← Reshaping Workflows with Dell Pro Precision and NVIDIA RTX PRO GPUs
Reshaping Workflows with Dell Pro Precision and NVIDIA RTX PRO GPUs
How ARC is Reshaping Filmmaking with AI
00:36:55
ARC is turning phones into cinematic tools. Todd, Andrew, and Jeff share how AI and Dell tech are revolutionizing previs and 3D storytelling.
Loading summary
Transcript64 lines
- [00:00]
Todd
Foreign.
- [00:05]
Andrew McDonald
Welcome to Reshaping Workflows with dell Pro Max PCs and Nvidia, where innovation meets real world impact in high performance computing.
- [00:20]
Logan
Hello, welcome to another episode of Reshaping Workflows with Dell promax and Nvidia RTX GPUs. I'm Logan, your host and we've got something very, very special for you on this episode. You know, kind of through, you know, the podcast, you've seen all about, you know, the different workflows, you've learned about, you know, the new Dell Promax portfolio, you've learned about GB10, GB300, you've learned kind of the horror portfolio, Blackwell, Nvidia, Nims. But now we're going to show you something really cool that I think will really appeal to those that are in the kind of M and E world that are doing any sort of kind of 2D to 3D workflow and how do you really accelerate kind of the process. But I'm probably getting ahead of myself. I've got three guests with me today, so we're going to stop pause. Todd, Andrew, Jeff, please take a second, introduce yourself and then we'll talk about Zugazi and arc. But let's start with introductions a little bit on your backgrounds.
- [01:15]
Jeff Renfro
Thanks for having us, Logan. Appreciate it. Real quick, we're working to revolutionize the 3D visualization space by making tool sets that are more streamlined and more accessible to creative professionals of kind of all stripes. You could sort of think of what we're doing for 3D as what Figma did for two dimensional UIs and design design there. And everybody comes from a media entertainment background that we think that the applications of what we're doing can extend, you know, well beyond that. But that's what we're working to tackle to begin with. And background wise, I'm a multi time entrepreneur as well as a former studio executive. I used to work at Disney and did a lot of product work there and we all dove into this project a couple years back and we're enthusiastic about what the future holds for us.
- [02:07]
Logan
Awesome. Let's go with Andrew next.
- [02:11]
Andrew McDonald
Hi, I'm Andrew McDonald, I'm one of the co founders of Ark. I am film and television professional for the past 20 years now my background is in, in shooting factual so directing and shooting as a cinematographer in factual. And the concept of this came with Jeff and another colleague of ours. We all went to film school together here in Toronto and they had reached out and knew that I've been since in my film career I've been sort of on the leading edge, working in interactive digital media, more recently producing VR content. And Jeff has followed me since school and knew I was sort of in this high tech area. And they had a brilliant concept to take the idea of using the phone on set, which is what all DPs are doing these days. It's replaced the director's viewfinder and adding to it augmented reality. So now we can pre visualize and position things in, in front of the phone. And had reached out to me and I put together another person who's in New Zealand, who's an expert developer who was already down this path. And we ended up forming a company right at the convergence of AI coming into play. So it was exceptional timing because now we were able to bring in all these other APIs into the app we're building and have built this true Swiss army knife of pre visualization that we are having tons of fun playing with and we can't wait to get it out to everybody else.
- [03:33]
Todd
Hi, I'm Jeff Renfro. I'm one of the co founders. I'm a director, producer, writer, filmmaker. And yeah, I went, I went to film, film school with Andrew and one of our other co founders. Feels like 50 million years ago now, but back when music videos were cool and we were, we were doing some pretty interesting stuff in music videos through the late 90s and early 2000s as I started getting into movies and then into into series work, doing episodic, kind of reconnected with, with Chris and we, we both found ourselves on the Chucky series. For the last couple of years we essentially were horsing around with AR out in the desert. It was an app that was, I guess designed to, for a sculptor to kind of see a naked, you know, model basically. And so we, we, we situated this naked model in the desert and, and you know, scale it up to like 100ft. And we were just laughing our asses off and then kind of the light bulb went off where we're like, well, if we could kind of, you know, if we could choose our focal length and if we could actually focus this lens and if we could start to, you know, treat it like a cinema lens, wouldn't wouldn't that be cool to have a digital camera in the virtual space looking at these digital objects. So it really started from that kind of simple perspective. But now, you know, as we gathering momentum and we've got this great team. Andrew, we brought Todd on, we've got this, you know, insane. We call him Jimmy Page. His real name is Dan Monahan, but he's the Jimmy Page of, of, of writing this code for us. And yeah, we're just feeling kind of silly, like a rock band. Kind of got some great hits. Right now we're having a lot of fun. As Andrew said, we've been using it on the Chucky series, Chris has used it on a, on a Super bowl commercial. And, and now we're finding some really wild workflows that push us into the animation sector in a really interesting way, which we can talk about later. But that's, that's, that's the macro.
