Latest / And Now For Something Completely Machinima / S6 E205 Machinima News (Dec 2025)
Transcript
- Tracy Harwood 00:42 hello and welcome to this week's episode of And now for something completely machinima, I'm joined here today by Phil and Damien Ricky's not with us. Where is he this week? Phil? Phil Rice 01:15 Do you know he is on an eight hour he's on an eight hour bus ride from from Tucson to Las Vegas. Tracy Harwood 01:28 Oh, yeah, preparing for Phil Rice 01:31 himself, just driving on empty bus. Yeah, Tracy Harwood 01:35 excellent, excellent. No me either. Absolutely, I would not envy that at all. Anyway, this week we're here, we're going to be talking about some news items that we've got, and I'm going to be kicking us off with something that I've picked up and and then Phil's got a few things that he wants to share with you. So let me start. Then it's, it's really about YouTube and the changes that have been revealed in how it is prioritizing its own platform experience and increasing privacy according to what it's saying, and giving creators more visibility on their own channels through embeds, albeit with less control for third party sites, such as those you might embed into places where you might want to stream, such as Second Life and What have you so through an API, which now don't seem to any longer work well, which is a bit of a problem if, if you're doing a show through a second life venue, for example. Anyway, crucially, I think, from what I have understood of what they're doing is that it's how it's responding to the slop that's been dominating the channel for the last few months. Now, I found a really good summary of what the changes have been implemented by YouTube. Apparently, around mid November, some of you may already have noticed that quite a few of the slop channels have already been removed, which is a good thing, but what's interesting is the way that that it's been done. It's now using Gemini to check the content and whether that's human, whether you know whether what you're doing is human, derived and original content, and basically how the content that you've put out has actually been created and described, implying that you now need a little more thought in the way that you then present the videos you produce on YouTube. And as I understand it, it talks in this video that I'll share with you about what's going on. It talks a bit about something called trust scores for specific channels. But these are actually not things that we can access and assess in terms of our own kind of content or your own content. It's it's basically just a sort of an internal metric that's that's used by them to evaluate your content. So I thought I would just take a little look about how a trust score might be derived for our channel, and therefore how Gemini might actually describe our channel so that I can really get a bit of a handle on whether it's, you know, trustworthy in any way, I suppose. Anyway, what it did when I asked it about our channel was described us as a passion project and a podcast of in. Individuals with extensive experience in the field, which suggests a high level of expertise in our specific niche, although we are a small channel now, this channel is a distinct entity from the original, now defunct media machinima Inc, channel, which faced significant controversies and a poor reputation for its practices, and it goes on to describe some of the practices that machinima Inc was known for. We are not that channel, and we never have been that channel, although we had various associations with the early machinima.com and and the website, long before it got taken over and became what it became as a multi channel, network, multi network, channel even. So, yeah, so it, I think, from what I understand, really is it quite accurately describes who we are. It talks about us having sort of over 80 years of machinima mastery combined between us. That's Phil me, Ricky and Damien. In actual fact, I'd probably say it's a lot more than that. Now, since we've been doing this for five odd years, it's Damien Valentine 06:16 interesting to stop and think about that for a moment. Yeah. Tracy Harwood 06:21 Yeah, yeah. Well, combined, add it. Well, that's how I did it anyway. Interestingly though, it also picked up our current subscriber number at the point that I did the evaluation, which was, which was kind of interesting, because most of these tools do a retrospective, and they're at least few days old, but this didn't seem to be doing that. It definitely distinguished us, though, and it was quite interesting in the way that it did it. So, so there you go. I think it's, it's interesting that the algorithm is advancing in a way to address some of the real, significant concerns that we've raised here on this channel, in terms of the the impact that all this swap is going to have on Discovery of content, quite how that will pan out over the next few months. Because, I mean, there are literally millions and millions of channels that have emerged with totally pointless generated content in in the last few months alone. So it's a massive task they've taken on. So it'll be interesting to see how that unfolds, and what the impact of that will be on on genuine channels producing their own genuine content, such as ourselves. Anyway, that's all my news for today. I'm going to hand you over to Phil. Phil Rice 07:38 Now I'm curious what made you draw the connection between the Embed problem on Second Life and the new content policies? Because I'm not seeing a connection between the two other than maybe a coincidence of timing. Tracy Harwood 07:54 No, I think it's, it's to do with the way the whole thing has, haven't had an over Hall in that period of time. And the key thing that I saw was that one of the ways that they're tackling it is to sort of stop people just embedding stuff, and