The Beginning

I have started an environment project as I know I desperately need to add environment work to my portfolio! For this project, I have chosen to create a temple inspired by Al Khazneh Treasury in the Jordan Desert. This amazingly beautiful and ornate temple is literally carved straight out from a sandstone desert cliff in Jordan and is amazingly intricate in details. I think the most interesting and awe-inspiring part of this place is the pure scale! Just looking at the reference images, you can see how behemoth this temple is compared to the people standing at it's base. It's amazing how any civilization were able to create this with minimal technology. 

Initial Blockout

  Starting this project, I first opened up an Unreal Engine 5 project. The reason I am choosing UE5 as my game engine for this is because I want to use Epic Games' most recent technology in the lighting system Lumen and the virtualized geometry system Nanite to achieve the greatest levels of detail possible, whilst also getting myself accustomed to the new engine that I'm sure game studios will be switching over to from UE4 once it leaves early access. These systems in place allow me to achieve realtime global illumination and essentially unlimited geometry with static meshes for my environment. 

I recently purchased a brand new Huion Inspiroy drawing tablet too as I've started to pickup ZBrush again so this project is a perfect subject matter for all my needs at this point in time. I will be creating temple assets whilst also using Quixel Megascans library to achieve the rest of the environment.

This is the current progress of the blockout for this environment. The main goal of this blockout was to get the general scale of the architectural assets setup and ingrained in my mind, ready for when I start modelling and importing assets. It's not perfect. I already see issues with the scale that I am going to fix but that's the entire job of a blockout. Comparing this to the original reference, I think I need to make my temple slighly less wide. The columns are too far apart at the moment and the inside needs some rearranging too.

De-rustifying My ZBrush Skills

For my first asset, I thought I'd start with one of the main columns. Luckily for me, this column is repeated throughout the entire temple so I will only have to sculpt one and use it modularly. For this and the walls, I want to make modular to save time and for efficiency purposes. In environment art, the importance of modularity is paramount. 

Beginning the column, I started off creating the basic primitives in Blender, getting the main shapes modelled and then exporting these to ZBrush where I would add all of my intricate details and layering. For the most part, it was creating insert mesh brushes and using curves to creating the vines. Then dynameshing and sculpting out the details of the rest, such as the large leaf ornaments pertruding from the surfaces around the column and the flower heads that looks like bells. For the vines, I just used snakehook brushes to create the thorns. I took some creative liberty compared to the original references to make it my own, but I wanted to keep it clearly inspired. After modelling everything, I added surface noise and damage to the column to give it further detail. Sandstone over time wears away. Usually this would be caused by errosion in more humid climates, but in a desert, I'd expect this would be due to sandstorms and abbrassive sedements in the wind that gently wear away at the surfaces over long periods of time.

05/08/2021

Redesigning Blockout

Blockout version 2. Going back to what I previously said about getting the scale of the temple right, I went and redid the entire temple blockout to a higher standard. I think I am now closer with the scale and I prefer this overall composition for the temple.

I will be continuing to make changes scalewise to the blockout until I start implementing the final version of the assets into the engine as it's KEY to get that right in this project.

19/08/2021

Creating Overall Blockout

I have created a new quick blockout for the cliffs. I just dragged some alphas over a speed sculpt of the general shape and composition I wanted for the scene, decimated and exported the mesh straight into unreal engine. At this stage in the project, I am researching a lot into the best workflow I am going to use to create super highly detailed cliffs. I was going to use purely nanite meshes with vertex colour painted in ZBrush as textures but I have decided against that as I cannot get the detail and precision I want with handpainting these gigantic cliffs. It's too time-consuming and inefficient, therefore I think I am going to approach this by making a 'low' high poly mesh with UVs, that I can take into a texturing software such as Substance Painter or Quixel Mixer. From there I will add rock materials that I will have lots of quick procedural control over and I believe I will achieve a mucher greater result. Especially if I export the mesh with tesselation using the materials I've applied. This way, I should be able to achieve the amount of details I want, efficiently.

Once I have the cliffs modelled, sculpted and textured. I will then begin to work on the landscape material and material blends to get greater variation and detail. I will continue to work on the temple after I have the initial scene in place. Start with the large objects first and work my way to the smaller details later. That's my plan!

