Transcript
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8" />
<title>Office Hours 2024-10-14</title>
</head>
<body>
<h1>Office Hours 2024-10-14</h1>
<p><strong style="color:#72B372">Eliot:</strong> <span style="color:#808080">[00:00:00]</span> Hey, hey, Kamar, how are you doing? Fine, how are you? Very well, very well. So I actually just, Greg just tested on 18. 1 and it worked for, I mean, he did a brief test, kind of waved the phone around with his hand and it worked, uh, there, so I'm just, I'm trying to figure out how we should test to try to track down what you're running into, or do you, did you find anything new?</p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> Yeah, I have some news for you. So, it just happened a couple minutes back for me. </p>
<p><strong style="color:#72B372">Eliot:</strong> All right. </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> So, yeah, a couple of things actually, uh, let me see if Vikram is in already, not yet. He must join anytime. Okay, I have a question for you. Uh, so the <span style="color:#808080">[00:01:00]</span> origin setup, not just the position, but orientation solution to this.</p>
<p>What does it cost you? </p>
<p><strong style="color:#72B372">Eliot:</strong> The, the, the, you mean the adding the origin adjustments? Yeah. Oh, yeah. That's, uh, I mean, we're going to be working on that. That's, that's one of the things that we, that's coming up, uh, coming up soon. Okay. </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> What if you have a solution? </p>
<p><strong style="color:#72B372">Eliot:</strong> Um, let's see, you mean for which part of it, for remote control, for, uh, manual control, which, which one?</p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> Yeah, I would say remote. I mean, you're not touching the phone. Nothing. No marker. Nothing. </p>
<p><strong style="color:#72B372">Eliot:</strong> Yeah, we don't have that yet. We have, we, we're just wrapping up, um, one portion of the, uh, We just finished up our re rendering workflow and have a <span style="color:#808080">[00:02:00]</span> couple of things to patch in Autoshot And then we're going to be jumping back into, into Jetset.</p>
<p>Do you have stuff? What's, uh, what's going on on your end on that? Just so I understand. We </p>
<p><strong style="color:#583E31">Vivek:</strong> did something for the, uh, yesterday, so, yeah. </p>
<p><strong style="color:#72B372">Eliot:</strong> The mic is, um, your, your mic is super faint. Am I good? Yeah, it's better. The closer, the closer, the better. </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> Yeah. Okay. But I hope now. </p>
<p><strong style="color:#72B372">Eliot:</strong> Yeah, better. So what's, what's going on on your end?</p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> Yeah, we were working on this, uh, yesterday and, uh, uh, yeah, we, you know, we have something. But yeah, of course, we still need to run through the entire pipeline, which we haven't done. Uh, one moment. </p>
<p><strong style="color:#583E31">Vivek:</strong> I'm just checking.<span style="color:#808080">[00:03:00]</span> </p>
<p><strong style="color:#C50052">Joe:</strong> Hey, Joe. </p>
<p><strong style="color:#583E31">Vivek:</strong> Hey, everyone. </p>
<p><strong style="color:#C50052">Joe:</strong> Hi, Elliot. Hi, Kumar Swamy. </p>
<p><strong style="color:#583E31">Vivek:</strong> Hey, Joe. How are you? </p>
<p><strong style="color:#C50052">Joe:</strong> I'm good. I'm good. </p>
<p><strong style="color:#583E31">Vivek:</strong> Yeah, all fresh. </p>
<p><strong style="color:#C50052">Joe:</strong> Yeah, just, uh, actually about to, I just wanted to pop in and just say hello, because I, I'm actually heading to a virtual production shoot now for that, that one music video we're doing now and like soon.</p>
<p>So just trying to get some breakfast in before, but, um, one of the things, so I just wanted to, sorry if I interrupted anything, I just, I have to bounce soon, but I just wanted to say, so Elliot, this is kind of what we were talking about, which I just thought I wanted to share before I left. So unreal is the whole 800 pound baby analogy proving very correct.</p>
<p>It's. Um, and so like I was even told by the producer, <span style="color:#808080">[00:04:00]</span> Hey, we need a green screen backup just in case this bullshit doesn't work. It doesn't like play they want. And so, you know, um, yeah, I just, I thought you would find that. Well, wonderfully endearing. Um, yeah, I know I did. I was like, I can't wait to share this with Elliot.</p>
<p>Um, so I'm actually today where we're going as part of the backup. I was like, I may have a plan for you on this Jetset today. So a quick question that I had, I tried doing the lens calibration. Like this kind of low level PC laptop. And when I did the lens calibration, I got like a CUDA error. Is that just because the graphics card isn't good enough?</p>
<p>Is it, is that just the problem? </p>
<p><strong style="color:#72B372">Eliot:</strong> Graphics card isn't cutting it in the phone windows. We, we, we, we look for like a moderate, a modern NVIDIA the acceleration libraries are called CUDA. So it sounds like the laptop didn't, I mean, try updating the drivers, but if it's an old laptop, <span style="color:#808080">[00:05:00]</span> it's probably like, No, it's a new one.</p>
<p><strong style="color:#C50052">Joe:</strong> It's a new one. It's just one of those like touchscreen, like, like, you know, uh, tablet, it's, it's not very powerful. It sucks to be honest. It's just, it's the Intel is sponsoring it. So that was the laptop that, so I was trying to do it on their machine, but it's just not even good enough. Um, yeah, we're, we're going to try, I'm going to try testing it out today.</p>
<p>Um, because one thing I learned, and then I promise I'll let you get back to it, KumaSwami, is, uh, that you can do green screen in the frustrum of Unreal, which I didn't know. Like, the inner frustrum square that you have? You can just have that be green and have the rest be the set. So you still get some of the like light reflections, but when shooting with, uh, like when shooting with our talent, we, we, we, we, we, I get a green mat from behind him in camera so that I could just easily pull a mat from him, separate him from, from that, and then come and back onto the background.</p>
<p><span style="color:#808080">[00:06:00]</span> So I'm hoping. That works. Um, I'm hoping that it works out well, and that's I'll keep you posted. </p>
<p><strong style="color:#72B372">Eliot:</strong> Yeah, I half understand what you said, but I'll have to like, see, see the image a little bit better. </p>
<p><strong style="color:#C50052">Joe:</strong> So they needed us to add in things into the environment, like the envi This is part of the thing, was like, they weren't able to get like the electrical effects and all of those Particles like baked or prepped in time.</p>
<p>And they're like, you have to comp that in later. I was like, what the whole purpose of this. And so, you know, I said, I was just like, without the light in her, I was like, well, what am I doing? I could just shoot this in front of a green screen, but it's just going to be a flat, like background in that sense.</p>
<p>So yeah, so that, that was interesting. So the idea was to like still try to take some of the benefits. So, you know, the, the, the frustum that is recording that, so they can just make that green with track markers on it so that when I'm shooting the talent. He said in green, but the rest of the outside, the frustum is still the environment.</p>
<p>So you're still getting some kind of like light <span style="color:#808080">[00:07:00]</span> fall on it. I don't know how well it's going to work, but it's interesting in theory. </p>
<p><strong style="color:#72B372">Eliot:</strong> Yeah. Okay. Now, now I understand the selective selective green freshmen. Yeah, I've seen, I've seen that. I've seen that. It's a, but anyway, </p>
<p><strong style="color:#C50052">Joe:</strong> that was, I didn't mean to interrupt you guys.</p>
<p>Uh, </p>
<p><strong style="color:#72B372">Eliot:</strong> I just look at that and go, we have a, you know, a several million dollar, uh, reactive lighting system, which you could achieve with, you know, a couple of programmable aperture lights, the green screen. And, uh, and an </p>
<p><strong style="color:#C50052">Joe:</strong> app. I'm telling you, seeing them realize that was an experience. Like, I'm dead, I'm dead serious.</p>
<p>Seeing them be like, you know, like just trying to live in the market. It's like, so we can't like change this or do this. They're like, well, no, we got to bake it all in. So it plays well in the end display, you know? And you're just like, Well, then what the fuck is the, you know, it's, you just have to have a lot, a lot of prep time, like way more than people even like thought.</p>
<p>So yeah. </p>
<p><strong style="color:#72B372">Eliot:</strong> Yeah. Way more than people are used to for, especially visual effects. You're like, we'll figure it out later. You're like, Nope, can't do that anymore. </p>
<p><strong style="color:#C50052">Joe:</strong> <span style="color:#808080">[00:08:00]</span> Exactly. </p>
<p><strong style="color:#72B372">Eliot:</strong> All right. Well, that's fantastic. Well, good luck. Let me know. Let me know how it goes. Ping me if, if stuff comes up and, uh, </p>
<p><strong style="color:#C50052">Joe:</strong> yeah, yeah. I was like, if I have any questions while we're on set today, I'll, I'll, I'll hit you up to try to, because I, I, it would be such a funny way to Trojan horse.</p>
<p>Just love it. You know what I mean? I'm like loving it. I'm I want to show them. Cause I think once they see the camera dropped in unreal, they're gonna be like, Holy shit. Right. You know what I mean? Like once it's processed and just copy and paste it, I think they're all going to be floored. So, </p>
<p><strong style="color:#72B372">Eliot:</strong> okay.</p>
<p>Fantastic. Fantastic. </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> input for you. I mean, you may be already calibration. You could still use a MacBook. If you have around, uh, it works and it would just do the unreal elsewhere. </p>
<p><strong style="color:#C50052">Joe:</strong> Okay. Yeah. Well, because that was my question was, if we did this and I recorded it to a separate PC, would I be able to just copy that data?</p>
<p>On to another machine, or does it like, have to be on the same machine that I do the <span style="color:#808080">[00:09:00]</span> Unreal work on? That was something I wasn't totally sure on. </p>
