 Welcome, everybody, to another exciting episode of Logic Live. My name is Andy. I'll be your host for today as I'm the host every day. Today, we are going to have a fantastic session. We're going to be covering CG compositing and flame plus color management with our good friend, John Ashby. Before we get rolling with the slideshow here, there is a poll that we've put up on the meeting today. So if you go to the bottom of your Zoom interface there, you'll see a button that says polling. And we just kind of want to gauge everybody's exposure or, you know, comfort level with color management to kind of gauge how we go today. So please fire that up and enter your answer into the poll. Oh, good. Thank you, Quinn. Appreciate it. All right, we know that color management is a wide ranging topic, and we're going to try to tailor it to the audience level for today. Let's get underway. Logic Live is proudly sponsored by our friends at Cinesis Oceana. Cinesis has been my reseller for about 15 years. I love these guys in addition to taking care of us lively. They've always been great supporters of the logic community. They host the user groups, the flame user groups in New York and Chicago, Toronto, Atlanta, Dallas, and they've always supported our one frame of white contest. They always sponsored prizes for the parties that we've had. So huge thanks to Cinesis Oceana. Cinesis provides solutions to keep teams connected and working. Find out more about the remote workflow solutions at Cinesis.io. Cinesis Oceana supporting flame artists since 1997. All right, I'm sure you all saw on Facebook, we had a fantastic first logic fest, logic fest 2020. And with 13 entries and and thank you everyone who contributed thank you everybody who voted. There were five, five entries that got the most votes, and you can see them all on logic you can see them on logic dot TV. And of course, the five, five winners there each received a one year license of mocha and sapphire courtesy of our good friends at Boris effects. I hope you all had a chance to check out their live streams their virtual NAB this week. They had they showed off new versions of mocha and then of course, silhouette paint, which was fantastic so definitely head over to Boris effects.com check out their offerings if you haven't yet, and I want to thank them for their continued support. And this Tuesday, April 28 at 1pm Eastern time is the auto desk up close and personal with flame webinar. We already have a ton of people signed up it's going to be fantastic. Please if you haven't already registered, go to the logic dot TV and click on the events page and sign up. Will Harris and the rest of the flame, the flame development team are going to show off what's new in flame 2021. The five winners from logic fest are all going to be presented some are going to present live. Others are going to show video presentations but it's going to be a fantastic webinar, lots to learn. Lots to see and enjoy so definitely sign up and tune in on Tuesday, April 28 at 1pm. All right, let's get started with the main event I'm going to stop sharing my screen and bring up everybody's friend. John Ashby, if you don't know john you should. He's always been a great contributor on logic. He's an absolute visual effects wizard. He's the master of everything. He's also a good friend, and it's always been very generous with with his time and with his knowledge when I've needed help with stuff. So I want to welcome john Ashby to logic live how's it going john. Hi everyone. That was a bit of an intro. So should I call out the poll. Thank you for filling out the poll. It gives us a good idea. I guess of what spectrum. Ha, we're dealing with. Hey, what gamut you're dealing. And. Yeah, so there were some brave people on there who said what's color management. And that's awesome. I think we get the idea where the majority lies here and that's pretty cool. So it's really a trick question because I'm not going to change the presentation regarding results but it is pretty good to see this because at the same time, I think there's going to be something for everyone in here. So, without further ado, should we get started. Do it. So color management and see if you can positive thing in the plane. And I guess how they're related. So, this is a disclaimer beforehand. I think it's important to let everyone know that this is my own opinion. It's based on my own experience my own way of working. It's definitely not the absolute way to work you can always find a better way and technology will always improve and you reserve, we all deserve the right to change our opinion in the future as new technologies and workplace become available. As always, perform your own independent research. This is the first safe harbor state that we've ever had on a logic live I'm very excited. Well I feel like a lot of this stuff especially when we deal with things like color is so much of it is subjective and like what is a good image what is a pretty image and you're going to do all this work and it's great and the client's going to say needs more yellow There's no yellow in this question. It's just the way it goes. We've all been in the day. Yeah, the client is the one who makes it in the end and you're like, well, it's not as good as version nine but. So, first and foremost, I guess we've all heard this or thought this. Why do we need to work like this. It's so much more complicated and everything work fine before. And that is anonymous frame artist. We could do another phone to see how many people can relate to this statement. They will remain nameless. Yes, it did. It did work before. Kind of. And lots of films did get made before. And we also used to work at one K or even SD resolutions. And now the world is different technology evolves. And the thing is, it's like, I just bought a TV from the $2,000. This television is ridiculous. Thanks to the flame community for suggesting the. But I really wish we'd gotten a kickback from LG. Everybody come logic recommended that even I bought it. Well, I know. And that's the thing is that we're all even like my brother in law in Florida, they have a 75 inch HDR television is that you can buy this stuff. We wrap up the pace of technology now is is beyond what we're kind of used to keeping up with. So again, I want to learn about CG compositing. What does this got to do with color management. And to our implicitly linked. And I guess the analogy that I'm going to use is that vegetables, the color management for CG composite also known as cake. You'll eat your greens and you can appreciate the cake more or you can look at it as that you're not going to get your cake until you get your greens. That doesn't apply to people who have been locked in their house for 58 days. By the way, that's not the quarantine diet as far as vegetables right now. Let's talk about