Starfield Movie Trailer Reveals Cool New Technology

It’s fair to say that Starfield is one of the most anticipated games and it’s not hard to see why – for all its flaws, Bethesda has built its empire with massive open-world RPGs. There’s a reason games like Skyrim are still popular today: the carefully designed worlds and sense of freedom appeal to the imagination. On paper, Starfield looks like a logical conclusion, a game that extends beyond a single planet through the tributaries of space. I thought it would be fun to dive into Bethesda’s show and see what we can get out of the game – from the basics like image quality and performance to the overall approach to technology and design.

Let’s start with the resolution: The trailer is shown in native 4K resolution, but the footage varies in resolution. Interestingly, the gameplay sequences seem to lack any kind of complexity, so you get very sharp edges with aliases visible everywhere. Conversely, TAA cinematic footage is used in a similar way to Fallout 4, which should be more in line with what we’ll see in the final product.

In addition to subtle subtlety, we can get a sense of the development team’s design goals by looking at how Starfield deals with open areas on the planet, interior spaces, character displays, and ultimately outer space. For example, in an outdoor scene, we can see that the game has long-range shaders, which is critical for preserving distant details. This is one of the major issues we identified with Halo Infinite and it’s great to see Starfield have a solution.

Starfield also seems to have a system that displays the amount of local mist in the valley fissures, which sounds pretty cool. In general, the presentation of the atmosphere looks very powerful from what we can see in this demo. What I still don’t understand is the air system – it sounds promising, but given the low bitrate of the trailers we had to watch, it’s hard to tell if we’re looking for a good volumetric air system or a simple air dome. Either way, it delivers attractive results – we just need to see how dynamic it will be in the final.

Everything is then tied together by the terrain system – it is possible to build planet surfaces and structures using a combination of procedural generation and manually placed assets, a common approach today. The terrain that presents itself is similar to previous Bethesda games, but the popup is kept to a minimum and details appear from a distance. While it is attractive, the presentation features do not go beyond any limits – which is understandable given the large size of the game and the long development time.

Inside, things are different: the diffuse forms, once restrained and loved in the outdoors, are now clearly defined internally. This section evokes an atmosphere similar to Doom 3, with direct lights cutting through the darkness while highlights appear on the surfaces. Compared to Fallout 4, the subtle jump is significant, as that game features rudimentary interior lighting and a distinct lack of texture and object detail.

This raises an interesting omission: the lack of reflections. In the original trailer, we almost saw RT-esque reflections, but in any gameplay sequence, there’s no evidence of any reflections in screen space, let alone RT reflections. At best, we see basic cube maps. For a setup that flushes with metallic surfaces, this feels a little weird, and reflections in screen space will go a long way to improving the overall picture cohesion.

There are also many positive elements here. For example, weapons look great. I’ve never been a fan of the designs in Fallout 4 – the models and animations left me cold – but Starfield offers weapons that look elegant and powerful. Enemy animations are generally better too. As an RPG, you still seem to drain more of the life bar than you do direct damage, but the reactions are vastly improved. The only thing missing is the motion blur of each object on the weapons and enemies.

Character rendering has also greatly improved since Fallout 4, especially if you look beyond character creation screens and focus instead on the actual in-game look. Underground scattering, absent from all viewers, can make things better, showing exactly how light interacts with the skin’s surface. It’s on the ears in the photos we’ve seen, but it doesn’t apply to the rest of the leather that highlights regular cards. Also, the geometry of the tear duct is so luminous and traps light that it almost appears to glow. In addition to these small points, there is a significant increase in the quality of the animation. The conversations in Fallout 4 have harsh and even ugly animations, while Starfield looks more elegant by comparison.

Starfield’s last major placement is space, and while we only get a brief glimpse, effects like lasers and explosions hold promise — definitely a step above low-resolution smoke when landing on a planet. The big question I ask regarding space travel isn’t about the visuals, but more about the possibilities – I’d like to see ship management play a role in the travel. Imagine rising from the captain’s seat to explore a ship and manage both resources and systems. I think this could make the interplanetary journey more attractive and challenging. It is not clear if this is an option, or if the player “becomes” the ship in flight.

Some other noteworthy technical criticism is the game’s indirect lighting. This has become a major focus in recent years and is key to rendering realistic – simulating the phenomenon of photons bouncing off one surface and indirectly illuminating another area. The problem now is that areas that aren’t directly lit in the Starfield will turn out a uniform gray that doesn’t match the lighting results you’d expect. Global ray tracing lighting works well here, but it has a high performance cost. The baked solution can also work offline using probes, but with so many planets the GI data is likely to be quite large. This is a difficult problem to solve when building a game of this size.

Then there is the performance. Our trailers are encoded in a 30fps container, which limits the amount of analysis we can perform. However, there are still issues worth reporting, which is the fact that all gameplay footage shows significant performance issues and regularly drops below 30fps. This isn’t unusual for a game at this point in development, but Bethesda’s record of extremely variable playback performance on console gives me pause. It’s the most obvious flaw in the show and I hope performance improves at launch, but we’ll have to wait and see what happens.

The other aspect I’m curious about has to do with cities – in previous Bethesda releases, larger cities were usually divided by loading screens, while smaller cities were seamless. Can you land on a planet and go to a big city without loading screens? We hope to find out soon.

But while I feel bad, Starfield is still Bethesda’s most appealing game to date – most of the ugly parts that plagued Fallouts 4 and 76 have been stripped away and left with some beautiful environments to explore instead. Starfield also displays structures and dimensions unlike anything they’ve built in the past. The entire ‘1,000 Planets’ feature seemed silly at first, but you can imagine that the major planets were carefully built and designed, relying more on procedural generation to handle the rest. If the gameplay structure supports this well, it could be great. Even if I’m a pretty exhausted person in open world games, I’m pretty fascinated by Starfield.

All of this means Starfield will be a tough match to analyze when they kick off next year – but I’m looking forward to the challenge.