Captains Table Environment - Texturing in Painter (Week 7)

Work In Progress / 27 November 2022

Full Title: Game 3001 Professional Brief 2 ) Captains Table Environment - Texturing in Painter (Week 7) (14/11/2022)

Baking and Setting Up Masks In Painter)

By using my setup texture sizes I had planned previously I was easily able to make scenes and add my low poly bulk exports for baking. I simply had to enable vertex colour painting in the bake settings and toggle the far side distance for the camera projection. Unfortunately baking is not without issues and for smaller props I realised a lot of detail was lost if not cut out like the typewriters metal hammer head. Some geometry was failing to bake since there was a missing plane below to try and capture the floor detail so this fix was relatively simple. I went back and re-exported the high poly with a different “opacity colour” plane below and cut out the value when baking by making a fill colour mask. Set to the appropriate colour for opacity. Furthermore I resolved my normal issues by painting the value out of my low poly. To do this I enable ambient occlusion as a fill channel mask and set both normal and ambient occlusion to replace. This means I can drag and drop the masks into my scene with a white mask and paint out the black values of where baking artefacts may occur.

Texturing and Alphas - Larger Props)

Fortunately the main setpieces of my large prop export, like the corkboard and paintings had alphas I could easily curate from the episode. As per the nature of the public domain licence I could easily manipulate and creatively change artefacts/shots from the episode to adapt them as alphas for my texture. For the Mrs Melville painting, as a lot of the original episode was in relatively low resolution I had to use AI to upscale the original image before adding a fill content around the outer edge. I then lightly blended the painting together along any sharp colours to try and make it more natural and extend the border edge of paint since it would be relatively zoomed in. I also additionally made a height map by converting the painting to grayscale and adding a filter to mimic the effect of oil paint when it dries. 

Furthermore I repeated similar steps to my photographs to make them into polaroids but curated a base height/roughness map since the gloss of the photo would be evidently more than the white base. Elements of the old style Los Angeles map remained unchained as it would be odd to change the reality of urban planning however I did take a newspaper article from the time and reframe a shot from the episode of a “magazine issue” to make it more believable as an interview from the paper. Also changing the time on the newspaper in the process to help fit the time period the murder would take place in the 1970s. 

Texturing and Alphas - Small Props)

Evidentially I had a lot more alphas to curate for my small props as a lot of my props were of varying accounts of micro detail/labelling. For the easiest alphas I took to Photoshopping shapes of the hole indents on the rotary phone dial alongside the numbers/cutout hole combination. These worked well on Substance Painters fill layers as masks and let me add the normal detail in without generating further masks. Meanwhile the champagne bottle I made from scratch and tried to mimic an old timey french brand with its make of label. I did however choose to base it more on an easter egg as the brand is called “One More Thing” one of Columbos famous lines. 

Another detailing was the font work, as the labels for crates, typewritten script and evidence stickers were all closely matched across references to make for believable labels. The scribbles seen atop of these are also drawn by me in real time as I chose to take a thick marker and sign as Columbo filling out the evidence boxes as messily as I could. I then took these forward into Photoshop as pictures, using curves and masking to get rid of the page. I had the most fun making these alphas as I could really go crazy hiding details in dates, signatures and writing to make the narrative feel more lived in.

Texturing and Alphas - Typewriter)

Following the import I had the most issues texturing this asset as the low poly variant had very messy topology. I had originally attempted to combine planes and flatten the circles on the sides out to the rest of the mesh to save on geometry. This was a horrible choice as the bake took to making everything but spherical shapes. So I kept the geometry in and textured very roughly as a smooth typewriter would be. But ended up once again running into further issues as the keys were all sharing the same UV space. Meaning the individual keycaps could not have differentiating letters. At first I figured I would simply use a decal, but this could prove very expensive I later uncovered. Since there are over 20 keys and 20 decals on 1 mesh would be costly. Let alone the mesh holding at around over 8k tris. 

So I went back to the drawing board and realised I would have had the top of the keycap be unique on a separate sheet with all the variants of the possible keys. However to texture this, I would also need a lot of alphas. All 23+ for each key and button. Exported individually. Manually. It took a day and was very draining but the result came out great so I’ve learned crucial lessons in choosing my retexturing battles for the future And realise this would also be the least performance costing measure.

Texturing and Alphas - Books and Pages)

For this sheet, I needed to design the books the “Mrs Melville” series curated. As in a primary shot near the beginning when Ken (the murderer) meets Columbo for the first time he cheerily stacks all the books on him for reading material. Colymbo of course does the reading and soon realises that Ken did not write a single book with Jim. So with a single shot from a side profile I went to work making fictional book covers. I did not want to start from nothing however since I had not experienced such graphical work before. 

So I investigated some series of books now in the public domain due to their over 100 year old age and took the art/stylings for my own. Replacing the name, changing colours and key focal elements and wrapping it around a banner. I made 6 of these titles, which are your favourite? I also took to personalising the arch binder margins after making a fictional replica of one titled “Flare”. The scribbles once again coming from my notebook in the similar Columbo handwriting style as I would imagine even the messiest and outward all over the place detective would keep records someplace. In a somewhat readable fashion. 