color. We've all seen this. This is a chromatic diagram and it's a 2D version of a 3D version, which you realize that color is actually a three representation and we've seen things like scopes and RGB waveforms as well as people raise as ways of representing color on the screen and then you've got more abstract concepts of color of the idea of complimentary colors and primary tertiary analogous. You know, it's we're getting the hang of all this and then all of a sudden HDR comes along and then there's more kinds of HDR and then HDR and don't be vision and you think like, OK, well, I've got to get the hang of this. We've got the audio desk help page I'm going to see and we've got, OK, what is the color space and it's like color space allows people with software to communicate colors and ambiguity using a numeric representation, which isn't helpful. So you're like, OK, well, I'm just going to open flame and just go for it and you're the color management page. You're like, oh my God, what's going on. And then you open up the color management note and you're like, oh no, this is crazy. I need to do some more research here. You go online and you find out what all these cameras do and you were like all these cameras, great colors that are all they all have a completely different way of representing what they're recording. And you realize then you start looking for this thing's like all possible colors that you never thought were things. So then you get there and someone says, well, you know, you should look at ASUS ASUS is a great way of explaining this. So you sort of do some more research like this and then someone at your office says, hey, can you explain this to me? And it's basically, this is you at a certain point. So it's like you, it just becomes a bit too much. I think that's what I'm getting to. And we did this to ourselves. That's always the thing that you know color is actually quite complicated. That's the reality. And it's complicated to understand. And it's complicated to explain. And as the good news is that as the effects artists, we do not need to know or totally understand everything about. I understand probably about 10% about but I understand the important 10% that matters to us. Yeah, let's get on to that. So color gamut and why they went up. Basically, modern cameras are capturing far more data and good fit in the rec 709 color gamut. So this is taken from the ASUS page in terms of taking from the color science. But as you can see here, you've got rec 709 at the bottom. And that's all of the colors of the visible spectrum. And then you have P3, which is more rec 2029% we're getting up there. And the Sony and Alexa. And then we have ASUS, which is a color space designed to cover everything as well. So terms like gamut gamuts, one deluxe 3D lutz. And the way to see in linear was to de gamut the image, which is also known as a 1D lutz. Locked to Lynn or SDV sRGB rec 709 to linear version. We've just taken the camera. This is fine as long as you're coming from it 709 and staying in rec 709. The newer cameras are coming with color data that's greater than rec 709 in color. So we need a way of getting that color data into a smaller container and the way that keeps those colors. So we've got examples like 3D lutz where you have like log C to ASUS CG. And I'll cover this better in a way to explain that just now. What about these different color spaces. Hey, John, sorry, just just just one quick thing. It's, I think it's just important to point out to everybody that working in scene linear is the kind of Holy Grail of compositing and the, the, especially when compositing CG. And the challenge has always been how to get from what was shot to scene linear. Am I correct. And yeah, and I'll show. Well, there's a couple of slides of the nature of that, but I'll definitely show that in the plane. So a system to deal with all these color spaces. It's quite simply, you can just look at ABL or always be linear. So that's it just being linear space. It just takes care of everything else for you. All the transforms have been done for you. That's all you need to worry about. It's simplistic, but that's where I do it. So here you've got a workflow of how linear works with ASUS you start here. Excuse my mouse pointer. Yeah. Yeah, so you have to input color space, which is your camera color space. This is coming from maybe ARI log C, red logs, sony log, et cetera. You then transform this image, this data into a really wide color gamut, which is ASUS. This is sort of just the one that covers everything. And then from ASUS you go to ASUS CG, which is our working space, or you can go and send me an X709 or ARI linear white gamut. There are lots of different flavors of linear, but I'll explain why. You should stick with ASUS CG. And then on the output end, you just go from ASUS, you bounce that out to ASUS and then from ASUS to X709 or P3, about 20 years. The idea being that your transfer out of this is that everyone's dealing with, you're able to have a chain of responsibility of color. So you're working in seeing linear and you're seeing the result on your screen that you know kind of what it's going to look like. So if it then gets put into all these other color spaces, these are just simple trim passes that colors are doing to get to these different outputs. So how does this relate to CG composing, which I think was your question. So first and foremost, light is additive. Okay, physically based renderers, Arnold, D. Ray, Mantra, Renderman, they all look to replicate real life. So in CG terms that means you take your lighting, you add to your reflection, you add to the refraction, you add to the specular and you get your final image. Simple, right. So of course there was an aspect there, there's always exceptions. So tracking seems to work better in video space still on plain. It will change eventually, but there are always things that text module as well. I don't know where that's intact, but it wasn't. Random tab. Anyway, the text module, if you switch into, if you're in color space, when you're going to edit it, it will go dark. It's because it has the older interface on there. We all know that. Keying log, you can sometimes get better find it, hair detail in log. And I would say that in Keying, all bets are off. If you're finding a better key in the chroma channel, then just use that. Yeah. It was once explained to me, I think it was our last week's guest Alan Latteri explained to me that some of these tools are designed to work in the, almost like in the same space that you see them in, you know, so like your motion vector tracking or the motion vector won't work necessarily in scene linear. So you can flip it over to rec 709 to video to what you would normally see. That's how the algorithm is is engineered. And that's why you get better results for some of these tools. Exactly. And