23/08/2021

Implementing My Unique Nanite Meshes

I have finally found a workflow for creating my own nanite meshes. It involves sculpting the mesh using dynamesh, getting the final highpoly details, however many millions of polys I want. You then duplicate the mesh and using 'Remesher', remesh the high detailed sculpt to something more manageable for UVing. Around 40K was where this cliff piece was remeshed to. Then, using polygroups I used 'UV Master' to create UV islands from the poygroups. You can also use the ZPlugin 'PolygroupIt' to create polygroup islands quickly. After UVing, I then subdivided the mesh to around the number of millions of polys I originally had on my detailed sculpt. Maybe a little less so it could be imported into Substance Painter or Quixel Mixer. After this, all you have to do is reproject the details of the original high poly sculpt onto the UV'd remeshed subtool that has been subdivided and BOOM! All the details are back onto the UV'd mesh that can now have materials applied or be textured in another external software. I have currently opted to export this mesh straight into UE5 and plan on using the megascan materials there. I might decide to change my mind and use or texture it using Mixer or Substance later on for more control. 

At the moment, this mesh is just a smaller section of the vertical wall, but after figuring out this workflow, I think I am going to create the entire right angle section of cliff as one mesh, almost like a mega structure, then use the same worflow on that. This will just require a bit more attention to the UV islands orientations and seam placements, but otherwise, will be the same thing.

HERE is how I learned to create my own nanite meshes, UV and texture them.

24/08/2021

Creating The Composition

Made a bit more progress on the environment. I have imported and set-dressed some cliffs together. Also finished creating the main cliff piece around the temple. Starting to come together nicely. The main challenge at this stage will be creating the ground as I do not have access to tesselation or RVT landscapes (virtualised textures) in UE5 which means I cannot get the details I want solely from a material or any procedural blening between assets and the landscape. I will most likely use a mix of megascans ground assets and sculpted terrain such as small modular windswept sand pieces along with larger boulders to poulate the ground. I know I also want a puddles of water in the sand to give reflections in the valley too. and break up the roughness valleys. 

Another thing to consider once I have the majority of the composition down is the weather and lighting I want. This stage will need a lot of research and development until I'm happy. I will probably create a colour swatch and get some inspiration from desert photography. Maybe stormy deserts too since I want water in the scene. This includes post process effects and even particles like sand blowing through the valley.

25/08/2021

Starting the Ground

Time to tackle the ground. This is a very important process in this project as I have a lot of ground to cover....literally. All jokes aside, without displacement being available in UE5, how was I going to get the amount of fidelity and detail I wanted in my terrain? I can't quite use it in the usual methodin-egine, but, I found a workaround that yeilds significant results by using Blender's 'Displace' and 'Smooth Correction' modifiers onto a subdivided grid and applying the material displacement map onto it, I could then export that plane or mesh and apply the material in engine and I get a tileable, modular ground piece that casts shadows and has all the geometry as a nanite mesh. 

I will use this, in conjuxtion with quixel megascans assets to create the valley. I will be able to break up and hide any seams where I change materials, etc with rocks, boulders, etc. If the displacement map used to create the meshes geometry is similar enough in noise to another material, I might be able to get away with using material blending to break up the tiling effect. Luckily the material I am using at the moment tiles extremely seamlessly and you can barely see the repetition. Esp[ecially once I add assets over it.

HERE is the link to the tutorial I found to create this. 

I thought I'd challenge myself to model and render this hiflofiltro oil filter in a day. I am an avid motorcycle junkie and personally use hiflfiltro products on my motorcycles and I thought it would just be overall, good practice. I am planning on using this scene to learn some basic animation and fluid simulation to try and put together a cool little advert or something. 

IMPORTANT: This is NOT a paid or sponsored promotion for Hiflofiltro. Purely for my own educational use and purposes.

After some experimenting with mantaray's fluid simulations in Blender, I was attempting to achieve a more interesting composition for this project. My vision was to create oil flowing out or into the filter in an artistic fashion. I am now considering learning how to bake and animate this fluid simulation so that I can create my own little comercial for the product as practice for rendering cinematic shots. Could be interesting!

I have started another hard-surface project whilst I wait to get a new pen tablet to do some more environment work. This is an FN P90 I am currently modelling. It is my all time favourite submachine gun. I love the unorthodox approach to ergonomics, the general looks of it and the design for the magazine and feeding system.  I tried modelling this 3 years ago when I just got into 3dart and was learning Maya. That was a bit ambitious at the time to say the least! Now I have returned to model it properly. 

With my current skills and knowledge or 3d workflow after graduating, I am a lot more comfortable at tackling the challenges that come with this model. The trickiest part of this model is definitely the flow of the body and all the intricate curves that flow into larger and smallers curves which then flows into more curves! This model to put it in current-gen terms, is THICC and getting those smooth curves right is important to getting this right. 