<p><strong style="color:#72B372">Eliot:</strong> No, you can copy over the project. You can copy over the project folder. I would say you'd want to generate the scripts. On the machine you're going to be processing on because it's when it generates all the scripts, it has paths and those paths will be machine dependent, but the overall project folder with like, you know, the takes and stuff like that, the takes are just data, you know, it's just data in a, in a folder when you start building the, the, the, the scripts and the image sequences and, and then, you know, of course you can, you know, okay, always rebuild them or like, you know, reset the links and stuff like that.</p>
<p>It's just, it'll be a lot easier to. To generate the takes on the machine you're on. </p>
<p><strong style="color:#C50052">Joe:</strong> Got, right, okay, yeah, because that, that machine is going to be like their Unreal machine. So it's like I could try to process it and then save the Unreal project file, but I just was like, in case I can't do that, can I take that back, take the Unreal project file to my, and then like just put them both on my machine at home.</p>
<p>Totally, </p>
<p><strong style="color:#72B372">Eliot:</strong> totally. If you have those take files and you put them in Autoshot, you can just regenerate the, uh, regenerate the script that drops it into Unreal locally. You know, all the data is in the takes. And then everything else is downstream <span style="color:#808080">[00:10:00]</span> from the takes, you know, so you can recreate, you can recreate all those unreal files on the fly, wherever you're at.</p>
<p>Yeah, that's that's the whole design of it. </p>
<p><strong style="color:#C50052">Joe:</strong> Oh, man, I, I, you know, you guys may be my heroes today, so I'll keep you posted on how this all develops. But yeah, we're definitely going to show it. Everybody's going to see it. And, uh, yeah, so I just wanted to, again, sorry, sorry for the interruption. Oh, no, no </p>
<p><strong style="color:#72B372">Eliot:</strong> worries.</p>
<p>I look forward to the, to the, the producer moment when he realized he's burning, you know, X tens of thousand dollars a day and, and, uh, Yeah, reactive light and </p>
<p><strong style="color:#C50052">Joe:</strong> quoting so much money. It's crazy how much money all these places are trying to quote and do that. I'm like, yo, we're not Disney. I'm not sure what, what kind of blood from a stone you're trying to get from us.</p>
<p>Trying to be like, I need 30 to 50, 000 per scene. I'm like, get the fuck out of here. Yeah. </p>
<p><strong style="color:#72B372">Eliot:</strong> Yeah. It's </p>
<p><strong style="color:#C50052">Joe:</strong> making these environments. 50 minute on YouTube videos. I was like, there's no way I'm giving you 40, 000 for like, you know, and I even <span style="color:#808080">[00:11:00]</span> saw like the work that they did. I was like, I could have done this. I was like, I'm not a good unreal artist.</p>
<p>And I'm like, there's like 20 ass, there's only like 20 things that they kind of duplicated and just kind of spread around. I was like, this is, I was like, what a racket. So, yeah, it's, </p>
<p><strong style="color:#72B372">Eliot:</strong> and it, and it's. The fundamental problem is you're trying to get things to render in synchronized. In a synchronized manner, you've got 42.</p>
<p>7 milliseconds. Hard stop at when you're doing 24 frames a second. There's no way around it. And, you know, that was, I mean, I spent 10 years hitting my head against the wall and decided, this is crazy. You know, we just, you know, get a good approximation and then we're going to post render this stuff and, and it's going to be fine, you know, and, and you, you take off that 42.</p>
<p>7, you know, hard stop and everything's good. It gets easy, you know, then you've got the lighting on any whatever you get, you get it. So, all right, well, so, all right, let me jump back in with tomorrow and, and we'll start working through this, but I can't wait. Can't wait to see what crops up. I very much look forward.</p>
<p>Look forward to seeing it. So come on. What's a. So yeah, tell me where, where are we at? Cause I, I, we did <span style="color:#808080">[00:12:00]</span> a quick test. Let me check with Greg on, um, the key question is whether, whether you're getting, I know on 18. 1, you had some problems, he tried it, he ran a test build of it and it was going into Unreal.</p>
<p>So the question is, are you able to get, were you able to get a green light, uh, of the data into Unreal yet? Or, or where are we at? Oh, there's a green light. All </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> right. </p>
<p><strong style="color:#72B372">Eliot:</strong> My day just got better. </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> I'm not really sure, you know, what I did, because I was struggling with a couple of things for, you know, just to see, uh, but I feel it was to do with the, uh, frame rate primarily, um, which here it was 24 all the time.</p>
<p><strong style="color:#72B372">Eliot:</strong> Mm hmm. </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> Then I pushed it to 30 by, you know, looking at, uh, </p>
<p><strong style="color:#583E31">Vivek:</strong> the, yeah, </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> the project frame </p>
<p><strong style="color:#583E31">Vivek:</strong> rate on the phone. </p>
<p><strong style="color:#72B372">Eliot:</strong> And, and just to make sure I understand, is the Um, for what, oh, for <span style="color:#808080">[00:13:00]</span> this test, are you taking video from a cine camera or are you taking the, the, I'm sorry, are you taking the video from a cine camera or are you taking it from the phone or where's, uh.</p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> Right now with this phone, I'm not going to take it to the, and I can, yeah, this happened just before the meet now. </p>
<p><strong style="color:#72B372">Eliot:</strong> Oh, okay. Yeah. </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> Just before the meet. I mean, I actually wanted to show you something else. But yeah, this happened. So good. Yeah. So realize, you know, what exactly it is, but yeah, I believe it is to do with the frame rate.</p>
<p>Uh, I will just reconfirm that, but in any case, I have Greg now to catch on in case who has corrected. But yeah. Okay. So this is one part of it. Now. Yeah. I have this million dollar question for you. And maybe Joe, you're listening, the adjusting the origin and orientation without touching the phone. </p>
<p><strong style="color:#72B372">Eliot:</strong> Remote control.</p>
<p>Yeah, we so we've we designed, um, we've <span style="color:#808080">[00:14:00]</span> designed the pieces of it. We haven't have not implemented it yet. That's that's on that's coming up fast. Um, just in terms of what what we're working on. But the next. Um, we're, we're, we're fixing a couple, we just did a push to release of Autoshot with the re rendering.</p>
<p>That was the big thing we've been working on to get, to get out because it's a whole new workflow. Uh, and then we're, how are you, are you running this thing into this where you have, you're running into the problem on set right now with remote control? Is that the, is that what's, what's you're running into?</p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> So it's a simple adjustment. Uh, we tried it yesterday with one of the, uh, User. User. </p>
<p><strong style="color:#583E31">Vivek:</strong> So your, uh, yeah, I have a huge issue with my mouth.<span style="color:#808080">[00:15:00]</span> </p>
<p><strong style="color:#6600CC">Kumar:</strong> So we just made a dummy, uh, body </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> to which we made a live cam as child. Mm-Hmm. And, you know, I just, in the scene and then, you know, things start from there.</p>
<p><strong style="color:#72B372">Eliot:</strong> I see. Okay. Okay. So, so you're parenting it to that so you can move through it. Okay.</p>
<p>Right now you have it on the parent. Oh, there, um, so if you click on the, the parent won't, won't be animated, um, it'll be the, uh, <span style="color:#808080">[00:16:00]</span> okay, you can set your offsets of that, and then. </p>
<p><strong style="color:#6600CC">Kumar:</strong> Yeah, I mean, of course, the scene is not </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> something really that, uh, small for you to quickly realize, but I think you get the.</p>
<p><strong style="color:#72B372">Eliot:</strong> So is, is, um, okay. So, is it all working now? Is that, I just, I want to make sure I understand what I'm looking at. </p>
<p><strong style="color:#583E31">Vivek:</strong> Yeah, one moment. It was supposed to be </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> working.</p>
<p>I don't know, where do I find the body? Somewhere here? </p>
<p><strong style="color:#72B372">Eliot:</strong> Looks like you're pointed at a cone. </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> Yeah, I asked the guy to make it as a camera, then he said, too many cameras.<span style="color:#808080">[00:17:00]</span> </p>
<p><strong style="color:#583E31">Vivek:</strong> No, nothing changed.</p>
<p><strong style="color:#6600CC">Kumar:</strong> Oh, yeah. I think you can see the theme changing. </p>
<p><strong style="color:#72B372">Eliot:</strong> Yeah. Yeah. So as you, as you are moving the Jetset Live Link parent, the camera, the parent is, you know, it's taking the, the, the phone along with it, or yes, taking the camera along with it. </p>
<p><strong style="color:#583E31">Vivek:</strong> That </p>
<p><strong style="color:#72B372">Eliot:</strong> should be correct. Right? You want it. If you want to parent the feed, then that, that should be correct.</p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> Yeah. Sorry. I have a buggy mouse. So it's really troubling me. I have really lost the. If you don't mind, I'll just do a quick restart. </p>
<p><strong style="color:#72B372">Eliot:</strong> Yeah. </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> And hope I get </p>
<p><strong style="color:#583E31">Vivek:</strong> back to that. And this also.<span style="color:#808080">[00:18:00]</span> </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> Yeah. This is what I observed, you know, the frame rate. This was hitting at 30 all the time. And then I could see. Yeah. Under settings, it was 24. </p>
<p><strong style="color:#583E31">Vivek:</strong> Which I changed to 30.</p>