the thing is, it's like, the great part about this is you can just have a separate chain that goes off, putting it into rec 709. Do you're tracking there and then just blow away that and keep tracking it back in your other action or whatever. So it's not, it's no big deal you get used to it. It's really great doing this is a fantastic slide because I know there are some people myself included when I first started working in scene linear, and the tracking didn't work I was just like well the hell with this I'm going to go back. And the answer is that there's, that's why I think it's a great slide. Yeah, there's no rules, just do what you need. But the thing is, it's like in scene linear, you can deal with some really crazy value. So there's never going to be necessarily an algorithm that works for everything. So sometimes you just got to know where gremlins are going to be lurking. Yeah, scaling, sorry, image scaling can be clean ass in log. If you do, if we have time to do a test, you can do it yourself. Take my word for it. What about viewing spaces. So a viewing space is just a chain of transforms that convert from the working space. I mean you can work in linear, and then your monitor is representing the final image, or as a variety of the final image, but the great thing is that you can work in scene linear, your computer's in scene linear, but you're looking at it as if it's rec 709 or whatever and you're, you're parrying on as if you're working before you don't really notice the difference. So an example of that would be going from ASUS CG to ASUS to HCU rec 709, or in a film pipeline, it could be something like this, but you're going from ASUS CG to ASUS, ASUS back to log C, log C where you're applying the shot luck and then after that show up. And yeah, it sounds hectic at that point. But if you think about this part, this is the bit that's happened. The reverse of this is what happened on set. So they've done their CDL is a sort of more non-destructive transformer. I would look at it like a sort of color correction where they're balancing things that they're quite useful CDLs because they're easy to pass around versus a lot which can contain a lot of crazy information where CDLs can be very lightweight. And yeah, and then the show luck can be anything. It can be a variation of log to rec 709 or it can be a sort of film effect luck or whatever, but it's generally what they were looking at through onset. So yeah, it's a lot of luck. We're still considering legacy CD linear. Do we really need to consider using ASUS? So I guess short answer, yes. Long answer, it depends. There are certain age cases, and it's important to know when and how to fix them. And that's kind of all you need to know, I think. So yeah, or management's like the English language, you know, there are a million exceptions to the rules. Well, yeah, there are some examples saying that like all the English language, like there's like 1200 words of being used in the New York Times over a spectrum of months or something like that. So, yeah, stop it with the color space analogies. That'll be our next contest will be the best color space analogy. So I'm just going to stop the share a second and get into playing. Meanwhile, there are lots of things in the chat. I haven't been following those. Everyone says you look fantastic. You have wonderful depth of field. There was just one about shrinking actually just the Alex did that help. I think that helped bring the chat window up. It's gone. All right, cool. Thank you. That was the voice of Alex Arce. Thanks for being credited for this at the end. If you'd like to book him for voiceover work he is available. His agents information will be in the chat. Sorry, go back to you, John. Anyway, yeah, so on with the cake, linear curve, sorry linear gradient, we've all seen this hilarious. We've never seen it through zooms compression notes. It's like two bit color. So if you're taking your linear image that comes in, and you're just converting it. And for the sake of understanding off the viewing lines, there's no other things happening. So if you're working in video. Like using the legacy the video method that was, what was in our claim. If you do things on master grade after that, you can see what's happening is that it's really like we're doing a game which is a straight multiply. Everything flying off at the top, but you can see it's going to be really hard to control this part of the image. This is I'm in linear. And I've got these two mimic links. But you can, you can see the controls a lot more manageable here. So if I just want to brighten the image slightly. You have to control stuff whereas if in video, that same those same values give a completely different image. You know, we're going to just make it obvious 1.4 1.2. Okay, in video if I if I'm in video that's what I get. If I'm in linear. That's what I get. So is and the curves comparing the two. One of them is all over the place and one of them is a lot more under control. So that's the basics of again, not particularly practical example, but here is one I promise you a practical example. Here you go. So what have we got here this is again bypassed so there's no funny business. This is a high dynamic range. It was taken with my DSLR. It's just one of the faces for a spiritual image. You can see some numbers here are nuts. We've got up here. So it's 50 somewhere up here. There's 10, 10, five. So all these, and if we were in old rec 709, we'd be looking at essentially everything gets clamped above here on the way out. So we're going to we know it's going to get clamped to make the point. So everything's there clamped and, you know, this is now everything's there ones, but we've lost all that detail. And of course you can come in here with a grade, and you can try and get all this back. And you can say, okay, well, you know, we've never sent that out like that. The reality is we bring this all the way down and you can see how far we've got to get down to be able to see these buildings and you see how saturated the buildings are coming down. And so we've got a lot of conversational gathering up, but we're now getting just washed out tone mapped image kind of thing. And so now I'm going to come back to the lift and I'm going to bring that in. You know, and all of a sudden you're fighting so hard to keep some sort of image happening here and like, yeah, we've kind of got there that it just looks at a certain point you know shadows now getting crushed. If we do an ASIS transform where we're going from linear x709 to ASIS, like I showed in that diagram, an ASIS HD video on the way out. So here, for example, it's taking care for me, like if all these values just in that single model, all these values are back in spec and we've got a beautiful sort of image here and it's all there. We've come from I've