[WIP] HRUNTING/YGGDRASIL Mark IX Mantis (Halo Mantis)
Hey Polycount!
It's been a while since I posted on here - the last project I worked on was a KOTOR Star Wars apartment that I abandoned after having burned out on 3D. However, I'm back at the poly-grind now, and currently studying my Masters to refine my skills and to get me back into the flow of things.
I actually meant to make this thread a few weeks ago, but I got caught up with the modelling and other assignments, but I'm looking to start this thread in the hopes of some feedback and critique as I go through the process of this current assignment.
What is the assignment?
The assignment, as far as modelling, texturing, and presenting is concerned, is to create a Mech, Mecha, or spaceship to modern modelling standards. For me, I chose to do the Mantis from Halo. There is also a part of the assignment to create alternative designs for our chosen project - for me, I'm creating three alternative arms for alternative deployments of the vehicle:
- The original Mantis design
- A Flamethrower version
- A Civilian Fire Suppression version
- A Heavy Lift version
Reference:
I have a PureRef folder filled with other reference images of the Mech, but I don't want to fill this thread with screenshots of screenshots.
On top of the gathered reference images, I've also sourced the original model from The Master Chief Collection to help me break down the model - how it was built, what techniques were used, and what considerations were taken for its animation rig.
What surprised me, more than anything, was how low-poly it was compared to modern vehicles in games - particularly ones to the same scale as the Mantis is in-game. Then again, this is the Halo 4 model, and that shipped on the Xbox 360, so maybe I should be too surprised.
Initial progress:
I blocked out the model fairly quickly, starting with the original, intended design. As it's only the arms I'm swapping out physically, the rest will be the same and so it made the most sense to me.
I played with the idea of several techniques, including the mid-poly workflow 343Industries used for their production cycle on Halo Infinite, however, in the end I chose to do a high/low workflow for information in the texturing phase and utilise Floating Mesh Decals to further add detail.
For the high/low workflow, I'm using the Dynamesh Polish workflow, creating a solid base in Maya, and then kicking it through Zbrush rather quickly - it's fast, iterative, and I'm left with some decent results.
Below I'll post a series of screenshots just showing the initial progress thus-far on the project:
In the last few screenshots, I've added some rough approximations of colour to the individual pieces of the Highpoly in Maya to help me visualise what is what, and to give me a general idea of what the final thing would look like.
When it comes to breaking up the model for texturing, I plan to have the legs, body, arms, and cockpit as separate Material IDs to keep things clean.
I plan to move onto the main body next, though if anyone spots anything that stands out to them, I'm more than happy to address the issue and make the model as good as I possibly can.
I would appreciate any any all critique and feedback!
Replies
Update!
I've started on the base mesh for the hip itself, though I haven't prepared it for Zbrush quite yet. A lot of the finer details, such as minor bolts and various insets all over the model will be added via Floating Mesh Decals once the bakes are finalised. I also plan to mirror the joint to the leg to save on some UV space as it's only semi-visible to the player at any given time.
Update!
I've continued chipping away at the torso to finalise my base mesh, and generally speaking, I'm fairly satisfied with where it's at currently - though there are a lot of pieces to it! I'm definitely looking forward to reducing that down significantly once I've done my high-poly.
Some of the shapes for the actual head were a little tricky, but at this point, I'm not worrying myself about topology - what's important is the silhouette and the proportions (which I've been testing, as you'll see) with the bits and pieces from the original mech.
Next, I plan to generate the high poly for the hip and torso, and then move onto the cockpit. I'll just reiterate, the topology that you see in the above image most certainly isn't the final topo - it'll all get reduced to game-resolution once I've made all the parts.
As always, any critique and feedback is welcome and most certainly appreciated!
Update!
With the torso high-poly sorted as of the last post, I've decided not to move on to the cockpit just yet, instead choosing to get the base arms for the mech done first - this way, I have a base to work with for the alternative designs, and at least one design that resembles the original silhouette in a recognisable way.