<p>Maybe that was the reason.<span style="color:#808080">[00:19:00]</span> </p>
<p>Okay. Report.</p>
<p>Thank you, Peter.</p>
<p><strong style="color:#72B372">Eliot:</strong> Okay, so I see it. Um, </p>
<p><strong style="color:#583E31">Vivek:</strong> so maybe I to, uh, try to pull back. Maybe I rotate further, but</p>
<p>changing. </p>
<p><strong style="color:#72B372">Eliot:</strong> Yeah. I mean, is it, looks like it's working. Is there, is there something broken? I don't, I don't see it yet. Because as, as you change the, the, uh, the, the location of the <span style="color:#808080">[00:20:00]</span> parent, then the, the tracking, I mean, you have the, the camera parented to it. So that, I mean, it looks basically correct. Is there, is there anything, is there something broken?</p>
<p><strong style="color:#583E31">Vivek:</strong> No. Yay. Fantastic.</p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> Yeah. I mean, it looks like it's working actually. So just do a </p>
<p><strong style="color:#583E31">Vivek:</strong> rotation. Could it be</p>
<p><strong style="color:#72B372">Eliot:</strong> that you had one or the other hooked on your travel router and the other one wasn't like your, your, your PC was on your travel router and the, and the iPhone was on a different IP address. Uh, no, no. Okay. I wonder what that was. </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> Oh, you, you, you're talking about, uh, the Andrea, uh, sorry, the iPhone three, the <span style="color:#808080">[00:21:00]</span> 200.</p>
<p><strong style="color:#72B372">Eliot:</strong> Well, I was just curious why, why it didn't work last week. Okay. </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> Okay. You, even now I do have issue with, for example, okay. I did load this model. We did not have a graphic. I did load some other model which had issues. Uh, let me try that.</p>
<p>Okay. No. Okay. Now I think the app has an, a hang of it. So. It is taking the model, </p>
<p><strong style="color:#72B372">Eliot:</strong> so,</p>
<p><strong style="color:#583E31">Vivek:</strong> yeah, so, yeah, now </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> that, you know, I think both things are sorted, you know, the phone work, plus we have some solutions with the repositioning remotely, so you don't have to, </p>
<p><strong style="color:#72B372">Eliot:</strong> ah, good, good, so you can adjust it, adjust it on the fly, okay, okay, good, so then, then I think you can be, <span style="color:#808080">[00:22:00]</span> that will be a good, Temporary solution, uh, for the set to parent, parent the data.</p>
<p>Okay. I had </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> asked for your developer, if we can just, you know, write this, uh, script wherein he makes one dummy parent for this, you know, just like what you're generating in the Jetset. </p>
<p><strong style="color:#72B372">Eliot:</strong> Oh, </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> one dummy element. Which could be, you know, </p>
<p><strong style="color:#72B372">Eliot:</strong> Let me, let me think through that. Let me think through that. By default, what we do is we write in, we, we write in the name of the, of the scene locator in the scene.</p>
<p>And then we, we actually, actually, oh, that's a different script. So in a live render preview, I'm just thinking through this, um, JetsetLiveLinkParent.</p>
<p>Yeah, we, we could do that. <span style="color:#808080">[00:23:00]</span> For most people, I think that's pretty, pretty simple though, right? You just made a static mesh. Yeah, yeah, </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> yeah. You know, I mean, it just makes it, you do nothing, just, you know, </p>
<p><strong style="color:#72B372">Eliot:</strong> Yeah. Yeah. Yeah. I'm just, I'm thinking through it. Just said LiveLink parent. I have to be a little bit careful about it because as soon as you put something with offsets in it, um, this it's very easy for people to, to break things and not realize where, where, where things are not lining up.</p>
<p>So I, let me think about that just a little bit. Um, yeah, </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> yeah. No, we, I don't want to run through this with the entire, uh, post, uh, Pipeline, so that, you know, just to make sure things are all fine, which we haven't done. Uh, we'll just, you know, yeah, whatever I'm showing is what we have done so far, but yeah, just this in principle, this is all the latest data, nothing should really be, </p>
<p><strong style="color:#583E31">Vivek:</strong> um, going wrong with this right now.<span style="color:#808080">[00:24:00]</span> </p>
<p><strong style="color:#72B372">Eliot:</strong> Good. Good, good, good. Okay. </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> So we had this. Yes. And we, uh, okay. I just wanted to check with you on one more stuff with, uh, the. Yeah, override factor of the length to the app is the same as I can do on the web with, uh, the, uh, where do you have the, this is as good as overriding? </p>
<p><strong style="color:#72B372">Eliot:</strong> Uh, different, different. So the, that is the assimilated focal length.</p>
<p>What's that? Yeah, I'll, um, yeah, I'll get it for you in just a second. Hang on for just a second.<span style="color:#808080">[00:25:00]</span> </p>
<p>Okay. So where are we at now?<span style="color:#808080">[00:26:00]</span> </p>
<p><strong style="color:#583E31">Vivek:</strong> So just asking you if, uh, </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> the, </p>
<p><strong style="color:#72B372">Eliot:</strong> oh yes, yes. </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> The lens is as good as the orator. </p>
<p><strong style="color:#72B372">Eliot:</strong> Uh, there's two separate things. The simulated focal, and that's, that's a good point. Cause that could be confusing. Um, the simulated focal length is just when we have normal Jetset running and you tap the button to, to increase or decrease your focal length.</p>
<p>Um, you know, uh, um, the </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> rectangle that you see on the phone, you want to </p>
<p><strong style="color:#72B372">Eliot:</strong> It's, it's more of when you, um, are, when you're not using Jetset Cine and when you just have the, um, uh, when you're, when you are tapping into simulate the field of view that you would see on the, on the, on the, on the iPhone, that's what it changes.</p>
<p>It changes the simulated focal length. Um, the Cine calibration <span style="color:#808080">[00:27:00]</span> override is a separate, a separate sort of thing. Uh, what that is, is, is taking an existing Jetset Cine lens calibration and using it to extend. Uh, into ranges that normally you couldn't calibrate with, with optical. Um, so it's two, two separate things, but it's, that's a good point that it's a little confusing in that.</p>
<p>So I'll have to think about how to, how to describe that. Um, </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> I think one of the things we did, uh,</p>
<p><strong style="color:#583E31">Vivek:</strong> yeah, I think Justin </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> has some strange behavior with the phone. I'm trying to. You find, uh, it's </p>
<p><strong style="color:#72B372">Eliot:</strong> going, it's, it's AI ing in the phone.<span style="color:#808080">[00:28:00]</span> </p>
<p><strong style="color:#6600CC">Kumar:</strong> Hi, Elite. </p>
<p><strong style="color:#72B372">Eliot:</strong> Hello, let's see who else is here. Oh, is, uh, Kumar, there's Bikram. </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> Yeah, yeah, that's Bikram. </p>
<p><strong style="color:#6600CC">Kumar:</strong> I was there in the last call with Kumar sitting beside him. </p>
<p><strong style="color:#72B372">Eliot:</strong> Oh, good. Good. It's exciting to hear what you, what y'all are working on, uh, with the productions, uh, productions in, in Bangalore. Did I get that right?</p>
<p>It's a, that's the location. </p>
<p><strong style="color:#583E31">Vivek:</strong> Exactly. </p>
<p><strong style="color:#72B372">Eliot:</strong> That's exciting.</p>
<p><strong style="color:#583E31">Vivek:</strong> background. Okay.</p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> So is that any good? </p>
<p><strong style="color:#72B372">Eliot:</strong> All right. Let me pin your, uh, uh, pin what you're doing. Okay, there we go. So I'm seeing that, okay. </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> I'm trying to define, alright. <span style="color:#808080">[00:29:00]</span> Right, but it doesn't take me there. </p>
<p><strong style="color:#72B372">Eliot:</strong> Um, you probably need to, do you have a, do you have an existing calibration loaded? </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> Yeah. I mean, there's one already. Not really.</p>
<p><strong style="color:#72B372">Eliot:</strong> But, um, let's see, actually, you know what, tell you what, let's, let's do a QR code so I can see your screen for real, just a second, let me, uh, let me pop that up. Like, we have a system for this. This is why we do this. Uh, QR codes. Alright, let me share the screen.</p>
<p><strong style="color:#583E31">Vivek:</strong> You need a pre calibration. </p>
<p><strong style="color:#72B372">Eliot:</strong> There we go. So yeah, go ahead and, uh, exit Jetset and turn it back on and point the iOS camera at that and I'll bring up our Aha! </p>
<p><strong style="color:#583E31">Vivek:</strong> There you go. </p>
<p><strong style="color:#72B372">Eliot:</strong> Did it work? <span style="color:#808080">[00:30:00]</span> Okay. Fantastic.</p>
<p>Okay. So, so is it working now? </p>
<p><strong style="color:#583E31">Vivek:</strong> Yeah. One moment. Sorry.</p>
<p>Hey, you are sharing my screen? </p>
<p><strong style="color:#72B372">Eliot:</strong> Yep. I see it. </p>
<p><strong style="color:#583E31">Vivek:</strong> Okay.<span style="color:#808080">[00:31:00]</span> </p>
<p>I don't see.</p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> Okay. So what do you want me to do? </p>
<p><strong style="color:#72B372">Eliot:</strong> Uh, let's see. Uh, all right. I just, I think we broke a web certificate on a screen share. So we're going to have to do this manually. Greg is, Greg's fixing it now. So let me, let me, uh, pin your. Uh, pin your screen, uh, spotlight for everybody, right? </p>
<p><strong style="color:#583E31">Vivek:</strong> Uh, okay.</p>
<p>There we </p>
<p><strong style="color:#72B372">Eliot:</strong> go. Okay. Now I can see you. Okay. And of course, the screen is a giant, </p>