done nothing by taking this out to prove a point. And I'm always a fan if there's a short on short ways to final take it, you know, oh sure. So, does that make sense? Oh, yeah, totally. I mean, I'm sure that every single artist on here has has gone down the path of trying to like brute force, you know, color correct something to get it look right to get it to look right in this weird like color environment that I have to deal with like a client's giving me a piece of a piece of artwork to use as a logo, or a sign that I have to put on a you know on the on the bus, whatever that thing they call them now the link NYC if we had to replace that screen you were given, you know, like a PNG or a JPEG or something. We've all gone down the path of trying to brute force color correct that to make it look right in this color space as opposed to converting it into the proper color space. And the great thing is here you've got, again, a chain of responsibility, because this you knew you matched to this like this was our color this this was our viewing transform which I'll show you in a second, which is how my viewing transform set up. I'm looking up. Well, it's not specific, but I'll show you what I do. And that is that I can do all my color correction inside of here. You know, look, while looking at my viewing line. And then I just turn it on exactly what's doing my color management and I'll just do. I think actually set up here. Yeah. So, yeah, I'm just going from seeing the New York 79 to ASIS ASIS to ASIS CG. So I've got my image here and I'm coming in the master grade. And I know I'm just going to maybe add some contrast and a bit of saturation for hell of it. And I know that I'm still giving these ridiculous values, I haven't lost anything. But I know that whoever does this on the other end can, and in this instance it's going to be ASIS CG to ASIS ASIS to HD video. John, there's a question from the chat. Are you are you working in a legacy project. Yeah. Okay, so the idea of a legacy project, see linear project, I think it's a little bit confusing. It doesn't matter what you're working in the, and I'll show at the end of this time how I have my project set up. Just because it's legacy or linear in here color management. So that's how I have my project set up that have none of the stuff. It's very simple. I have something the log and just because I want to see log like a log image. And this is the thing about viewing rules is that actually viewing rules can kind of catch you out. The legacy project we'll have is this set to unknown, I think action set to unknown. This is all the same. And it will be a long list of stuff. So I generally delete all that and just keep my default, which is my ASIS CG to ASIS ASIS to HD video. And then I'll have, yeah, it keeps over the normals just because we used to seeing normals like that kind of pale blue thing for the map and something for video. And yeah, that's it. Thanks. So in here, for example, with the, I guess the concept of the legacy project, you know everything's getting tagged as unknown and stuff so you can get, I think it's better. Um, I like to know, I like to control what I'm seeing. I like to know that I should be seeing what I'm seeing and if something looks wrong, but I want to be able to catch it. So I guess a good instance of that is, yeah, if you have a color management and you convert something to log here and just do I'm just for the sake of it, I'm going to do just simply log to Lynn or Lynn to log classic. I want it to look like log. You know, I want to see it and I said I see that and my viewing much sharing it and I've got here and I tagged it as log. That's where it is. I want it to look like log because I don't want to accidentally have this being represented as a wreck 709 image on my display and be working in a lot. I don't want that cat called out like that. I want to see that and go, Oh, something's wrong. Does that make sense. Totally. Yeah, so gremlins to be aware of they're always gremlins. So let's take a very colorful image. This is, again, I'm going to turn the viewing law just to make sure you're seeing what the screen is so or what the data is so here I've got some very saturated colors and you could see some of these lights or something and then we're going up to one here up to one here up to one blue here and red is saturated. So just going to create a pseudo log C image. So we're pretending it's from a log camera, an area log C camera here so we've got all this is now much flatter. So if I go from that, using what we just talked about in ASIS of going log C to ASIS, ASIS to ASIS CG, you can see some crazy stuff happening. So, for example, here we're going over one. On the green negatives, we've got negatives on the green and the red here and we've got red seems to have survived but what you'll see in if this is a real image this would manifest itself in kind of like weird halo around things like exit lines or Yeah, at least likes and they become a tech fix essentially that comes back to the effects to fix. So we just have to know how to deal with that and know that we are not going to have multiple flavors of linear to work in because I know you can work in area log C or you work is very wide gamma area but it's not really practical to have all your textures. All set up a log C pipeline or sorry a wide gamut pipeline also up for an ASIS CG pipeline also up for a scene linear and so on you just want to have everything in one unified color space and know that when you have these problems know to deal with them. So what's interesting is if you go from log C into scene linear area, we don't have these negative values appearing. It's able to contain them. So what's a good approach is to when you get one of those shots just ask for the log C plate or the red red log plate and just paint through the the bit so do the transform in one and transform on the other the ACCG and then just paint through the bit in the camera space. And that's how we do it. Good tip. So why custom. Yeah, so you may have noticed that I always seem to use color transform and trust him. So, again, gremlins, because if you use things like the view transform that like instant classification they do take care of the tagging for you, you can go from log C to into video but if you look at the transform some of them do some stuff that I'm not doing like it's going to XYZ here. And so again, chain of accountability, you know that you want to be able to hand off an image or work with a set of lots that everybody else is using from new to resolve to my so if everybody's using the same LUT chains then things are just going to work. So in this instance, I've heard it go from in this instance you ask log C to ACIS to HD video just because I know what it's doing and at the end I'm taking it and again it's as a slight price to pay of convenience