Grade School: Grading an Entire Timeline with Just Three Knobs (Desert Island Edition)

i think i'm live am i live can you hear me youtube are you guys out there i'm gonna trust that y'all are i see some folks finding their way in here welcome everybody welcome to grade school i'm really really excited about what we're doing today because uh i just think it's gonna be a hell of a lot of fun and i think we're all in for uh some surprises and uh we're gonna learn a thing or two myself included before i dive into what we're actually going to be doing today i just want to sort of point out on a broader conceptual philosophical level the power of limitations any of us who've got any kind of experience uh in any sort of craft or uh creative endeavor have had the experience at least once of like oh my god like whether it's a deadline to submit something or whether it's like i remember we did an exercise in film school or they're like you have to make a film with just one cut like whatever the limitation is it can be completely arbitrary but man it just can set your imagination free sort of as a paradox there so that's very much how i feel about the exercise that we're going to be doing today which is saying all right you get three tools and you're gonna grade everything with just those three tools so i'm gonna see how i do today i've picked out three and i'm hoping that i can live by them and you guys can can watch me try to uh for the next couple minutes here i'm really glad you guys are here everyone's having an awesome week and awesome december i'm having a great week uh i was just telling my buddy gedali who's joined uh me today to help moderate the chat that uh as soon as time opens up in one area of life i i don't know what it says about my personality but i just tend to fill it with something else so i've been on a mad tear of coding all this week and uh figuring out some new stuff learning some new stuff but uh all that free time that i uh thought i might have at the beginning of the week has somehow evaporated into the midst of uh trying to solve coding problems so that's uh i i i don't know if i hope that's what's happening with you guys or if you're actually wise enough to just enjoy having bandwidth if you do but that's where i'm at and i'm excited to get away from staring at uh line after line of code and trying to figure out what i'm doing wrong and actually talk and speak and interact with you guys and go back to the creative side just a little bit and i see already uh alexander is starting to uh take stabs at what the three knobs are you guys can feel free to chime in at any point and make some educated guesses as to what three tools i'm going to choose for this exercise now just to sort of like frame the exercise and give it a bit more specificity before i dive in here's the way i'm thinking about it so the first thing is i'm not including color management as one of my picks because color management is not a tool of color grading color management is what surrounds our color grading okay so we're going to be working color managed as i always do so that doesn't count against my three okay and uh from there we are i am going to pick out three and i'm going to aim myself at the task of grading up this entire project that i have pulled in here let me flip onto it this is a short film that i graded earlier this year really beautiful film called the laundromat of malibu it's all set inside of this laundromat here and the right now we are looking at only my color management so just to kind of catch us up on that front if i go to my project settings these are not like magical project settings these are not like there's nothing sacred about these let me actually get these turned on since i am finally learning from experience that you guys can't see what the hell i'm looking at by default here let's do [Music] this guy there we go so here's what i've got for my color management uh right now i'm doing davinci wire rgb color managed i'm not doing the new 17.4 automatic color management which we can talk about in a future grade school i'm doing a full custom color processing mode and i'm doing it in this waste because it's actually despite needing to set all these parameters in a particular way it nets out to a really really simple color management pipeline that basically says hey all my stuff's airy don't input map it i'm going to work in airy and then at the very tail of the stack take assume a full 10 000 net maximum luminance uh image and compress that down to a rec 709 gamma 2 4 100 nit image so that's what my color management is that's how i have it set up here and a bit about why i have it set up that way you can rip this off if you want to if you want to see what it looks like but as i said there's nothing particularly sacred about it it's just the way that is going to work nicely for this project now let's take a look at this timeline and i'm going to kind of work in reverse order of i'm going to talk about the tools in a different order that i'm going to use the tools so i'm just curious let's see if we've got any other guesses on tools yeah so does offset and rgb offset count as three different tools no so let me let me kind of finish uh framing up my uh assumptions or my rules or my boundaries for this exercise i'm assuming like offset whether i'm using it to uniformly manipulate red green and blue which is what the offset wheel does or whether i'm using the offset ball and individually manipulating red green and blue that's the same tool that's just do you want to manipulate all three of your channels at once or do you want to only manipulate one or two of those channels so that's the same thing that's fair play we've got some good guesses coming in here the very last comment there about using lab or yuv color models other color spaces those are fair play those fall under the heading of overall image pipeline to me so i can move in and out of those things as much as i want to and those don't count as tools and i can get different results out of my different pieces and i'm already seeing a couple of good guesses come in here and i'll sort of frame what i see as the fundamental i'm getting ahead ahead of myself i'm going to give you guys tool number one the most essential tool to this process the tool maybe not the most essential but the tool that i'm going to use more than any other tool of these three that we're going to talk about here today first one is going to be offset okay now if i were in a linear domain if i was working in linear which i sometimes do as you guys know that would instead of instead of selecting offset i would choose gain but there's for points of for for purposes of today's discussion there's really no difference i'm picking offset for a couple of reasons first of all i need some way to move exposure you guys know how much affinity i have for exposure as the foundation of all good color grading i need some way to directly and cleanly manipulate my exposure and i think offset is a really good way of doing that i'm also choosing it for a bit of a sentimental reason because if we look back at the history of motion imaging and we think about the people who preceded us as colorists the color timers or color graders of yesteryear who were working on films you guys probably already know this because i talk about it all the time do you know how many tools those uh men and ladies had uh when they were like exacting the creative vision of their clients on uh the film print they had one tool and that tool was offset literally plus red plus green end or plus blue or minus red minus green and or minus blue that was all they had so i want to uh give a a nod to our four bearers and uh use our our offset for that reason as well but that's going to be my anchor thing that's going to be the first thing that i go to on any image okay so if we look at where i'm at with this timeline right now i've got my color management turned on which as we've just decided doesn't count against me and if i did nothing but properly color manage and work my offsets i could start to get decent results i could at least set exposure like if i if you were to make this desert island challenge even more difficult on me and say you only get one tool it would definitely be offset because at least i can sweeten exposure on each of these frames so