I started on the most complicated part of the arms, deciding to get that out of the way first. I used the original mech to help me proportion-out things like the gun barrels and give me an initial idea of the detailing, as seen below:
I wasn't a fan of how stubby the chaingun was - it felt too short, and far too small for it to be visually effective. To this end, I leaned more into modern chainguns, in particular those that could be carried. Through these design changes, I've made the barrels longer, more numerous, and added more detailing along the barrels.
As with all the other pieces I've made so far, I wasn't particularly concerned one way or another with geometry, just as long as the smoothing wasn't unduly affected. With using the Dynamesh/Polish method, much of the original topology was scrapped in any event. The only times topology really came into play was if a piece of geometry was too small to effectively be Dynameshed.
With the base done for the chaingun arm, I began to focus on the missile launcher on the other - this, as one can imagine from the simplistic design seen below and on the original much earlier in the thread, is much simpler with far fewer parts.
With that modelled, I then moved onto the arm - singular, as I plan to mirror the final result for all the arm variants where possible. The geometry, and subsequent details which I plan to bake down, were simple, with only the proportions finding themselves to be much of any challenge as much of the pieces are rotated at non-standard angles in the reference model.
With everything modelled for the arms at this stage, I took all the pieces through Zbrush to generate the high-poly models - the base meshes above, as with the other pieces, set aside to later be optimised and re-used as the low-poly models. I also took the time to colour ID the pieces in accordance with the rest of the model.
Here, an example of all of the base pieces for the mech can be seen - the arms are offset from the rest of the model simply to keep me from accidentally attaching them to any of the other pieces, which are all organised through parent-hierarchy's within the Outliner.
Going forward, I'll begin to work on the alternative arm designs, and their appropriate attachments before tackling the cockpit. The cockpit is a stretch goal, which I plan to complete should I have the time, though I'm currently aware of the work that still needs to be done before this model can even be considered remotely close to finished. At the very least, the first milestone has been reached.
Update!
So, I meant to do an in-progress post of all of this, but I ended up getting swept up in the process...
I've low-poly'd out the mech as best as I could, and performed several test bakes to see where I could get away with reducing it to the lowest possible amount that I could. I gave focus to curved surfaces and the silhouette in particular to avoid the common mistake of reducing something so low that it begins to chamfer.
To this effect, I have several screenshots of the individual pieces below, separated via their planned material ID's. I would also like to preface that several areas may appear to be rather dense due to thicker wireframe lines, but that is because I have Border Edges set to a thickness of five in my preferences to highlight open faces in my models.
Leg Wireframe:
//staffs-daforum-1.s3.eu-west-2.amazonaws.com/monthly_2022_04/image.png.1f51088d25ff2610394f2e94237da4ad.png
//staffs-daforum-1.s3.eu-west-2.amazonaws.com/monthly_2022_04/image.png.be798f692b7d929e7129a4377bb805d2.png
//staffs-daforum-1.s3.eu-west-2.amazonaws.com/monthly_2022_04/image.png.5f7648d59d9fe1d1207d3d969b475901.pngLeg Bake (4K, no AA Sub-Sampling):
Torso Wireframe:
//staffs-daforum-1.s3.eu-west-2.amazonaws.com/monthly_2022_04/image.png.e23c1b6a0946150c979ffea86ea7e098.png
//staffs-daforum-1.s3.eu-west-2.amazonaws.com/monthly_2022_04/image.png.94150ef327113b2a396a23824aaf1277.png
//staffs-daforum-1.s3.eu-west-2.amazonaws.com/monthly_2022_04/image.png.d5168790274d60594214b559d318217c.png
//staffs-daforum-1.s3.eu-west-2.amazonaws.com/monthly_2022_04/image.png.ba57282dc38471592644544ccbc30767.pngTorso Bake (4K, no AA Sub-Sampling):
Arms and Weapons Wireframe:
//staffs-daforum-1.s3.eu-west-2.amazonaws.com/monthly_2022_04/image.png.07e70e18dad97224370ffe34feb40f55.png
//staffs-daforum-1.s3.eu-west-2.amazonaws.com/monthly_2022_04/image.png.9a13595c169c134ed6ab9773b91b8aba.png
//staffs-daforum-1.s3.eu-west-2.amazonaws.com/monthly_2022_04/image.png.abbe95df8314fabe47c54e9f215e52f9.png
//staffs-daforum-1.s3.eu-west-2.amazonaws.com/monthly_2022_04/image.png.e00af548ac361fdd3280ef6847785538.pngArms and Weapons Bake (4K, no AA Sub-Sampling):
Next, I plan to work on the flamethrower assets to create the variant - I will block these out properly the next time I come to work on this project. I plan to reuse the arm meshes I've made so far, while everything else such as the weapon, the fuel tanks, cables, and a potential shield will be unique and have its own Material ID - the arm pieces will be baked onto this new Material ID to avoid adding another potential ID to the model when imported into an Engine within a production pipeline.