Captains Table Environment - UV Unwrapping and Vertex Painting (Week 6)

Work In Progress / 27 November 2022

Full Title: Game 3001 Professional Brief 2 ) Captains Table Environment - UV Unwrapping and Vertex Painting (Week 6) (7/11/2022)

UV Unwrapping Process)

A lot of my geometry was unfortunately cylindrical which did pose some issues when unwrapping. The new version of 3DSMax has two unpeel modes, the default being “Unfold” and the secondary (possible legacy) being LCSM. Throughout my entire unwrap the first and default mode is horrendous to geometry and does little to flatten the UVS with constant warping even on square faces. I have lost so much time going back and forth between collapsing the modifier to check on geometry and save it, to reopening and not realising I am on the default Unfold mode that proceeds to pack, detach and ruin my pre-existing UV islands. I have tried to swap the default but to no avail and apart from that I didn't have too many issues with the unwrapping process. I tend to unwrap meshes separately and then after resizing the islands, putting the entire set out of bounds in a clockwise direction so when combining I can easily redrag it in. Resizing out of bounds is also much easier when combining a lot of UVS since it means geometry will not overlap and cause having to detach and move it safely out of the way should it be symmetric or sharing UV space for whatever reason.

UV Unwrapping Final Pack)

After unwrapping all my objects separately and proceeding to combine them I started to pack that of my unique geometry first. The typewriter was relatively easy since a lot of faces could be symmetric minus the top cover since it had a unique emblem of text on the far right side for the company logo. Models such as the desk were very simplified as geometry was already squarish minus the bevelling that I did have to flatten down and may be slightly wonky. Since a lot of drawers are the same I can easily copy and cut down on UV space by having a few and duplicating later. More complicated packing was encountered during my smaller props since I had to consider what deserved more UV space. A lot of my geometry was quite modular but I knew I would accrue some minute detailing on the champagne bottle and rotary phone by scaling them down so I ensured they took priority. I also tried my best to flatten these cylindrical meshes since I knew applying alphas and texturing patterns were more complicated on UVs that were not square. This brings me onto my final sheet of books and pages that I did not unwrap completely square as the unwrapping gave up. I kid you not I had so many issues flattening and then copying this geometry across that I figured I would tackle this in texturing via tri planar and means to fix it further there. A lot of space here would also be highly reused so I scaled paperwork higher in terms of UV space to ensure it could be readable. Considering my previous budgeting sheet of 2K for texel density.  

Vertex Painting and Combining for Bake)

Vertex painting was quite quick to exact since following the unwrapping process I had packed and maintained sheets of similar assets already. All I needed to do was select individual meshes and ensure I  filled colours that were contrasted in values and different across the board to break up areas of varying materials such as the wooden chair vs the filing cabinet. The inclusion of the chair on this sheet was also last minute as I realised its importance in the scene I could too have given it another unique sheet. However I knew since a lot of its geometry was symmetrical I would be withholding a lot of UV space on this unique draw call and instead opted to simply add it to the large props sheet. 

Weekly Evaluation and Fixes)

Overall this week was quite intensive as I had to finalise, fix and unwrap my models. However as I did quite a lot the week prior I was able to catch up following some unproductive days and busy schedule. One caveat I did encounter was the unwrapping process felt a lot more tedious on this updated version of 3DSMax. Since the default failed to unwrap correctly I don’t understand why it's even an option if it's marginally worse. However on the flip side I did uncover new ways to make spring like coils which I will definitely take forward to further models since it's a detail I would hate to sculpt manually and may be put off retopping as well.

Captains Table Environment - Modelling Large and Small Props (Week 6)

Work In Progress / 27 November 2022

Full Title: Game 3001 Professional Brief 2 ) Captains Table Environment - Modelling Large and Small Props (Week 6) (7/11/2022)

Modelling Larger Props)

To begin modelling the remainder of my props, I chose to tackle the larger ones first since similar to my table greybox, the geometry was already close to final. For my painting and corkboard I simply had to bevel the edges and create a distinct cutoff between the centre and inner of the board. However, for my corkboard I also added some additional planes to try and map out with colours how I could arrange pieces of paper, clues, photos or maps to make an interesting composition. I opted for a large spread with a lot of lines interconnecting where I plan to add Ken Franklin's badly buried clues that cropped up over the episode. Alongside Mrs La Sankas unfortunate demise at the lakes. Further on the crates and blinds were modelled off of boxes with the blinds using align tools.Align tools are great for maintaining consistency amongst spacing meshes, and by making a spline that went down horizontally I could generate even distributed blinds. For my final model however I did delete these meshes as I did not want to unwrap them all individually when I could simply reuse one.

Modelling Small Props)

I had a lot of small props, primarily a lot of modular ones but a few unique instances. The most unique of the bunch had to be the coffee mug of which I took reference from my own as I frequent tea brews during modelling. Luckily a lot of these smaller props were cylindrical and symmetrical so I was easily able to extrude and delete unseen geometry below and shell anything seen to thicken geometry. I did encounter some minor issues when modelling the ashtray as I wanted to reduce geometry as much as possible however when adding a turbo smooth modifier to add a higher poly variant issues cropped up. 

The geometry did not deform well to my optimised triangulated angles to save faces between the centre and outer edge of the ashtray. I ended up having to remake this geometry specifically for my high poly and symmetry the shape back to as close as I could in position back to my low. For meshes more aligned with realtime ref however I was able to pull off the champagne shape quite well by extruding upwards for the body and scaling it to a real time reference. I knew a lot of surface detail for the label and uppermost packaging would be much easier applied in texturing so I did not model this in. 

Modelling Rotary Phone and Lamp)

During the modelling and blocking process for my rotary phone I struggled to get the side shape and profile of the main body. I ended up having to accumulate some reference on planes but also realised the complexity might end up making this a more intensive prop than I had realised prior. As the glass sits on top of the numbers with distinct holes drilled in of which you would spin the rotor to dial. To pull this off in the engine I would have to design a glass shader and place it onto the geometry. So I modelled the overlaid geometry on top and split it off to the side to ensure my bake would not accumulate further issues as I have in the past with overlapping geometry. 