you know on any one of these i can just kind of dial that into taste however i have two key challenges uh that that uh sit in front of me i've got one challenge that i one more thing i'd really like to be able to do on my individual clips but there's another even bigger challenge that if we again think about color timers and the way that they worked what is the fundamental ingredient that facilitated that single tool that allowed them to get away with using a single tool when they were grading does anybody have a guess you can uh sound it off in in the chat there i know i'm a a couple minutes behind but i'm actually curious if anyone can guess at this one to repeat the question what is the fundamental other ingredient in the overall process of taking a film negative and making a reproduced image out of it what is the other ingredient that allowed a color timer to get away with doing nothing but adding or subtracting red green and or blue what what facilitated that anyone have any guesses how did how how does to catch a thief look like to catch a thief with only plus red plus uh plus red minus red plus green minus green plus blue minus blue they really just did that with with offset how do they do that printer lights is a good guess printer lights actually is exactly what i'm talking about though when i talk about offset this is the i i might actually experiment in the future with doing i think there's like a lower lag form of live stream that we can do it's kind of fun interacting with you guys as you're you're there's some some smart folks out here and some good guesses coming in already yeah i see printer lights pouring in printer lights is a a great uh guess but printer lights actually is what offset is quite literally that's what i mean when i say offset i'm going to give it away do you know why apocalypse now looks like apocalypse now close encounters the the third kind looks like close encounters of the third kind it's not because the color timer shaped that with their offset it's because they were feeding into a film print film prints facilitated the overall image to be able to look like it did to have the polish that it did and the palette and that sweet contrast ratio and color separation and hue sweetening all that stuff that we love about filmic image reproduction that's really facilitated and standing on the shoulders of the film print which is a highly engineered substance okay so that is what i am most missing right now in my stack oh i see jamie did get it print print and neg that's exactly right so i could roll the negative into this conversation as well although the negative was uh traditionally aimed at a much more neutral capture and more of the color character is really uh meant to be imparted into the print and the system itself is designed to sort of like so those two elements could balance one another out against the things they couldn't iron out in one of those elements individually anyway the film print is the piece of character that i'm missing from this equation right now so what do i mean when i'm talking about when i'm talking about that right now we're in a digital box we're not printing film obviously what i'm talking about is look dev i cannot build a look with offset i mean i could try but it's not going to be very excuse me it's not going to be very good it's going to be i'm going to be very limited in terms of what i can do because that's a situation where the simplicity of offset and or printer lights to whatever you want to call them is actually working against us because all we can do is add or subtract red green or blue we can't shape or bend or remap or contort or refine things at all all we can do is really crudely move things around and the reason why that works so well on individual clips whether you're a color timer or a color colorist working in a proper setup here in 2021 is because is to the extent that you are working underneath a really good look that is doing more complex stuff on a consistent basis downstream of their kind of under the hood on your behalf without you having to revisit those behaviors shot after shot after shot so that is what i'm missing next and that is actually going to be like i said push come to shove there's other stuff i'd like to be able to do with the clip level but push come to shove i can grade this timeline with just i can grade the clips with just my offset i'm pretty comfortable with that i need i need tools for uh doing my look dev so i'm going to use two additional tools for that this is what the other two tools in my stack i selected because of what they can do for me on the look dev side okay and they will have some benefit one of them anyway at the local shot level but not not as much as what they're going to do for me at the overall level so i'm just curious let's see if we had any guesses out there okay no no guesses on my my other two i'm going to give you guys just a second i'm going to hold you in suspense and in the meanwhile i'm just going to kind of demo out like you know what it is for me when i'm sweetening exposure this is by the way an incredibly valuable exercise for you to do tape your color management and just buzz through see how much you can make yourself like your shot by just adjusting your exposure you will be surprised at how far you can get when you force yourself to stay with that decision for just a little bit longer you'll you can get further than you think just by lingering on that decision and refusing to move off of it until you've got things looking really good so that's kind of how i would start to begin to work things if i was doing my offsets some of these clips have had their input color space stubbornly left in the wrong place so i'm going to periodically have to do that move right there to get the proper color management but that's kind of what i'm seeing with those and again as i mentioned earlier i'm not thinking of offset wheel and offset ball as different tools they're the same tool of the same operation being applied to the red green and or blue channels of the image it's just a question of do you want to apply it to all three to just two or to just one so in all these cases after sweetening that exposure i could do a new node and i would start to work with my rgb offsets as well and probably have my vector scope there as well for that just to kind of get a reality check on how i'm doing and same kind of thing here this is such a powerful exercise like think about things like okay now that i'm balancing i can think about balancing as a sort of mechanical exercise of like oh i need to make sure i'm getting a more neutral reproduction like okay fine and if there's a white thing in the frame or something that i can use as the basis for that that's fine a lot of the time what i'm looking for when i'm balancing at least to the extent that my skin is already looking reasonably good maybe start with skin that's probably the most important and but the the next thing that i'm often looking for kind of subconsciously when i'm balancing is i want to max out my color separation and you guys can see like if i go to this wide frame here this one here is a great example i'm going to do a new i'm going to go a little bit less moody on my offset here and i'm going to go to a new node now and i'm going to start to work my balance okay and as i do what i'm going to look for on my scope and in the image is where am i maxing out my separation where am i getting the greatest feeling of color contrast between say the floor and the washing machines and the walls and the uh talent's wardrobe where am i getting like the maximum separation there without having to do anything fancier than offset that's a great place to begin there so you guys can see like that's not bad right there i'm definitely getting more color separation here at this position than i had before my colors are popping more against one another so all stuff that we can do just with our offsets now let's see if we had any uh good guesses out here let's see here all right so we got offset curves and color warper we got log wheel lift gain contrast pivot jamie colloid print node i wish i could but i did feel like for this exercise that's not fair play i am that's one of my limitations i'm going to limit myself to uh the resolve film print emulation great guess as well i