the other way is all this content. So I I see it as a completely rounded solution that they've addressed, they've used to address the problem set of problems. Phil Rice 08:21 Basically, I know part of the Embed issue for them is that some of the certain types of embedded browsers could circumvent their ad delivery, yes, and so they basically cut off a whole category of browsers and kept it to a kind of a controlled set, and it second life's embed browser just happens to fall into that category. Yeah, yeah, but it's that for that particular move, I feel like is not as much about the overall AI initiative, which is big, that they're trying to get control of that and clean it up. But also, I think mainly that one is more of an ad revenue thing, the article that we'll include in the notes about the second life thing, there are some very reasonable workarounds, including, you can still use a YouTube video. You just kind of have to do a little bit of work around to you're actually, ironically enough, you have to embed the video on a website, and then you can embed that into Second Life, and it will work. Apparently, you can also use Vimeo, but Vimeo has become a, to me, a less attractive platform because of the pretty significant price hikes. The only reason I keep my Vimeo subscription anymore is because, mainly because of the Machinima, machiniplex collection that we've hosted there and put there if I were to discontinue my. Subscription, they would take down a good portion of those videos, like my overall allowed storage for even past videos would be like shrunk down, and they would just get rid of stuff. It's really that's lousy to me. It doesn't really cost them anything too significant to host videos. But, you know, I mean, they're a business. They got to make money, so I get it, but Vimeo used to be a place where, even as a free creator, you could go and easily host quite a bit of content. And they've just there's been some decisions that they've made that have made it a little bit less friendly to just general artists and more about being a prestige platform, a paid prestige platform, which, if that's what you're after, it's great, because it gives you a lot of control over how you embed and things like that. If you're a paid subscriber, they've lifted a lot of the file size and storage limits for paid subscribers as well. Like there used to be some kind of a limit. I want to say it was five gigs a month. For the basic plan, you can only upload five gigs a month, which seems like a lot, but if you were uploading regularly, you can hit that pretty quick. Or if you're doing 4k videos, for example, you can hit it pretty quick. They've, I think they've all but eliminated that monthly limit now and then, I recently got an email saying that they had greatly increased the amount of total storage that I could use to something just astronomical. That move makes sense to me, but the for a user who's not willing to pay them a regular subscription, it's not so much a useful place anymore. Tracy Harwood 11:37 So it's locked you in then, hasn't it? Really, for the Yeah, it kind Phil Rice 11:41 of has, I mean, my renewal happens in November, so I've got 11 months to figure out, yeah, what I would do about that. But the problem is, we've, you know, I put all these bitly links out there and stuff like that. Okay, all this, this is where the machiniplex collection is like. So it would be quite an endeavor to move the stuff, but we'll see. Tracy Harwood 12:01 Could it not be pointed to the archive, Internet Archive? It? Phil Rice 12:06 Could I just find the viewing experience over there? Yeah, I agree. Pretty, pretty terrible. It's about, it's about 10 to 15 years out of date in terms of the viewer functionality and things like that. It's a great resource, and certainly is generous in all the ways Vimeo isn't anymore, definitely, but the actual viewer experience there is a little bit lackluster, needs some updating. So okay, um, yeah, my news, so I've been dabbling a decent amount in the past month or two with Blender. Now, I've dabbled with Blender over the years and used it mainly just as a gateway to get models into iClone and things like that. You know, sometimes you have to download a model in Blender format and then export it. It's just basically been a pass through tool for me, but there's some stuff that I'm working on right now where I'm actually building and rendering it in iClone, excuse me, in blender. And so it's kind of gotten me a little bit more exposed to some of the tools and community of freely available add ons that are available for blender. And there is a there is a blender, I clone pipeline. I just have not done anything with it yet, so I'm not sure how that part of it works. So right now, the work I'm doing in it is purely just, it's all Blender. I haven't even tried to connect it up to iClone yet. But during the course of researching that came across some interesting things. So one of them is, it's called the third person kit. This is made by a person named Riley Brown. They are Riley B, 3d on x, if you're on the X slash Twitter platform. And basically this flat on turns blender into a third person character controller. So if, if you've played video games, you know what third person, you know view is, if you've used iClone, it's like motion director, that type of thing. But you can actually, you can actually manipulate a character in blender in that real time format. And I thought, well, that's as that's as machinima as things get. You know, that's that's really a core thing. So you can customize it with your own animations, or you can pull them from available libraries like Mixamo. And I think where it might be really interesting is if someone was working with the iClone to blender pipeline, you bring in a character from character creator four, you bring in, maybe bring in