but the control is that your image doesn't get screwed. Yeah, I think it's a really good point to make is that although there are other ways to do that conversion in in flame if you go the ACIS route, it's going to look the same for anybody else in any other application in any other department that is also working in the ACIS world. So if you're if all if your stuff is going back and forth to nuke or back and forth to resolve. You're kind of like you said Maya, you're guaranteed that they're going to see the same thing on their side, which is vital is the last position you want to be in is to be that guy in the chain who screwed up the color. So that's it. You know, it's embarrassing and it's something you will do but no facility wants to be the one who has to explain to the function that they've got to re-render all the shots with colors from. Correct. And that's and it's something that gets caught very late because all of a sudden everyone just expects them to work with the LUTs. So especially on a big show where they're going to color correct hundreds and hundreds of shots. They're going to drop yours in and drop somebody else's in and then they just work out or hang on a second. We have got this LUTs that's working for these shots and these ones aren't working anymore. And then they're going to find out this because Muppets decided to work in a clamped REC709 space and totally. But it's also, I think it's a great answer to the question, why ACEs? Well, because it's universal, it's again. Yeah, I think it's accountability. I think it's color accountability. That's what I like about it. And again, I can show the example here as well is that here I'm going, I have a TG aircraft. Is this cake? Are we re-entering? Are we starting our dessert phase right now? Well, that was like a taster of the cake and now we can get like a little bit more cake and then we can get to the final bit of cake. Cake overdose. And then so, yeah. The chat room is ready. They're saying forks at the ready because people have been trapped in their houses for 58 days. Is anyone like asleep yet? I hope we're keeping. It's difficult. You've got to keep people. You're sitting at the table, man. What's great is I just realized Doug Walker is here and I'm curious to see what he voted for in the poll. Oh, no pressure. Yes. Well, hopefully he missed the first one. No, he said he's agreeing with things. Yoda, he is. Yes, he agrees that Doug Walker is his personal Yoda. Thank you, Doug. I think with all this, the thing is the point, if there is one takeaway from this is it is super complicated and nobody expects you to know everything about it. But I think what all you need to know is how to get yourself into trouble and how to get yourself out of trouble and nowhere trouble is going to look because, you know, we are working with some crazy value sometimes in CG. And it's just because in cameras, normal camera world is its own thing. But when we start converting from computer linear generated images, because you can, we have infinite power obviously in the computer. Some of these HDRs can be monstrous. So we really need to know where to problem spot this stuff. So here I'm going from. Again, I think I'm just doing the, the example of old. So we have our tagged images linear. And what we want to do is here we've got our viewing luck. Which is basically I'm going to turn off basically going from linear to rec 709 and then clamping the output, because this is what's going to have to happen when it gets into commercials. So here, I know I want to make this color this image work and look at my look at through the context that's the word I was looking for. Now, here we know what we want to do. Obviously, we can see it's a bit flat. We want to add, we want to make the reflection really bright. But then we know we've got to add more contrast to it. And, you know, what we end up what ends up happening is we get just to get our values down first, because we are up. I just do this one. Just to show you how far up we are. We're above one, above one, so we still want to come down. And yeah, now we can get this somewhere and we still now lose contrast and, you know, and then this starts happening and starts looking dirty around here because what too much contrast and you're really fighting against keeping something in one to zero. And also what we don't want to happen is we do a bunch of color correction that we know that it just gets clamped in the color suite and we're like what do you do in my image. That's not what I made. And we've all been there. So what I really like about ACES is that, you know, we've got everything is under control. Yeah, so what you're showing here is the same exact source image. There's linear EXRs from CG, but just converted to ACES and from ACES to REC709. Yeah, exactly. So if I just come in here now, and what's interesting is you can show, I'm just going to. Hello, this is a lesson lesson. Yeah, you don't want to make sure that when you're, you want to be careful to put the, the map in because you don't want this to happen. Yeah, when you're color correcting that's going to cause trouble because it's a pre-multiply image. So just make sure you do that. Woo, lesson kids. All right, so here we've got the great thing is here where you can come in here to our, I've got my viewing luck on here, which is here. It's the same thing. And I can really push this all the way up. And I'm still, I come in and bypass that. We're still below one and we crank this up to, I mean, we're up to 0.9. So again, caveats, sometimes you want to blow things out, but it makes it a lot harder to do it in HDR, but you're also going to burn people's eyes out. So if you care about that. All right, more fun cake. Let's talk about what that meant earlier of the lighting. So here, I'm put my lot on work down inside, brought in my image. I've converted it into ASIS to my, my beauty pass. That's the basic render that we've got. And then here's just an individual pass. The reflection. And yet you can see the HDR used for the reflection. The whole runway. And we have a specular or a fraction and the specular refraction is just going to be those little bits of glass that you have in there. So, yeah, if you take your lighting, you add on your reflection, you add on your reflection and reflection, a specular, sorry, you're going to have your final image. So what's great about that is that if we want to do. If we want to do what we talked about earlier, where we're saying, okay, so we really want this reflection to stand out, we really want to see this. I think it's called a belt line, but it's very synonymous with when they're doing CG cars and CG planes they want to have this strip goes all the way down as you can see how that just see how great it is. So, in this instance, we can just excellent sleep pass. Yeah. So in this