would love to have a solid fpe lut or my colloid tools that i could drop in but again i feel like that's kind of out of bounds i'm actually going to try to do everything inside the box here uh log shadow log highlights good guesses as well channel mixer color warper all good guesses so let me turn on my scopes here as well since i see that question and i'll put it to you guys this way both of my next two picks are in the list that we came through my next two picks drum roll please are my curves my custom curves specifically and my channel mixer or my rgb mixer and here's my reason why the curves i can use for the main thing that i am hurting for still at the individual clip level here which is like man i really want to be able to adjust my contrast so before i get into the look dev side of things let's just hang out here at the shot level for a little bit longer and i'm going to show you guys how i'm going to cheat a little bit and use my curves before i even use them for their more complex application that they're really going to help me out with in a minute i'm going to use my curves to build a contrast note because i don't have a contrast knob contrast is like one of the main tools that i use when i'm grading and i need a contrast knob so we're going to build a contrast knob real quick here now here's the fun thing if we think about needing to build a contrast knob what's the the sibling or the companion to the contrast knob it would be pivot right however if i have offset i don't really need a pivot because i can simply move my offset upstream of the contrast and change how that contrast is being distributed accordingly yeah it would be nice to have the two but i can get most of the benefits of a contrast pivot simply by properly setting my offset upstream okay so what we're going to do is we're going to make a contrast knob i'm going to go over to my custom curves here and because i want to make a neutral contrast knob that borrows equally from the bottom and from the top and either increases or decreases contrast let's just check out how we can do that so we're going to make a negative contrast knob first and then i'm going to make a positive contrast knob so all i'm going to do is take my upper control point here and i'm going to drop it down so that it's online with this first horizontal line down from the top and now i'm going to do the same thing at the bottom here and i'm going to bring this up from the bottom to this first horizontal line down from the bottom so all i've done here is create a contrast adjustment that's compro compil compose of 50 gain and 50 lift and a lot of it right like i don't anticipate ever needing to go this far and at any point in this timeline and that's the whole point because what i'm going to do now is go to my key output gain which i'm not going to count as a tool because all that i'm doing here is saying how much of this node am i allowing to actually impact the image this is just transparency that i'm doing here this is not an actual tool this is just saying how much of it do you want okay so if i now go to my key output gain i could label this node decon or whatever i want to call it and on and i could leave it at a unity position of zero okay so i can make this part of my stock node graph and now if i want to make a decontrast adjustment on an individual clip i would just go to this node and increase my key output gain and as i do i'm getting less and less contrast and i'm borrowing equally from the bottom and from the top pretty cool right same exact thing for a positive contrast if i do a new node go to my custom curves and go one click one vertical line to the left and one vertical line to the right i'm now increasing my lift and gain in equal measure in a positive direction label this con like so let's do the key output gain to zero on this guy and let me set that to zero over here and these two things would end up becoming part of like my template node tree for this desert island challenge so as i'm going in through here if i want to do positive contrast that's hitting equal parts lift and gain in the bottom and the top i could do it like so right and again i'm just talking about very simple linear adjustments that i see myself wanting to make at the clip level so i now have uh global exposure i have global offset or rgb balance and i have negative and positive contrast all with simply using my offset and my curves and by my tally which i think is fair i hope you guys will agree with me is a fair uh tally and i'm not bending the rules too much i'm only two tools in and i can reuse these curves in a really significant way at the overall look level for this project so like i said i'm kind of going backwards because i wanted to list out the tools i'm using from the simplest and most intuitive and work out to the more complex ones but the way i would actually use these in this desert island challenge is i would probably set up this node tree like so so this uh exposure node or maybe i would just do exp i would just label it and leave it empty and know that that's where all of my offset wheel stuff is going to happen next node i would do bal for balance and know that that's where all my individual offset rgb stuff is going to happen and then i would have my templated decon and con nodes built out here and know that this in the desert island challenge are the four nodes i'm going to use on every single shot to get the clip level adjustments that i need okay now let's save this as a still or better yeah let's just ripple it i'm just going to select everything i'm going to hit command a and i'm going to middle click on this shot and now every shot in this timeline has this template node graph okay no shot has received any grade yet because these are all dummy nodes at the moment but i have a template that i can work my way through in the same way that you guys have seen me do for other kind of grade along videos so with that in place we've got our template node graph we know what we're going to do how we're going to do it at the individual clip level and we've already got our color management in place the next thing i'm going to do in our sequencing today is actually what i would be doing first in this process which is like okay i need to get my overall look built out before i go any further now i'm in resolve color management so i would want to let's actually just do a quick test here i don't actually know if my timeline stuff happens downstream or upstream of my output transform okay so that's that's uh upstream good so we're gonna build out our look and we're gonna build out our look with just two ingredients okay those two ingredients are going to be our curves and our rgb mixer like we talked about before so curves nothing earth shattering here i'm just going to build out a nice tasty curve like you guys have seen me do before the only curve ball to make a really bad pun is i'm going to go for a moment over to this look dev timeline that i've already got built out and i've got a chart prepped and all i'm going to do here is use this chart as a reference for grabbing my middle gray which in this little plug-in that i have built for this purpose i've got middle gray set to 18 i've told it that i'm in log c and i've got three total steps and in the middle is my middle gray so this is pure perfect 18 gray in log c so if i go to this empty node that i just made downstream and i tap that middle gray patch i've now got a control point for middle gray in my curves and if i hit command c to copy this node and if i go back over to my timeline my main timeline and back over to the timeline section of the node graph where i'm going to build this look and i hit command v to paste that node i've now got a template for my creative curve that is going to force mid-grade to anchor where it came in so it's going to refuse to let mid-grade be moved up or down and then from here all i'm going to do i'm actually going to for reasons that will become clear in a moment i'm going to paste two copies of this node we'll deal with this one number two in a moment node number one i've got that good anchor point and i'm going to turn all of my channels on simultaneously and this is where i'm going to draw my overall creative contrast curve now this does actually feel like i am upstream or downstream rather of my