I had a lot of fun in particular making the cable as 3DSMax has a spring spline tool that I attached to another spline with path deform to easily stretch out copious amounts of coiled cable. I then duplicated this spline stack and pathdeformed a cylinder of similar proportions for my low poly variant. In conclusion, my most satisfying model to date. Further onto my lamp, luckily my greboxed simple geometry was there partially so I simply had to remake the base with more detail in an extruded cylinder and shaped the upper holding segments using Path Deform. I manually extruded out the geometry to make spherical orbs and used smoothing groups to break this apart from the main mesh more. I had debated whether to model the inside of the lamp but through some further research realised the outside may be green, but the inside is white. This is due to how bounce light works and if it was to be green the light emitted may be affected by this tinge in hue. Additionally white does not absorb as a colour and a glossy overpaint will help tremendously in spreading light. So I shelled this detail in.

Captains Table Environment - Texture Planning and Further Research (Week 5)

Work In Progress / 27 November 2022

Full Title: Game 3001 Professional Brief 2 ) Captains Table Environment - Texture Planning and Further Research (Week 5) (31/10/2022)

Texture Planning)

Considering the briefs not so strict budgeting guidelines I decided to myself set a few boundaries to ensure I get the most out of the 1024 texel density. The primary points being that more important assets/complex props will get their own sheet to maintain detail and quality. This can be seen here in my planning, as the wooden desk has a 2K sheet to itself alongside the typewriter at 1K. In general I believe the typewriter may need another decal sheet/channel on the pack to have all the keys be separate considering the nature of a typewriter's ability to write over 20 keys. However I also realise I have a lot of small and large props planned accordingly like the painting that could easily be unwrapped and textured on the same sheet. To conserve on draw calls and to make organisation a lot easier in engine.

Modelling Table & Chair)

As my greybox was modelled off of a photo reference from varying angles, the silhouette already matched the real time variant I wanted to replicate quite well. So I took this mesh further and refined/fixed its drawers by extruding out the faces and dragging them out around with a shell to make individual drawers. I then bevelled the underside of the table and the very corners to make for more interesting surface details with the engraving upon the wood. The drawers however I realised were very similar in design and width so I only had to model 4 of them and I can easily duplicate the meshes later on to fill the remaining 3 drawer slots. One issue I encountered however was partway through I lost some of the underside of the mesh as I had deleted this geometry thinking I would not be seeing if the chair should be tucked in etc. So I bridged this back and welded the vertices correctly back.

The chair on the other hand was a lot more difficult since the low poly greybox variant did not match the silhouette whatsoever. It was a lot more blocky and simple so I opted to start fresh and take my referencial plans into 3DSMax. I quite enjoy this method of modelling as I have recently gotten into a habit of prioritising references that can help me model accurately in this manner. Further on I use squares with missing front and back planes to help drag the silhouette from a sideways angle and further refine it from the front by alternating my point of view. For more complex geometry such as the armrests on the chair I used splines with the path deform modifier to curve the angle of a cylinder that I later simplified to a more squarish shape. I did this by removing the edge loops and the centre to make a rectangular armrest following the curve. I realised that luckily I can maintain only having to model one since a lot of the chair is symmetrical in nature. And used some more extruding and cylinders to form the bottom side of the chair and chair legs accordingly. 

Modelling Typewriter)

The typewriter was a very late night stretch project which I do not recommend making in one sitting as it definitely did damage to my sleep schedule. However, back on the modelling side, it was very satisfying to shape the plane from a side on view and further extrude it onto the front. The shapes of typewriters are very well optimised and broken up into the top head that hides the internal key switches and stamping portion, the side and back that house the joining of these two aspects and the top that holds the paper. Primarily when modelling I was able to get away with using planes to shape to my photo reference however I did have to bring in cylinders due to the nature of  both keyswitches, the rollers and clips for the paper side of the mechanimal equation and the small pipes connecting the keycaps to the stamp. 

Whilst a lot of modelling here on the hard surface was smooth sailing by alternating, swift looping and cutting to my reference planes I did encounter a few issues. This came in the department of micro detail such as the back plates having vented holes and screws. Additionally the topside had some text that I was unsure if I should accost to modelling or simply alpha on in the texturing process. My final issue came in the cost department as the asset ended up being over 20k tris for a mid poly. With mid polys I duplicated the mesh and allowed there to be an even higher smooth variant for a true high poly and moved onto optimising, retopping and fixing the lower poly to a much cheaper cost. I knew that whilst leaving more budget for hero props was not outrageous amongst assets over 20k is still beyond the boundary for reasonable assets. And would have to half this to even consider putting it in the engine.

Weekly Evaluation)

Conclusively I’m very happy with the progress I have made. Unfortunately due to the scope of this project I have had to put some of my personal work on hold and know my schedule is more busy for this month than the previous project so will have to knuckle down. However I am also a large Columbo nerd and my enthusiasm powers me to work at ungodly hours and get stuff done. So stay tuned!

Captains Table Environment - Blockout Lighting and Iterations (Week 5)

Work In Progress / 27 November 2022

Full Title: Game 3001 Professional Brief 2 ) Captains Table Environment - Blockout Lighting and Iterations (Week 5) (31/10/2022)

Research Columbo and Narrative) - Table and Chair

After some constructive Columbo bingeing (I swear it was for research) I settled on the first episode as previously discussed the week prior. However, I realised that as the project centralises upon a table that narratively belongs to the protagonist/antagonist this will too be my greatest challenge. As Columbo is a detective who is always working in the field and very little is shown of Mrs Columbo or his home life. And so there is a lack of what Columbos desk would really resemble should he even work from home. So I took to imagining this aspect by researching various hardwood furniture that matches the office styling of the 80s. I knew a lot of hardwood was popular, however I didn't want to choose anything too fanciful since Columbo is known to get by on what he’s got for too many years despite any wear and tearing (like his coat). And so with a large messy desk in mind and a collective worn chair that matches the wood styling of the desk I also thought it best to look into his signature coat. As despite having no knowledge of modelling or constructing clothing I figured it would be a fun stretch goal to achieve and put in place this project since it would be highly distinctive of Columbos character.