<p><strong style="color:#6600CC">Kumar:</strong> right? This is a lot of reflection. I think, uh, maybe, </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> uh, yeah, you'll <span style="color:#808080">[00:32:00]</span> see anything. </p>
<p><strong style="color:#72B372">Eliot:</strong> Uh, okay. There we go. When it gets close, I can, I can see it. So. We go to Cine Calibration, and Focal Length Override. Huh. Alright, so let's go back to Recording, uh, and let's, uh, Oh yeah, there's no Cine Calibration loaded, so let's go to, uh, </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> Okay, I don't have any calibration on this, I believe.</p>
<p><strong style="color:#72B372">Eliot:</strong> Yeah, you'll need a calibration loaded, and then the, I think the Focal Length Override will work. </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> Okay, it's only after that. </p>
<p><strong style="color:#72B372">Eliot:</strong> Yeah, yeah, because it needs that calibration to basically do a piggyback off of that. </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> Yes, and then you, yeah, I mean, since the tracking just worked for the first time, so. </p>
<p><strong style="color:#72B372">Eliot:</strong> Oh, good, good.</p>
<p>I like that. </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> Okay, Vikram, it's news for you. Uh, the tracking on my phone is working. </p>
<p><strong style="color:#72B372">Eliot:</strong> Okay. Okay. <span style="color:#808080">[00:33:00]</span> Fantastic. What was the issue? </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> I'm not really sure. I think it was to do with the frame rate of the stuff, but I just tried to play with this now. I changed it to 24 and it looks like it's still working despite that.</p>
<p>But yeah, not 100 percent sure. </p>
<p><strong style="color:#72B372">Eliot:</strong> Well, that's good. I mean, they're going to release 18. 1 in two weeks, so it's good to, good to thresh out the bugs beforehand. </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> But yeah, you, you, you need to extend my, uh, pre version. It did not work for me. I </p>
<p><strong style="color:#72B372">Eliot:</strong> wonder what it was. </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> Yeah, I'm not really sure. </p>
<p><strong style="color:#72B372">Eliot:</strong> I also did </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> handset reset.</p>
<p>But yeah, right after that also it did not work for that matter. Just happened just before the meet. I just, you know, yeah. Around. </p>
<p><strong style="color:#6600CC">Kumar:</strong> So I think, uh, regarding the focal length override, uh, we had one question <span style="color:#808080">[00:34:00]</span> about it. When we do the override, uh, so let's say our base length is 24, um, um, and then we overrode it to, let's say anything like 35, uh, in, in Unreal.</p>
<p>Um. We don't see that, uh, uh, change being affected, uh, with the Jetset. Oh, let </p>
<p><strong style="color:#72B372">Eliot:</strong> me, uh, let me, uh, let me double check that. Okay. </p>
<p><strong style="color:#6600CC">Kumar:</strong> Yeah. If we use a scenic calibration, then it works, but if we do a overwrite, then it's not updating. </p>
<p><strong style="color:#72B372">Eliot:</strong> Okay. So let me, let me make a note. All right. Okay.<span style="color:#808080">[00:35:00]</span> </p>
<p>Okay. So let's see. What do you, um, thinking how do we, so you had this thing in the focal length override and you would have depth. Okay. So let me check to see if that's, if that's a real thing. Okay. All right. I sent that to Greg to check it out. That'll either be a thing or not a thing. Um, okay. We'll look at, we'll look at that.</p>
<p>That, that feels like something that no one ever tried before. So that, that, that's a, that'll be a fast fix. If that's the case. </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> He had a 125 mm. Brought in. For which, of course, we only had one lens calibrated for that. Right, </p>
<p><strong style="color:#72B372">Eliot:</strong> right. I know, I know exactly how that goes. You had, you know, <span style="color:#808080">[00:36:00]</span> you wanted the time to calibrate the lens and you didn't get all of that, so now you have to do everything manually.</p>
<p>But as long as you got the offset and things like that, then you can make it work. Um, yeah. I, yeah. Yep, I've been there. </p>
<p><strong style="color:#583E31">Vivek:</strong> Bye bye. Okay.</p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> So, yeah. Now, next, we need to do one rundown with the whole post actually, so we're waiting for our camera to be, uh, Uh, so once we have that, I think we have full kit. </p>
<p><strong style="color:#583E31">Vivek:</strong> Yeah. </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> So now we have this</p>
<p><strong style="color:#6600CC">Kumar:</strong> camera, but that, that doesn't have support for, uh, time code And external <span style="color:#808080">[00:37:00]</span> timecode. So that's where we got stuck. So we have a blackmagic camera, but it's like out of stock everywhere. So we have to wait for a week, I guess, now. </p>
<p><strong style="color:#72B372">Eliot:</strong> You can test the whole Autoshot workflow with just the phone. You can test, test, uh, you can just shoot, you know, just shoot video on, on, uh, on Jetset regularly and transfer that, uh, over.</p>
<p>And all the workflows were built to work fine with either just, just the phone video or the Jetset Cine workflow. It's, it's almost identical. Yeah, yeah, </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> yeah. Yeah, we are targeting the VFX guys or the movies directly. So we want to, you know, have them Okay. Uh, go through that pipeline, which they would, you know, </p>
<p><strong style="color:#6600CC">Kumar:</strong> yeah.</p>
<p>So we are, we are, we are able to do the, </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> uh, </p>
<p><strong style="color:#6600CC">Kumar:</strong> apart from that, we are able to shoot. We are able to test it out. So, uh, I, I think last time we talked about that, how do we do, uh, you know, those minute, <span style="color:#808080">[00:38:00]</span> uh, settings, uh, or my new changes so that we exact, we get the exact framing, how the director wants. So we did some tests or we did, we tried some few, uh, few things where we can handle that in Unreal itself.</p>
<p>So we figured out what we can do is in the Unreal, uh, Kumar, can you show open that scene, okay. Can you share that?</p>
<p>No, we can only see your video. Oh, I've, I've got him, uh, </p>
<p><strong style="color:#72B372">Eliot:</strong> let's </p>
<p><strong style="color:#6600CC">Kumar:</strong> see, I've got, </p>
<p><strong style="color:#72B372">Eliot:</strong> I've got screen shared. I can see it. </p>
<p><strong style="color:#6600CC">Kumar:</strong> Oh, I, okay. I got it. I got it. Sorry. I wasn't, I had to switch it. So, okay. Uh, so what we did is we created a parent, just an empty actor and, uh, we parented our Jetset <span style="color:#808080">[00:39:00]</span> live camera to that. With that, what we can do is we can move the camera or the Jetset camera to the desired location and, uh, log the frame.</p>
<p>And then the camera can move around on its own, like as a child. And, uh, that way we figured, I mean, we do get some test shots, uh, in the office and we figured out this was. Working quite well, uh, for us so that, you know, if we want to, instead of, uh, doing those, uh, changes on the app, uh, if we do this parenting, uh, we were in in, in Andrea itself, we were able to move the camera, we move the parent to where we want, and we set the camera.</p>
<p>Or we set the Jetset app at origin. Uh, uh, you know, the, uh, when we <span style="color:#808080">[00:40:00]</span> set the scene locators. So instead of choosing any scene locators, we set it at origin. And, uh, so that there is not much offset between the parent. And the Jetson Jetson live camera, there would be some offset, obviously. Uh, but the lesser the offset, the better, the easier it is to actually move the parent and position the screen.</p>
<p>So this is what we tried, uh, as a, you know, as a way to do, uh, to, to have the framing, uh, uh, In, in, uh, unreal itself rather than going to the Jetset app and doing those minute, uh, changes. Uh, so this is one thing, uh, we tried. So just wanted to run, run it by you. I know there would be a problem when we, there could be a problem when we, uh, do the syncing of the videos.</p>
<p>Uh, <span style="color:#808080">[00:41:00]</span> with the, uh, you know, with the Jetset, uh, the, the data, which is there in the Jetset, the depth data and all that. And then we get the, uh, output, uh, the sequence, uh, output to unreal. So there, I think we'll have some problems because the camera might not be at the correct position. </p>
<p><strong style="color:#72B372">Eliot:</strong> So if you're going to do this, if you're going to, um, and I think it will actually be okay.</p>
<p>Um, if you, you make a parent object. Um, and, um, and have, have, you know, the, the Jetsa data linked underneath it. Um, I think what you'll end up doing then is, uh, you'll want to save a parent object. Uh, I see you're using a static, static, static mesh actor. I think that will be fine. I think you'll be able to type in the name of that, uh, static mesh actor.</p>
<p>Uh, I know it works with a, with a, just a generic actor with that's how we build our scene locators in <span style="color:#808080">[00:42:00]</span> unreals with a, you know, a base generic actor. Um, but you'll be able to type in the name of that into Autoshot when you process it, and it'll look for that. That, that, uh, that static actor, and then it'll, it'll correctly parent it, um, in the, um, uh, uh, in the generated level sequence.</p>
<p>So I think it should actually work fine. The one thing you're going to have to do is you're going to have to keep track. Cause the director is going to be moving it around on lots of different shots. You're going to need to, you know, keep making new, uh, new, uh, uh, Jetset Live Link parents. And renaming them per, per every time you move it, uh, you're going to need to, you just keep track of that, you know, you, you, you know how it works.</p>
<p>Yeah. </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> Yeah. </p>
<p><strong style="color:#72B372">Eliot:</strong> They're seeing locators. </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> You just copy the same locator and no name then, I mean, look at us, you would have anyway. </p>