output transform so i'm gonna wipe these out and i'm just gonna i'm gonna do these things uh let's do them in a group okay so i've wiped that stuff out of the timeline level i'm gonna add everything into a new group we'll call it main because everything's going to live in here and i'm going to go to my group post clip which is going to allow me to do what i was just trying to do at the timeline level but i think it's going to get me some better results no maybe not maybe it was actually fine at the timeline level and i was just being overly sensitive to where those uh shifts are in the shadows are taking place so i'm just going to draw a curve here kind of by feel okay something like this and then i'm going to do the same thing up here in the top like so so now i've got my very humble beginnings of doing the job of the film print okay i'm just getting some creative contrast into the mix and that's already not looking half bad right and i know that i've preserved middle gray so even though i don't have my colloid tools and i don't have a solid film print emulation lut like i might normally i know that i've got a nice looking curve and i know that it's preserving middle gray so that i can live with i'm going to call this curve next i'm going to do my split toning which you guys have seen me talk about or heard me talk about a ton for this i'm going to un-gang my channels and i'm just going to do a sort of stock cocktail of a little bit of blue into the shadows a little bit of green to go with that so that we're getting a little tealy down there and then up in the top we're going to do a little bit of red little is going to go a long way as it always does a little bit less of my green and then i'm going to back everything way off so i'm going to say remember these sliders the from 50 to 100 is the is how how much strength is being uh that we're using from these uh is the overall strength of this uh adjustment that i'm making on my blue channel if i go less than 50 i'm actually going to be inverting for example this rainbow on the blue channel don't want to do that but i do say want to cut it in half so i'm going to say like 75 for my blue because that's half same for my green same for my red let's see how that's feeling let's give this a label and then uh i will zoom in a little bit since i know you guys can't see it when i go full screen and flip that off and on so that's nothing crazy and that's desired i don't necessarily want to do a big crazy look i just want to get a little bit of kind of film print vibes and i like the little bit of color separation that i'm getting from tonally separating shadows from highlights and cooling shadows and warming highlights okay so i'm kind of on my way here in terms of setting up an overall look now here's the biggest thing that is missing i've talked with you guys about this a little bit before but i actually really like doing this look dev exercise solely within resolve because i know there's been a lot of questions about it and it's a really good thing to discuss you should be able to do what i'm doing right now just within resolve if you're if you're using luts or using like my products or any other trustworthy like third-party plug-in to get your look it should only be after you've done uh the homework and gotten comfortable with doing this yourself and realized the walls that you're going to bump into there and seeing the benefit of those other products otherwise you're just turning what could be a transparent and learning process into a black box that you're like okay cool like footage that looks okay comes in and footage that looks better goes out like that's great but if you're not getting better at your craft by being exposed to a tool and working with it i think that's a pretty huge compromise so that's just kind of an aside but what i was driving in a moment ago is we all at once have like uh way more than half of what this look is going to like ultimately like the the impression of it that i'm gonna get when i flip it on and off we have way more than half of that right now at the same time we don't have nearly all of it because all we're doing is manipulating the contrast of our red green and blue channels we're not doing anything interactive we're not doing anything cubic we're not actually creating a color palette we're letting all colors pass through in a technically neutral way that's great as a baseline for accuracy but that's not really a great like creative thumbprint to impart on this image in almost any image i could imagine us wanting to grade there's some benefit to be had from sculpting down what we could think of as the creative gamut of all incoming imagery into a smaller more harmonized gamut that's designed for that project and that's indeed what's going on in a film print it's a huge part of what we like about a film print because it's pumping all of this contrast and separation and sort of flavor into the image but then it's simultaneously it's being it's very much limiting the gamut of colors and confining them to a smaller palette of uh possible colors that really starts to glue things together and keep things from getting garish i mean even like this frame that we've been hanging out on for a while now like the thing that catches my eye immediately is like these red washing machines that's i don't want to see a red that bright in that primary really ever certainly not in this project so that's where our rgb mixer is going to come into play okay so let's go over to the rgb mixer and what i'm going to do here this is again kind of a limitation of our exercise today my temptation was to take you guys into fusion and show you in a 3d space what the rgb mixer does but we're not going to do that today we're just going to look at what it does on our 2d image which doesn't give us the full story but it can give us enough to understand what needs to happen here so the rgb mixer for those of us who aren't familiar with it is simply a name for a tool that we find all throughout image science and all throughout mathematics and linear algebra in general which is a three by three matrix and what the rgb mixer is saying is for a particular input pixel value which is going to have a red channel a green channel and a blue channel what we are saying is to get the output from that input how much of the input red will go into that output red how much of the input green will go into that output red and how much of the input blue will go into that output red and then we repeat the same three questions for our green channel and for our blue channel and as you would expect by default to get a unity position of nothing at all happening my red output is composed of 100 red no green and no blue and my green output is composed of 100 green no red and no blue and blue you guessed it 100 blue no red and no green so right now my rgb mixer is obviously doing nothing at all but we can do all kinds of interesting things in terms of bending around the hues of our image by introducing say some green into our red channel or even extracting some some uh red from that same channel so you can see i'm uh getting this sort of change in the character of my uh vectorscope here and i'm actually even getting a little bit of a sort of hue rotation along the way here so this is something that if you're not previewing it like in a fusion like looking at a cube type of thing you're just going to have to feel your way through to get to a good result but it's definitely something that i would do and definitely like as a starting point if i were to say like all right i really want to pull the overall amount of like red in this image quite a ways down like let's see what happens if i go to 80 of my input red and then 20 of my green that might be maybe i could do 80 of my input red and then 10 and 10 from my blue and my green you can see what's happening there on my vector scope and a good rule of thumb here is if you want to preserve your neutrals in an rgb mixer then you need the sum of each row to be one so point eight plus point one plus point one equals one just a geeky thing that doesn't like you don't have to think about all the time but it's a good good thing to be aware of so you can see i'm definitely bringing that these reds down on this and maybe i wouldn't want to go this far maybe i just want to go to like