NVIDIA’s GeForce Beyond GTC event featured some expected announcements (the new GeForce RTX 4000 Series graphics cards and DLSS 3) and a completely unexpected reveal: RTX Remix.
Out of the proverbial left field, NVIDIA unveiled a free modding platform capable of remastering DirectX8 and DirectX9 games to include AI-upscaled textures, ray traced effects (even highly advanced ones), and support for technologies like DLSS 3 and Reflex. To demonstrate the potential of RTX Remix, they showed the RPG classic The Elder Scrolls III Morrowind looking almost like a modern game.
The game choice was likely no coincidence, as Bethesda’s titles are the most popular on modding communities like Nexus Mods. Following this mindblowing reveal, we reached out to some of the most prominent modders to get their perspective on RTX Remix.
Today, we’re publishing an extensive conversation with Rebelzize, the lead developer behind Skyblivion, the unofficial remake of The Elder Scrolls IV: Oblivion in Skyrim’s engine. We talked at length about his take on RTX Remix and also got an update on the total conversion mod project, which is now closing in on the final stage of development (volunteers can still apply here). Enjoy!
What did you first think when seeing the RTX Remix announcement from NVIDIA? Did you already speak to your team from Skyblivion about it?
Funnily enough, we were actually in Discord hanging out, we made it like a group viewing thing. We were watching the event as it was happening live with a good few people from the team.
We were a little bit shocked, to be honest. I don’t think anyone was expecting NVIDIA to come up with modding tools. Especially the demo from Morrowind, it was really, really cool and exciting and just kind of mindblowing to see what this tech would allow us to do. Although it’s still, at least to me, a little bit unclear how this will exactly work. I haven’t really had time to read up on what these tools will actually do for us, but at least from a marketing perspective, they clearly did a good job at blowing people’s minds.
Everyone from my team was really, really at least excited or intrigued by the software and by the pretty impressive visual overhaul that RTX brought to Morrowind in lighting, the AI upscaling of textures, and of course the remeshing of assets. Or at least they’re replacing assets with similar counterparts, which was very cool to see in real time.
From my perspective, the ability to capture all of the assets directly from the original game seemed the most impressive RTX Remix feature. Do you agree?
Yeah, this is definitely the feature that intrigued me the most as well. It looks like you were basically exporting your scene from Morrowind, importing it into RTX Remix, and then it would allow you to export the new assets basically to replace them in your game. If you look at a project like Skyblivion, remaking the assets has been the biggest time sink. For many reasons, it’s probably one of the biggest aspects of the project. There’s tens of thousands of assets that need to be remade. Some are small, like a little pot or a table, and some are really big, like a castle or a house. Seeing RTX Remix, at least for a lot of the smaller objects, pulling that out of whatever resource it’s using behind the scenes was very impressive. It makes you think about how it could impact mods like Skyblivion.
At the same time, to a certain extent, I do think that AI will have certain limits when it comes to artistic styles. Especially when you look at things like architecture, it’s important that everything feels like it’s part of one cohesive whole. When you let AI run loose…. with the Morrowind example, I did personally see assets that I thought didn’t quite fit into the scene. Some things were pretty much a perfect replacement while other things just looked a bit out of place. With a small pot or a candle, it’s hard to go wrong, but when you work on more intricate assets, it might still be required for someone to do the work themselves.
And I think on top of that, something that’s very powerful within video game development is to try and optimize your game as well as you possibly can. For a project like Skyblivion, this means that we have certain texture materials, like standard texture materials that we have in the game. For instance, we have ten different wood textures ranking from exterior to interior, lower class, middle class, upper class, and a few mismatches that can be used in between. We do the same with bricks, mortar, and cloth materials, and you try to reuse those as much as you can, especially for smaller assets that don’t need unique textures to save on texture space. I’m not sure with RTX Remix if the AI is able to take that into account because, as far as I know, it really needs a manual touch to put all the assets on the same trim sheets. But again, it’s hard to know how the software and how the AI will develop.