<p><strong style="color:#72B372">Eliot:</strong> Yeah. </p>
<p><strong style="color:#6600CC">Kumar:</strong> Do you think this is this, this will be a good, uh, workflow or do <span style="color:#808080">[00:43:00]</span> you think there could be issues later on? </p>
<p><strong style="color:#72B372">Eliot:</strong> I think it should be fine. I mean, what you're basically doing is.</p>
<p>You're just leaving Jetset at, at origin mode and, and parenting the data inside unreal and whether you, you know, w and we w we do with it, we'd do the same thing if we were, um, I mean, we still record inside Jetset. We still record the data in origin mode, you know, at origin. And then, then when we put it into post production, we reparent it to whatever scene locator there, there is.</p>
<p>So it's really the same thing. You're just doing it. Uh, instead of doing the real time parenting inside Jetset of the, the, you know, the scene locator to the tracking data, you're just doing it in Unreal, that's fine. Uh, I don't, I don't see a problem with that, um, you know. Now, it may surprise me later, but, uh, I, I don't, yeah, it's just 3D parenting, so there's a limit to, uh, what can go wrong, you know.</p>
<p>Um, I think, I think it should be fine. I think the trick will be keeping track of, okay, we're on this shot, we're on this scene. This name scene locator. So I <span style="color:#808080">[00:44:00]</span> would just be very careful with your naming and just keep copying them and name, name them very precisely so that you can go back when you're like, Oh, I'm on, you know, shot one Oh five, you know, and you know, okay.</p>
<p>I need to be on scene locator one Oh five, just, you know, you'll think you'll figure out a naming convention that works for you. Um, but yeah, yeah. I mean, it's just, it's just 3d data, you know, and parenting. Yeah. So exactly where you do it. I don't think you're going to break anything. </p>
<p><strong style="color:#6600CC">Kumar:</strong> Yeah. Yeah. Yeah. I mean, this, this seemed easier workflow on, uh, on a production set because we can handle everything in unreal itself, uh, then going to the, I mean, we didn't do need to set the origin that that's there, but that's just one.</p>
<p>Like we have to do it only once for each shot so that we, you know, like when the camp because generally we saw that, uh, the cinematographer will try some, uh, like test shots <span style="color:#808080">[00:45:00]</span> and then they will set up a starting point. From where they will, uh, start taking the shot that will be locked once they're ready.</p>
<p>Once the director is satisfied with the framing. So I think then we can ask them to, uh, do the, uh, then we can ask them that we need to set the origin and we can use the printed, uh, I think the printed, uh, um, yeah. So use that and set the. origin near the camera and then we can do the adjustments. Yeah, Kumar.</p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> Just have a quick, uh, input or, you know, ask, uh, yeah. Um, see the marker that we use, uh, pointer, you know, I want a slight adjustment wherein, you know, maybe you can have versions of this wherein I say, uh, this is away from, you know, half a meter or a meter from the camera. You know, uh, what <span style="color:#808080">[00:46:00]</span> happens is, you know, the camera locks in a position of that.</p>
<p>Okay. And I don't want to be asking them to move back just to get the origin. </p>
<p><strong style="color:#72B372">Eliot:</strong> You want the new markers? Uh, the new markers are, are more than twice the size. They're on the downloads page. </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> No, even those are going to put the location, the origin to that, uh, location. I want that to be offset by a distance. </p>
<p><strong style="color:#72B372">Eliot:</strong> Ah, that's an interesting question.</p>
<p>So what, what distance would you want it offset by? </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> You would say a meter is something ideal. So that, you know, wherever the camera is, you know, I, I have a good room of a meter or a half meter, at least so that, you know, </p>
<p><strong style="color:#6600CC">Kumar:</strong> and also keep in mind Kumar, keep in mind that there would be, uh, there will be different lenses, right?</p>
<p>So </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> a meter is a lead, or I'll say maybe two meters, <span style="color:#808080">[00:47:00]</span> maybe if accuracy comes under question, then say, okay, let's stick to a meter where, and you know, from there it goes to the camera because I cannot ask the camera to move back and then into settle again. </p>
<p><strong style="color:#72B372">Eliot:</strong> Yeah. Let me, let me just, let me think, think through this carefully.</p>
<p>Let me think through it carefully. Um, can you, that's, um, let's do a whiteboard so I can, I can understand this. All right. So there's a whiteboard. Okay. So, all right. So if we have a top view, so here's our camera and then somewhere over here is, here's our, our, our target.</p>
<p>And right now, and here's, you know, here's our phone, the Jetset, just running on it. So right now, the way we, the way it works <span style="color:#808080">[00:48:00]</span> is we detect using LIDAR. We detect the distance from the phone to this, uh, and then we recognize this. And then we drop our, for the vertical origin. Then we drop our, uh, we drop our origin, right, you know, right, right below the center of this.</p>
<p>So. Where are you thinking the, uh, and we don't, uh, we don't pay attention to the orientation of this. Um, we don't look at that. We, it just does a straight shot down. Um, so what, where are you thinking the origin would be? </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> Yeah, no, no, back to the camera. The origin now should shift to the camera because that's where the camera starts.</p>
<p><strong style="color:#72B372">Eliot:</strong> Interesting. Because. You know, at that point, so that what you're, what you're saying there is that you would want the tracking origin under the camera. Interesting. </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> I mean, that's the origin, right? What, what the underpinning of origin is that <span style="color:#808080">[00:49:00]</span> or implication of origin is that.</p>
<p><strong style="color:#72B372">Eliot:</strong> Okay. Cause at that point, what we could just do is have something where we do a straight shot. Down from the phone. Like we don't need the target, you know? Um, </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> Yeah. I tell you why again, because you know, the cameraman that's the camera, for example, you know, that's the shot, you know, he's ready for the tape now.</p>
<p>Okay. Uh, he does all the checks, so possible angles and position and he settles down. Okay. And he, he puts his focus also, uh, because you know, there's a need to do Uh, setting. So they all lock it. I cannot then go and ask him to, you know, just, you know, uh, offset his, uh, stuff for the </p>
<p><strong style="color:#72B372">Eliot:</strong> I see. Because if you, if you haven't put your target here so I can recognize it, now your origin is here.<span style="color:#808080">[00:50:00]</span> </p>
<p>Um, and then so you, and, but he can't, this, this is fixed and you, and tell me again, just so I understand why, why do you want the origin? It's just a new way of, of, of working at it. And I'm, I'm used to seeing the origin. Out somewhere. We're here where, um, you're going to hook a piece of CGI onto a live action object.</p>
<p>Um, so I'm used to that workflow. I'm just coming up to speed on and, and it's be, is it because with this workflow you're doing, you, you mostly have, you want to be able to adjust in unreal the position of, you know, the, of, of where the camera is and it's, and it's, it may, it's easiest to do that when the, Well, the origin is basically right underneath the camera because that way you're not not dealing with a weird offset where the camera trend, you know, okay, okay.</p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> What we read as origin from the lens <span style="color:#808080">[00:51:00]</span> plane or say the phone plane or the phone lens plane, uh, tend to act better in the sense, you know, it starts to move. Your FOV or the frame is all right there. </p>
<p><strong style="color:#6600CC">Kumar:</strong> No, and, and the other, other thing is, as other thing also is there, uh, because we also need to figure out. How far our talent should be from camera and if, if the origin is, uh, right underneath the camera.</p>
<p>So, uh, you know, it is be, it is easier to calculate, uh, that how far I see from the camera, uh, our talent is and other, uh, you know, like. If there is something, some CG element, which needs to be near the talent or, you know, that calculation becomes easier because that's what we figured out. In that </p>
<p><strong style="color:#72B372">Eliot:</strong> case, what we probably want is, is we, I mean, we want to have a button that drops, <span style="color:#808080">[00:52:00]</span> drops the origin right at the lens nodal point.</p>
<p>So if you're on Jetset Cine, then it does the offset calculation to the, the, the frustum, um, nodal point on Jetset Cine and then drops the origin right there. Um, Okay. So in that case, um, okay, I gotta, I gotta think through this a little bit. So, uh, would we want, where's my eraser?</p>
<p>Okay. So in that case, Would we now want the origin to be relative to the camera, right? So this is plus X. We, when we, when we drop the origin, we usually have the X axis going off to the right. So do we, so, uh, in this case, we would do not, we would not keep the originally set origin distance. We would actually just have it relative to the <span style="color:#808080">[00:53:00]</span> camera.</p>
<p>Is that right? </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> Uh, could you come again, uh, Elliot, what is the last line? </p>
<p><strong style="color:#72B372">Eliot:</strong> So, for example, um, all right, if you, if you, um, when we place our, the origin, we have to put our position, but we also have to put the orientation. And so my question is, if we do this kind of drop, drop origin, you know, drop a nodal point origin kind of thing, Normally in Jetset, when you turn it on, you tap an origin.</p>
