what you're going to see in any film system now there's a trade-off there and there's a there's something to be navigated because what is true is that film Engineers were actually trying to get neutral shadows mid-tones and highlights they just couldn't because they're working with organic materials that are imperfect that have certain tolerances that are unstable so the best they could do was get something reasonably neutral somewhere near the middle of the tone scale and then let things swing cool and warm on either end of it but no neutral Shadows are and in fact the opposite of something you would see with the film look and so that's in fact something you want to avoid you want to get some tinting in your shadows if you want to successfully emulate that aspect of a film print and then in terms of tinting Shadows or correcting Shadows at the individual shot level which was the other question do I do that how do I think about that I don't think about either very much at all to be honest with you again I operate my base assumption if I have I've got kind of my pre-flight checklist that I could go through here we've got good color Management in place okay we've got a good looking place that embodies my creative intent that's clean works for all my images I'm now grading at the individual shot level my assumption unless the image proves otherwise to me is that I have reasonably linear balance from Shadows up to Highlights so that's why in my balance node when I'm color grading I am color grading using my offset generally speaking because I just want to affect the bottom top and middle of the image with the same brush now I'm not going to say this is the only adjustment that I ever make or that I never see the need to make a different adjustment in the bottom than in the top I do but my Baseline assumption is that I don't need to go looking for instances where that's necessary because I've properly managed a camera using what the manufacturer describes as its response or its behavior in response to physical light so I'm not going looking for those non-linearities I'm waiting for them to call themselves out to me and when they do I just know like that's a that's something I can't do on every shot I can't go in and critically evaluate are my shadows mid-tones and highlights perfectly balanced on this shot I can't do that for 2000 shots I can only wait for that to make itself obvious to me that oh those Shadows are feeling kind of weird are those highlights feel really warm or whatever and then I'll go in and treat those things a la carte but it's not something I'm proactively looking for and same thing I would say in terms of tinting Shadows no because I'm trusting my system to do that so I'm you know I've Got My overall color management I'm trusting that the camera was a reasonably accurate capture of scene data I'm trusting that when I balance the top it's going to balance the bottom as well when I'm doing my shot level color grading and I'm trusting that the bias or the tinting that I introduce into Shadows or highlights in my look if I am introducing those things is introducing what I would expect or what I would want to see in those areas so my default assumptions there are I'm good if I do those things everything is happy and I'll wait to be bothered or shown that is not enough those assumptions are not holding in the case of shot a b or c and then I'll spot treat those shots but it's a difference of being aware that that can happen versus proactively looking for how I'm going to make those adjustments in every single shot which is more the way that I was taught color grading honestly is kind of to play more defense like all right new shot assume the shadows mid-tones and highlights all have different balance or bias and you need to neutralize them and then assume that you're going to need to reintroduce your tint in order to get an acceptable result my opinion is kind of what I just outlined here is a much more fun much more efficient approach that leads to better looking images a question about development and if you are creating a look whether there is a power grade or a local table how would you go about testing testing the the grade to make sure it doesn't break or creates any artifact in your routine Lane oh this is a good question too so there's lots of different ways you can go about this I'll show you guys one you know like my my two go-to's that I will use to assess how clean or messy a look is are going to be a cube and a ramp you guys see me look at the ramp all the time here I'm resolved because it's a bit easier to pull up going to go to my grayscale ramp here this is not a end-all be-all but if I look at this and I bypass my color management so I'm just looking at the effects of my look which right now is it's very simple right it's a very simple look all it is is this custom curve here and in fact this waveform is mirroring what I see here in my custom curve but if I see breakage or odd behavior in this waveform that is a good predictor that sooner or later an image is going to pass through this look that is going to break it like I'll give you an easy example I've done this in the past like go to your log Wheels here and let's spin the bottom end here or let's even better let's do the the mid-tone like you see that if I look at this I I don't even need this graph because if I look at the image the image is going to look super weird by the time I turn my color management back on like that feels very strange obviously but for the sake of a dramatic example I'm going to say maybe I don't notice that visually but if I can see that if I see a weird Behavior like that in my ramp clear cue to me that sooner or later the right shot is going to pass through there