Research Columbo and Narrative) - Lots of Props and Clues

Through a lot of rewatching I had been able to identify key props and bits of evidence that cropped up throughout the episode to lead to Ken Franklin's demise. Columbo ultimately stumbled on a lot of these too or made note by conversing with Ken who as the episode went on was way too consoled and calm for being a possible hit suspect for murder when his co author Jim Ferris dropped dead. The first of many being the note that was left strangely folded and typed from Jims typewriter regarding a list of gangs. This was falsified evidence by Ken and its folded state in the drawer was strange since if it had been typed in the office of which Jim was taken its state would not have been so. Considering this I believe modelling both the typewriter, the note and the later secondary murder weapon of the champagne bottle are crucial. 

The champagne bottle had battered the poor witness of Mrs La Sanka. Who may have been able to assist Columbo should she too have not fallen for Ken's schemes and instead opted to blackmail him at a dinner party of which she later got intoxicated. The rotary phone is also key as whilst not direct evidence as the one Jim used to phone his wife claiming he was at the office, was located up at Ken's cabin. During the 80s it was possible to place phone calls that could not be directly tied to a receiver and as such hidden from possible police investigation as the outgoing address is never stated. I also believe the design of a rotary phone will date the detective's desk heavily since mobiles hadn’t existed fully yet and will bring about nostalgia to those who maybe owned them later of the period.

More smaller assets I plan on curating to fit to Columbos character are coffee mugs, boxes of evidence and old timey cigars/ashtrays. Nearly every episode I have watched (up until season 9) there is frequent drinking and smoking of cigars. Whilst Columbo doesn't drink as much his pose and entrance to an episode is always accompanied with a small cigar in his mouth as he sniffs about the crime scene. The effect of smoking would also impact other assets as the fumes would eventually dye plastics and other materials nearby. Furthermore by modelling this in with some light VFX for the smoke I can pull off the 80s LA detective aesthetic through wear and dirt accumulated on the soundings due to the conditions of smoking. 

Some more generic detective oriented props I decided to take a look into were that of the typical evidence filled corkboard and the folders/files. I realised to make a messy scene I would need a large amount of paperwork and books but with the time constraints I would need to plan this efficiently. Looking at the construction of folders, the arch binders and general paperwork I figured I could make a few assets and interlink them to make more complex meshes. However the corkboard has a lot of these papers and more detail of points added on top. I realised if I replicated this method I might end up making the environment look too similar so I will plan accordingly to have the corkboard be more unique. I definitely want to highlight aspects of the Los Angeles area since that is where Columbo is based and a lot of evidence surrounds particular locations of Jims office and Ken's cabin in which I can map. 

Research Columbo and Narrative) - Room Structure and Lighting

For arranging and finding out what props to curate I looked into some shots of police precincts/offices of the time period. Alongside looking at fellow dioramas and how they are structured/composed to give that detective feel. Overall I found that a lot of materials are limited to carpet, wood and paint plaster for a more home office feel. However wood was very prominent in choice for chairs, doors, cupboards and tables with a dark oak colour to it. Since I didn't want the values to read poorly in possible dark light conditions I will be very cautious when deciding the colour of my walls and floor when compared to lighter folders/case papers.

Modelling Blockout)

After getting my preliminary research moodboards done, I moved onto blocking out my greybox meshes. These are highly low poly and focus on shape for composition alongside scale. I knew as a diorama that might be highly modular I would want to spread assets evenly across but still balance the usage of modularity so as to not feel overused. Considering these factors I used a lot of block squares to quickl;y model out and swift loop cupboards, filing cabinets and the general diorama space. It would be a cut in of the room and as such I wouldn’t have to worry about making a door or any real ceiling as the lighting condition would let me hide this factor. However I wanted to split the values a bit more as the blocky square meshes felt very compact so I used cylinders to make bins, champagne bottles and the rotary phone. I wanted to focus on arranging the mess as well, as after utilising MassFX Rigid Body I found out how easily you can simulate a scattering effect without manual arranging. However this modifier was not without issue and would clip meshes with some protruding too far downwards under the floor so I had to readjust accordingly. Overall however the silhouettes of my meshes are there alongside the blinds that would serve as a focal point of lighting as I went to export it into UE4.

Setup and Lighting in UE4)

The import was luckily without problems however I did realise the scale factor was quite small initially for the scene and I had to ensure all models were scaled up by 9x. An oddity that had occurred to me in the previous project however after I scaled things up I began to add some simple master materials for colour, emission and roughness. My primary objective was to help get the closest values that I would use in base colour texturing so I can play around with the lighting setup and try to closely match it to my reference. Overall I used 3 main lights; one shining through the blinds to the desk to cast lined shadows, rectangular light to make the interior brighter and a final one around the outside to light up the false city view. Due to the nature of rect lights this flooded the light complexity into the red which is not ideal but not entirely negative to performance on such a small diorama scene. However I realise at later stages I would most likely have to play around with more post processing to limit the cost of lighting as my geometry will no longer be as low poly. 

Feedback and Changes)

I was able to get some insightful feedback from my peers after my greybox. This primarily revolved around how dark the outer edges felt and how the entire scene overall needed more lighting as a lot of meshes were difficult to see. Additionally the strong yellow lighting was a bit off putting so I will have to play around with value later on to make for a more warmer but distinct detective piece. Some more concerns cropped up around the coat since I have never modelled clothes before. The zbrush work and construction will definitely be a challenge and highly time consuming alongside the typewriter so I will most likely have to cut it, but have put it on the list as a further stretch goal since it means a lot to Columbos character. Overall however I’m pretty happy with how it is going and looking thus far, and after adding some textures onto my final meshes I can play around with fine tuning the mess of the office and pushing narrative aspects.