<p>The X axis is pointing off to the right. That's just kind of that way standardized on how computer graphics work. So if the camera's pointed over, over this way and we dropped the origin, we probably rotate the X axis along with the camera. Is that, is that right? Okay. Okay. </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> In this case now, instead of we dealing things with the Unreal, uh, for, you know, uh, positioning, Uh, for us to get the origin, uh, to the camera, uh, lens, uh, <span style="color:#808080">[00:54:00]</span> plane, should be good.</p>
<p>You know, it's a decent, uh, uh, input to begin with. The rest is all happening in, until now, in terms of orientation or, you know, change of, uh </p>
<p><strong style="color:#72B372">Eliot:</strong> I see. I see. So the, the position is more important than the orientation. Because in Unreal, you can just rotate it. </p>
<p><strong style="color:#6600CC">Kumar:</strong> Yeah, right, right, </p>
<p><strong style="color:#72B372">Eliot:</strong> right. Okay. Okay. Because </p>
<p><strong style="color:#6600CC">Kumar:</strong> it makes, it makes our life easier, uh, uh, in the calculation part of things, because what we notice that if the, if, if we don't place the camera at the, uh, in unreal at, at a proper distance from our, uh, supposed talent, then we don't get the correct parallax effect.</p>
<p>Uh, because, you know, the distances. Not the same. I see. So, in some regards, the reflect effect was broken. </p>
<p><strong style="color:#72B372">Eliot:</strong> I see. So, in this case, if you had sort of like a, uh, an image plane background, um, that you're going to, you're going to <span style="color:#808080">[00:55:00]</span> drop in, that's not, not pure 3D, you want to have it, that image plane move correctly with respect to the camera.</p>
<p>So, it helps to be able to just drop the. The camera at the correct distance to the image plan that you already measured manually. Uh, I see. I see. Okay. No, okay. This, this makes, this makes a lot of sense. Um, what I'll do is, is I'll, um, uh, we're, we're writing out the design for, uh, to be able to adjust both the, you know, the origin adjust and both in the phone and remotely.</p>
<p>And so we want to do this as. All kind of as it all together so that it's, it's a, the interface matches, you know, is as close as, as we can get between the, um, between the app and the remote control. So what I'll do is I'll, I think this is a good idea. I think this will come up many times and I can call it something like drop the origin at the nodal point and everybody understands what, what that means.</p>
<p>You know, that's, that's then it's, and it's the same thing between the iPhone and the Cine camera version of it. So there's, <span style="color:#808080">[00:56:00]</span> you know, if you're on Cine, it's going to drop it at the Cine, Cine origin. iPhone, it'll drop it at the iPhone origin. Okay. Okay. That makes sense. That makes sense. Um, okay. Yeah, I think, I think we can do that.</p>
<p>I don't see a problem with it. It's, uh, it's going to take, you know, we have, we have a couple of weeks of work to get, to get all these, these pieces in there. So just, just to kind of set, set expectations, it's going to, we got to get, we got some work to do, but I, I don't think that one's going to be a big, a big ad.</p>
<p>I don't think that's going to be huge. </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> On the same line, Elliot, uh, the beta versions or the alpha versions of what you're doing. Can we get access to the test flight so that, you know, we don't have to wait for the </p>
<p><strong style="color:#72B372">Eliot:</strong> Yeah, that's a good suggestion. That's a good suggestion. I think we can do that. Um, and I'll check on and how we get the invites and stuff like that, but it should be pretty straightforward.<span style="color:#808080">[00:57:00]</span> </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> Okay. And, uh, I, I, this is going to be a repeat of my question, but yeah, you, you were totally exhausted the week before. Uh, but I ask you again, uh, the plate, digital plate, how critical is this for us to have in the whole workflow? </p>
<p><strong style="color:#72B372">Eliot:</strong> The, um, uh, I'm sorry, the, the, the, which, which plate? </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> Digital plate for the sync generator, you know, that, uh, the timecode.</p>
<p><strong style="color:#72B372">Eliot:</strong> Oh, slate, slate, got it. Oh, um, let's see. Um, it's still, The way it works is, what the slate is doing is it flashes a set of codes so it automatically identifies things. And that's, that's a big help for, for identification. You can do it manually, you know, so the, the, I don't, you probably saw the, the tutorial I go through where I just load the clips into resolve and I, and I work it out and stuff.</p>
<p>You <span style="color:#808080">[00:58:00]</span> can do all this stuff manually. It's kind of a pain. Um, and so we are working on. Um, doing it automatically with time code. We don't have it yet. And, and we're actually probably, probably. We have so much work to do in Jetset to all these, the remote control things. They were probably several weeks away from being able to get that automatic sync in, in Autoshot.</p>
<p>Uh, so the answer is yeah, you can do it right now. It was just manual and you're going to be sitting there looking at the takes going, okay, is this one, this one, this one, you know? Um, and, but it's, it's all, it's all doable manual manually. It's just a pain. </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> Yeah. I mean, I'll explain you the real time issues with this, but yeah, the question is, I think you did try to mention this, uh, the timecode sync that we're using, uh, isn't that helping in any way, uh, to get this, uh, you know, sync right now?</p>
<p><strong style="color:#72B372">Eliot:</strong> Let me make sure I understand. So the, <span style="color:#808080">[00:59:00]</span> are you speaking of in the real time, uh, synchronization in Unreal? </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> The typical thing that we're using. Okay, because that is embedded into the iPhone, as well as into the Cine, uh, file. So you have a common factor to, you know, seek and then align. </p>
<p><strong style="color:#72B372">Eliot:</strong> Yes. </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> Uh, so we know why we still do need a visual sync app.</p>
<p><strong style="color:#72B372">Eliot:</strong> So the answer is, is. Um, is that we have not written for, for real time to get the synchronization into Unreal. We still, we still need that embedded timecode. That's for the real time part. For post production, um, we don't use the, the timecode yet. We have not written the code to identify the timecode in, uh, the Cine file and match it to the timecode in the Jetset data.</p>
<p>We haven't written that code yet. Uh, we, we plan to, you know, it's, it's very much on the list. Uh, but <span style="color:#808080">[01:00:00]</span> what, what usually happens is that we, um, we've pushed, we focus more energy on getting the Jetset pieces, uh, refined because that's why it's on set, you know, and that's the, you know, uh, yeah. </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> Okay. I mean, okay.</p>
<p>A quick question. For example, if, uh, for some reason, the, um, the shooting setup has their own time code, uh, uh, set up in place already. So can we then just, you know, make use of that to think with, uh, that's it. Or we still need our own, </p>
<p><strong style="color:#72B372">Eliot:</strong> still need the Tentacle. Um, but the tenant in this case, what you can do is.</p>
<p>Is you can use jam syncing, right? So you can take the Tentacle connected to their timecode source and have the Tentacle read that timecode so that it's in sync with the timecode and then switch it, and then, you know, that's, that's <span style="color:#808080">[01:01:00]</span> fine too. Yeah. But the, right now the Tentacle is the only device that we can talk to on the iPhone.</p>
<p>Um, it has a Bluetooth SDK that is that's, that's why we use it is we can talk to it. </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> The Deity isn't, uh, compatible or you didn't, uh. </p>
<p><strong style="color:#72B372">Eliot:</strong> We haven't seen an SDK yet. Um, and maybe there is one, you know, I know several people have asked about it. Um, we, the Tentacle worked so well and was so inexpensive that we, you know, we just implemented it and went on, went on with other things.</p>
<p>Uh, you know, it's a, it's a great little device and it just works. So. And, and we can use it with other time code systems by taking the, you know, jam syncing the, the Tentacle to that, to that time code. Um, and then, then we're reading the same time code. </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> Yeah. Perfect. Okay. The reason, you know, uh, for the, uh, or the digital, uh, sync, um, through the flash is, is really a big challenge <span style="color:#808080">[01:02:00]</span> in, in, you know, in production.</p>
<p>Uh, again, the reason is, you know, uh, the syn game has to roll. First, and that is good 5 to 10 seconds,</p>
<p>uh, you know, before, uh, they are ready for the, and they cannot afford to cut and then, you know, go for the shot. And it so happens that, you know, they're ready for the shot. And for some reason you are, you know, you're like, you know, you're pausing or, you know, cancelling and then retaking. And we cannot really go every time flashing, we think, uh, we take good five seconds to </p>
<p><strong style="color:#72B372">Eliot:</strong> Yeah, yeah, I, I hear you, um, it is, uh, </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> And then the focus factor is also a thing because they, they're all over the site.</p>
<p><strong style="color:#72B372">Eliot:</strong> So do they have a slate that they, they have, you know, do they have a, you know, something where they, they have a, a, a practical slate? </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> No, they don't, <span style="color:#808080">[01:03:00]</span> they don't use it anymore. </p>
<p><strong style="color:#72B372">Eliot:</strong> Okay. They're not slating takes? Wow. Um, </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> okay. There is blanking happening now, so they know all the shortcuts. Everything is, you know, before they wind up, uh, not even wind up, you know, after every shot, they are, you know, lining up the whole, uh, frame, you know, the sequences.</p>
<p>There is a </p>