Update!
So I meant to post this earlier, but unfortunately baking the meshes took a little longer than I'd initially planned due to an unforeseen glitch with the arm high-poly...
When I uploaded the last post, I had everything but the flamethrower attachments concepted and made - I've since then done a loose, simple concept, played with placement for said attachments, and actually made the meshes, but I'll go through it all as it happened.
The whole purpose of the flamethrower attachments, as with the other attachments I theorised, was to create a variation on the Mantis design in-keeping with the design and deployment philosophy of the universe in which I took it from: i.e the United Nations Space Command (UNSC). Said design philosophy is utilitarian, adaptable equipment that is flexible in multiple arenas of combat. A solid example of this is the sheer number of variations of the iconic M12 Warthog, or the M12B (deployed from late 2552).
Taking the M12/B as an example, there are the following variations:
M12/B Warthog:
M12/B M343A2 Chaingun Warthog:
M12/B M555 Gauss Warthog:
M12/B M80B Rocket Warthog:
M12/B FAV Flame Warthog:
Despite all of the variations for deployment and actual use on a battlefield, the clear design variation is a difference in the kind of turret it has mounted on the back, or even if it has one at all. This allows the UNSC to quickly swap and change quickly only the fly, allowing for maximum flexibility in their arenas of combat. This isn't simply unique to this one vehicle either; there are a number of similar design philosophies applied to not only vehicles, but to armour (MJOLNIR and ODST armour, for example), weapons, ships, and everything in between. With this in mind, I wanted to keep that going with the redesign of the Mantis, even if I knew I was making my life harder than it needed to be.
To do this, I started playing with the various components of the original mech, trying to decide what worked and what didn't. From the get-go I knew that the cockpit would have to stay as it was - if I ripped that out, then there was very little point in me still calling it a Mantis. I briefly played with the idea of putting wheels or treads on the mech instead of its iconic legs, but then much of its iconic animations would be lost, and again, it came down to keeping the original silhouette and what initially made the Mantis what it is.
That left the arms. They were attached to the mech originally in a way that made sense to be modular - they would simply be detached and replaced by a large machine, and then everything would be good to go. After this was decided, it came down to what should be on those arms - I ended up leaning more on the Flamethrower variation rather than the other two I concepted with simple block-outs early on in this project (i.e Civilian Emergency Services and a Heavy Lifter).
Initially, I thought to have a weapon on one arm, and a protective shield to prevent any ammo containers (propane) from igniting in the event the Mantis has its energy shields overloaded by any weapons-fire (Plasma, Ballistic, Energy, etc). This actually showcased several flaws with this idea when I began to block it out:
As seen in the first screenshot, using the original mech from Halo 4 as an example, much of the iconic silhouette is lost due to the nature of what a riot shield (or a Tower Shield) is. Not to mention that the shield would have to be so large that it would be rather ungainly to actually use functionally. In addition, in the second image, the tower shield would actually have to be much further out from the mech for it to actually walk - the mech takes quite large, loping strides, and as a result would require the shield to be extended quite far out from the body.
This would actually create some issues of its own besides creating quite the ungainly large arm. All the weight distribution of the mech is toward its heels, and its wide legs give it quite the solid centre of gravity. By essentially adding a heavy slab of metal in front of it (most armour in the UNSC on weapons and ships is heavy Titanium), you completely change that making many of the design elements of the base model null and void - something that a modular attachment can't really afford to do.