Final Greybox Lighting)

As seen here are my typical camera layouts. I tend to freeze the transforms here so I can do a before and after nearing the end of the project. And helps keep consistency in my screenshots. One main detail I wanted to highlight in my screenshots was a different focus per the camera shots to highlight something new of interest. The first angles highlight the entire environment as a whole, with the following being more focused on the mess on top/under the desk. Final angles revealing the corkboard and painting that hangs behind the chair. 

Captains Table Environment - Brief, Research and Concepts (Week 5)

Work In Progress / 27 November 2022

Full Title: Game 3001 Professional Brief 2) Captains Table Environment - Brief, Research and Concepts (Week 5) (31/10/2022)

Writers Note) 

Hello and welcome back to my first blog post of the secondary professional brief projects! Due to my previous weeks being dated from 1-5 I have decided to keep up these conventions to showcase how I worked in consideration of my planned timetable. I have since attached my project for formative submission and have been in the process of getting feedback of which I will later act upon nearing the final weeks. This is a personal decision due to wanting to focus on another project and in the case of learning new methods of fixing, modelling or more streamlined pipelines I would prefer to polish at the end rather than the beginning.

Brief Breakdown)

As the first week for the new project I decided to tackle my chosen brief, Ubisoft’s Captains Table with great caution. This brief has a generally creatively open theme basis for the environment, simply having to fit as a diorama for a “Captains Table”. The table and its contents being the primary focus in the scene and henceforth of higher detail. There are some recommendations for texel density in terms of matching the 1024 per metre quality that many games of the previous generation are developed towards. I will be taking these challenges on and will explore making more higher polygonal assets that have been optimised to a greater standard thanks to this open ended brief. 

Time Management)

With the nature of the brief being very considerate of concepts and themes that would apply to a captains table I also budgeted an extra week into my timetable. I realised that I would inadvertently need more time for research and concepting as the weeks come on due to the nature of having to research everything from scratch and bring a table/room diorama to life from my own design. However I too realised that this coming month would be one of the busiest for myself personally due to varying events, personal commitments and holiday season that may cut down on how much accessibility I have to my computer/labs working environment. Considering all these aspects in mind I went to replicate a more open ended timetable from my previous project leaving additional days for missed work and shall design my asset repertoire accommodatingly for modularity. To ensure I get the most out of my assets and time spent in development. 

Research and Concepts - Captain of the Miners Brigade Workbench)

My first concept came about from creating the most nitty gritty but contrastingly glowy diorama, centred around a captain glowstone miners workbench. A more rough and tumble design, framed under a set of lanterns and glowing ores preceded by an eternal dark mining hallway. Assetwise it would be very simple with blueprint plans, pickaxe and glowing ores scattered about to dictate the captain's rank and expedition. This concept sets upon a high fantasy setting and would be a primarily texturing challenge to create realistic but convincing gemstones/ore. With the provided additional setup in the engine to make glowing effects and setting a more whimsical scene. 

Research and Concepts - Head Librarians Illustrious and Dangerous Tome Study Table)

My secondary concept revolves around a more darker and fantasy based narrative of a head librarian's illustrious library of hidden tomes. The individual being a more apprentice in the means of the magical/darker arcane arts would have a series of floating and demonically inscribed items around their desk. Depicting scrolls and spells with instructions laid in the aged pages of forgotten tomes. Overall a much more particle based challenge compared to my miners workbench as a lot of assets would be animated to have floating if not VFX based visuals to give off the magical aesthetic. Furthermore I can imagine blending the realtime technologies of the early 20th century to add more real life flair to the librarians possessions with a gramophone. Playing ambience into the scene.

Research and Concepts - Lieutenant of a Cyber Noir Alternative Future Work Desk)

Moving onwards to my third and final concept, I tried to distance myself from the more fantasy based ideas and create a concept sounding around hard surface sci fi elements. Depicting a police captain, heavily inspired by the genius homicide detective Columbo on a case in an alternate future. Whilst taking palace in the noir like era of the 70s and early 80s the technology would have a more holographic means to blend with the environment. Adding a lot of light and narrative challenges to redesigning pre existing technology such as the typewriter. To fit a more alternate future should holograms have taken root in screens and replaced paper for keeping records, maintaining evidence and notes. I thoroughly enjoy the detective and crime solving genre and feel this would also be the most interesting to narratively build as the desk would be flooded with clues and bits and pieces surrounding a mystery the player could resolve. 

Final Concept and Narrative Changes)

After getting some feedback from my peers I realised with the realistic timescale of the project I would not be able to fully delve into developing/conceptualising alternate tech should I go with the detective concept. However I also realised that this was the most narratively compelling concept and might prove easier to develop than the miner workbench design due to its evidence based points. Additionally I have over the growing  months grown to love the hit 80s detective series “Columbo''. Which has in recent years become public domain, meaning utilising episodes and shots of characters/video are free provided some transformative nature. And so I have concluded that I would love to pay homage to Columbo by transforming my final concept from the complicated and possibly head scratching alternate sci fi detective themes, to the realistic diorama of one smart headed homicide lieutenant of the LAPD.  The following board depicts many assets I would be excited to tackle alongside evidence and clues I will further research into this week to make for an interesting diorama.