- [05:37]
Logan
I appreciate it. I mean it's a good overview and I think you all kind of hit it and I want anyone can answer, you know, feel free to chime in, but kind of talked about the whole idea of kind of the intersection of kind of 3D spatial computing. How do we in theory put a camera, digital camera in a space to be able to visualize, mock up, et cetera. So but for our, you know, there's some petal people listening, there's some people that are not like maybe give two minute kind of overview of ARC and what it's specifically designed to do and then after that we'll get into some of the actual workflows, Jeff, that you talked about that are kind of interesting where ARC makes sense and what you're all doing makes sense to, you know, really accelerate the creative process.
- [06:18]
Jeff Renfro
At its core, Arc, ARC is, is the, at the intersection of spatial computing, 3D, AI workflows and mobility. But it really sings when you begin to have it interface with AI workflows. It's great on its own and it has lots of benefits. But the timing kind of couldn't be better in terms of what's been going on with AI because even with what ARC does, you still need. ARC is sort of a lego lego for 3D, right? If you have existing models working within. ARC is, is just incredibly high speed relative to what used to be the case with 3D. It's really incredibly valuable in terms of being able to craft scenes and then even more so be able to capture those scenes in a cinematic sort of way for whatever, you know, tv, film, commercial project you're, you're working on.
- [07:11]
Andrew McDonald
It's. There's this sort of watershed moment where a lot of these 3D workflows have become accessible and it's mostly due to being advent of, of, of AI in the sense that now there are sites out there that you can AI prompt into existence a three dimensional asset and have it rigged and animated, which Is something that up until about five minutes ago you needed ten years of schooling. You need to be familiar with some very sophisticated software programs and have a lot of skill in that world. And now these, these workflows are becoming accessible and the timing that we've had with this I think is pretty perfect because designed from the phone up, designed to be used. This is designed as a camera, not as a, as a, as an app. So because we're using the camera on this and now we. It's an, it's an aggregator because now anyone can almost generate any kind of three dimensional asset or scan an environment and put it in the phone almost instantaneously. And then you're operating with just the phone, which everyone's familiar with and has in their pocket in the boardroom, on the desktop, able to go pre visualize and film scenes and, and the product that you get out of the app because we've approached it from, you know, as Jeff was mentioning, like the prime, we want to see a prime lens on there. We want to see a shallow depth of field, we want to see a cinematic look. With this stuff you're, you know, it's designed as a previous tool, but we're beginning to discover that you can almost get finished content out of this, especially when you bring it back into an AI workflow for an AI video to video post processing.
- [08:40]
Jeff Renfro
So yeah, when we started we, it kind of seemed like it was, it was all about Previz. But it was revealed to us that it was really a 3D visualization tool and that as we'll. We'll show in a bit too that we could be. You can do a lot of great Previs stuff in it. Just like Previz accrues value to not just the camera department, but lots of departments across the production. ARC does that too. But the other piece of this that you know, seemed pretty evident but became really clear to us over time is that you could get certain levels of decent quality 3D content out of this. And again paired with 3 with AI workflows like BE able to generate things that, you know, would have taken, you know, weeks and weeks and lots of resources. And you can, you know, put together a pitch that like Jeff will show in a, in a bit that you know, is pretty compelling and pretty decent quality and that you'd never been able to execute before as easily. And you know, in a streamlined of.
- [09:42]
Logan
A way it makes sense what you're saying, right? I mean, and I, after seeing the technology, after working with you all for a while, you know, it, it is very compelling kind of what you can do. So I want to, I'm just going to be two part question before we kind of get into Andrew's and Jeff's kind of your workflow and then we'll get into some more technical questions. But you mentioned, we've mentioned, we've heard the word AI quite a bit. We'll touch on that in a second. But give me a traditional workflow that ARC is really enabling and let's say go back, let's say five years from that workflow. How long would that have taken? How much tools would you have had to use? You know, what technology would you have had to enable it to? Now, what is possible with arc? You know, running off a Dell Promax workstation kind of enabled by, you know, Nvidia RTX Blackwell or RTX ADA gpu. How quickly is that now able to be done? And then let's start with that in.