In addition, I also experimented with the placement of fuel cannisters mounted against the cockpit, though there are actually very few places where they could go that wouldn't interfere with the cockpit hatch (almost the entirety of the top and back of the cockpit) or block any exhaust vents.
Alternatively, I thought of putting weapons on both arms - the firepower that the mech could deliver to a combat zone would be doubled from the outset, and it could rely on its energy shield which is, admittedly, rather resilient. While this didn't change the silhouette initially, it did actually work more in favour with the design philosophies I've been following - they would use the base arms I created for the original design, and minimal changes would be required for the rest of the body for animation, if any at all.
Now came the tricky part: How do I get fuel from those tanks, into the weapons themselves? Originally, I was going to use tubes to deliver the fuel, but honestly, they really got in the way of the functionality of the arms and the rest of the cockpit. In the end, I decided against the additional tanks and actually incorporated them into the weapons themselves.
When blocking out and playing with the design of the weapon, I very much leaned into the design of the turret for the Flamehog, and also some of the concepts for the base turrets from Halo Wars 1 (I apologise if the quality isn't all that great, this is a screenshot from my PureRef document).
//staffs-daforum-1.s3.eu-west-2.amazonaws.com/monthly_2022_04/image.png.2e4278ff67dd8b54ac4221ce1356b05b.pngIn particular with this design, I took special interest in the barrels themselves rather than the housing. I did at one point play with the shape language of the turret as a whole, but it was too large and ungainly next to the Mantis - another possible addition to the design was the Flame Mortar, but the point of that quickly became moot due to the fact that Mortars, by design, are intended to be stationary rather than moveable. I did briefly play with the idea of something like Napalm missiles, but moved away from that as it would have looked a little too close to the original silhouette.
The actual barrels inserted into the base of the guns themselves came from the propane tanks attached to the Warthog in the concept art for the Flamehog.
//staffs-daforum-1.s3.eu-west-2.amazonaws.com/monthly_2022_04/image.png.530aa26b04fbc560dbe05d6292bf5655.png
//staffs-daforum-1.s3.eu-west-2.amazonaws.com/monthly_2022_04/image.png.f3a626e21f86c1f569020b5dad8f63ba.pngWith these, I began to worry about their security within their mountings, and ended up putting some solutions to those in the form of a pair of steel pipes that would lock in with each successful reload (which would have to be done manually - there was no getting around that).
In the end, the final design looked rather mean, and dare I even say successful? I briefly blocked the two weapons on either side of the portions of the mech that were complete (just awaiting texturing) to get a general idea of how it would all fit together, and I have to say, I was rather happy with the outcome.
While the changes aren't quite as drastic as I had initially hoped for, I find that it's a happy compromise between the established design principles that 343industries have already established, and creating something new. Though I'll fully admit, I completely understand why 343industries have yet to actually introduce any significant variations on the Mantis design thus far; though it has been an interesting learning experience as hard surface vehicle and weapon art isn't my typical area of expertise, seeing as I'm traditionally an environment artist. It's been fun to expand my skills with hard surface and the changes to the design that I've made thus-far.
With this section done, I've taken the time to properly bake everything out today at 4K resolution, and 4x4 Sub-Sampling; the results being some slightly cleaner bakes than before, but now I'm ready to begin texturing. I plan to create several Smart Materials that I'll be able to apply to the various parts of the Mech that have been Colour ID'd out already (including the Flamethrowers) for maximum consistency across the Material IDs. Though the plan is to create a much more 'personalised' texture variation for the Flamethrower Mantis to reflect its imaginary pilot, in-keeping with actions often taken in modern militaries, and also seen throughout the media of Halo. Once these are done, I'll create my Decal sheet for minor details and then proceed to present it in Unreal Engine V.
I'm also unsure as to why half of the images in the last post have failed? I'm not sure if that's something on my end or not.
As always, critique and feedback is appreciated and welcomed!
Update!
I've started building the base for the Mantis Smart Materials, to make sure that everything is consistent across all of the Material IDs. I began by starting with the Torso, as it had the largest surface area of them all, so I figured that if it works on that, then it would work for all the other parts as well.