Maybe, at some point, they will actually be able to take that over. What was potentially done in the RTX Remix demo is that they were just taking the original textures and AI upscaling them, so that might also just be how the results were achieved, and I think that that is how they did it. But that for me raises another question. For instance, Bethesda, I’ve been in touch with them about Skyblivion quite recently. I asked them if it would be possible for me to take some Oblivion textures and use AI upscaling to essentially remake them; I asked if that was allowed or not and they told me then that this was not something that they would allow us to do. It’s up to them. But that also raised the question of, if this was the response to my request to AI upscale textures from Oblivion to use for our own mods and that was not allowed, will it be possible for other modders to use RTX Remix to AI upscale Skyrim or Oblivion or the next Elder Scrolls game’s textures en masse and upload them? I’m not sure legally how that would work, but that raises an interesting question at the very least.
Interestingly, I’ve been in one of the RTX Remix press presentations where someone actually did ask about potential copyright issues and what developers would think about this. Well, NVIDIA said they already talked to many studios, and as far as the reaction goes, it was really good. It didn’t sound like they had any issues with RTX Remix so far. That was interesting to hear, but of course, the AI upscaling textures tool seems to be pretty central to RTX Remix. If it was true that Bethesda won’t allow it, then that would be very disappointing as their games are a big part of the mod community. Maybe NVIDIA can get them to relent on this topic.
Yeah, or maybe it’s also just a matter of, in this case, a bigger brand, a bigger company approaching them and asking the same question; maybe that also makes a difference. We don’t know what the politics are behind the scenes. Whatever the case is, I accept that, it’s not a problem. I love Bethesda all the same. It’s not a deal breaker for me in any case.
Another key goal of RTX Remix, according to NVIDIA, is to simplify things. Once captured, the assets will be converted to the USD format. That allows them to be tweaked in many tools, including NVIDIA’s Omniverse. They’ve said it will be possible for one modder to be working in Blender while another is working in Adobe Substance, and their work can then be synchronized pretty much in real time. That sounded like it would be really good for huge projects like yours.
Yeah. I think a big part of modding any game is that oftentimes we don’t have a lot of tools at our disposal, or at least we don’t have the tools that the studios had internally to make the games. Oftentimes these tools are then made by community members, which can sometimes take years and often also don’t work as well as they should. So any help you can get to make it easier to export models and remake them in the case of Skyblivion, I think that that would be a huge help. On top of that, I work with modding tools, but the definition of what modding is really depends on the actual mod. I don’t remake assets, I’m not a 3D artist. When I mod, I make maps, I make interesting locations, I make quests and other content within the game. While other people take it upon themselves to remake all the assets of Skyrim, for instance, because they were very low resolution or the meshes themselves were very, very undetailed, which is a completely different skill set.
For instance, using these RTX remix tools to help me replace some meshes or maybe even make new ones for my own mods. I think would allow a lot more modders to potentially enter aspects of modeling that have previously been very difficult for them to get into. I mean, even now, you can get free assets on something like Sketchfab to use in your mods because there are some resources out there. But again, RTX Remix might simplify that process by a whole lot and give you a lot more freedom to experiment. At least, that’s what it looks like. But time, I guess, will tell.
We’ll have to wait for its release to know all the details. Coming to the next question, do you fear that projects like Skyblivion, Skywind, or Beyond Skyrim might be rendered obsolete by NVIDIA’s RTX Remix?
Luckily I don’t think so. There’s one thing that RTX Remix can do that we cannot, which is the fantastic lighting. That’s been its strength so far. We’ve seen that with Minecraft as well, for instance, and that’s really cool. But other than that, I think projects like Skywind, Skyblivion, and Beyond Skyrim are more than just a coat of paint, which I think RTX Remix is to some extent. It allows you to really spruce up the older DX8 and DX9 games. It allows you to AI upscale your textures, improve the lighting, and replace or move models within the game world. And that’s great. But I think the strength of projects like Skyblivion and Skywind is that they go a step further than that; they actually take the old games and put them in a new engine. With that, you get improved mechanics and a much more enjoyable player experience. Even with the best mods, Morrowind still plays like Morrowind. It doesn’t change the game or the mechanics. It’s easier to try and remake mechanics in a newer game. Skywind has an armor system, for instance, which allows you to separate your armor sets into all the different pieces that you’re familiar with from Morrowind. Skyblivion has the armor degradation system from Oblivion remade as well. We also have spellcrafting in both Skyblivion and Skywind, so all the mechanics that people love from the original games are there.