Fruit Cart Environment - Final Renders for Formative and Project Write-Up (Week 4)

Work In Progress / 28 October 2022

Full Title: Game 3001 Professional Brief 1 ) Fruit Cart Environment - Final Renders and  Project Write-Up (Week 4) (24/10/2022)

Fixes and Feedback)

This week as the final week before my formative deadline (for harsher feedback) I took to polishing the project the best I can. Thanks to my peers and tutors I got a lot of mistakes pointed out to me which I promptly went on to fix; wood grain wonkiness, alphas popping out with the flag and much more. My hardest fix was to go back to 3DSMax, re-unwrap the wheel flat into any available UV space and retexture it before combining it with the original material. This took me the longest as I suffered  a lot of baking issues but was luckily able to keep two separate files to prevent any crossover damage to my original cart painter document. My flag on the other hand just required some manual repaints of the edge border and contrasting manipulation in the engine to have more predictable damage/tears without floating geometry. Items such as the lantern had more nodes added in the blueprint to speedup my timeline and with light tweaking to light values/glowing I developed it into a more consistent flicker effect. 

Renders)


Material/Lighting Breakdown)

Modular Breakdown)

Video Fly Around)

Project Conclusion and Evaluation)

Overall as the first brief I feel very refreshed in my knowledge of both texturing and modelling efficiently. I feel I could have added a lot more normal detail however in bakes and that by avoiding Zbrush I did more damage than good to my project. But at the same time I was able to efficiently manage my time better than before to create assets that fall tightly upon the budget both texturally and model wise. I had the most fun with lighting and modelling my windows as setting up the glass shaders and playing with the cloth modifier taught me some nifty tricks around more complex shaders. I will definitely take this knowledge forward onto the second brief as I want to make more sculpted means and try to use 3DSMax modifiers to their potential to simulate more complex meshes. However I do also note that as formative the project is subject to feedback and there is still work to do. So this isn’t the end just yet :D 

Fruit Cart Environment - Decal Making, Material Fixes and Niagara VFX (Week 4)

Work In Progress / 27 October 2022

Full Title: Game 3001 Professional Brief 1 ) Fruit Cart Environment - Decal Making, Material Fixes and Niagara VFX (Week 4) (24/10/2022)

Making Decals)

Luckily from my summer personal projects I already have some experience with developing and applying decals and so was able to use a complex mask setup shader I made to stick a bunch of my decals onto a master decal trim. I chose a bunch of old war posters since the time period being only 20 years young of WW2 I figured some might still be about or as seen from their condition, very decayed/worn. It also sets the mood of London in the coming years, a steady victory in a country still recovering from the ill effects of war. A grim reminder if you will. However, moving on my other decals were a lot simpler such as the puddle decal which had mainly roughness and normal values. This decal works best at lower camera angles as at the end of the day it can only fake that water detail. My book also due to its symmetrically unwrapped nature was missing text and so I added some diary entry alphas to hint at Old Man Morris’ week (spoiler there's cake and candles). 

Applying Decals)

Some may think decals are as easy as dragging on the applied deferred decal material into the world, however as usual the engine takes out extra time to make things difficult. A lot of my decals spawned facing at odd angles and hence were deformed so I had to manually adjust these values in scale and rotation. I then played around a lot with depth since the freedom of using the deferred decal workflow allows light fade effects that worked a charm for my posters. My signage was also a more difficult fix since the decal was slightly off shade so I had to edit it in the engine with a Material Instance to get the correct ground values and hide the symmetry. The effect the decals provide definitely hides a lot of symmetry and makes it feel more lived in.

Fixing Texturing Issues)

I realised a bit too far into the exporting process that my scene had a lot of weirdly missing roughness values. So I had to go back to Painter especially for assets such as the fruit to paint more value into the base colour and roughness settings. I blame this predominantly on not setting a base background roughness before layering my mask editors on top as it must've taken that value as flat and made a very basic roughness map atop the entire fruit. Luckily my crates were not affected so it took a short time to paint in. My flag also as previously told in feedback felt too clean so I painted in a lot of wear on the middle and bottom since those are the segments of the flags that would be most blown about/worn with time. Also subtly included some gunshot marks in the centre where it is dark to see those alphas. 

Fixing Lighting and Using Level Sequencer)

Another issue I tackled from the previous week was just how dark the scene was in total. I realised I was being very sparse with my lighting and limited so I went the nuclear route, deleting and starting from scratch again. By adding a volumetric fog this time I realised just how much I can gain from scattered light and worked to add more rectangular lights. By setting these rectangular lights to a very wide setting and not casting shadows I could light up the left side wall a lot more without casting odd shadows as a predominantly “fill” light. I also added a beam of light coming from the window as a spotlight to give that extra focal direction to the cart since that was the primary source of regular light in the scene. With a subtle direction to cast the harshest shadows as a over and up moonlight source of light. Level sequencer also came in handy to set up the stills engine of capturing the moment of my animative cloth/glowing lantern. And as further described below rendering out some shots of my VFX.

Wind in UE4)

I realised this week that whilst discussing the setup of wind I hadn’t included it in my blogpost of the earlier week and so want to document it here. The setup is fairly simple, by applying an alternate material ID to the moving element e.g. the flag you can import the document as a skeletal mesh. This renders it not static unlike the other standard import and assigns it a physics asset. In said physics asset you can make out some collision boxes for sections you don't want the flag to clip through like the middle pole or the very top of the banner. Back onto the skeletal mesh you can select via the material ID and create clothing data. This lets you vertex paint clothing values from 1 to 100 of how little/much the cloth should respond. There's also some more settings regarding the strength of the cloth, tearability and very physics oriented presets for things like leather vs cotton on flags. But since it's all manual editing for this I felt the default worked well enough.