point .85 and then do i don't know 0.08 and .07 on the mixer and maybe that's starting to get me a more reasonable sort of overall gamut and like this is the fun part about look dev guys like i feel like i could talk about this for the full hour i may feel that oh i've gone too far with this and i want to back off of this adjustment or i might say hey with that gamut thing happening downstream maybe now i want to take my split toning intensity back up to 100 and see how that stuff is feeling which in this case is not necessarily driving me wild so i think i indeed will go to about 0.9 in my red and then we're going to say 0.05 for our green 0.05 for our blue 0.9 plus 0.05 plus 0.05 equals 1. so i'm preserving my neutrals but i'm just bringing down my reds i'm sculpting my gamut just a little bit on the scopes but let me go to this view here but like this to me from here to there makes a difference it feels like i'm in a bit more of a tucked in kind of world and again what i could do here with this or anything else i'm going to do in my rgb matrix is i could go to my key output gain and say just give me 65 of that gamut limitation and pull things back even more without having to endlessly fuss over that rgb mixer there so now i'm at 50 of that adjustment just a little nudge down here on the scope just a little change in the image and maybe that's all that i need last thing that i'll say about the rgb mixer that i can use either here at the look level or at the individual clip level there's another thing that the rgb mixer can do for us i'm going to do it in a new node here let's look at what happens if i have my luminance preservation turned on which i'm generally going to want to do and if i go to 0.5 red 0.5 green and .5 blue okay my exposure hasn't changed has it what has changed saturation i've got a saturation knob here so this is another thing that i could actually add to my template node graph if i wanted to kind of find the top floor for like what's the most i'm ever going to want to desaturate things and i could copy this over to my clip level side of things give this a label call it like dsat go to my key output gain zero it out for now and then let's turn that dsat operation off here at the timeline level for a moment or at the look level and if i want to play around with this now i can if i want to desat my image i didn't have to burn up one of my choices on a sat nob and i can just increase the key output gain on my dsat node like so and same thing would work in reverse if i add 1.5 red 1.5 blue 1.5 green i will have added 50 saturation to all of my channels i could save that as a separate node called sat i could set the key output gain to zero and then simply increase it to get saturation again without ever having to step out of my matrix or break out another use up another of my tool choices on that task so just another bonus of that rgb mixer there but back to our topic here let's look at this look that we were cooking up here so that's kind of a good start here and this is something that again because i don't like my normal look dev process would be to look at like all right what's happening not only on an image but on like an rgb cube and give me because i have a sense of like what a film gamut will look like on an rgb cube i can kind of aim at that as well as a particular visual reproduction on a particular one particular image at a time since i don't have that this would be a very exploratory thing that i would be continuing to refine as the process uh went on and uh just playing around with like you know how could i um sweeten my gamut at the overall look level by uh playing around with these parameters and by the way everything that i just talked about that's all a fairly color sciency way of thinking about it you can also approach all this stuff as like i'm just going to start twisting some knobs and see if i get some interesting results like that's totally valid and the last thing that i will say about the rgb mixer that i think makes things even more challenging when we are in when we don't have like the fusion 3d cube view as part of our tool kit is if i right click and i flip my gamma into linear just to give a nod to that good guess that was made earlier today i could do an rgb mixer in linear which is going to give me quite different results than i got a moment ago now just fair warning rgb mixer in linear generally doesn't like negative values for your uh any of these channels although this preserved luminance thing is going to help but generally you want to be working on the positive side when you're in a linear domain but again like we could talk at length about the sort of mathematical and sci and image science uh ramifications of what it is to work on an rgb mixer in log or in linear but you can also just like take my word for it that there's gold in these hills and if you wanted to do look dev with nothing but offset and curves and an rgb matrix you could do it uh just with these tools and in fact in my case like this is going to be my starting point i would even call this like my gamut maybe and like at this point i would feel pretty good about going to the top of my timeline starting to buzz through images and do what i sort of started doing a few minutes ago like all right let's sweeten that exposure let's go for some positive contrast by going to my key output gain and pumping that up a little bit and if i look at the net of my image here that's not half bad right on to the next one i'll probably be super lazy because this is all set in one environment and i'm going to paste the grade from this time from this prior shot and see if it holds i think it does maybe we want to go a little bit warmer here to match that previous shot like so on to the next one this one seems fine let's add a little bit of contrast in here and keep going until i get a better result and by the time i've added what i feel like is the right amount of contrast i now feel like my i would normally use my pivot since i don't have my pivot i would just go over to my offset pick things up a little bit okay now i've got a pivot let's go over here let's steal the grade from the wide see how close that gets us along the way here i'm going to start to pull references this wide frame is a great one and we'll wipe to that check this out i don't know if this one needs this warming operation or maybe just doesn't need quite as much of it so i'm going to back that off a little bit on to the next paste that grade and we're off to the races here guys and as i start to move into different points in this timeline i may decide that i want to revisit uh the behaviors of my creative gamut or i might decide here at the local shot level now that i've built that uh little saturation gizmo that we talked about a moment ago or at least i like i said i didn't save that off into all of my uh uh into my node tree and then ripple that out everywhere but let's just pretend that i did i'm going to do 0.5 0.5 0.5 set this to 0 and then like in this shot where i feel like even with my gamut control happening here i would like to attack those washing machines a little bit further and for that i would just reach for that key output gain and go a little bit further with it like that that's not bad right so i hope you guys can get a sense for like how i would start to operate on these shots now and just work my way through you know like let's look at some of these singles here let's maybe open up exposure a little bit let's go over to our decon here and decrease some contrast kind of see into the bottom of things a little bit here's kind of an opposite thing we're like now i feel like my i would want to pivot so that i'm getting uh a slightly different pivot point for that decon since i don't have that i'm just going to revisit my offset adjustment here and now i'm getting that same result in essence just using different tools so i'll probably grade up a couple more shots uh well as we're talking here today but uh i've been talking for a while and i'm sure we've got some questions stacked up so i'll take a breather here and see what we got there good dolly okay chris is asking um uh to the earlier point uh why gain in linear versus offset uh i think i know it has to do with offset additive function versus gain multiply function yeah exactly right that's just