It’s just the experience of playing it is a whole lot more comfortable and more modern than Morrowind can ever be, even after an RTX Remix remaster. Because of that, I don’t think that these mods or modern projects will ever be obsolete. If anything, it might help more people get into the original game and then, as a result, also make some of those new people hyped for the project that we are working on. I don’t think that we’re threatening one or the other. I think we can coexist and maybe in some ways help each other a little bit.
That’s true. RTX Remix is mostly a remastering tool, whereas in your case, you are remaking Oblivion, but in a newer engine that can do a lot more things than the original one. So that’s obviously going to be a better experience in many ways. As you said, unfortunately, RTX Remix is currently limited to DX8 and DX9 games that use fixed graphics pipelines. During the press Q&A I mentioned earlier, NVIDIA said that they are researching the possibility of expanding compatibility to more recent APIs. Let’s say that they find a way to support even newer games like Skyrim. Would you want to use it to speed up the process of making assets, improving lighting, etc.?
Hell yeah, if it’s available, I would want to use it 100%. I think, though, that the problem is, for one thing, the lighting that comes with RTX is only compatible with RTX GPUs. The problem there is that you’re locking features away for certain users, which I think is fine as long as it’s optional and you don’t force people to own an RTX card, which would be a bit sad. Other than that, in terms of the assets unfortunately or maybe fortunately with the current state of the project the only assets we really have left to do are clothing, architecture, and the occasional weapon. From what I’ve seen RTX Remix is not able to create these assets. We try and make, as I mentioned earlier, cohesive weapons and armor and clothing. They require good concept art and certain rules to ensure that the clothing feels like it’s all part of the same game world. We want the people of Cyrodiil to have a very distinct look and not look anything like Skyrim, for instance.
Cyrodiil needs to feel warm and colorful and multicultural, whereas Skyrim is more for the Nords by the Nords; it’s more gray and cold, etc. So I think it’s hard to portray the art style in the more intricate and difficult assets. If the RTX Remix tools could make architecture that I would 100% be excited for. While it’s pretty easy to make architecture for a product like Skyblivion, since the architecture in Oblivion itself was already really good, giving the perfect cheat sheet, it still needs to be done. And because they’re bigger parts and not as interesting as a weapon, it’s very hard to find artists to do them. If for those kinds of assets you could use a tool to help remake them, that would be fantastic. I would 100% love to use that. But again, until we actually use the tools, we won’t know their limits and how powerful they are. But yeah, should it be possible, especially for the tedious things, I would love to try and use some AI to assist.
That makes sense. With regards to the RTX tidbits, the good news is that while you actually need to own an RTX graphics card to make mods with RTX Remix, to actually play them you just need a graphics card that supports Vulkan ray tracing, so one of the latest from AMD or Intel work as well. We’ve talked at length about RTX Remix, but what can you tell me about Skyblivion today? It’s pretty far along now, isn’t it?
Honestly, it’s come a really long way, especially in the last year, year and a half, I’d say we’ve gotten a lot of milestones checked off. Recently we finished the last creatures for the project, we have been able to do at least half of the architecture that needs to be done, which doesn’t sound so exciting, but as I mentioned earlier, architectures are really, really, really big 3D workloads that not a whole lot of people are interested in, so that’s been really good and pretty much all the clutter has been finished, for instance, which was also a really big 3D task. In terms of quests, I think we have around 2/3 of quests implemented at this point, which is also really huge.
When we go into the game these days to test things out, it looks and feels like a proper game. A lot of things are final and actually playable. The biggest problem that we have, and every project like us has, is that generally, people stick around for a limited time. It’s a volunteer project, so there’s very few people who have been here from the start and are still here. I’m a bit of a loner in that regard as a lot of the people that I used to work with when we just started have since moved on. That’s their right, I’m not upset about that. I’m super glad that they’ve been able to do what they have done. But as the project grows older, some senior people who have been helping out with it at some point move on. Perhaps they move to a new country, they get a very high-end job, they get married, they have kids, stuff like that, which makes it sometimes a bit difficult to keep the momentum up and to keep the project moving forwards at the pace that’s been moving forward. Because of that, we try and work on developer diaries, we try to give people an update. We’re currently working on a new video project that we hope will be a bit more mainstream; with a bit of luck, it might go a little viral, and we’ll bring in the last volunteers for this.