animations somehow, from I clone and then control it in real time inside blender. It's an interesting possibility. In that same vein. This hit my feed just yesterday, so I barely got it into the news here, but someone by the name of Daniel byts has released a freely available plugin for blender that is a full on cloth builder and simulator. Wow. Yeah, the the comparable software for that would be some, I believe it's called Marvelous Designer. There's another competitor that's a 3d coat, I think is the other one that does that. Marvelous Designer is kind of the gold standard, though, for that type of thing where you can actually design cloth with physics, design whole outfits and stuff like that. Well, this person has come up with a way where you could do that all within blender. Now, some questions arise for me as somebody who blender is not my primary animation platform, and that would be, would I be able to take what I create there and bring it into iClone, because I clone has an amazing soft cloth physics, you know, side of it. And when clothes are designed with that in mind, they can be they can be really remarkable. And that comes from a lot of exposure to clothing for I clone that's available, that's maybe not as well designed, or is designed for an older generation and doesn't quite fit the characters well. And Damien, I'm sure you've dealt with this some. It takes some work sometimes to really get it to take on the shape of the character and to not have the character break through it. And, oh yeah, there's a lot of tweaking involved to get that right. Sometimes you don't notice until you're already an iClone, and then you have them do some kind of a stretch they didn't do, and it when testing and character creator, and you have to bring it back over and fix that somehow. There's a various, various approaches to fixing it anyway. This is that cloth design type approach with physics and everything. There's a really nice video that that demonstrates how it works in a tutorial. So we'll, we'll include that in the link, but it's a really interesting and encouraging development. Probably not for the people at Marvelous Designer. They're probably not Harold, happy about it, but there is a market of people who can't afford, what is it? 300 bucks, I think, for Marvelous Designer. And I'm not even sure that that's a perpetual license or not. These softwares are priced very reasonably for creators or studios that have a decent budget. But there's a whole range of people who either create or want to create, that do not have that kind of money, that it was all they could do to get a PC with the right graphics card, you know? And they can't fork over all this money for all these different softwares to get it done. And if they do, maybe they pick one thing to invest in, like a real illusion package or something. And then they can't keep buying Substance Painter and world builder for terrain generation, and each one of these is 300 bucks a pop minimum. So I like the fact that there are options out there for people that have less of a budget to work with that they can can delve into territory like this, and clothing is one of the hardest things for me. I think in iClone that if you can't find it, like, if someone hasn't already built it, man, it's tough. It's really tough, like you can't typically, for most types of clothing, you can't just go download an FBX model and bring it in and it's going to perform like clothing does. Hair is another one that's really tough. If you're Damien Valentine 18:45 after some clothing that's very specific, like with the Star Wars stuff I'm doing, I wanted the uniforms to be like they are in the films. Of course, no one's done that, or the x wing flight suits. Again, you can't just put somewhere an orange outfit and it passes. It needs to look right. So I did mess around. So my solution to that was I couldn't find any models to import, other than full characters that someone ripped from one of the Star Wars games. But then you can't change the faces, because that's part of the model, and you can't have all the officers looking identical, because some of them are actual characters. Like, it's fine for the background, but then you have, if I did that, then you'd have background characters with the really good, accurate uniform, but the ones who are at the front of the camera are the ones that don't have that. So I did, I cheated, and I found a shirt and I modified it so the buttons were invisible and changed the texture. Phil Rice 19:40 And I've done that very thing so many times, yeah, absolutely, yeah, yeah. And the thing is, is, this is, this isn't a slight against real illusion that they have some incredible tools to bring in content you find somewhere else. Accu rig is, it's just absolutely incredible. That tool that you can take, you can. Find any humanoid 3d model, I think even for non humanoid, you can do it, but I've never tried that. But any humanoid model you can bring in, and if it, if you can get it into a T pose, you can rig it like with relative ease, and it will work. But when it comes to fine tuning and customization, like, like, what you're talking about, Damien, like, you helped me find some models that I needed for that hair to the Empire spoof that I did a year or so. But, like, yeah, there were challenges like that. Like, I found a Darth Vader that someone had exported from a fortnight model or something. But then I bring it in, and there's no physics on his cape. It's stiff as a board, yeah? And I, so I had to kind of figure out how to add that myself and make it not just fall off his back and yeah, it's, it's, it's a it's, there's a whole science to to that. So anything that lets you design stuff from scratch, I think is appealing. Now, it doesn't mean that you won't need skill to get it done. You will. You still need to kind of know what you're