- [10:37]
Todd
Production, we use it every day, all the time for a bunch of different reasons. What started as a blocking tool has become, okay, here's, here's a place where we have to shoot a poor man's process with a vehicle in a quantitative hut now the DP's already squinting at it and trying to see where he's going to put things. Basically we, you know, we, we use ARC to bring in a digital car into that, anchor it into the quantit hut, move it around. And now we can all sort of stand back and look at it and go, hey, there is enough room for a drop back there, you know, or another one was classic on, on Chucky, you know, last year Chucky was at the White House, right, Shooting in Toronto. There's no exterior White House location. So we end up building a set of stairs and a door essentially just surrounded by green screen, right? Just walls and walls of green screen. And you know, so we take this really insanely textured model, a digital model of the White House and we decimate that and get it into ARC so that we can then situate it out there in the parking lot and go, oh yeah, there's some pillars there. Wouldn't maybe the shot's better if it's over here. And, and the other thing we can do is start to plan crane shots by scaling the White House down. So on a conference room table, we'll scale it right down and talk about a crane shot that starts 20ft up. But we're just moving, you know, we're moving through, we're moving ARC on the phone through space and showing Everybody in the conference room right there, live render, there's no wait time and you're able to completely visualize in this three dimensional.
- [12:25]
Andrew McDonald
Space the extension of that, that's, that's happening pretty much five minutes ago. As everything in this industry is, is that we're also working with a couple of production companies here in the city that are producing shows for major networks. And as I said, within the past couple of months the, the floodgates have been opened in the sense that major networks are now comfortable with and see the legal pathway to use Gen AI in finished broadcast product. So there are now commercial data sets for image generation that are coming online that are being offered in an enterprise scenario. So they're in closed instances. And that's, that's where powerful GPUs and workstations come in quite handy because the more secure you're, I mean it's all, I'm beginning to learn all of this process in the past like two months. But what we've discovered is the more secure your image generation is in the sense that if you're running comfyUI on GPU with a crap ton of VRAM in your own room and you don't have anybody polluting it with any other inputs, then you are the most safe in the sense in the, in the legal terms of, of being able to, you know, redistribute the ip. And so this floodgate now has just been opened and the major networks are literally commissioning shows saying go ahead, use Gen AI. And so what we're finding that we're plugging ARC into some workflows and I can queue up a video right after I explain this that demonstrates this where with Gen AI video you can prompt a video into existence with little control. It's, it's again, we all know Gen AI is a bit of a roulette wheel. So the more you can reduce the numbers on that roulette wheel, the more accurate you're going to get with landing on number eight or wherever you're looking for. So what we're finding is that people, creative types with the skillset to be able to run image Gen AI, which is a creative task, it's your ability to describe and illustrate to an A.I. what you're looking for can be augmented with Arc, which is simple to use in the sense that you can create assets, you can scan rooms, you can put things, your own IP on the table and then you can bring ARC out and start filming it. And now you've got a video guideline that you can now Feed back into AI to style on top of. And it might be a good time to queue up this first video on. Yeah, I was going to give an example of like, I mean, there are some shows we're. We're working with arc on currently that are for broadcast that I can't really show footage from. But what I'm going to show is a bit of a demo because what we've been doing is looking at commercial animatics and just taking those as they were, hand drawn commercial animatics. And we've been kind of like just riffing off of them to see what we can do with ARK to replicate the same kind of thing. So this is just a tiny piece from, I think a car insurance company.
- [15:10]
Jeff Renfro
So just to provide a little context in advance of taking a look at this, the one of the things, as we were saying earlier on, we were, we understood the previs opportunity here. You know, with arc and all the value that accrues in production to being able to make previs more accessible. It's not that Previz hasn't existed. It's that it's. It's existed, but it's expensive, it's kind of. Kind of slow. It can be tedious. And it can also make filmmakers feel like they're working through, you know, being disintermediated and trying to find their creative and share their creative intent. So those are all things that are really valuable in terms of arc, you know, as a. As a fundamental tool. But one of the other really exceptional things here when it comes to what arc does is provides a reference that can be fed into AI. So the intentional, the intent of the filmmaker around a camera position, around the way that the camera moves, moves within space and a whole lot more gets to be sustained in the end. Creative. Whereas the text to video, text to image, AI is, you know, as Andrew was saying, a bit of a roulette wheel.