and it's going to break so how hard is too hard how harsh is too harsh that's something that comes with experience but that's my first line of defenses I'm looking at the waveform to look for too much complexity for Sharp corners for Sharp edges those are the kind of things that may not create artifacts in every shot but that will produce problems at some point on some shot so that's the number one thing that I look at and then the other thing that I look at is my Cube this is something that you kind of have to come up with a clever solution for here inside of resolve but the one that I've come up with I really like I use a product called omniscope to do this let's go to my miscellaneous power grades here and this is a tool that a member of my community actually developed for me that just produces every color in an RGB Cube and then we have our omniscope output here if I open this up and I connect to resolve I can actually see the result of my Cube let me set this up like so I'm going to say stay on top and now I can look in real time at how my adjustments are affecting my overall Cube so this is the type of cube that I would be perfectly happy seeing if I start to see hard corners or harsh edges or breakage in the continuity of my points that's my other key cue that there's something messed up in my look that sooner or later is going to bite me in the butt and I'll go in there and look for what is producing that and adjust the component accordingly so those are the two main things that I look at and then on top of that I just do lots and lots of testing and you know like I will beta test looks especially if they're you know if if they're for filmmakers I'll I'll uh you know work with some of my team members here at Cullen Kelly color and just have them all test things out on real world images if it's for a product like a release for you guys like the Voyager pack I'll again have members of my team or members of my community beta test stress test run it through lots of images make sure there's no breakage no artifacts and make sure that it consistently produces nice looking results on an image that's being properly fed into that look so that's kind of my Approach there I'm not a big believer in so-called stress test images those like synthetic images that will show you breakage I don't think there's anything wrong with those inherently I just don't think that they are I don't think they tell me anything that these first two types of testing don't tell me and I and I think sometimes they tell me things that are a bit Irrelevant in my experience meaning that I can see things in a stress test that'll make me go oh wait I need to redo that when the reality is I may not need to redo anything at all so some of the ways that I think about stress testing and evaluating is this look is this Lut ready for prime time and is it going to be reliable and robust for its lifespan uh Funk listening here have you ever tried a regrading a project that hasn't really been released just for fun or for practice I'm going to go about doing that yes I do all the time I you know like lots of different forms that that can come in um it's funny that it comes up this week I'm actually right now in the process of re-grading a film that I graded about a year ago uh and it's really cool actually the filmmaker came back to me and said hey you know like I think we achieved the exact look that I asked you for but I think I asked you for the wrong look essentially like I think we could do something a little bit more bold so like a little bit more eye-popping and we'd like to regrade the film so we've been working on that project for the last week or two and really having a blast with uh in that case doing something quite simple of saying hey a lot of the shot level grading is pretty good most of it in fact is really good we just need a stronger look so that ties into this question and also to our theme of look Dev because my original look was well built because my grading decisions underneath were well founded and because my look Dev practice is in a pretty mature place we were able to go in and revise the look on top of the grades that we'd already done and then really all I did was take a trim pass through the timeline and look at the graded images through this new lens of the new look and make sure things hold up and in cases where it's like ah that got a little bit oversaturated or that feels a little bit crunchy or whatever just going in and trimming Things based on the lens of that new look so yes that like that that's something that comes up for me in different forms all the time that's one form there where clients will actually ask for a regrade another way that it comes up all the time is when I do directors Cuts or even DP cuts of commercials or films or whatever usually commercials where we've got the agency cut that's what actually goes to broadcast that's the one that the advertising agency and their client approved and then we might do a slightly different version for that director's reel or for the DPS reel where we will kind of reimagine either a little bit or a lot what was done for the broadcast grade that's another way that it can come up a lot and then a third category is I'll do that all the time just for exercise just for fun just to see if I can rebuild a better car um and sometimes I'll get to the end of it and be like ah I just redid everything and tore it down and tried it again and it's no better than it was before and sometimes they'll go wow look at all I could have done for this project if I had done this the first time that can drive you a little bit crazy but it's also a really really good exercise to just tear it down and rebuild it and just see if you can make it look five percent