doing. You need to learn to know what you're doing. But the alternative of trying to do it without a software that's designed to make that clothing. Yeah, your options are pretty limited. You've got to, got to kind of carve it out of what's already existing. And it's, it's, it's quite labor intensive. And often the, I think, with my Imperial soldiers in my short Damien, I ended up that they don't move very much because of that, the fact that I couldn't really customize much of anything. I managed to separate the outfit from the person, but, boy, I couldn't. He wasn't going to be doing any jumping jacks or anything. So fortunately, the script didn't have any jumping jacks my Star Wars film. So I got lucky there. Yeah. What else speaking of revolution they have. I was just talking about Accu rig, that is one of the most amazing tools. I mean, it's, it's technically, it's really only useful if you've got the iClone and or CC four suite. But it's amazing. Their lip sync tool as well, Accu lips unbelievable. It's so amazing. And now they've come out with yet another kind of stepping into really forward thinking kind of stuff. Now the people over on the iClone Facebook group will complain about this because it they say that the biggest problem with iClone is the rendering. Of course, like, no matter what real illusion does, they can't win because it's not the rendering. So anyway, there are solutions for rendering, if you you know, if you just apply yourself, you can find them, but video free solutions, yeah, absolutely so. Video mocap. This is something that's it has been around as a thing for a while, basically something that can take video footage from your iPhone, from a webcam, from any source, really, and it attempts to interpolate from that footage to create motion capture data that's then usable. And there's lots of tools out there. Some of them are even free. Some of them are paid Rococo themselves, even though they sell the big motion capture suits and all that, but they have a video to a it's aI driven motion capture from video. And those tools work fairly well, some of them better than others, but they can be useful if you don't have the the pretty expensive gear to get it done in for real. Well, I clone or real illusion has released their own brand of this tool. It's video. They call it video mocap, and it is the functionality of those tools of you bring in video and then it basically can extrapolate the motion data, but then it's like immediately right there, starting to integrate with the tools that make iClone what it is like the it, as far as I'm concerned, it's the leader in making character animation accessible to people right away. You can optimize clean up. You can capture multiple characters from the same video. It's amazing. It's an absolutely amazing tool. And these things aren't free to provide for them, so they're not free to use either. But the way. They've got the pricing set up. It's using kind of a credit system, very similar, if you've used any, any AI tools at all, a credit system that you can buy a bank of those, and it's a very reasonable amount to capture 60 seconds of motion and also clean it and have it basically be immediately ready for use with within iclo. In fact, you can, even as you're auditioning and seeing how the motion capture is turning out, you can pick which character it's applying to, so you can see it in real time. Like, what is this actually going to look like on my character of choice? It's a brilliant, brilliant innovation, super timely. It's just just tremendous. So that came out within, what is that Damien, maybe two months ago at the most. I think it might have been even just November, yeah, I think it was November. Yeah, it's fantastic, really, fantastic tool I keep looking at. Damien Valentine 25:58 I think I need to try it, but I haven't had time to. Actually, I was going to sit down. But the point of it is, you're not supposed to sit down. You're supposed to do something. I've tried other video motion capture tools before, and this was in the early days of it, so it was very rough. So I'm hoping this is going to be a lot better. Phil Rice 26:19 Yeah, I've tried some of those tools much more recently, and it's still pretty rough. Okay, now some of that has to be there are certain optimizations when you're doing video for mocap optimizations in your filming environment that are going to greatly improve the result. So first of all, getting yourself fully in the frame, having a contrasting and unbusy background. These things aren't mandatory, but they really, really, really help. Damien Valentine 26:48 The other thing go ahead, well, the other thing is mandatory is space. Because you're you do need the full body to be visible. So if you've only got tiny rooms where you can, you place the camera right up against the wall, and you're on the opposite wall, and you can't see all of you because you just isn't space to get yourself in the frame. That's not going to work. Phil Rice 27:07 And speaking of which, one of the nice things in video mocap that is available in some of these other video to mocap tools, but not all of them, and that is, you can actually set the mode to worry just about the upper half of the body, okay? Well, a conversation, or just someone talking and only worrying about capturing the upper half of the body. It that's that's a mode that you just flip a switch and it will do it that way. So, and the results from the the tests that I've done, I haven't done it extensively yet, but the tests that I've done. It's tremendous. It's really tremendous. So, and I think when you get it, they give you, they give you a certain amount of credits or a couple free passes to play with that you don't have to spend anything just to, just to audition it and see how it works, and stuff geneticism as well, because, yeah, and there's very