the way that the math shakes out in those two uh domains like if i want to uh it's it it it's uh just the way that that like doubling of the image works in a log versus a linear domain in linear when we multiply stuff that's the formula for uh like photometric exposure that's like the way things work out in the real world like if you think about linear light we're actually in linear light right now we don't perceive it as linear cameras don't often don't uh encode it as linear displays don't reproduce it as linear but out here and the reason we call it linear is because that's what it is like a objective measurement device would record a doubling of the amount of light in this room as a doubling in the intensity of the signal so that's if you think about what a gain function is it's a multiplication so if i wanted to apply a one-stop push in a linear domain that would be take your input and multiply it by two to the power of one so that you are getting a one-stop increase it just doesn't work quite that same way in offset when we do gain and offset you guys have all noticed before i'll break out of my desert island prison for just a moment when i'm in a uh excuse me when we do offset in log is what i meant to say just a moment ago we don't get exposure change we get contrast change so if i add gain right now i'm picking up the top but i'm not affecting the bottom of the image am i if this was in a linear domain i would be moving the entirety of the image like so what else uh alexander is asking could you not almost get every tool by changing color space on the node and using curves yep absolutely alexander very good insight and i i uh for the sake of uh sort of limiting ourselves to a reasonable portion for the hour that we have together here today i decided not to do it but like we could do all kinds of like more complex forms of what i could call gamut sculpture like i started to do with that rgb mixer as you said by moving in between color spaces color models tone curves and then essentially creating you know like a sat versus sat or like basically we could recreate any of our versus curves uh over here and then some simply by moving in between different uh color models and tone curves and using the custom curves very good insight and i've got a little bit of an answer in the chat but uh sasha was asking is the rgb mixer kind of the same as printer lights no the rgb mixer is interactive so it is we in the same way that like think about it this way that this might confuse the issue more for some of you guys but for at least some of y'all it will illuminate it any of us who know the difference between a 1d and a 3d lut that's the difference between printer lights offset and the rgb mixer so think about in a 1d lut i can do whatever i want to with the red channel and whatever i want to with the green channel and whatever i want to with the blue channel but i cannot make them talk to each other i can't say i can you know in a 1d lut if i like represent it with my custom curves here i can say oh i want my red channel to to have this shape to it as it climbs from zero up to one but what i can't say is oh at this point i would like to add a little bit of the blue channel into it i don't have any means of doing that this curves tool has no knowledge of anything but the curve that it's currently operating on i cannot make those things interact at all with an art and that's the same uh with the printer lights so i can add or subtract red green or blue but i cannot add green to my red for example or subtract blue from my green i can't do those things for that i need an rgb mixer and that's exactly what it allows me to do is it is technically applying rather than offset it is applying gain but in an interactive fashion so that i can say how much green is in my red and how much blue is in uh my red and how much green is in uh my blue and like all these different combinations again if we just look at the it's pretty intuitively laid out how much red how much green and how much blue are going to go into my new red channel that's something that we can't do with an offset all we can do is say how much of the original red how much of the original green and how much of the original blue would you like but you can't like change those contents around except by linearly bringing them up or down okay and we've got a question from gabby asking what is the difference between the highlight slider and the regular log highlights tool i seem to be prefer preserving more highlight detail when adjusting with the highlights slider oh that's a good question and i think a a fun thing that we actually haven't talked about in grade school before so the most important thing we can we can answer that in two ways the the easiest way this is a bit of a tangent but i think it's a good one to go on today if we because like a a big part of being able to navigate this challenge well is having the ability to meaningfully evaluate tools right so if i look at my good old friend the waveform let me move my face out of the way so you guys can see this thing oh wait now it's actually tucked in there perfectly great and let me just for a moment bypass my color management so i'm looking at this linear ramp okay and it looks like i've got just a little hair of my positive contrast note or something turned on here so the first objective way to evaluate that question is to say like what happens if i move what happens to the curve when i move my highlights up or down we're like okay that's what happens at least just in terms of the tonality of the image when i move that highlight knob and i'm sorry goodalli remind me we're talking about log highlight was the other highlight knob we were comparing to right okay cool log highlight if we go over to the log wheels whose little icons have changed in this whatever this most recent version of resolve is and really threw me for a loop and we take a look at these highlights very similar range right although we can notice that the transition has kind of a harder elbow to it right and of course the other benefit of this highlight wheel or that what i just identified is not a benefit i would count that as a con because it's more likely to introduce weirdness into your image but one of the benefits of the highlight wheels we can actually set this range and say no don't start don't crook that elbow until higher or lower in the chain so that's the what i just showed you guys that's the best way to evaluate pretty much any contrast or luminance or exposure adjusting tool throw up a ramp kill everything that's giving you other than a pure linear climb from zero to one and look at what the things do and compare them to each other stage a shootout essentially there's a bit of a curveball here with the highlights the reason more than likely that you are perceiving that you are recovering or preserving more detail with your highlights is there's actually a texture operator going on inside of that highlights knob so as you bring highlights in you are adding mid-tone detail is as well so that's can be a reason to use or to avoid the highlights knob i generally since i know that that's in there prefer to like do my tone mapping and my highlight recovery or remapping or whatever i'm doing in a clean uh more universal operator and then if i want to try to pump some texture into the upper end of the image do that in a separate layer so that i can control those pieces independently that's just a personal preference but it is very helpful to know in addition to what range am i affecting with the highlights and how soft is the fall off into it having no influence at all but also to know that that highlight knob has an effect not only on the tonality of your image but on the spatial properties on the texture of the image as well which like i said is probably why you are perceiving a better job at highlight recovery with the highlight wheel because if you think about it if you pump whatever available texture is into that area as you're pulling your highlights in the idea the logic behind that behavior is that you will get the perception of additional or recovered detail that was previously clipped you're not actually doing that of course because you're just sharpening what was already there but it can lead to pleasing results if that ratio of sharpening to reduction