When it came to reference, I paid a lot of attention to the Mantis in Halo 5, and also the Warthog in Halo Infinite. Personally, I wanted to find a nice mid-point between the last time we visually see a Mantis in the Halo games, and the new direction of the art style that Infinite has gone in.
From the get-go, it's clear that there isn't a whole lot of colour variation going on in the Halo 5 Mantis, no doubt mostly relying on the Roughness, various skins, and in-game variables to sell it visually. With the Warthog, however, there are a lot more subtle variations in the paint - darker and lighter discolorations, with additional dirt and subtle scratches all along the body. This is what I wanted to try and emulate.
To start with, I began grabbing a bunch of reference images to add to my PureRef file, and I'll include my most-used ones down below. These helped me come up with a lot of the surface definition and really helped with those initial stages of building the bases for my materials.
As a result, the first pass of the Torso resembled this:
After chipping away at the textures a little more, I think I'm at a stage where I can comfortably call it a first pass:
The colour variation helps to break up the entire model just that little bit more, and definitely helps to reinforce the fact that this is a military machine. While the roughness is quite varied across the surface (the light isn't quite catching it in these renders), I'll definitely continue to tweak it.
Furthermore, I also plan on adding localised areas of interest, such as oil stains and the like once I'm happy with the other materials. This was also helpfully pointed out by Tom Woodward (from the 021 Discord) when he offered a bit of feedback:
I'll continue chipping away at the texturing and continue to post regular updates as I go!
Update!
Since the last post, I've put a lot of time into establishing a base-pass for the mech overall. My thinking has been that if I can get the groundwork established for the texturing, then I can adjust and tweak as needed and continuously improve as I progress to the end of the project.
This began with setting up the Smart Materials for the arms and legs of the Mantis; each 'section' of the model uses the same materials, so it was just a case of applying them and then reducing the amount of edgewear and other such material build-up to make it all look appropriate. Don't get me wrong, I still need to pull a lot of it back and adjust the materials more, but for now, I at least have an established base to work from.
With these all textured with my established base, I decided to bring it all into the engine to see how it all looked in comparison to Painter - after all, using the ACES LUT or not, it's always good to double check the end result with what you're working on. Please forgive the default lighting of Unreal Engine V.
It was around this time that feedback began to trickle in about material build-up from other sources than previously mentioned. A previous co-worker suggested:
The advice was solid, as it is far too uniform at the moment, and is in-line with my plans going forward and how I plan to develop the material build-up. Before getting to that stage, however, I wanted to have some kind of visual reference for myself of where the Decals would be, and what kinds of Decals and details I would need to make considerations for. So, I set about making a really very simple Decal Sheet.
On the original model, the Mantis makes use of two separate Decal sheets; one for the stickers and detailing, and another solely to create the shape on the sensor node on the cockpit. Both of these sheets are full of details used throughout the rest of the game and weren't really applicable for this project. As a result, I took what was on the Mantis as reference and began building a disgustingly simple Decal sheet.
I knew I wanted to incorporate the Normal Detail for the few insets and such on the Mech that I didn't bake down, but I also wanted to be able to use the full extend of the Albedo, Roughness, Normal, and AO for details like the stickers, logo, and text. As a result, the Decals ended up looking something like this:
The logo from the UNSC was sourced online and is used within the games themselves, while the stickers were built by myself, based on some initial templates I bought in a pack - I used the original Decal sheet to create rough approximations of what they would say as the originals were too low-res and small to actually be readable. The weld detail I grabbed from Quixel and made some small adjustments to for readability when applied to the mesh in-engine.
Throughout the process, I iterated on the design and overall layout of the sheet as I went, but in the end, the model currently appears as:
An example of the weld-detail having been added to the model can be seen below. This was originally suggested by my friend in a previous post to avoid the appearance of the geometry crashing through one-another. This is mapped to a strip of polygons that matches the shape of the intersection.
Going forward now, I plan to continue refining my textures and improving upon them. Once I'm happy with the olive-base variant, I'll begin working on the materials for the flamethrower variant. I believe I have the large-strokes done, and now all I have left is small adjustments here and there. Either way, I'm pleased with the direction that it's going, and I'm hoping that I have the time to set the two of them up in a cool little diorama to present them both!