nice tutorials as well, like absolute step by step how to do it, that are very, very well produced. Tracy Harwood 28:07 What were you saying? Tracy, I was, you know, because we've done quite a bit of mocap testing, but we end up using Connect cameras. And the problem with those is, well, you need more than one. It's old tech. Now you can't actually buy connects. You know, it's, I don't know how many years out data is, and, and there's a lag that's so noticeable if you're going to try and do real time, you know, real time puppeting kind of thing. It's quite, quite problematic to use. Do you get a sense this is being pitched at puppeting, real time stuff on a, you know, stage performance, or, I don't Phil Rice 28:57 think so you think, because of the way that it's so deeply integrated into iClone. I don't think it's intended for real time stuff. In fact, I'm I'm not certain whether or not you can feed a live camera into it. I haven't tried that. Okay, they get once a video file, right? Okay, yeah, so and I clone itself really isn't structured to to to support that type of real time. Yeah? So I don't think that's their target. I think it's more it is animation, yeah, video, yeah, video to animation type of thing. Yes, interesting, very interesting. Tracy Harwood 29:33 Yeah, and it does fingers too. I'm definitely going to be taking a look at this, because I think this has got lots of potential. Phil Rice 29:40 Yeah, it's really sharp looking. Yeah, all right, speaking of sharp looking, Unreal Engine, there's some new animation features in the Unreal Engine. 5.6 there's already talk online about 5.7 is 5.7 already out now, I don't even know it's that they're really. Just they're not slowing down one bit over there. I didn't dig in a whole lot to the details, like a bullet list type way of the features, but the video demonstrating is animation. Just to give you an example, it's there's a giant octopus crawling through the street over cars, an octopus, maybe like the ultimate animation challenge, you know, and it looks spectacular. So, yeah, Unreal Engine. It just keeps, just keeps pushing things forward. My feet is constantly full of people just raving about the features. There's Tracy Harwood 30:42 eggs. Exactly the same, Phil, I just wish we would see more good quality films come out with it really, yes, there's no excuse not for the content not to be there, but it just isn't. Phil Rice 30:55 Yeah, a lot of their marketing is apparently they have determined that their biggest potential market is makers of games, and there's a lot of competition for that right now between them and unity and Godot, they're all kind of fighting over wooing development teams to to use them to make their Games, which is fine, you know, I mean that there's gaming is a multi billion dollar market. Of course they would, they would be interested in that. And let's face it, the filmmaking market isn't, isn't, isn't in its best place right now, you know? So I get it, but the potential for film there is, is terrific. And if you want an example of that, look at the film that we at the film that we talked about a couple weeks back with the James Bond film by Ben Tuttle. He He did the animation in iClone, which makes sense, but all of the sets and everything and all the lighting and all the rendering and all that are done in Unreal and it's gorgeous. It's just a gorgeous platform. And Damien pointed out, when we watched that film, that what was nice to see from that film of Ben's is that it not that we don't enjoy watching glitzy fantasy sci fi stuff. We like that, right? But it was nice to see somebody do one that's more of, more of an approach leaning towards modern day and realism. And they and he pulled it off. You know, the look, as Damien mentioned, he was London looks. It looks exactly like what London should look like in that context, you know. So, yeah, I wish, I wish more people would would delve into that. I'm sure there's a learning curve challenge that that keeps people at bay. What I've been discovering because, yeah, I mentioned earlier this episode that I've been taking blender a bit more seriously for some some work that I need to do, and what I've been surprised by is how much of the knowledge that I've gained from working in these other platforms over the years, movie storm less so, but iClone in particular is how much of that knowledge translates Like how many of the concepts with regard to lighting, with regard to key framing, with regard to, you know, setting the stage for a scene, it all translates. And I have to think that that's going to be the case when I dive into Unreal Engine as well, which is in my plan. So, yeah, I would encourage you to give it a try. There's a lot of great tutorial content out there for these platforms. There's some paid ones as well, which I don't I've never done any of those, so I don't know what the quality is, but, man, there's people selling unreal courses just left and right. You know, you become a professional filmmaker like me, this person you have never heard of or recognize, but I'm a successful Hollywood filmmaker, and I could teach you how to make this Godzilla VFX scene, and in just 30 days, a lot of it just feels like just, just hype, you know? So I don't know, but hey, if you don't want to plunk down money for a course, just there's a lot of do it yourself. Learn as you go. Content if you've got any kind of foothold in creating mission, a minor virtual production, my recommendation wouldn't be to go find some beginner tutorial and then work through 30 hours of course work, because there's just going to be way too much stuff in there that you already know. It's just not going to be interesting. I think a more useful approach is go in there and see what you can figure out. And when you find something that you're stumped by, or I wonder how this works, this