works for you in that stock ratio let me know yeah sorry go ahead yeah alexander uh is wondering how many episodes are you doing of masters of color there's going to be eight of those and then uh i i don't know what the future holds after that i would love to do more of them there's definitely more than eight colorists and filmmakers that i'm uh keen to spend that time with so maybe we'll get to do some more time will tell i'll keep you guys posted um and if we're kind of caught up on how we doing on questions do we have more about the the desert island challenge here good dolly are we getting kind of caught up i think that's what we've got for now okay great great yeah so i'm going to keep grading and we'll see what comes up along the way so oh yeah i turned off all of my i was like wait where did all my pretty look go i forgot i turned it off here so let's just keep flipping through and grading some shots now let's talk about like if i could get one more what would i pick next if i could get a fourth what i would pick next in this situation for sure would be something spatial like if we look at a frame like this where like it's a really strong composition i really like the the headroom and the the way that the characters are weighted in this frame i definitely would want to take some heat off of this portion of the image here in a way that i just don't you know i can't make there's i didn't select anything that allows me to make spatial control here if i had this would be you know like in this very first pass of the uh project this would definitely be where i would just take a circular power window flatten it out to like a horizontal grad go super super soft and then do either an offset or if i always wanted to do a linear thing a linear gain and just kind of like you know treat this like a flag and walk it in try to feel like i'm not whacking my talent too much there and you know that would be my my net on this guy if i had that tool available you can see boy that just does wonderful things so turns out three tools is not enough to get all the magic we normally would want but i'm going to stick to my rules for now and keep cruising through here that's really the only thing that i see in this image and you know that like like i was saying guys at the beginning of our time together limitations and necessity can be uh really really great catalyst for creativity i might in the absence of that just kind of bias my exposure a bit lower without like letting my talent completely sink into shadows and that's not a bad compromise i like would have used the window of course if i had it but that's not a bad compromise to me to just kind of bring that in a little bit this side we need to go a little bit warmer right let me get my vectorscope back up down here i've always got one eye on my scopes and one eye on the image itself and i'm always flipping off and on on say a balance operation like this to see where i started and where i'm netting out like so and a lot of the time i'll just kind of like spin stuff and see try things a few different ways and see if i can't get a better alignment that's starting to feel pretty decent that's a you know just side note matching wise it's a tough match because the frame contents are changing radically there are very few common elements in these frames the floor is not common washing machines are a different color there's only a little bit of wall here that there is a lot of here really the only common sort of hues are my talent's wardrobe and uh you know like whatever i'm seeing of that wall there so there's just a good side note about like sometimes it can feel like oh man that match isn't great like that's it still bumps my eye these still trip me up when i'm grading we want to do our best the best we can to make them feel contiguous but like they're never going to feel super flowy because they share very few overlapping frame contents um and the best thing we can do in this situation is actually to go back to the histogram and kind of make sure that our uh overall like signal intensity is somewhat aligned which it is so we got that these are like uh v effect shots and little edit cut down so i would like grade this shot once the way i did over here and then i would just shift-click all these and then middle-click to apply that same matching grade on these i don't see that being a shot we have to go through many rounds of notes on i'm going to ripple that same adjustment that i made before i've got to say like i'm really happy with that that dsat tool that we built here let's even go ahead and label it let's go over here same thing we just pull in on our exposure again if i i would love to have some control and be able to knock this down up here just a little bit just pinch it in so we're kind of gently nudging the eye more toward our hero but i can live with this here maybe i'll back that off just a little bit and by the way like there's a we can be reminded as i go through this that like it's not an accident that i'm starting with exposure and then i have balance and i'm often not even moving my balance exposure is more important than balance our eyes are more sensitive to changes in luminance and contrast than they are to changes in color i'm not saying color is unimportant color matters of course but my priority is always going to be on exposure because that's going to be the where the biggest difference is going to be made to uh the end viewer if i can get that right and if we look here you know like i've i once again like this is more just a general grading thing but once i've got my hero here i'm going to keep referencing back to that for any frame where it makes sense and a lot of the time i'll just look at it and be like i wonder what if i steal that what if i middle click that wide and just see how that does for me that's not bad we just need to bring exposure further in like so and then so on and so forth if i look at this side it can probably match that same level of exposure and i think this film is starting to look pretty decent uh with the desert island challenge that we've undertaken here and you guys can see let's see how many shots there are total in here okay so there's 160 a lot of those are some of those are like those repeated clock shots that won't necessarily take same same amount of time as a normal shot but i would feel very comfortable ripping through the rest of this timeline in this course of the next maybe hour you and limiting myself to the tools that i have yeah i would probably most often uh bristle it like oh man i need a power window i need another power window but other than that i'd feel pretty comfortable playing around with this thing and i might hit some instances where i'm like oh i really want to desaturate uh my reds without affecting the rest of the image so much but like one of the other like lessons that i think we can take away from this desert island challenge is to kind of like wrap us up for the day is it really shows us the power of broad tools and of forcing ourselves to think more broadly and simply for as long as possible like i hope every one of you guys have been watching closely for the last uh hour or so and noticed me land on a shot and had an insight or an observation of like what does that shot need and then watched me solve it in a broader way than you would have done if you were not doing a desert island challenge and just uh doing your normal grading process a great example of that might be here if you have that same sort of like hit that i do like ah that red's a little dominating back there you would probably your go-to if you're like most colorists your go-to would be like a qualifier or you know like a loom versus sat or a hue versus sat if you're more in my school of thought of trying to do things more softly and gently and in a uh you know like that kind of fashion but i just want you to notice like this actually did a really good job like the only dsat that i notice here like if you were my client i wouldn't do this of course but if you were my client and i told you oh i'm going to qualify the reds and desaturate them you would not like disbelieve me if i flipped this adjustment on and off and told you that's what was happening so there's this really good lesson hiding in there that like when we are looking for complexity and we are