- [16:21]
Logan
I agree. And I think right before we show the video, one thing that I, I love about this is that you've kind of called out the point is that, you know, AI when it comes to generative AI, you know, whether it's video or text, it is a bit of a roulette wheel. Like creating a great image or great video. Not hard getting what you want in terms of character consistency and all of that, even more difficult. And then you add the scene on top of that, the shot angle, all of that, it's even more difficult. And that was what I think I love about arc, which is so great, is that you can make a consistent character. But the question is, is how do you get the scene or the color of the car or the style of the car, any of that Correct on model, consistent throughout. And by being able to bring arc in and basically do, you know, reference 3D scan of that, bring that in, you're eliminating one more, I'm not going to say point of failure, but you're eliminating one more potential variable that you have when you're trying to create really anything.
- [17:17]
Andrew McDonald
Yeah, you're taking numbers off the roulette wheel.
- [17:20]
Logan
Yeah, you're taking numbers off the roulette wheel. So with that I'll get off my soapbox and Andrew, I'll let you roll through your clip because this is really awesome.
- [17:27]
Andrew McDonald
Okay. So again, this is like just a, a copy of an existing animatic for a. For a car commercial, for an insurance commercial. Want to go?
- [17:41]
Jeff Renfro
Want to go, bro?
- [17:43]
Andrew McDonald
Hey, don't mess with my discount. You could save up to 30%. That's a short little clip, but yeah. So if you saw in the bottom corner, I'll just replay it with the volume off. So in the bottom corner here is arc with these establishing shots. And the interior car guy, I'm moving around him so it's picking up that camera move. And then I'm able to send that back into. I think I was using Sora here and just restyle the, the look, tell.
- [18:11]
Logan
Me from start to finish creating this shot, how long did it take you?
- [18:16]
Andrew McDonald
How long did that take me? I. Half a day, I guess. A couple, like three, four hours. So what I. The process was on the workstations we were using. There's many apps, but I've chosen reallusion because it's sort of a low barrier to entry version of Maya where, where you can do 3D animation with characters. The characters themselves are prompted into existence using another AI image. Three dimensional AI asset generator. I think we're using meshi for that. So meshi prompts in some characters. I dropped them into reallusion. I instantly rigged them and stick them in a car. And then I've got this asset that the two cars pull up and the two guys talk to each other. And I've just got this quick little animation that took a couple hours to build. And then I export that as an fbx and that now is a super lightweight file that I can give to arc. And now it's in Arc. Anyone can put it on the boardroom table, my living room out in the street. On the actual location, if you want to see how the cars Fit on that spot that you were going to film it in kind of thing. And then you just film it in arc. And then once you've got your, your footage, because it just records into the phone video library, you've got the footage, you can do what you want with it. So you can cut it raw from arc or you could take a clip like I'm showing there and just feed it into a AI video to video. So yeah, it is truly high speed ideation in the sense of it. In the sense that if you compare to what it took to do this kind of workflow not five years ago, it was enormously more complex.
- [19:47]
Jeff Renfro
Two years ago.
- [19:48]
Andrew McDonald
Yeah, two years ago.
- [19:50]
Jeff Renfro
Some of the things are like in the last five months, I mean, the facial performance and lip sync. Jeff, do you want to talk about that? You've taken a dive into that facial performance and lip sync and focal capture a lot.
- [20:01]
Andrew McDonald
Yeah, that, that facial performance in that piece I just showed was what you're talking about too, Jeff.
- [20:06]
Todd
Yeah, it's a great workflow because, you know, prior earlier this year, even the big comment was how can we get more expression in the face and how, how can we bring these sort of video gamey looking avatars and give them real personality? The workflow now is, you know, when we, we bring in these avatars into arc, we situate them in augmented reality or virtual reality, we hit record, we shoot our shots in arc, we bring those to workstation, cut it and we actually cut it first. And then once we know what shots we're using, we, we go in and we use one of these face sync platforms to, to drive those individual shots. As you saw Andrew, that was Andrew driving the guy's face there with that vocal and performance. So. And it's really fun to do as a director who directs actors a lot, to actually get in front of the camera and do some of these performances has been super duper fun. And it's fun to realize how you're puppeteering this digital character. And you know, just every blink and every eyebrow, it's all captured. It's really fun to do.