didn't then go look for something. On that. Go look for instruction on the thing that you don't know how to do, and you that's how you can fill in your gaps you know, on what what you know and what you don't. So to me, that's that's less daunting an idea. Then, okay, let's, let's take a college level course. You know, the equivalent study level of that to learn something new. It's like, Man, that's a big time investment. It's very intimidating, and frankly, it's probably not going to be as interesting as you hope. If you've got any experience at all, then you already know how to do more than you think you do in an engine like Unreal. So that's my Damien Valentine 35:40 advice I'm going to give is, I have a film I made with movie storm, called it alone, and I made it, it's like seven or eight minutes long, and I made it specifically to learn how to use movie storm. And then when I switched over to icon, I thought, well, I need to do something. I want to make something, to learn how to use it. I remade that film. And then when I wanted to switch to video Omni do and video Omniverse rendering with icon, I remade it again, and I just remake the same. It doesn't matter. It's remade. It teaches me everything I need to learn, like how to move cameras, how to do the lighting, how to do some basic effects work. And so for our listeners, if you've got a short video idea that's something like that, just make it learn. It'll teach come up with an idea. It'll teach you what you need to learn. And then you can apply that to other things. And if you do switch software in the future, remake it because you know that film, you know how it's going to look. It doesn't matter if you're you're remaking, you're not doing like a reboot B master Hollywood kind of thing. You are teaching yourself how to use the skills that you need. And like the Phil said that a lot of the software, the skills do overlap, so you probably won't necessarily need to do that again, but if you do that's a good way to do it. And you know, if something like icon gets a huge update that completely overhauls the lighting system in the future, go back to that project. Relight it, because it'll teach you how to use a new system. You can get it the way you want it to look. So it's really smart Phil Rice 37:21 eliminate extra variables. You know, you don't have to rewrite it. You don't have to, yeah, you can focus on that one segment. I took a very similar approach when I was learning I clone, I didn't remake the same short only because I didn't think of that. That's a good idea. I should have done that instead. Instead, I picked I set up a series of short film ideas where I would specifically focus on getting good at one particular aspect of it. So that's why, and that progression still continues to this day. There's a series of shorts that I did called Ralph and Chuck, which essentially there's very little animation in that it's two guys standing in a bathroom talking to each other, because I wanted to focus on getting face and lip sync and mood and all that with close ups. I just wanted to focus on that. So that's the whole reason, production wise. And I did that series well, then the next one was called odad, and that was two people sitting in a car, again, very little animation going on, but taking what I learned about the face and all that, but now I'm also I'm trying to improve compositing skills, bringing in footage sources from elsewhere, making it all look like it belongs together. And so that was a great kind of petri dish, petri dish for me to do that. And then now, now I'm in production on something called Lance and Nance, and that's like adding even more to the mix. Now I'm going to change settings every episode, and now there's going to be a little bit more action going on, and there's going to be background characters and all this stuff. So it's the same, same basic idea that you were talking about Damian, of you pick stuff that you want to make that also teaches you while you're working on it, you know? Yeah, so it's Damien Valentine 39:15 relatively small, so you're not looking at you're not teaching yourself how to make a four hour film. Great. That's, don't do that, something that's just a few minutes long, that gives you enough to play around with, and you can teach yourself exactly what you need to make it like I wanted, obviously, characters moving around and talking. So that's the lip sync and creation animation. There's some flashing lights. So it taught me how to animate lights. And do that I wanted some smoke to appear, so it taught me how to use the particle effects to make the environment fill up with smoke. You know, things like that. It worked, and I'm sure them again at some point. Phil Rice 39:59 Yeah, that's. At some point you know, you have to release a reel that shows them all in sequence. Damien Valentine 40:04 Would be great. That's a good idea. All Phil Rice 40:07 right, let's talk a little bit about AI, because it tends to come up here from time to time, the world of AI music has had some interesting legal developments just within the past month. The two main AI music platforms, one of them is, I think it's pronounced Udo, or Udo you do, and the other one is suno. They have both made settlements with the major record labels. There was, there was legal action basically threatened against both of them, and because there's this whole question about, you know, the sourcing of AI training, and there's been some interesting arguments made, particularly with regard to music that well, one person, I thought, put it really interestingly that that, like suno, for example, is not actually using like samples of music when it's When it's creating something, it has, it has, basically, it's a machine has studied certain vibes, of music, certain styles and and learned to emulate those sounds, but without actually pulling