As always, any critique and feedback is welcomed and appreciated!
Update!
Following on from the last several posts, I went back to refining my textures and really pushing the materials further. My first hurdle to overcome was the texel density; with having split the model into four Material IDs, I found that I was lacking the resolution to really take full advantage of what Painter has to offer.
As a result, I split the model into a total of eleven Material IDs to achieve a nice balance between global uniformity and individual resolution; with this being inspired by Halo, it had to stand up visually from almost every angle, as it would be viewed in-game.
After setting the Material IDs and baking the model out again, I went about rebuilding my Smart Material and really pushing the albedo and roughness maps in particular. I added a little more colour variation, and layered my paint wear and tear using Anchor Points and varying height on the layers. For example, the underlying damage on all parts of the mech has a lower height value than the paint layer - this gets chipped into the paint and works to emphasize the scratched and dings that would come from combat and general use. The damage layer above that has no height, however, and this is much higher in frequency than the underlying and is used more for implying flecked paint chips.
The images below show some of the Work-In-Progress renders taken in IRay using the ACES LUT to highlight the previously mentioned painting techniques and to show some of the material definition gained as a result of making the changes to the model as a whole.
Clearly, the textures are much better than the first pass - texture resolution has been improved across the entire model, and I have been able to take full advantage of that through the use of the generators, which I then controlled via painting into black or white masks, depending on the context of the various generators used.
I continued on with the rest of the model, and it currently looks something like this:
One huge difference to the model is that many of the decals no longer exist as floating mesh geometry - instead, they are painted directly into the texture. While the position is locked to being placed in Painter, the decals are masked out with the appropriate mask textures exported for adjustment in-engine on the fly to iterate colours (hence them being white to override within the material shader).
I've also taken the time to create an alternate set of textures for the flamethrower variant, taking inspiration from the varying elements of fire. The base coat of the paint is based off of the UNSC Razorback from Halo Infinite (see below).
However, in-keeping with the theme of fire, I went for a subtle gradient mask (layered and blended with several more to break it up and control it more accurately) focusing on the top-down and front-to-back based on settled ash. I adjusted several of the decals to work within this colour scheme using deep reds, and the adjustable decals will be a warm, burned orange - these bold colours will contrast the pale blue-grey of the underlying paint and really pop visually. This colour scheme followed onto the flamethrower weapon itself, and the results can be seen below.
For the propane tanks, I took a lot of inspiration from the Flamehog (pictured below) - in appearance they look to be titanium barrels within a stronger protective shell, which I integrated into the model of the flamethrower. With the expected lifespans of the fuel within each tank over the course of a deployment, I would expect them to be changed out and swapped fairly often, so they are obviously a lot less worn and damaged.
My next focus is to rig and pose the mech as the textures can then be easily adjusted based on the lighting and setup inside Unreal Engine 5.
Update!
A friend of mine was kind enough to sit down with me for a few hours and show me how to approach Cat Rigging for the mechs. With the way I had prepared the model prior, it was a simple process to skin each element to its assigned bone for later posing. An example of the established Rig can be seen in the image below:
When constructing both the individual parts of the mech, as well as working with my friend to create a rig for posing, I looked at as much reference for the Mantis actually moving as I possibly could. With it being very mechanical in nature (in comparison to, say, an Anthem Javelin), it meant that I didn't have to worry about deformation - more, the concern and focus was pushed more to pivot point and what was linked to what.
Some good reference was found on Youtube of gameplay of the Mantis itself, which helped to break down how it moved, which joints it relied on the most, and what really drove its movements (please refer to the below videos for examples of said movements). Beyond those things, it largely came down to how much freedom I wanted to give myself when it came to posing - I elected, in the end, to rig everything but the canopy as I didn't plan on having the interior show; if for no other reason than I never met that stretch goal.
https://www.youtube.com/watch?v=RY_UtLHWDSw
https://www.youtube.com/watch?v=cl9ZJqWln6s
While the rig was made in 3DS Max (a program I haven't used in years), it transfers over to Maya seamlessly - though obviously the two programs have two very different methods of displaying the rig.