instance, it means all we have to do is come in here and modify the reflection. What I'm able to get is a lot more of a sleeper looking image without having expected the color or, you know, so it's, this is why I think it's better to work in passes and the thing is how this relates to seeing this video is, if we do the same thing in video space, so we convert everything to video, we take our video passes now. This is the simple line from legacy to video, take the light, and we add the reflection to that. I mean, it all looks right. Okay, it looks correct doesn't it. But when you add that on in video space, weird stuff starts happening. And you can see this isn't this isn't anywhere near what we're expecting. And likewise, if I do the opposite of taking subtracting the reflection off the main image, weird stuff starts happening. So it's like, just being linear, this stuff works, it's meant to work in linear. There's no reason not to be. Oh yeah, I mean I remember being trained that like, when of course when only working in video space that yes if you get a spec pass you either add or screen that on whichever looks better or the same thing with a reflection you know you end up going through all the different transfer modes or you always multiply your your shadows like these were all gotchas go arounds and compromises as opposed to like you said in that slide at the beginning, like being an additive process you can add and subtract these things. And they will behave as we're supposed to. Exactly. And the thing is it's like we don't want to be. We don't want to be that person who screams their reflection pass on top of the image. So what does that mean in terms of. So yeah, I mean, I, if anyone's ever had the fortune to work on one of my comps for they'll know this is never how neat. It normally looks like someone's just thrown up a bunch of notes everywhere, but you did comment on the states. So I've tied it up just for you. So, yeah, this is just going through what can be done with these passes so you've got a CG render we've got to look strange because when we haven't tagged it as seen linear, we haven't tagged it as anything. I don't really like to tag my stuff here, because if you're caching the media, when you tag it, it's going to be cache. So you, I prefer to have my tag out here, along with the color conversion that needs to happen, just because it's just it's easier to bring everything in and not worry about things hidden inside of things. So I find this is a, again, this is an interesting gratification thing but when you're trying to troubleshoot a comp and find out why something's wrong or not working. So having something inside of here or here is a classic for me, because it's just like, oh, really, you know, so it's, again, it's an extra step but I think it's ultimately a lot more robust way of working. So yeah, first things first was to fix an alpha. So this is something that's going to happen in CG, you're going to see some things creeping through sometimes. And the first thing I do is always a slap comp of just bringing in the plate, which has gone through just putting it into a CCG. Here's the map being up here. And yeah, and I can see immediately where are things that I want to fix and I've got like a, yeah, going to fake sky background on there just something to not be looking at black. And yeah, you can see some funky stuff happening on the lights. And that's clear. We can see it on the windows you like. Yeah, exactly. Like what's going on there. So bit of homework and come on sticky whack on sticky whack on. But you know what that one 46 minutes I gotta say that that was pretty good for the welcome. And yeah so just a simple gene mask. Nothing, nothing fancy that that fix it. And when we come in here. Now, no more funk, great. So a funkless, funkless composite. That's what you want again at gremlins. There's always gremlins and the thing is it's like embrace the gremlins just know how to swap them. So, yeah, again, taking my colors and I have to say a caveat to all this is when you start breaking down somebody's work into their component pieces, like, in this instance, we've taken our beauty parts and we're just dissecting them into the individual render elements and then combining them again, do be a little bit respectful because obviously the lighting artist has put this together. So if you then to start dismantling everything they've done and just stopping scraps sometimes it can be a little bit offensive so have a trade with care, especially if you start breaking the shaders and they're like why is the reflection terrible you're like, because I separated it and ended up by 400. So, I always use the algae of like, CG's job is to make it look photo real comps job is to make it look like a camera shot it and just start with that basics and then work out from there. And obviously, some things are going to get in the way and change all that as always they do. First things I noticed about this image that I want to fix. First off, feels very flat. It's very good problem with this shot because the client wanted it shot isometrically or they wanted it shot flat to the image. They wanted to see it to be, I don't know why it was a choice. I don't get that. But typically that's why you always see planes from above, just because it's just a much more pretty way of showing the scale of an object and the problem you've got is when you do this all of a sudden it kind of looks a little bit like a toy. So, we had to combat that we need to add some dimension to it and also what we've got is another challenge of reflective objects is that, especially metal ones is once you get to these extremes. So we're creating basically a mirror. And, you know, the computer, it's a, what's the word, physically rates renderer so it's going to try and do what happens in real life and that's what you get. But this instance it's just not particularly what we want because we want to control this a bit better. We want to add these highlights but we want to maybe take them away first and bring them back. So here we can see that the paneling here versus the paneling here. This is really sharp and crisp. This is, we're losing this a little bit here. Also, the, it's just, it's just feeling a little fun. So we'll try and fix all this stuff with as little work as possible, which I always think is the best way. So, where to start. So let's start with the nose. We haven't added an occlusion pass. And I'll go through these passes, but essentially this is to bear in mind this is version 56 I think of the renders. So what we started off thinking we were going to do is something different to how we eventually put it all together. And we've realized some passes we didn't need or some passes, you know, no, don't give us that give us this one, but we didn't name it differently. So it's, yeah, this thing