VFX - Fireflies)

I was luckily able to bring back one of my favourite VFXs I made in the second year and edited the properties down to a more minute spawn. The fireflies are also a bit larger to help visibility at longer distance shots e.g. the diorama view. And using post processing has a highly glowing effect. Some settings I did have to play around more prominently with however was just how many spawned initially. I strongly felt that not enough spawned from the original settings but also didn’t want to overencumber the scene with lights all the way from the left when the main focal point was the cart.

VFX - Steam)

Another highly useful and reusable VFX asset I made prior within my summer projects was the smoke/steam effect. It uses a simple low res smoke map that animates to a less visible puff with a range graph controlling its scale over time. Through the end of its lifespan after spawning the steam gets more see through and fades quite smoothly. Like my previous effect, the fireflies, this too had to have a lot of internal tweaks. Since I used it in my summer projects to project a lot of steam atop buildings like smoke from chimneys. However by removing a lot of value in the original spawning box settings I was able to narrow down a section in which it spawns only directly up. And by further narrowing down a smaller pool of smoke spawning I could get a consistent steam effect of which is close to that of a small steam vapour from hot tea. In which I hope this effect showcases subtly considering the thermos nearby and the nature of the thermos keeping drinks hot for way too long. (Personal Note: add ice cube to your thermos to cool tea down)

Blueprints - Glowing Lantern)

One of the most fun things in the totality of my project was setting up this simple looping blueprint to add more charm and realism to my scene. It runs on an eternal loop that defines itself on a core boolean value of IsColoured. When false it will wait a few seconds before running through a timeline value graph. This timeline graph increases the float value of the scalar within the material overtime (adding more and less glow). However as it comes to conclusion it also sets the value of isColoured back up to true which makes it wait further to run it back and reset the original value back to False. This infinite loop is engine efficient and gives me plenty of control of how fast the fade occurs by simply speeding up or slowing down the timeline.


Fruit Cart Environment - Cleanup and Export to Unreal, Shader Setup (Week 3)

Work In Progress / 27 October 2022

Full Title: Game 3001 Professional Brief 1 ) Fruit Cart Environment - Cleanup and Export to Unreal, Shader Setup  (Week 3) (17/10/2022)

Cleanup For Exporting to Engine)

With all my materials made, models textured and made I went back to give the final polish before export. This was in the form of delegating appropriate material IDS to all, with most meshes having only 1 draw call material upon them. However with cases of the flag (which I want cloth engine applied to) I would need 2 draw calls. Furthermore my window material needs both glass and the back glowing pane so I need another 2 draw calls to separate  these materials. I also used this time to quickly stick together some meshes I pulled apart for the baking processes like the lantern similarly akin to Lego. My final pipeline of preparation for export is to reset all points of rotation to the centre of objects apart from taller bits of geometry such as the banner and cane. By setting it to the floor the general mesh can be manipulated much better in the X plane of rotation. I also removed any unnecessary data in XForm and set all meshes to point 0,0 as is standard for exports.

Setup in Unreal Engine)

Thanks to setting up my lighting earlier in blockout and general arrangement I already had a lot of my scene ready to drop in my models. However I messed up the scale somewhat as everything in the engine was 2.5 scaled up and the meshes fell short. I had to readjust this manually per object however the remaining process of material setup was very smooth. I was able to play around with varying shaders to bump up the quality of my windows and fake that refractive effect when moving side to side. Through a “skeletal mesh” I was able to add cloth to a physics object copy of my flag by vertex painting points of stress onto the mesh and adding “Wind” to the scene. However upon initial inspection I uncovered I still had work to do as my meshes on the cart had a stage lighting effect where the centre was way brighter in base colour than the exterior. I uncovered this was due to some open vertices and after welding the issue was resolved.

Unreal Engine Creating Some Shaders)

As previously mentioned I went on to make a glass shader for my windows as the regular opacity was quite limiting in faking that glass look when viewed from other angles. This shader was relatively easy and I need to thank Jovan Tomasevic for his tutorial series on Glass Shaders on Youtube(Tomasevic, 2022). A very insightful look into how the refraction nodes are assembled to have a convincing lighting effect and quite easy to follow along too. Moving on I experimented with parallaxing my brickwork and stone trim respectively however uncovered that there was little effect to the eye in difference and distance from my chosen diorama angle. The engine cost would not be worth the shader and so I instead opted to randomise the brickwork instead to add variety to the tiling. This worked out great and I highly recommend checking out Prismatic Devs video series on 5 minute shaders (Prismatic Dev, 2021) such as this that can add polish to any material.

Final Thoughts:

Overall I quite like how the models look in the engine however through some insightful feedback I uncovered some more points of polish to tackle in the coming week. Some feedback from my tutors pointed out, the flag is very clean and polished so add more wear and tear. The area appears quite bare so decals would spruce up the sounding quite a bit. Additionally the area is very dark too dark for a regular street and so adding fog or some directional lights might help and provide better shadows. With this feedback onboard I’m confident I can polish and add more in the coming week. A note to also make that whilst it meets the university brief quite snugly, visually I can go more ham on my personal ArtStation presentation since I will not be limited in technicalities there. Eg adding more fruits, upscaling my textures so on and so forth.

Bibliography:

(2022) How to create a GLASS Material in Unreal Engine. YouTube. Available at: https://www.youtube.com/watch?v=dQehkg9VHJo&t=106s&ab_channel=JovanTomasevic (Accessed: October 20, 2022). 

(2021)Texture Variation Node | 5-Minute Materials [UE4] (2021) YouTube. YouTube. Available at: https://www.youtube.com/watch?v=yMmUdoAVfOs&t=228s&ab_channel=PrismaticaDev (Accessed: October 20, 2022). 