- [21:21]
Jeff Renfro
And while some of that stuff is processed after it's captured, I mean really key element of ARK is that ARK is the renderer is working in real time. There's no kind of offline render time here. We're setting up the shots, we're, you know, cueing animations, what have you, and then capturing it in the in camera. Letting a camera be a camera because it's real time. There's all sorts of benefits to that. And then because of AI workflows, there's things that would have taken days and weeks like the, the expression and animation of face and you know, you can get that done, you know, in a day or a couple max.
- [22:01]
Andrew McDonald
The other thing that, that's kind of cool is like as the fact that all these tools again are accessible to everyone and you can, you can kind of with this, with any. I think there's over two dozen apps now on the iPhones and the Android phones where you can walk into a room and scan it and end up with a pretty decent three dimensional scan of a place that, you know, prior to that it was go get a ferro lidar LIDAR scanner and spend $20,000. Now you can do it with your phone. So the fact that you can do this and then have it as a scan and then again, three dimensional asset generation can create a character. This is a little quick video I want to show. This was just yesterday. I was on a location scout for a spot that's being filmed at the Hockey hall of Fame. If you're a hockey fan, you, you'd know what that is. It's a V mecca for hockey. It's here in Toronto and this is spot being filmed there. So it's in the Hockey hall of Fame. There are these big sculptures of some of the more famous hockey players that it's like a night at the museum. They come to life. So we want it. We have got 3D assets that we're looking at putting in an environment of the Hockey hall of Fame. So I was there, I scanned the actual rooms in the Hockey hall of Fame and I prompted up template stand in character of Rocket Richard, the famous Montreal Canadian player. And I put him in the dressing room. That's a, a mockup of the original dressing room for the Montreal Canadiens. So I don't know if you want to look at this video. Yeah. Okay. So yeah, this is sort of like a quick example of the workflow. So.
- [23:27]
Jeff Renfro
So this is this screen capture right off, right off the mobile device.
- [23:30]
Andrew McDonald
Yeah. So this is arc. Yeah, this is ARC with its UI interface. So basically I've dropped my asset in that I've made of Rocket Richard. I go to the scan in your.
- [23:38]
Jeff Renfro
Kitchen, in my kitchen, on Andrew's kitchen.
- [23:42]
Andrew McDonald
And I bring in the scan of the actual dressing room. So this is something, you know, as, as a director talking to the production designer or the CGI guy who's gotta do something or, or anybody creatively on the team. You in the boardroom can quickly Put this together right now. I've just mocked. That's a 360 HDR background. Because as you can see, the scan needs some set extensions. Right. Cause I don't see out that door. So now I can just make it visually look really good. And then I move into lighting. So ARC now has a single dynamic light. It's pretty pretty much what the modern phones can handle right now. But pretty soon we can get a couple of dynamic lights in there. So that really helps blend the set together and also really helps when you're sending this as previs information to an AI because now it really gets the coherence of it all. And now I can float around in the scene like I'm on a techno crane or a Steadicam and find different angles. I'm going over here. This is. This is Rocky. Richard was number nine, was this famous jersey number. So there's the actual jersey in the. That's from the scan of the actual room. So, yeah, you frame it up, find focal lengths.
- [24:45]
Todd
You can.
- [24:45]
Andrew McDonald
You can previs your depth of field, all this kind of stuff. And then when you hit roll, you can roll and you get clean 4k video here. And so this is good for, you know, illustrating your concept or as I said, feeding it back into an AI for video to video AI post processing. And then here's. Right now there's a. Next is an example of. I just quickly ran it through Sora about five minutes before this phone call and got this.
- [25:16]
Logan
Incredible.
- [25:19]
Andrew McDonald
Yeah.
- [25:19]
Jeff Renfro
It's important to emphasize that your intent, right. How much your intent and what you wanted to shoot in that Andrew was preserved because of Ark being in that mix, that trying to prompt that to life could be hours, hours of work that may not be fruitful in the end.