happens in production, but we've got some of the other passes that we all know. So this is, again, focus amount you can ignore these some of these are be raised specific passes that we didn't use, but they just got left in there. We actually preferred the image with the green in than the D noise image we felt there was a little bit more detail in there so didn't use it. We have a diffuse filter. So these are again these are the raw passes. If you really want to know about the raw you can look into it and they'll explain the map is slightly different it's using multiplication but it's all on their website. And you do get a more granular way of fixing shots that nine times out of 10, we just did the answer. Next, again, these are all these are our multi mats are famous. We know that these are glow illumination we didn't use in the end. We started with it, didn't add anything useful. So in production, you're often going to end up with a render and you're like, well, that's all this stuff that we don't need. And it's like, well, that's why it just happens. Normals again, I don't really get into the process of relighting everything I know you can do that just because you can doesn't necessarily mean you should. You can throw a normal passing action you can do some incredible relighting but again by everywhere you can break an image and none of the lighting you'll ever do will look as good as what can be done with Arnold. So, Kevlett just because you can doesn't mean you should as something as a concept my clients have never really fully embraced. Yeah, no, I mean it's it's great like what you can do in action the light relighting of stuff and but I think often you can do that exactly the same way with tech passes from CG, and I'll show those. So, here again reflection filter this will be able to pass again. Top down this is a great pass this is local wide actor so from the top and from the bottom. Fantastic passes really good for cars really good for CG basically use them as a map. Next UV didn't use that world normals. But in world coordinates. Again, I don't think we use this. My favorite of all the passes the fruitiest for all the passes by just bring this down you can see so what is a position pass for those that don't know. What it allows you to do is it is something everything on the on the object has a value color value. And that does not change throughout the life of the object. So, what's great about that is that it means using this wonderful, wonderful map position map plugin from Lewis Sanders. It's amazing. And I just got to show you what is so matchbox. These are just interesting. So, yeah, just popping in the foreground, the map and I need the position parts, of course. So, position past just divided by the alpha so to make sure we've got an undefined position part so not getting stuff in transparency. That's why the game down here. Yeah, results, Matt, and the mat is a mat output for you, not the mat input and the position passes. So, what we can do is we can say, I want the tail. Okay, we'll just increase the tolerance. And you see it's drawing a map on the tail, and the thing is that sticks to the object all the way through. So great. Oh my God. So, and the wonderful thing about it is, as you'll see here, it takes a clear of the occlusion for you as well, which is huge. So, if you think about what you wanted to do in with just straight up G masks and stuff it's just, it's a godsend. You can say how many times I've done that with straight up G mask. Yeah, so in this instance, like I knew I wanted to use the pass of Matt to add definition. So I can just color correct down just in those the plane. So I created my mats. Just nothing fancy here, dividing by the first. I'm getting my mat, mat edge to really make sure I'm getting that. And then I use the position map to just extract the nose. And as you can see it's taking care of the propeller for me. And then yeah I multiply that I've got a map of the nose with about five minutes work. Again, laziness is key. And laziness is like the, the, the thinking man's efficiency that was just a side to that. So we've done that that's my excuse anyway. So that's one map sophisticated efficiency. Exactly. Top down passes, bring them in. We've got, I'm just crunching them a little bit. I'm going to use those. And I'm just adding two months together and climbing it up to make sure we've got no funky stuff happening. And then I'm feeding them in as Max into action on top of our four layers here. This I'll come back to. So here we've got our comp. This is pretty much the same as I slapped right now. And we come in here and I love this new way of working in action. It's fantastic. First layer. We've seen that adding that one on adding that one on adding that one on. Let's just start what we did out in the action. The person here on my, I'm adding a master grade to my reflection. So just adding a touch of contrast. Now, one of the problems I've got is, again, these are now getting really hot. It's kind of distracting seeing those two. So what I'm going to come into is, again, isolating the highlights of the reflection past, we know the troubles in the reflection. Again, just getting another map to get them a bit more visible and then have a position map, just to isolate the root of the plane. And the nose and those stick to it. And I'm just going to multiply those maps by my isolated map. So I've got the most egregious parts in a map and then I'm just going to say to the demo and just show you them. I don't know if you can see it overdone it here just so you can see it because it's very subtle, but essentially I'm just not like that. Definitely not like that. And then the difference is that now these are a little bit more under control. So yeah, I'm able to tone get these reflections being a bit more exciting, but those ones are not so next up is to really shape the dimension of the plane. So first thing is first is getting the nose where we want it. So if I come in here, all I'm doing is adding a selective master grade. Well, it could be anything but at camera effects. This doesn't really matter, but the point is that you're adding the. I'll do it with a selective master grade and selective master grade. And I'm coming in here and I'm choosing in the controls my input map, which is in this instance number five. And I'm just saying that to alpha. I can see my map coming in and this is a really cool way of working now I find compared to like having lots of nodes all out in three, because you can switch stuff on and off a lot easier. And yeah, so if I just come down to that just means I contrast in this one little piece and you gotta be careful with this because in the array, the ambient occlusion pass is actually called a dirt pass you a dirt pass because it does it's like the more you contrast in those pieces the more it will begin