making the base assumption that complexity is what's needed that assumption proves correct right and when we are looking for simplicity and we are making the assumption that simplicity is usually what's needed that assumption also proves correct so it really comes to like what lens do you choose to look at things through and that's why i choose to look at things through the lens of simplicity because it's faster and it's never going to break i don't have a qualifier that i have to babysit here this is going to travel this is the ideal way if i can do it in this way this is ideal because it's big and broad and soft and it's doing less to the image than any other approach i could potentially take so there's some really good lessons here like i said with the exercise of limiting yourselves to a desert island challenge like this it will show you more power and more and set your creativity free of like wow look at all the fun stuff i was able to do and make out of my curves node and my matrix and i'm using my matrix to sculpt a gamut and to pull or add saturation and i'm using my offset to manipulate exposure as well as balance like you can use all these tools in more ways than you've probably thought of before so i would encourage you like on your first pass of your next project give it a try just try a desert island challenge and even if you know that okay i'm gonna have to deliver this at some point and the job deserves my best work and it doesn't deserve deserve an artificial constraint so i'm gonna have to remove that constraint at some point it's still valuable to say hey for the next 20 minutes or the next 20 shots or for this pass i'm going to limit myself to these tools because even if you're eventually going to take the leash off along the way you will probably make some pretty interesting discoveries about what works and what you like and the way you grade that you wouldn't have otherwise made so that's my sort of case for doing your own desert island grading challenge whenever the opportunity arises uh and i can take another question or two before we wrap up for the day if we have anything out there yeah got a couple quick ones um first from chris asking how did you quickly bypass your colored men the one time you were doing that uh so i just right clicked on the clip i was working on the the thumbnail for it and i said bypass color management now is that showing up please tell me my contextual menus show up not seeing that man uh okay so if you my highlighted thumbnail thumbnail number 24 there if i right click on that let me see can i potentially get this to appear i just feel like talk is cheap for these things oh it's not even going to let me do it you'll just have to take it sight unscene from me right click on a thumbnail and toward the bottom about two-thirds of the way down the context menu that will come up there's an option to bypass color management and that will flip everything off all right and ec sanka is asking about some advice on working with versions and also um how to deal with them because they reset uh when you're going between mobile and remote yeah i mean i i i'll put it this way i see the utility of versions in a workflow they are not something that i use and maybe that's in part due to that observation that you just made that like oh they reset in a non-intuitive way when you're flipping between local and remote which i do many times in any given project so i i use the hell out of remote and local grades i will like the other thing that i will i will do in lieu of what i see some people using versions for is like in this project for example in any project after i've finished this first pass the very first thing i do before i kick out this uh quicktime for viewing assuming that there's going to be like a a review link on frame i o or something i'm going to make a stills gallery and i'm going to call it first pass and i'm going to right click i'm going to say grab all stills from first frame and let those populate in and no matter what i won't touch this album except to refer back to it and see where i was and what i did and then going forward i don't need to up version my clip versions because i have a saved instance in time of where we were at when i kicked out that viewing so that's obviously not the only reason why you might want to use versions but that's a common thing people use versions for that i prefer this workflow for just because it's easy for me to easier for me to track what i'm doing than when i'm working with versions all right and last quick one uh jamie's asking about thoughts on texture pop of x who likes the idea but has been getting mixed results i'm with you jamie i i i like the idea like i i uh it's it's on my my list of like tools to contemplate of like you know we can think of the most organic way for me to think of like sharpness or softness in our image and uh you know sharpness in the rough and broad all the way up to the like tight and uh tiny details of an image is in the form of an mtf modular transfer function curve so i've actually had the thought in mind for a while like boy wouldn't it be cool to have to be able to shape an mtf curve as a tool inside of resolve so if i uh manage to invent that in my uh diminishing spare time then i'll let you know but i'm with you for sure like the the base light texture tool i feel like uh even though the form is so similar between them base light gets a little bit better texture you can get optimal results if you make a simple adjustment and do it in a linear domain you're going to get better results just a quick tip there but i am definitely with you i i've had uh i i use it very very sparingly right now because i think it's easy to make strange looking images with it unfortunately and that'll do it for today guys i hope you liked the desert island grading challenge i feel like i went on a couple tangents and uh meandered around uh this topic a little bit but i hope we filled this hour with uh interesting uh like thought-provoking stuff and you've picked up a couple of ideas uh from our time together if nothing else go do your own desert island grading challenge and uh let me know in you know comments on this video if you learned something from it i learned a ton from i kind of was thinking about this all week to be honest with you guys it's like wow what would those be and these were not the first three that i thought of but i think i really could i could live and die by these tools uh pretty well and i learned a ton just from asking the question i knew and forcing myself into the limitation here in the grade with you guys today so i hope you have that same experience the next time you choose to make your uh own desert island grading challenge thank you guys all for joining me very much thanks to gdali for co-hosting with me i hope you guys all have an awesome weekend by the way we are going to be on our normal friday schedule for the holiday next week and the week prior um i know we'll all be hopefully on holiday and enjoying ourselves but i'd love hanging out with you guys and i'll be here at home anyway so i figured we'll put it together and uh if you're too busy doing family stuff to make it then we'll see you in the new year if you can join then i think it'll uh probably be a different sort of uh grade school vibe than we normally have and maybe it'll be fun so i hope to see you guys next week and on new year's eve if not see you in the new year when we have so much fun stuff coming that i know i keep teeing up but i'm not just bsing you we got an e-book coming out i've got all new looks coming out that are going to be uh sort of like set up in a different way than my colloid premium products that i think a lot of you guys are going to find really interesting a bunch of really new stuff exciting things coming out in 2022 at the top of that list as well it's just going to be a lot of new content for the channel here itself so i'm really excited to share what's coming with you guys uh in just a couple short weeks because this year's almost done man uh it's been a wild ride for me i hope it has been a great year for you guys as well everyone have a great day and i will see you all all very soon take care

Channel: Cullen Kelly

Views: 534

Keywords:

Id: k7qT3skmSFs

Length: 69min 15sec (4155 seconds)

Published: Fri Dec 17 2021