- [25:38]
Andrew McDonald
On text AI for me, there's all these tools that have come online. I can make a 3D object just by prompting anyways, or scan a room and standalone. They're neat, but what do you do with them? And that's why it's so exciting with arc, because now it's like this aggregate place. You can. All these new things you can do. You can put it inside a camera, a virtual soundstage, if you will. And you've got this virtual camera that you can now do something with. A really interesting thing is when. When A. When A.R. first came out and there was. It was used in marketing and you could like point your phone at a product and it would make an AR dancing logo on the product. And it was neat at first. And people were like, oh, that's cool, but it's, it's, it's high friction. Like holding a phone out like this as a camera is kind of high friction. And if you're just getting low reward for it, like some marketing thrown at you, it loses interest really quickly and people stop holding the camera out. But I find when you're creating, when you're putting all this stuff in front of the phone, it's like you're filming your family or something, you're into it. So you, you can hold the phone out like this for quite a bit of time and do stuff.
- [26:40]
Jeff Renfro
It fulfills a purpose, right? I mean, one of the things that we think we've been talking about is like just this, you know, entrepreneurs and you know, being in a startup from early, early on is what problem does it solve, right? And, and when you think about how valuable at one point storyboards used to be, and then storyboards got augmented by being, by pre visualization and that the very logical next step in that world, right, is that we make that more accessible to a, just a huge number of people in creative and design endeavors in order to, you know, make these powerful tools, you know, powerful for them and not have to do it through somebody who know, who spent, you know, four or five years learning all the components of these, you know, powerful desktop tools, but that there's a subset of it that can be catered to these sorts of audiences and really empower them in a, in a super meaningful way. Jeff, did you want to talk about, about one of the pitches that you've done and as an example of how.
- [27:42]
Todd
Some of that works, when we started to dabble around with just the full workflow of, you know, can we, from soup to nuts, get into a space where we're, we're essentially playing like a mini Pixar, you know, where, where you can create your own assets, your own environments, you can shoot it however you want with arc. You can bring it back to the workstation, you can drive faces, you can video to video, style transfer. And you know, that soup to nuts can all happen in the space of your office, you know, or in my case, it's like my office and my backyard has become my studio. We got this, you know, to test one of these workflows, we came up with this concept of robots who have, have been bad robots and need to go to therapy. The workflow here was essentially generating 3D assets through gen gen AI 3D asset creation. So we, we just generate these robots that are standing in, in the, in the T position each Very custom. And they all have their own personalities. And then we rig them and we. We use just a simple mocap again, through. Through the iPhone. Set up your iPhone on a tripod and. And act it out. So what you're about to see is. Is all me. I'm doing all of the movements through this. I'm puppeteering all these robots. And then we retarget that animation onto. Onto these specific assets, and then we import them into ark. Now I can place these robots around and in what we're about to show you. I did it in my. Just in my backyard. So I had them sitting around in my backyard. As a traditional filmmaker in the space, I'm now in this sort of Pixar world where I have these animated characters in my space, but with the camera in my hands, I feel like I'm on set. So rather than talk about storyboards and what's the shotgun to do, I just do it. I just shoot it. I just walk, physically walk over there. I get the idea. I'm like, no, actually, for this beat, I do need to be over this guy. Looking back onto this other robot. It's super fun to be working in dimensional space when the ultimate content is. Is. Has a kind of animated quality to it. And here, let's just roll the top of it now. Andrew and.
- [30:14]
Logan
Hi.
- [30:15]
Todd
Yeah, okay.
- [30:16]
Andrew McDonald
My name is.
- [30:17]
Logan
Mine's Rob. I had a little bit of an incident at work, and I have to attend group therapy now. Hopefully it'll go well. A little nervous, but.
- [30:29]
Andrew McDonald
Yeah, let's not give it a shot. How bad could it be?
- [30:35]
Todd
Okay. Okay, let's all settle down.
- [30:37]
Jeff Renfro
Sit down.
- [30:37]
Andrew McDonald
Sit down now. Sit down. I would like you.
- [30:42]
Todd
Essentially, what you're seeing here is. It. Are these robots that I've placed in my backyard. And if you were to come over to my house and see what's going on, you would just see me dancing around my backyard with. With arc in my hands, pointing at air molecules. But I'm really into it because, you know, I'm watching these robots just right in front of me. And to be able to, you know, get into that spirit of. I want to zoom in a bit. I want to snap, zoom. I want. I want a handheld feel around these digitally animated objects. It. It's. It's just so fun to work in dimensional space and. And to be. To. To feel like you are. You're shooting in a regular situation with a camera that's got all the focal lengths that you need and to be able to give it a look and then take it back to a workstation and give it this polished chrome look. Once we stumbled onto this and we're like, okay, that's the look for the show. And you know, we do that thing where we copy the seed. We found a look we like, we copy the seed and then the video to video style transfers. Once you land on something, they come out so amazing. So this is, and this, this one here is, is a virtual backdrop. You see the movement of the pool, stuff like that. It's just it, that's what Gen AI video to video is so, so good at. When you give it this, this starting, this input footage that is so dimensional and so custom and so articulated and all of the nuances of the gesticulation, really bringing these things to life, the.
