We've seen snippets of Halo Infinite over the last couple of years, but the recent reveal of Senua's Saga: Hellblade 2 was explicitly trailed as our first taste of what the next generation in console hardware will deliver from a visual perspective. However, confusion surrounds the nature of the content, which has been described as running in real-time and also as in-engine - two terms that are not exactly compatible with one another. Assuming this is much more than the standard CG trailer or the discredited notion of the 'target render', what does the reveal tell us about Microsoft's vision for the next generation of console gaming? It looks great, but can the advances actually be quantified and qualified in any way?
The use of the phrase 'in-engine' is a catch-all term that can mean just about anything, but based on established precedents, it indicates that we're looking at something closer to the real game than a CG teaser and in turn, it also suggests that we're getting a first look at what Unreal Engine 4 may be capable of delivering on the next wave of consoles. And while we're not exactly convinced it is running in real-time as such, there are imperfections in the presentation that seemingly rule out a pure CG asset.
For example, the depth of field effect in the trailer has standard problems we find in current-gen implementations of the effect, such as halos around certain objects. Similarly, tiny strands of Senua's hair pop and fizzle around the depth of field - something that wouldn't happen with super-sampling or a pure CG render. That said, these minor blemishes are literally all we could find in what is otherwise a nigh-on flawless presentation. Stacked up against Hellblade 1 on PC running maxed on an RTX 2080 Ti at native 4K, it's a night and day improvement in terms of image quality.
Our gut feeling here is that the Hellblade asset is exactly what it says it is - in-engine and likely not real-time. After all, based on when Microsoft first received Xbox One X silicon (early in 2017, around 11 months before launch), the Series X processor was likely only finalised very recently and creating an asset of this quality and polish in such a compressed timescale seems unlikely.
But what if it were real-time? The raw specs of the trailer file itself are intriguing. If you look at the metadata of the video - it is mastered at just 24 frames per second and the actual resolution of the rendered frame between the cinematic black bars is 3840x1608: so a little more than 74 per cent of a 'real' 4K in spatial resolution and at 80 per cent of 30fps. All together, that is 60 per cent of the amount of pixels pushed per second compared to 4K30 - and something that's decidedly easier to render than you may expect.
Even if this ends up being a pre-rendered cinematic or something running on a high-end PC, GameSpot's Series X reveal coverage suggests that the trailer "represent[s] new levels of detail, lighting, and rendering techniques we have yet to see on consoles." I was curious to see to what extent those next-gen techniques actually present in the Hellblade 2 teaser. Well, for starters, the increase in geometric detail is genuinely game-changing. Object density, definition and level of detail are frankly astonishing. Hellblade 1 is a beautiful game, but geometric detail is clearly concentrated on Senua herself as opposed to the surrounding environment. Make no mistake, that still looks great even today but the disparity is evident - and it is no longer an issue in the Hellblade 2 trailer, where everything is universally rich in detail across the board, right down to the tiniest twig resting under the perfectly modelled skull on the ground.
The tone and mood of some of the environments is similar to Death Stranding in some respects, making for an interesting comparison, but putting Hellblade 2 side-by-side with Kojima Productions' tech masterpiece demonstrates a proper generational leap here - as you'll see in the video embedded further up on this page. Where we veer into 'too good to be true' territory is in level of detail in that there are no visible limits to draw distance and zero evidence of any LOD 'pop-in', something that's very much present and correct in any current-gen game - including the original Hellblade maxed out on PC.
Normally, you can see distant geometry in-game either deform due to real-time tessellation forming their detail or you can see patches of geometry snap or fade into higher detail as the camera advances. Sure, next generation consoles will see new techniques for culling geometry, or in how they manage level of detail - but the perfection of geometric detail at all distances is an aspect of the trailer that makes me think it is pre-rendered and won't look quite like this in the shipping game. I can imagine the assets looking this good at LODs closest to the camera, but the perfection seen in the trailer would be a staggering achievement.
In terms of other techniques that may be representative of next-gen's capabilities, there is some evidence to suggest that ray tracing may be in effect here, along with a level of volumetric simulation we've not yet seen in a current-gen game. Outside of one-off PC implementations like Batman Arkham Knight, the volcano effect seen in the trailer presents volumetrics that move realistically, cast shadows onto themselves and indeed onto the terrain below.
This could be a taste of things to come, or it could simply be a singular, far-off effect achieved through cunning fakery. Rise of the Tomb Raider's avalanche effects do much the same thing by overlaying full-motion video on top of what is otherwise a fully real-time 3D rendered scene. Alternatively, it could be aping a volumetric look by layering 2.5D animated textures, which is how Epic's Infiltrator demo from 2013 makes its explosion effect look so fluffy. All told, what we're seeing in Hellblade 2 here is potentially something new, but we'd need to see more.
In that same vein, assigning the impressive lighting effects in the trailer as being as a result of triangle-based ray tracing is likewise not clear cut. For example, there are some reflections in the Hellblade 2 demo that could possibly be the result of ray tracing, but could also be traditional planar reflections based on the camera angle. The rest of the lighting found in the trailer gives no concrete indication of being anything other than what we have today running on a much more powerful GPU. For example, the softer ambient lighting from the sun could be derived from light probes or light maps - maybe it's using UE4's awesome ray traced sky lighting feature? It's present in the trailer for too short a time to allow us to offer any kind of firm conclusion.
Ultimately, there's no doubt that the detail level is immense here and there are tantalising hints in how 3D rendering could evolve on the next-gen GPUs. The overall asset quality in this trailer is obviously a generation beyond today's games and that includes Hellblade 1 before it. However, the level of its perfection in terms of distant detail - even given the constraints lifted by a 24fps frame-rate and a sub 4K resolution - do seem unrealistic, and I've yet to see research into techniques that could deliver the perfect LOD transitions exhibited in this trailer.
That said, on top of the detail boost, lighting and effects in general are indeed a healthy step beyond the best of the current generation - but we need to see much more from the game to figure out just how developers are seeking to exploit the big boost in GPU power that Xbox Series X represents. Hopefully we'll be very pleasantly surprised - obviously, once Hellblade 2 is available, people will be returning to this asset to see just how close Ninja Theory delivered on the promise showcased here - something that won't be lost on either Microsoft or Ninja Theory. Right now, the trailer looks wonderful, but I feel it says relatively little, but in that sense, it's a perfect hype-builder with as much substance as it really needs right now. Actual rock-solid takeaways are thin on the otherwise gorgeously detailed ground.
This may sound a touch pessimistic, but given what's happened in the past, perhaps the best thing is to remain a little sceptical and to manage our next-gen expectations accordingly while hoping that Microsoft and Ninja Theory can deliver. If nothing else, we now know that next-gen assets are going to be authored to an incredible level of detail - but I'd say that the level of dynamism and perfection of that detail across distances needs to be demonstrated, while we need to see more content to get a genuine feeling for what new rendering techniques are likely to be in place for the next generation of console hardware. And there's one more thing to factor in: while this asset is a mere 24fps, the base minimum for gameplay will be 30fps - and Microsoft itself is doubling down on 60fps gaming as the target for Series X. It's another challenge the developer faces (even 2x Xbox One X GPU power can only go so far!) but personally I wouldn't mind losing some fidelity if it means we get that 60fps experience.