I aim for maximum detail in this model which means creating an exponentially intricate  high poly to bake onto my lowpoly asset. I am currently using Blender and modifiers to create the highpoly which I will then duplicate and retopologize for the low poly asset. I want to model not only the base gun but some interesting attachments for the rail on the barrel  to add even more detail and contrast between the very smooth shell and the harsher, tighter details on the barrel and scope, etc. Current progress is going well!

Cycles Render in Blender with base material applied.

19/07/2021

Modelling the barrel was quite difficult as there are lots of curves that flow into more rigid and tight corners/lines. I took some creative freedom with the underside of the barrel and bits from different models of P90s whilst modelling. It's very interesting to see how the gun actually works just from looking and modelling from reference. The magazine was also a very complicated shape as there are lots of ridges and details that I will showcase later on after I have added the thickness to it. I still have the side rail to model and then attachments of my choosing which will be exciting! I was thinking along the lines of an ACOG scope or maybe an advanced/interesting red dot sight? But before that, I still have quite a bit of detail to do on the top of the body where the magazine sits above, and the magazine feed system, where the bullets rotate 90°C into the receiver. How cool is that!!!! One of the reasons I love this design so much. Such a unique feeding system that allows a lot of bullets to stack in just one magazine. 

04/08/2021

I have made a little more progress on the magazine for this gun. I've added thickness to it and I believe the feeding system for the magazine is now entirely modelled. The magazine ejection button is also modelled. Very soon it will be time to retopologize this high poly model to create the lowpoly mesh. The entire time I was modelling the high poly, I was using more divisions than I usually would if I was just going to make a high poly for the sole purpose of saving time in the creation of the low poly. This is because I already have most of the geometry I need for the low poly already. I just need to turn off modifiers and clean up edge loops I don't need to optimize the poly count of the low poly mesh.

 There are very few select areas where I wasn't able to add more divisions in the high poly where it would be too difficult to work with a dense mesh and achieve the shading I needed and those will be the areas I will need to maybe apply a few more divisions but other than that, shouldn't take too long. Before I get ahead of myself though, there are still the matter of attachments that need modelling. That would be the ACOG scope that will sit on the rail, and potentially any other attachment I feel like producing to the side rails. Maybe a flashligh or laser pointer.

11/08/2021

I have made a start on a red dot sight. Not too sure if I am going to stay with this choice. Compared to the rest of the gun, it's quite small and doesn't stand out too much. My other option I had in mind was a larger scope but I wasn't sure how it would look or fit with this gun. I might model both just to see how they compare or stick with the red dot sight for now whilst I start retopologizing for the low poly version. 

Instead of the more square shaped sight, I think a rpunded sight would be more fitting for the overall shape of the gun itself. Might feel more at home and less out of place looking than this current one.  P.S. I know, the sight is the wrong way around!!!

12/08/2021

Started modelleing a scope. I think this will be the final attachment I use for this project. I would like to incorporate a fabric/textile strap into the project too, as I think it will break up the materials abit from metal and plastic, showcasing that I can model and texture all sorts of media.

13/08/2021

Current progress on the scope. Just need to model the cantilever for the scope to attach to the rail of the P90 and then I believe I can start retopology for the low poly.

16/01/2022

I have returned! After a long break from 3d, I have returned to try and finish off this project. Since last time, I have baked my gun in Marmoset Toolbag, maknig use of the'paint skew' tool to correct the normal information around the scope details and screws, as well as texturing the entire model. During the texturing stage, I had a lot to consider and think about as the majority of the gun is made from a polymer which isn't very reflective, I needed to make sure to get believable and also interesting details in the roughness to make it pop and look good. I tried adding wearr, damage and fingerprints where I think it would make sense for it to go,which meant adding it around where you would hold the gun, the trigger area, the handle where you would cup, almost caress the gun into your shoulders when aiming down sights, and the eject button. 

Around the rest of the body, I added what I would call "battle damage" which could occur from debris flying around, dropping the gun, dragging it across a rough surface, etc. For the metal parts, I added edge damage, but I didn't want to go overboard which is very easy to do with procedural texturing, so I painted a lot of the damage out using a paint layer in the mask. Overall, I think it looks good. It's not perfect and I'm sure I could reiterate for days or weeks, but I think I have achieved something that looks relatively believable and aesthetic.

The details on the scope were done by adding fill layers and alphas I created in photoshop, sourcing images from the internet.I also used these alphas for the engraved/depressed areas where writing is stamped into the metal, also adding a height>normal filter so that I could capture that detail in the normal information instead of height. 

The magazine was a tough decision whether I wanted to make it transparent or not and in the end, I went with it being slightly transparent with bullets loaded inside, just to make it slightly more interesting and less bland. I also made sure to add roughness on the plastic too to break up reflections slightly and add more texture and story to it. This gun has been used for a while and it shows.