to look dirty and when you're doing a CG car or plane, you don't really want to do that. And also if you push these down too much it's going to look like the panels are sticking up, which you don't want to show. So, so that's what this guy is doing. And yeah, it's a little subtle, but it works. And it means that these now feel like these. Next up is top down passes. So again, same process bring it in, set to the media alpha. Check the output eight coming here and then I can just just gain the top of it and I'm just what I bring in as highlight across the top of everything that's facing up on the object is getting that top light brought to it and likewise on here. I'm doing the opposite and just bring in the bottom and all of a sudden, it's starting to feel a lot more three dimensional and you look at the difference before and after just that. I'd say it's nice and day. That's great that top down passes fantastic, fantastic concept and likewise I did something for the decal we had to just make it look a bit more like the color of the actual color is and that's the great thing about having a specific map of that. And then, yeah, obviously that the great thing is that without any without even thinking about clamping or clipping or color or any of that is I've got a really good looking object that I know by past my view and here I'm just doing exactly what the viewers doing, but like all these values are in check. So, the colorist has the flexibility to throw stuff out accordingly if they want to ruin your image or they can. I'm joking. Backhanded compliments that I love the most. You're giving them enough rope to hang themselves with essentially and then you know and likewise if they come back and you're like what the hell do you do in my image you can say, this is a lot like what's wrong. So, again, it's, I like it to train of accountability. And I guess. Yeah, it's been a final sort of thing that I did, which I really, really like. It's called a sweet and that, but one of the big problems with CG is things like this, where it's just a spec like that. And the thing is, cameras don't really do that. And they do what they don't. And what you want to see in these little bits is like some sort of chromatic aberration without it being like that classic thing of someone switched on the chromatic aberration plugin. So what I really like, so I've just taken here an iris. I think it's from lens, whatever that lens plugin is called fresh lift. No, the lens play one, the one that's called the Jane Abrams anyway, whatever. And so I just transformed it down a chrome or like that. And the new convolved thank you Francis for this is what you're able to do is add a very, very, probably that's the kernel, you're getting a much more photo like effect to those. That's great. Things that feel very CG, and that's one of my favorite things that we now have available in the new version of flame. Here and this, I think, I think there's much more I can tell you. John, that was awesome. Fantastic man. Does anybody have any questions now so far I'm just getting just getting accolades man. That was fantastic. And then yeah, you know what and the sweetener was definitely the icing on the cake. Does that. Absolutely. I guess it's like the aperitif afterwards right. Yes, sir. It was, it was wonderful. John, thank you very much. If anybody has any questions you can reach out to john on logic. I'm sure he'd be happy to help out. And really man that was that was excellent. Thank you very much. No worries. And likewise again, like this is not necessarily the best way to work, but it's pretty solid system and I would recommend embracing it. And just what I really like about the custom mode on the color correction is you really get to learn what it is that you're doing and it allows you to reverse engineer what you've got from editorial sometimes to say like hey what are these what's what do they do what order they go in and then if you think like oh well we know we're going from linear to camera space and then from camera space to CDL and then CDL the to the rec 709 love you know so if you think backwards from that console of the time you can get the color correct. Totally. Everybody know about the upcoming logic live sessions next Sunday is going to be the Andy and Fred show we're going to do some live Python scripting, which should be a whole lot of fun so if you've been curious at all about what kind of things you can do with Python and flame definitely tune in next week. It's going to be fantastic for the ad this is designed from the absolute beginner straight up to the overconfident intermediate which is how I like to describe myself. The following Sunday on Sunday May 10 we're going to do a neat video deep dive with Tim Chistikoff from neat video he is one of their senior developers over there and he's going to take us through really how to get the most out of neat video so many of us know it so many of us are actually own it and use it but maybe don't know all the possibilities that need video contains so definitely tune in for that. Then on May 17 we're going to do an interview with will Harris the flame family product manager to discuss everything that's new in the 2021 release. I think we're going to have some more of the dev team on as well and it's going to be a fantastic opportunity for Q&A. If you had anything you ever wanted to ask these guys there's any tools you'd ever like to see implemented. This is definitely the form for you. On May 24 we're going to do Maya for flame artists with Yuri Tempelski from Sao Paulo. Heading on closing out may rather we're going to do connected conform for social deliverables with Brian Bailey. I as I said last week I am one of those people who wishes I could social distance myself from social deliverables but if they're going to be a part of our of the new normal. We certainly come up with a few ways to make those go a lot smoother for you. June 7 we're going to do silhouette paint with our friends at Boris effects and June 14 resolved for flame artists with David Johns from LA. That's what's coming up on logic live you'll be able to find all the logic live sessions on logic dot TV, as well as a bunch of other great content. Please be sure to go over to our YouTube page and subscribe. Thank you again to our friends at sentences ocean for sponsoring logic live. Find out more about the remote workflows and any other solutions you might need that sentences.io. And finally, it's my cat Harrison's eighth birthday today so we just wanted to wish her a very happy birthday she was good. She didn't try to come in and and storm the session, which is a first for us so I want to thank her and wish her happy birthday but thank you john. Ashby thank you everybody on logic. Thank you Autodesk for for supporting us, and we will see you all next week. Take care everybody.