Game 3001 Professional Brief 1 ) Fruit Cart Environment - Making the Window, Walls/Floor Trim and More (Week 3)

Work In Progress / 27 October 2022

Full Title: Game 3001 Professional Brief 1 ) Fruit Cart Environment - Making the Window, Walls/Floor Trim and More  (Week 3) (17/10/2022)

Research Old Style Window)

Due to its complexity I decided to delegate a further week to my windows. For research I stuck pretty close to references of old Victorian windows with a squarish style. They were predominantly made of wood and their design and stature is iconic due to how symmetrical and stackable they are in the taller floors. Whilst they also were stylish of the period I have to mark how unsafe they can be as they only open halfway and I can imagine probably didn't provide too much ventilation in British Summers. Furthermore the design being square I feel offsets the rounded nature of my other meshes, with a cylindrical banner, rounded fruits and curvy cart made of polish wood. I believe it will be a nice contrast away from other forms in the scene and bring a lot of focus onto that of the cart in front of it.

Modelling - Old Style Window)


Like my other models using referential planes, I cut out the rough shape from a referential photograph of the window. I tried my best to account for possible skewing since a lot of references were from a top up angle even slightly. However with this shape of both the pane and the glass made I extruded the sets in varying lengths/heights to make it more convincing at a side profile. I then took these values polished up with bevel for the corners and added my swift loop + turbo smooth workflow as I have previously to my work to make the panes pop. I then moved onto the interior of which I had little patience to model anything complex that would be baked down anyways, so I opted for a drawn curtain scene. With this I was able to further explore the cloth modifier in 3DSMax that was able to simulate clothing ripples and forces. 

These effects relied on a very high poly plane that additionally was set with varying “node” points. These points coincided with vertices on the mesh that would deform as per outside meshes would too. I chose to settle on boxes and placed 3 to control the top, bottom and middle of the mesh for deformation. After a few attempts I got a slightly subtle side swept look and really liked how it came out so went onto arranging the window for baking. I knew I would have to somehow fill in the background behind said curtain but with vertex painting being able to easily differentiate the window from the glass pane, to the curtain and so forth I wasn't worried about this process.

Texturing - Fake Room Alpha (for Window)

After scrolling through my Photos Archive of various places and objects of my travels. I opted to use a bathroom/washroom photo from Warwick Castle in 2019 as my backing piece. I chose this one specifically as it had nice warm yellow lighting and had enough elaborate detail to be seen as a room, but also not enough to be seen as anything specific. Since the focus was not that of the room/window but the fruit cart predominantly. After making 2 alphas, one for my base colour and another for fake height detail I took these into the texturing process.

Texturing - Old Style Window)

After bringing my meshes into Painter, I first got to work on the window. I applied a slightly tinted white which would work as a rough colour base for an old wood paint job. I imagine that during the period as it is now, it's not the cheapest job to repaint windows and its slightly worn look would add to the pre-existing grunge and grime to the place. Subtly I also added cracks atop the corners and windows for slight wear and damage from possible elements/people/birds. Moving onto the window pane, since I knew this would be most difficult to set up and most likely semi transparent in a means I added just basic colour and highly glossy values. Onto my curtains, I wanted to highlight depth in my base colour mask so I added a few subtle streaks across the purple to mark what was closest/furthest in the material. Using my reference I was able to identify points of translucency to the background and ensured that the curtain blended subtly to said background alphas with a few masks.

Designer - Walls and Floor Trim)

This is definitely the mother of all my graphs, the greatest spaghetti I have cooked up thus far in Substance Designer. Primarily due to my very simple lack of understanding how imports work in graph to graph interfacing at time of creation (I know what to do now don’t worry). My trim graph has 4 different materials that all coincide and combine values to make the final output, henceforth why it is so ginormous and terrifying. To break down each smaller section in a larger “tldr” sense the top section combined the tiled floor stones into rows. It then randomly fills sections with colour for more variation amongst these stones, before applying height depth onto them against the background. 

I was able to get this to work on the lot of stones using edge detection before making further normal and Ambient Occlusions maps. The secondary smaller graph below is a simple blend of noises with masking to make a gritty pavement. I didn’t go too hard on developing normal or ambient maps here since this kind of the material is relatively flat with a few scrapes/cracks as per the nature of wear and tear. My other graphs followed similar structures for making a cleaner cut stone for the wall alongside a more heavily masked out street signage for that extra detail touch in the scene. 

Designer - Brick Tiling)

Starting on my brick graph, luckily Designer provides a plethora of tools for generating the basic shapes. However after some further look into reference I decided on a more longform brick and used transform nodes to get the best extended cut of bricks. I then used some light roughness to give a more roughened shape to the brickwork than standard clean cut rectangles. This roughness was able to carry over with further maps to curate the roughness blend that went over my entire brickwork. 

However I decided to take an alternate approach to my base colour within this graph. And for extra detail I added 2 variants with different levels of grime/wear for vertex painting in the engine. To add said grime and moss I simply applied some varying roughened noises with colour gradients that would blend with a multiply/add effect to give more/less of the moss across the bricks. For my noise however this remained the same for simplicity sake, and had been originally pulled from my pre roughened brickwork layout. 

Applying the Environment Trimm)

After exporting out my trim to appropriate resolution (2K) I quickly went back to 3DSMax to apply appropriate material IDs. For my wall I ended up having 2 materials in a larger sub object material node; the trim and the tiling brickwork. However for more simpler meshes such as the floor I only had to have one material set with the trim. This was to reduce draw calls which can be costly in performance and considering how not a lot will be culled away from view in a small short diorama every little helps. For my UV unwrapping in trims I handle the islands a bit differently. I end up flattening the initial UV down before applying a series of cutting edge selections. After the selections are made the UV can unwrap a lot more flatter with a simple “straighten” click. These longer versed UVS now tile vertically across my trim and can be dragged up and down into position alongside sideways distributed for that extra variety.