The beauty of the mech as I have designed it, is that the rig could be duplicated and then assigned to the alternate version, allowing for an incredibly quick turnaround in this part of the process. The end result meant that in-engine, using the Skeleton, I could create various poses and iterate on them in one place, depending on the environment as a whole. Early tests of this can be seen below:
Once I became familiar with this process for the extent of this project, I went about building up a small diorama using sourced assets from Megascans (mostly assets from the Burned Forest and Tundra collections) to populate the small area of the terrain (made in Gaea) that I made for presentation purposes. GIFs of this progress can be seen below, and afterwards, I will post the screenshots of the images that I have that show where it is currently.
The current renders are:
I based these screenshots heavily on the Halo 3: Believe trailers from Bungie with the table-top diorama, and more recently, taking inspiration from the Halo Wars 2 dioramas found in the main menu of the game. There are several aspects that I still need to work on and improve, but I think that much of the foundation is there to be built upon.
For example, to break up the flatness of the ground, I plan to very lightly add a slight undulation to it, and break it up with some more rocks. I also need to readjust the fires as they were placed with the trees in mind, which I have since removed.
I will continue to refine these shots before setting up my Sequencer render and experiment with other setups to get the most out of these models as I can - I'm optimistic about the direction that the project is traveling in, and I can't wait to see the final results.
Update!
I've created a secondary, and a third presentation scenario within Unreal Engine V with the focus being on creating several alternate types of renders for vehicles that one could come across on ArtStation.
The first render scenario was based on feedback from a friend of mine: Will Greenwood (Cloud Imperium Games) over a Discord call - he suggested a more concepted-approach, rather than the diorama-style that I had previously gone for. This involved adjusting the time-of-day lighting and repositioning several key elements within the scene itself.
Personally, I'm of the opinion that the first image is the strongest, but I've heard differing opinions to my own when it comes to the second and third images (other people like them, I'm not so convinced, etc).
There's some issues with the main shot - the DoF isn't quite right, so I'll continue to have a play around with that in my own time. A friend of mine suggested placing a fake focusing plane in the scene to help with the sensitivity of the DoF, so I'll see what I can do about that.
I continued to refine the original shots I established in the diorama. My original plan was to render out a sequencer video of them all frozen in place (including particle effects) however, since looking into doing that exact thing, I've learned that while Niagara effects can be paused, legacy Cascade effects can't be. Unfortunately, I don't have enough working knowledge of Niagara to recreate the packs in the new system, and with the deadline fast approaching, a decision had to be made.
In the end, I decided to forego the Sequencer render, and after some experimentation with the composition of the renders as a whole, I decided to remove much of the fire from the diorama - they were too bright and contrasting within the shots I had established and blew out much of the detailing on the models.
I did, however, take the time to add some emissive detailing to the mechs within Substance Painter, so there are two texture options for the nose and the weapons; these can be toggled on and off within the Material Instance using a Static Switch Parameter. For reference, I specifically looked at over-heated weaponry and took note that the colours were usually in the red-spectrum, though I added a slightly yellow at the very tips where it would be most hot.
My final rendering solution was a more traditional three-point-setup. Having looked through ArtStation to see how similar vehicles are presented, this became one of the most common. It allowed the details in the texture to be highlighted, and there was nothing else to detract from the models themselves.
For these renders, I opted for a warm light on the left, and a cooler light on the right - I had seen some examples of assets being lit in a similar fashion, while others opted for a more clinically white light to make their colours pop. In particular, I looked at examples of mechs such as:
Csaba Szilagyi (ArtStation: https://www.artstation.com/csabaszilagyi)
Mathew O (ArtStation: https://www.artstation.com/mathewo)
Ben Erdt (ArtStation: https://www.artstation.com/benerdt)
Aleksander Galevskyi (ArtStation: https://www.artstation.com/galevsky)
At this point, I'm calling it done on the project. I've learned so much from this over the last several weeks, and I can't wait to see where the next project takes me.
As always, any critique and feedback is welcomed, and the ArtStation link to this project can be found in the link below!
https://www.artstation.com/artwork/eJekwJ