better I think that's a really worthy exercise so very long way of saying yes absolutely that's a great thing to do it comes up as a need uh like as something you need to do as a professional and it's a great way to continue to grow and develop your craft as a colorist even if no one's paying you or asking you to do it let's go to a key question here how much of the film look can you achieve in resolve compared to something like scattered data interpolation and machine learning with thousands of data points this is a great question to wrap us up for today how much of the film look can you achieve in resolve versus more sophisticated methods like scattered data interpolation or you know other things that you would really have to do outside of resolve that's a difficult thing to quantify for sure because if you look at you know think about a really good emulation and again we've talked about my friend Steve yadlin a couple times today for my money his emulation of like a Kodak film system is one of the best that I have ever seen and it does incorporate some of his preference and taste so it's not a straight up one-to-one emulation but it's close it's really aimed at its foundation is emulate you know like this set of film stocks going through an idealized process and if you look at just tally up like okay you know Steve is doing his look Dev in Nuke for example and tally up the number of nodes that are being used to produce that final output I don't know how many there are I've seen the stack he's shared it and there's a lot but let's say there are 50. that wouldn't be crazy maybe there's 40 maybe there's 50 somewhere around that how many nodes if I you know like if this week's video wasn't a 10 or 20 minute video but a two hour video how many nodes in resolve would I have ended up creating not more than 10 to 12 because there's just not that much more that we can do within resolve we would have hit the upper limit pretty quickly right so if we compare those two 10 to 12 versus 40 to 50 the math there 25 right we can get 25 percent of the way there the bigger thing the thing that makes it a slippery question to answer is what is what perceptible difference does that make to the viewer I'm going to give you a bold claim here by the time you get a contrast curve by the time you get your color palette including like the color of colors so like yellow is going a little bit more gold to pull an example from this week's video by the time you get contrast curve color palette meaning the color of colors and the density of colors meaning how dark or bright are your red screen Blues science magentas and yellows and that split toning scheme of like cool shadows warm highlights by the time you get those you know three things in place with one of them having some subcategories you're like eighty percent of the way there that's the reality like you can get way more forensic than that and go in there and get details that matter and for detail attentive people like all of us in this room are real value ads that mean something and will likely mean something to our clients but in terms of like what does it actually mean to the viewer I would put that assessment at about 80 and when I say 80 I mean 80 in an idealized fashion so that's not just because you did a contrast curve a split tone thing and a hue versus Hue and some sort of like uh color density thing it's automatically 80 I just mean you could get to 80 with enough time and with the right application of the right skill right within resolve that's a bold claim uh and I can't prove that I'm right but you can't prove that I'm wrong that would be how I'd put that because I think there is a diminishing return as you add those little flourishes that we find so cool the difference to the end viewer becomes less and less significant that's not a reason necessarily to take shortcuts but it also is something help to be aware of when we are diving deep down the rabbit hole of look development which I hope at least a few of you guys do it's really fun down here but it also gets kind of lonely so come join us down at the bottom of the look development rabbit hole if you like and if not I hope I've empowered you with some good tools practices methodologies ways of thinking about look development right within resolve and even if you don't get anything else if you never develop your own look at all I hope that like in this week's video now in this week's grade school you've got some better means for evaluating looks that other people make looks that I make for you looks that you Source from other reputable places and saying well how did they do their split toning what am I seeing here in terms of like a Pew rotation or whatever the component may be and it allows you to dissect and think a little bit more critically about what's happening in those looks and evaluate what you love and maybe what you don't love and what you'd like to see change that in and of itself can start to make you more of an author and less of a passive consumer of looks so I I hope if nothing else that you get empowered with some of that knowledge and some of those ideas as you continue in your color grading Adventures um thank you guys for being here on grade school Friday always a blast hanging out with y'all I really love days where we get to talk about look development like today so I hope you enjoyed the session as much as I did and I will see you next week here on the channel for our next pre-recorded video and then for grade school next week and thanks to Rafa for co-hosting with me everybody have a great weekend great great weekend see you soon

Channel: Cullen Kelly

Views: 3,420

Keywords:

Id: rQvHziDg2oo

Length: 61min 42sec (3702 seconds)

Published: Fri Jun 02 2023