- [32:24]
Jeff Renfro
Workflow, you know, is essential and being able to figure it out for professional audience as well. Like it's nice to have access to some of these things, you know, in the cloud. But for professional applications, a lot of times or often the right answer is that you need, you need private deployments for the purposes of content security, for the purposes of, of all sorts of the IP dynamics and the rest. Being able to have the Dell workstation we have and be able to sort out these workflows and how they work independent of, of a cloud based service and how we make them all work together, you know, is a really essential piece of the equation too.
- [33:02]
Logan
Agreed. I mean I think, and we're, we're up against it, you know. I want to make a comment. Jeff, just random question. Were the cops called? When you were dancing around as a potential robot in your backyard, did you get in any trouble? Did the HOA get involved? No, I'm just kidding. I, I just think it's so interesting, right? It's like really the section, the intersection of, you know, creativity and you know, workflow optimization. Right. At the end of the day, like what AI is doing, I mean, I think what you've seen in this episode is AI. What is doing is it's enabling creative people to do more and also taking away the part historically maybe that was a little bit not the funnest part of the job, right. It's allowing you to accelerate the workflow. Key takeaway. It's not scary. I mean as artists you guys were still in full control. You had control of the shot, the way the robot looked, the design, all things you still had ultimate control. And I think that's the key takeaway that you're seeing from you know, obviously Zagazi and their ARC product. But also, you know, Other media entertainment folks we've had in the show. It's like you're still keeping creative control, but it's allowing you work faster, quicker, smarter and allowing you to be more creative than you really have ever been.
- [34:10]
Andrew McDonald
Exactly. It's putting the human creativity back into using AI. Like when it first came out, AI was in, in university labs and it was like literally ripping off other people's IP to demonstrate what it could do. And that gave it that scary like what is this? This is just stolen stuff. But some of the processes that we're figuring out, there's so much human involvement, we're just, we're just using better paintbrushes now, but it's still human creativity.
- [34:33]
Jeff Renfro
And I think it's one of, one of the real benefits of us being a, a largely filmmaker founded company too, that we're the users, you know, we're the users. We know how it gets used, we know the pain points and we can, we can dive into figuring it all that out as we're, as we're going on this journey.
- [34:50]
Logan
Yep. Makes total sense. So in closing, you know, Todd, Andrew, Jeff, appreciate having you on where can and maybe Todd, you handle this where if someone's out there watching, you know, the, the podcast or the webinar and they want to go check out arc, tell them where to go find it, how to download it, where to find you and then we'll close out best.
- [35:08]
Jeff Renfro
Way we, we're, we're in private beta now, public ba beta now I guess. And the best way to do that is by going over to Zugazi.com and sharing your info with us there. And we, we release additional distribution on that on a regular basis and always eager to get more professionals in the mix and getting all sorts of people's feedback. So that's been really good. And find, find me on LinkedIn easily and that's probably the best way to, to reach out.
- [35:39]
Logan
I love it. And like with every episode of, you know, reshaping workflows we try to bring in, you probably get sick of hearing from me and that's why I don't talk because I'm not really that interesting. If you want to hear about speeds and feeds, I can talk about all day, but that's why we bring the people in doing the actual workflows. Talking about how Dell Pro Max and Nvidia RTX GPUs are really accelerating the workflow process and being able to, you know, honestly, I'm not going to say invent, but I'm going to say reinvent the way that workflows are done in kind of, you know, the 21st century with the advent of generative AI. So with that being said, you know, Todd, Andrew, Jeff, really appreciate the time today. Until next time, we'll keep your workflows running and make sure they're running on Dell Pro Max and Nvidia. Thank you.
- [36:22]
Jeff Renfro
Thank you, Logan.
- [36:23]
Todd
Thanks so much.
- [36:24]
Andrew McDonald
Thanks.
- [36:33]
Logan
Do what you want.
- [36:34]
Todd
Do what you want.
- [36:46]
Andrew McDonald
This podcast was produced in partnership with Amaze Media Labs.