<p><strong style="color:#6600CC">Kumar:</strong> live, uh, editor sitting on the set who is basically lining up. </p>
<p><strong style="color:#72B372">Eliot:</strong> Who's that editing, uh, cutting it? Okay. How do they keep track of the name, which shot it is, uh, like which take it is? Is that a tail slate? Is it? </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> The feed that takes, uh, you know, the live editor takes, uh, has all the except data. So they know which shot to wrap up.</p>
<p><strong style="color:#72B372">Eliot:</strong> Okay. Okay. Okay. So the answer is, yeah, you can do this manually. It's, it's going to be, um, the problem is it's tricky. You have, you <span style="color:#808080">[01:04:00]</span> have like 500, 500 shots and there's no slates. Um, yeah, you're going to want the timecode data. Oh, boy. To get that to match. </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> Because, you know, this slate is highly impractical, because, you know, there are a lot of practical issues with this, as I, you know, mentioned to you.</p>
<p>One, their, you know, their role keeps building up unnecessarily 10 15 seconds easily, and then there's a lot of interval that, you know, we intervene in their shot. I see. </p>
<p><strong style="color:#6600CC">Kumar:</strong> No, I think the main issue, the main issue we saw was with the focus, because they generally for some of the scenes and uh, then they will have, they have to focus on the slate and then again, change the focus.</p>
<p>So change the focus. Okay. </p>
<p><strong style="color:#72B372">Eliot:</strong> Okay. So basically it would be very, very <span style="color:#808080">[01:05:00]</span> useful if we could do automatic take matching. With the time code, because then it's automated. It's just happening in the time code. Nobody has to get in front of the camera and change anything. It's, it's completely behind the scenes. Okay.</p>
<p>Okay. Yeah. </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> As mentioned last time, a copy of the Cinefeed, you know, in. </p>
<p><strong style="color:#72B372">Eliot:</strong> Oh, yeah. Yes. Well, we'll, uh, I think we can get there that that one's going to be longer, uh, because we need to first get the live feed in of, of the camera. Uh, compositing, right? And so, but the, the synchronizing, synchronizing with, um, with the, with the, the takes with time code is, is certainly on the, in the queue.</p>
<p>But this, this tends to put it up higher because you're, you're running into it. How is, are you in production right now with this one? </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> No, we are in still proving stage. Uh, I think, uh, uh, in a week plus we will have one more round of, uh, conclude them all so <span style="color:#808080">[01:06:00]</span> that, you know. Uh, I think next month, the first week or second is when they have a full schedule, uh, kicking off for the movie.</p>
<p>And, uh, in between we want to kind of, you know, uh, push this through as a standard practice, uh, for the movie. Okay. </p>
<p><strong style="color:#72B372">Eliot:</strong> Yeah, no, I, I, I'd very much like that. Um, okay. Okay. So it sounds like we have most of the other things. And so that the things that are remaining are, uh, the, the, you know, the drop the origin, um, and this post synchronization via timecode, um, on this.</p>
<p>Okay, that's, that's good to know. That's good to know. All right. Let me check on, on a couple of these things, uh, and see, um, because I know they're planned. It's just sort of what's, what's the order of, of where, when we can get each piece, piece of it in and I'll, I'll check on the, I'll check on the drop, the origin, uh, aspect of it.</p>
<p>Um, cause that would, we would need to put <span style="color:#808080">[01:07:00]</span> that in along with our, okay. I think we can do this. </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> Well, one second, Elliot, the origin is anyway is, you know, uh, read by the iPhone. Nothing to do with the cam in itself. Right, </p>
<p><strong style="color:#72B372">Eliot:</strong> right. Yeah, yeah, </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> yeah. So, okay. No matter what lens, uh, they have on, we are Okay. No, we can, you know, afford to Oh, </p>
<p><strong style="color:#72B372">Eliot:</strong> yeah, correct.</p>
<p>Yeah. Just </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> from the, you know, uh, the, the whole kit and then, you know, get that, I mean, since it's lidar, I think it'll also do good, uh, measurement, uh, evaluation so that we need not be so strict with. Where we stand. Okay. We just have to make sure it's steady enough while ready. </p>
<p><strong style="color:#6600CC">Kumar:</strong> Yeah. I think, I think, uh, uh, what Elliot, uh, might do is, uh, because he can, or, or the app <span style="color:#808080">[01:08:00]</span> knows where it is, so, you know, the app is able to locate itself.</p>
<p>Uh, so the app itself might be able to set the origin or like without the help of any kind of, um, marker. </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> Then you want to use the font zero, which is, uh, </p>
<p><strong style="color:#6600CC">Kumar:</strong> Yeah, I mean, I, this is not my domain, but I'm just saying that, uh, yeah, </p>
<p><strong style="color:#72B372">Eliot:</strong> we can set the, um, if we, if we have a floor plane, then it's pretty easy for us to set, to reset the origin, you know, whatever happens to be that, that's, that's okay.</p>
<p>That that's fine. The pieces that we still are going to need to do is the remote resetting of the origin. And so like they, they move this around, you completely broken your tracking. You want to be able to go and hit. Hitting and hit reset the tracking. So it detects the floor plane and we're planning to do that.</p>
<p>So it, you know, just, just like when you start up the app, it detects the floor plane and then it just picks an origin, you know, to <span style="color:#808080">[01:09:00]</span> start off with, just so you don't have to be, you know, sit there. Um, and then, and then you could, then you could remotely just hit it and have it drop the origin wherever the camera happens to be, and then, you know, then we're, then we, then we have a good remote control, uh, process, uh, for doing this.</p>
<p><strong style="color:#6600CC">Kumar:</strong> Right. Right. That, that's, I think that's should be our plan. Yeah. </p>
<p><strong style="color:#72B372">Eliot:</strong> Okay. Okay. Um, okay. In terms of priorities, just so we, um, would you put some of this, the remote control origin reset and the origin drop, would you put that ahead of the digital slate or would you actually put the, uh, would you put the, uh, the, would you put that ahead of being able to, uh, link ticks?</p>
<p>With timecode or do you think, do you think that the linking takes with timecode is actually going to be a more, a more critical aspect of this? </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> I would, uh, you know, since the origin <span style="color:#808080">[01:10:00]</span> management is relatively, uh, you know, uh, we have a fair understanding. This is just, I mean, I, I mean, in terms of doing this, uh, for sure, there should be a lot less effort related to digital plates.</p>
<p>Uh, but digital plates from production point of view, I would say that should be number one. In terms of, you know, that takes away a lot of intervening in the whole, uh, flow. And I mean, I see, I mean, I'm not really sure of, you know, how, for example, Joe or others, you know, do the production. I mean, see for a small, uh, commercial or a web series or anywhere, you know, you are doing a managed control, maybe good.</p>
<p>You know, you don't get the pinch of, you know, five seconds, 10 seconds. But it is again, not just a matter of time, but the whole rhythm of the, uh, there's a cameraman and the actor is what at the state when you have </p>
<p><strong style="color:#72B372">Eliot:</strong> a </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> repeat. <span style="color:#808080">[01:11:00]</span> It, you know, it just, you know, it's a big </p>
<p><strong style="color:#72B372">Eliot:</strong> problem. It's a big problem. Okay. I, okay.</p>
<p>I hear you. So what I'm going to do is, is, uh, so I have my notes on this. So basically, you know, drop, drop origin, um, and then, um, time code, timecode based.</p>
<p>So those, those are the, the, the two biggies, uh, that are, that are here. Okay. That's good to go. Let me, uh, talk to Greg a little bit and, uh, and figure out where we're at. Cause I, you know, with all these things, we've had a plan for a long time. It's just a question of what, what comes up when and how, and, and, you know, how, how we do this, but, uh, it may, it may just be time to do the time code sync.</p>
<p>Uh, just to get it. Get it working. Um, cause it's, as you said, it's, it's a huge production impediment, impediment. It's, it's a pain in the butt and, uh, we we've had a plan to, to get around it for a while, so maybe it's time to just <span style="color:#808080">[01:12:00]</span> do it. Okay. Fantastic. Um, okay, well, so that, so this is good. So you guys are up and running on tracking and it's, it's the, these are the main, main two pieces.</p>
<p>Okay. Um, I will, uh, I'll go check on that and uh, and uh, see where we, see where we can get, </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> and, uh, from the last week call, uh, if you recollect, uh, what we had that says, uh, you know, the one specific event that was there, that there week ahead. Uh, I think we too late for this, but yeah. We, me and, uh, Vira also did discuss some this, uh, further with, uh, the Army.</p>
<p>Um, so again, you know, you, you need to see how, uh, it could be worked out, but yeah, for example, just to give you a fair understanding of this event, uh, for the exhibitor, it was costing around, uh, </p>
<p><strong style="color:#583E31">Vivek:</strong> one and a half <span style="color:#808080">[01:13:00]</span> flat, </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> roughly, um, 2000 US d uh, for a stall, uh, which is uh, nine meters square. That's the size, uh, for three days.</p>
<p>Uh, that is just that cost. And, uh, yeah, putting up, uh, the show and everything else is other stuff. Uh, but yeah, we are keen to see how we can collaborate, but of course, you know, event cost and other stuff is not something really into our budgeting factor right now. But if you guys have anything, uh, you know, kind of put aside for the marketing, we would be happy to, uh, see how we can, uh, keep it.</p>