But I really think this is the last piece of Skyblivion’s development cycle. The amount of work that actually is left at this point is actually manageable, whereas two years ago, it still felt like we were reaching for the stars and that the whole project as a releasable product was still kind of a bit dicey and uncertain. We’re in a really good spot now where the last few quests just need to get implemented into the game by the talented team that’s currently in charge of that, the last assets need to be made, and then the last stage of the project will really be to stitch everything together, make sure it all gets in the game and works as it should. I think that that last stage will be interesting but also very time-consuming, but very exciting at the same time.
Game developers usually say the last 10% of game development is as intensive as the previous 90% because they have to, as you said, stitch everything together and polish it up. Still, I’ve been following and covering Skyblivion for quite a while, so I’m glad to see it progressing well. We can say it will be released before The Elder Scrolls VI, right? Because that’s quite far away.
I should hope so. It’s difficult because, on one hand, I’m a really big Bethesda and Elder Scrolls fan, and on the other hand, this project needs to be released before The Elder Scrolls VI for it to actually stay relevant. If Bethesda was to announce tomorrow out of nowhere that TES VI will be coming out this holiday season, I would be very conflicted. I would be ecstatic and stressed and sad because there’s no way we can wrap up Skyblivion in the next two or three months. That’s not gonna be possible, but yeah, hopefully we do come out on top before TES VI; I think we can do that. I never want to make promises, but I think we can do it.
By my reckoning, The Elder Scrolls VI won’t be released before 2026 or 2027, so you have time.
That’s definitely more than enough time. I really hope we don’t have to wait that long, though. That would suck.
On that note, I remember covering a statement of yours pleading to Bethesda to improve their engine before The Elder Scrolls VI.
That’s funny, the statement blew up; I got contacted about it a lot. I didn’t expect it to get as much traction as it did. What we’ve seen with the new Creation Engine looks very exciting. I saw a documentary from NoClip where someone was working with the engine. One of the things I noticed is that it uses very similar markers I’m familiar with from Skyrim. It’s still Creation Engine, but the big difference is that it looks very good. I think from a tech perspective there are also a lot more tools at their disposal. They have mentioned that the planets are procedurally generated, which I think is insane. From the Starfield gameplay demo, I noticed that the terrain is part of the land mesh. In Skyrim, if we wanted to create a mountain, you take a land mesh and drag it up and place rocks against it. Here it was one seamless landmesh from the looks of it, which means that the tools have taken a really big leap. Maybe it’ll be easier for someone like myself to start working on a planet and terraform it and make it into the biome we have in mind without spending hundreds or even thousands of hours. The lighting looks really good as well.
Indeed, Todd Howard from Bethesda said it’s the biggest technological leap from game to game for the Creation Engine. Also, there was a bit of a controversy when they announced the total number of planets. Some felt that they would be useless since they were procedurally generated, but as Todd hinted later, Bethesda may have done it to make Starfield a modder’s paradise.
Possibly. We’ve discussed it internally with the Skyblivion team, and there’s already a couple who are very excited to take one of those planets and turn it into something else entirely. It seems like they know what they’re doing. They’ve always given us modders the credit we deserve, we’re a really big part of Bethesda games. They’ve seemingly thought about us when it comes to giving us the space and playgrounds to do our own things, and that’s cool. Hopefully, as with Fallout and Elder Scrolls, the Starfield modding community will take off and have its own place.
Would you want to participate in Starfield modding, or do you plan to focus on Skyblivion?
I kind of have to focus on Skyblivion. I won’t lie though, I will certainly play Starfield because it’s a game I’ve been looking forward to for a very long time. I will definitely also give the modding tools a look to see what’s changed. But when it comes to actually making mods for Starfield, I think I’ll let that one slide and leave it to my fellow modders who are currently not tied up with larger projects.
Thank you so much for your time.