any samples. So their their point of view on it was so that's therefore it's not theft. I don't think it's that cut and dry. I think the the the training set matters, you know. But anyway, that was the crux of the beef that the record companies had with these AR music companies. Well, they both settled, and essentially what it means for both of them, and I can't remember which music company made the settlement with which thing, but both yudio and suno are going to be completely reworking their AI engine so that it only focuses on the record catalog of the company that they settled with, and of course, that is licensed, and there's a fee arrangement that's been taken, that's been put in place. All the previous music is just going to be for the purposes of the lawsuit and the legal action. All that is going to be completely disregarded, which is really interesting. So everything that's been created to date and up up through a certain time. Whenever the switchover happens, there will be no copyright claims or anything like that. But after that point, the only training database that it will use is the one that it's licensed for. And Apparently that was enough of a agreeable settlement settlement that the record companies agreed to it. So interesting. Development, fascinating, yeah, so it'll be interesting to see what the response is from creators, creators, particularly those in the kind of, let's say rabidly virtue signaling anti AI crowd, right? You know? Well, this, technically, is supposed to answer their main beef with it. So if they still remain anti AI, then they're going to have to express what the other reason is, which I suspect I know what that is. It's kind of a natural and understandable resentment of someone being able to just type in a few words and create passable sounding music, you know, so, and that's a harder, I think that's a harder battle to fight, because it's like, Who are you fighting against them? You know, you know, it's, it's a very interesting discussion, so, but yeah, that's what's going on with with suno. Will include a link to the suno one. If you were to Google, you could find the Udo article as well. But yeah, they both settled within the last month, and I think the switch over to the new thing is a few months away, because they have to basically completely rework the the AI itself, or the training, the training, which is, yeah, yeah, yeah. Tracy Harwood 44:36 So, so anybody that's it's sits outside of the catalogs as a creator. How does that work? Damien Valentine 44:47 Then the music would not be included in the catalog, so they wouldn't be their music wouldn't be used to train the AI, right? Tracy Harwood 44:57 But that still means it could be used, though, yeah. Uh, by what another AI? Phil Rice 45:03 Well, another one, yeah, yeah, but Damien Valentine 45:05 that's not these ones that are settled, Tracy Harwood 45:07 but, but even, but so presumably the ones in the catalog are only the ones that are licensed to be part of that catalog. But there must be 10s of 1000s of indie creators who are not in any licensed catalog that could be used Phil Rice 45:26 well, but by who? But by little AI company? Yeah, yeah, they could. But the thing is, is those little AI companies do not have the cash to fight legal battles for that. So all it has to be is for the record company to find out, and they can crush those and frankly, they'll have every incentive to do so now they have no incentive to settle with those little companies. Because first of all, suno is far and away the biggest platform for that. Like I think it's at some point, very recently, it was the number one downloaded app on the Apple App Store, not just in the music category, period, it's huge, just 10s and maybe hundreds of millions of users of that. Udo is large, but way lower than that, and then after that, it's just crickets. So it's a matter of, will there be some little company who will dare to poke the bear? If they're smart, they won't, because they'll just get crushed, like there's no financial incentive for the record companies to even settle with them. They'll just do cease and desist and try to stop them. So I think that will go away. Damien Valentine 46:43 So let's suno and udio. Their agreement is only use the music that they've settled with those record companies, so they can't go and scrape a load of music that indie music that people just put up on online without any representation, correct? Phil Rice 47:02 Okay, okay. And I would assume that, built into this agreement, they don't release all the details, of course, but I would assume built into this agreement is some kind of accountability measures to make sure that they're not doing that, you know. So think of it like the, you know, we send the UN inspectors in after a dictator has, you know, made a settlement with with the UN that, okay, I'm going to stop doing this bad behavior. And then the UN wants to send in inspectors periodically to make sure you're not really brewing up anymore. You're, you know, refined uranium. It's that type of thing. I think that there'll be some watchdog activity going on, because without that, who would ever know like you can't tell from listening to music made on suno, you can't tell which record, which records were used to to learn that you can't tell there's nothing there that tells you that. So it'll be interesting to see how it works with regard to the pricing of those models, because this is a brand new expense now for the company suno and for the company of Udo, could be a brand new expense. This royalty is something that they weren't having to pay before. Now they will. So the end user expense is going to go up, and it'll also be interesting to see how the specific licensing works, or songs that someone creates with those platforms, like up till now, basically anything that was created with