<p><strong style="color:#72B372">Eliot:</strong> So, and that's, um, I mean, that's in three days, right? That's the broadcast India. So is if we, if we can't do anything for this show, um, Is it, <span style="color:#808080">[01:14:00]</span> um, is there, are there other shows that would be coming up soon that would be a good fit? Or is, is this sort of, for example, in the U S the big one is NAB, right. In April in Las Vegas.</p>
<p>And there's, there's. You know, there's nothing else, anything like it, um, for the whole year. Is, is that, is this a similar equivalent in India or are they more distributed? </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> To be honest, I'm, I'm really not, you know, that is not something I've been following, but, uh, uh, these kinds of events are very, uh, less compared to, uh, you know, the West, uh, and this one is more specific and to the domain.</p>
<p>This is for sure. Uh, you know, uh, we do, uh, encounter a lot of, uh, hardware based exhibitions and stuff, electronics and automotive that are there, but, uh, coming to the entertainment or the broadcast, which is again, more technical than the entertainment <span style="color:#808080">[01:15:00]</span> side of it. Uh, I think this is one of, uh, kind for the year.</p>
<p>Uh, I honestly have to check these before I say no. </p>
<p><strong style="color:#72B372">Eliot:</strong> Okay. I guess the question is, do you, do you think it would be, because it would be an incredible amount of work for you guys to do this in three days. Um, do you, do you think it's a worthwhile thing to do? </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> I, again, you know, honestly haven't been to the past, uh, event of this.</p>
<p>They cannot really come, but they're, but just looking at their executive list and stuff, uh, you, you pretty much have, uh, from the, uh, dominant, uh, TV. Or the television, uh, crowd and broadcast crowd primarily. And then movie guys might be just, you know, chipping in on and off. </p>
<p><strong style="color:#72B372">Eliot:</strong> Okay. </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> But is broadcast any of your, uh, anybody is <span style="color:#808080">[01:16:00]</span> using?</p>
<p>I don't see broadcast guys really getting into this right now. </p>
<p><strong style="color:#72B372">Eliot:</strong> No, it's, it's, um, it's, it's not the best fit for broadcast. Um, the, uh, uh, it's a very, the tracking is very lightweight. It works, you know, work, works with it very quickly on a location suit suit broadcast you more or less is more of a fixed installation.</p>
<p>Um, and so it's, it's, uh, uh, and you can't, it has to work for like hours and hours and hours with no one ever dropping a frame and that's, that works honestly better with a dedicated hardware tracker, um, for what we're doing the, the, uh, The filmmaking kind of set of things. We have a post tracking solution that works really, really well in case you're going through a shot and something breaks.</p>
<p>Yeah, we can fix that in a post really easily. Broadcast, there's no post, right? So it's, it's just less, less of a perfect fit. So, um, but I, I'd be <span style="color:#808080">[01:17:00]</span> curious. What we would very much like to do is have. Uh, you know, we're, we're doing this first work in with, uh, with the, you know, the, the Bangalore project, projects, et cetera, here, what do you think is the best way to get the word out, uh, through the, the Indian production community?</p>
<p>What, what do you think is the, does, does everybody refer to a central source for some of this, or how, how do people talk about these things? </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> Uh, if you don't mind rephrasing, uh, the question, uh, Iliad. </p>
<p><strong style="color:#72B372">Eliot:</strong> Oh, sure. Sure. Um, what do you think is the best way to kind of, to get the word out through India, um, in the West, what we found is YouTube full stop.</p>
<p>I mean, you know, and we do like one trade show. We'll do NAB and that's it. That's it. You know? Yeah, </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> yeah. No, of </p>
<p><strong style="color:#583E31">Vivek:</strong> course. </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> Uh, see, uh, YouTube or social pretty much <span style="color:#808080">[01:18:00]</span> everybody is into it. Do you, you know, before anything, it is, you know, uh, Google search that happens. So having said that, uh, the virtual production is something, you know, for a serious player, they need to get hands on.</p>
<p>You can see all world, uh, you know, uh, worldy things on the YouTube, but yeah, you know, you still have to, uh, have, uh, for understanding for which in person hands on experience is a must for serious, you know, guys. </p>
<p><strong style="color:#72B372">Eliot:</strong> At some point, is there a production you would be able to have behind the scenes footage of you guys operating Jetset behind the scenes?</p>
<p>Would you be able to get some of that from this project upcoming? </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> Yeah, I mean, if things go well and we are continuing with this, of course we will have </p>
<p><strong style="color:#72B372">Eliot:</strong> some. That would be extremely, extremely helpful. Um, okay, <span style="color:#808080">[01:19:00]</span> okay, so this gives me a good sense of it. Okay, okay. Okay, guys, let me go look and see what I can figure out on, uh, on these, these pieces.</p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> Sure, and yeah, um, I mean, since we also have a 360 cam, I saw what you, uh, Joe was doing with the Insta360. So you guys, uh, kind of started working with Insta360 to do some, uh, development for the app. </p>
<p><strong style="color:#72B372">Eliot:</strong> I'm experimenting right now is, is, um, I've, uh, because there, if you're doing post tracking, um, to get a sub pixel refinement, um, if you're using, you know, a cine camera and you have a narrow lens, your tracking becomes very fragile, um, very fragile.</p>
<p>It's very, it's, it's quite difficult to get a sub, you know, it can work. If you, if you have a, you know, an 80, 80 millimeter lens. But the mathematics of tracking are working <span style="color:#808080">[01:20:00]</span> against you at that point. Um, and so even with our, our solution, um, that helps a lot, um, there are cases, or, or if you're, if you're moving the camera very fast, where the background is just a blur.</p>
<p>And so what we ran into with, um, uh, a couple of our users who are, uh, stunt people, um, where the camera, I looked at the shot and the camera is moving two meters per second. Well, the guy flies by him in the background. It's just like, there's, there's no, there's, there's no feature points anywhere. It's just a blur.</p>
<p>Uh, so we looked at that and I got interested in the 360 cameras because they, they see all directions at all the time. So you can generate very stable solves. And I've been looking through different ways of, of how to connect a 360 camera. And, you know, it's easy to hook the camera on the camera doesn't, you know, it's tiny, you know, I bought one, you know, <span style="color:#808080">[01:21:00]</span> this guy, it's tiny.</p>
<p>Right. And so the key, of course, would be to, uh, you know, run on a battery, it can run, you can plug a, um, you know, um, USB C into it, you know, so the camera team is, isn't going to care that much. It's not big, it's not in their way. Um, and it does a phenomenal job of capturing, you know, 360 video at, at 24, 25 frames a second.</p>
<p>Um, the question is, okay, how do we have that in a post production workflow? How do we automate it? How do we, um, put all the pieces together, uh, to automate the flow and synth eyes? Because. Um, well, as soon as you start dealing with multi shot workflows and synth eyes, it can, it can become very complex, uh, more complex than most people are comfortable with.</p>
<p>So what I'm looking at is how could we, um, how could we automate that process? And so I'm still, still thinking through it, how, how best to do it. Um, but it <span style="color:#808080">[01:22:00]</span> is wild. The, the 360 camera solves are boom. It solves really fast. Yeah. Um, there are YouTube art things of how they've done this for visual effects.</p>
<p>Then you look closely at how they, they have the 360 camera. So that's, that's fine. And then they, how they measured the offsets is they got a little, you know, little micrometer and they measured the, yeah, which again, it's okay. You know, it's better than nothing, but as soon as you, you have angular, they're having to dial in the angular offsets by eye and, and, you know, it's not going to work for lots of things.</p>
<p>So I want to look at something that is a process that uses the machine vision and solves it correctly. And I'm still working through it. You know, these are, this is the math, the math part of math part of it. So, um, but it's exciting and it's, um, it's not expensive and it's a very good data source and we can talk to it, right.</p>
<p>You know, on Bluetooth. So, uh, these are all the pieces that I'm <span style="color:#808080">[01:23:00]</span> weighing to find once I find a workflow that I, that I think is. Solid in the post tracking, then I can work, figure out the integration of it in the app. It's not, it's not that hard. You know, the primary thing is how do we use the data? So, yep.</p>
<p>That's, that's where I'm at. Um, okay guys, this is great. I'm gonna go check on, on, uh, on some, on some things, </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> oh yeah, sorry. But this in the whole loop. This is where we can afford a digital, a digital claim. Oh, so it's </p>
<p><strong style="color:#72B372">Eliot:</strong> over here. Yeah. Yes. Can </p>
<p><strong style="color:#DE4A1D">Kumar (2):</strong> have a last slate exclusively for that , you know? </p>
<p><strong style="color:#72B372">Eliot:</strong> Yeah, yeah. That, that's, that's a good point.</p>
<p>That's a good point. Well, the slate is really there only for the CNY camera. So we have burned in optical data in the CNY camera because Jetset, we already know all the information on the slate. Like that's where it's coming from. So it's really only for the <span style="color:#808080">[01:24:00]</span> syn camera. Um, but, um, okay. Okay. Okay, guys, I'll talk to you soon.</p>
<p><strong style="color:#6600CC">Kumar:</strong> Yep. Thank you.</p>
</body>
</html>