a free account on I'll use suno as the example, because it's the one that I played around with for that film that I made, Anything created in a free account with suno immediately goes into the YouTube content ID tracking system, and you will get copyright struck if you use that online. Anything you paid did under a paid subscription does not get treated that way. Now, how will that change as part of the terms of this deal that remains to be seen? You know, so, what is the viability of someone using these tools and and if someone were to use suno to create something, but then took it and did something transformative to it outside of suno. Then how would they ever track that? You know, are they? Are they going to track it the same way that that music companies try to track samples use of samples in other music? You know, there's, there's an audio detection system that that's built for that that's not perfect, but it it can work. So, yeah, it's, it's, there's a lot of lot of questions, but it does put at least those two AI music companies in a, in a, at least a legally stable place for now, you know where they can actually operate now. And suno is just flush with cash. They've got so much money, they've had a lot of investors, and they've made a lot of money. Tracy because, you know, people are paying to use the service like they're actually really profitable. And a lot of the smaller ones, you know, are just are not the computing power ends up being more expensive than than the revenue they bring in, you know. But suno is not that way. Tracy Harwood 50:17 So you're suggesting that the responsible music generator to use is going to be either suno or udio. Phil Rice 50:25 Yeah, after that, after that changeover happens, I think if there is one that's going to be, that's going to be it, yeah, maybe the record companies will surprise us, and they'll put out some kind of a template that other AI music companies can, hey, if you want to do this, here's how you sign up to license our material, and here's the terms of that and all that, I just, I just don't, I don't see any reason for him to do it, yeah, why kick the door open on that? We may essentially, it's not official, but they essentially own now part of those companies. Yeah, you know, those are vehicles for them, just like the radio stations, technically, are right. They can remember Damien Valentine 51:03 how tight the music companies were with people using music on YouTube and just releasing music online. There's no way they're going to put out a template. They might that would be too smart. Yeah, they may say that they'll agree to a licensing agreement the same way they have bucino and utio. Phil Rice 51:22 Well, that's what I mean. That's what I mean by a template. In terms of that the door will at least be open for inquiries to that, not that they would publish it online for scrutiny. No, they would never do that. But in terms of in their legal department, if you get any inquiries from AI companies, here's the terms of the agreement that we'll agree to, they just probably won't negotiate at all than they gave these Damien Valentine 51:45 guys right? Yeah, they'll do a basic one and a price that should be far too high for any of those companies to afford. So they won't put right, yeah, yeah. Tracy Harwood 51:54 That'd be really interesting to see how that unfolds. I think over the next few Phil Rice 51:59 months, it should be sometime in 2026 is when the changeover will happen. I'm sure you YouTube will be, will be the YouTube's already been flooded with videos of people talking about what might happen with it. So once it does, yeah, there'll be, there'll be people who talking about that. Suno is very interesting. They've actually got, they've for one of their paid tiers. They have a thing that looks like, from the screenshots I've seen them. It looks like multi tracking, where, basically you can create, instead of just having it generate the whole mixed piece of music. You could say, okay, guitars, you know, play guitars in the style of such and such, and then piano, add this, and then drums, heavy drums, and it will generate stems, basically, of music. And they call it sunio suno studio, I think so. It's pretty innovative, you know? I mean, if you're rabidly anti AI, then that's not going to move the needle for you, but it's an interesting development. And, you know, these, these are tools are going to be around, so I'm encouraged by any, any chance of of legitimacy and usage of them, you know. So, yeah, all right. Final note on AI, this is something that it's a little bit dated, but it's still enjoyable to watch, and since we're talking about AI, but neural viz, who we talked about on this show, he has one of his more recent creations is an actual character driven sci fi. It looks like it could be in like the world of like Star Wars or something like that, like this, this kind of a action adventure type of thing, with his trademark quick and witty dialog, very entertaining. And he's kind of, he's really pushing the pushing into new visual territory with his methods. And I'm sure that his methods have probably transformed completely since we since we studied them. So, yeah, it's interesting. It's called The Adventures of Remo green. Worth taking a look. It's a lot of fun. Brilliant. Okay, that's our news. Excellent. Tracy Harwood 54:18 And what a great set of things to discuss as well. So we'd love to hear what you've got to say on that, on all of the things that we've just been talking about. And you can get in touch with us at talk at completely machinima.com or you can just drop a comment below this video or on any of our social channels, we'd love to hear from you. So that's it for this week. That's that's me out. That's Tracy Harwood, that's Phil rice, and that's Damien Valentine, and we'll look forward to seeing you again next