Final Year Project - Week Four: (20th February - 26th February)
- Ryan Harlee Jones
- Feb 28, 2017
- 13 min read
Final Year Project - Week Four: (20th February - 27th February)
Monday:
Monday I went into my morning session to start the week with the anticipation of harsh and critical feedback. I felt like I needed some specialist feedback as well as some from my actual supervisor. My supervisor discussed upcoming deadlines, and made some comments about what my current progress image on my scene reminded her of, and that she really enjoyed the overall feeling and vibe that the scene was giving off currently. She mentioned one or two possible references which might be of use, as well as mentioning movies that the scene reminded her of. She also allowed me to vent a little bit to her, about some of my worries about how I was aware that I might be behind compared to other artists, and that I was comparing my work to other people very frequently. She listened to me, and put my mind briefly at ease saying that it is expected for this reaction to happen at this stage in the project, especially when I'm adventuring into new territory with the sculpting workflow.
Fig 01. Progress Image 01 (Jones, 2017)
After the session, I decided to go and speak to a tutor that was more specialised with the work I was doing so that I could get some technical feedback. The feedback as expected was fairly hard to listen to, not because I was doing badly, or anything like I had been slacking. But simply because there was a lot of advice that I could take into consideration, and some of it meant recreating sections that had already had time spent on it.
The tutor mentioned that the scene initially looked very scifi, and that it was very harsh on the eyes because of how complicated with detail the assets were. He discerned that with the way I was discussing my project and issues, and from what he could see, that I had an extreme lack of real world reference. Also certain things I was placing in the environment did not directly tie in with the art references that I was using. Personally he found that the complicated use of geometric shapes wasn't going to work very efficiently with the medevil style that Darksiders 2 displayed.
The tutor ended up advising me to sculpt an asset so that I could understand the workflow that was necessary, and to see if I could efficiently use the program to debate if it was a sensible course of action. He advised a brush pack that was used by a Blizzard employee called the Orb Brush Pack, which he believed was perfect for the style I was aiming for, along with hopefully making my use of the application slightly easier.
The rest of the day I decided to get straight into testing my skills in the Zbrush sculpting program again. The tutor had advised that I just create some simple blocks to dig into with the brushes I had been given. I took some time figuring out how to set the brushes and the program up correctly, along with understanding the use of alpha brushes, the new brushes that I had been given and various other tools.
Fig 02. Block Test Sculpt (Jones, 2017)
At the end of the day I had managed to sculpt into the blocks fairly efficiently, and the test had been worthwhile in the sense that I had learnt a lot from working on a test mesh. Personally I was pleased with the result, although I knew that I could achieve better on my next attempt because of the issues and techniques I had come across while going practising.
One of the larger issues I had noted at the end of this test, was on the subject of subdivisions and the poly count that had arisen while adding the detail. When I came to bring the sculpted asset across into 3DS Max so that I could high to low poly bake the detail, I found that the ZBrush file was very sensitive because of how large the file size was. This meant that it was very difficult to export and import the asset into another program without crashing either ZBrush, or 3DS Max. Because of this, I decided to take an active check on the poly count in future when going up subdivisions that probably were not needed. My research also lead me to be aware of the decimation master plugin which is installed on the university computers which help to reduce the poly-count while attempting to retain as much detail as possible.
Tuesday:
The Tuesday I began by opening up asset dump file where everything I currently had was placed together, I evaluated what the issues were and why certain pieces were harsh on the eyes and also what didn't work with the scene as a whole. I decided to work on what would be the "Boss Door" asset, as this asset would be the only doorway that would differ from the rest in the scene. I decided to make sure to create this asset with a much simpler design, while trying to retain the core shape language and feel of what I had in mind for the environment. I wanted to create at least one asset which I could be sure wouldn't need changing later, so that I could carry it through to and past the sculpting stage to make sure I could correctly accomplish the workflow that I needed to.
Fig 03. Boss Door Asset Images (Jones, 2017)
I made sure to think about the application of the door in the world, by segmenting the door into separate sections so that it could open via a local mechanism. The door is split into four sections so that it could open as needed. Although the geometry to allow the segmentation was added later after realising the door would need a way to be used, so the change was made to correctly allow it to function.
Wednesday:
Fig 04. Boss Door UV (Jones, 2017)
With the Wednesday I continued on with the same asset as the day before, I got on with unwrapping the asset so that I could prepare it for high to low poly baking once I had accomplished the sculpting stage. There were a few geometry related issues with the doorway as I went about unwrapping it, although nothing major. There were a few unwelded vertices and double edges areas, although overall the wire frame was rather neat and efficient.
After this had been accomplished I set up the ZBrush file ready to begin the sculpting process. The process progressed smoother than I originally anticipated it to, although I did have multiple instances where I knew what I wanted to achieve with my sculpt, but I had to take some time out to do some directed research. Examples of this were learning about things such as how poly-groups functioned, so that I could break the separate door sections into sub tools. Through this technique I could work on individual parts of the door without having to accidentally work over into the other sections, this also allowed me to have more control over the scene and the model to make my sculpting process more streamlined. Another very useful technique which I learned about was the use of the lasso mask, with the curve selection chosen, this allowed me to select harsh edged areas with a much wider degree of control than I was able to prior, the other selection modes for the mask were also helpful, but not as much as the one I had researched for me personally due to the nature of my assets.
Fig 05. Boss Door Flat Sculpt (Jones, 2017)
The process I decided to work from was to use the flatten brush to run along the edges of the chamfered inner blocks, this allowed me to smooth out the edges further inwards to the centre of each slab. By doing this I could create an uneven wobbly edge, which helps towards the illusion of thick rock slabs. After going over the edges I decided to flatten them out a little more using a combination of the typical smoothing brush, and the Orb brush pack rock flatten brush, to make sure the edges themselves were smoothed out correctly and not mishapen.
To help to not affect the rest of the geometry while sculpting, I also made sure to make use of the before hand mentioned masking tool, this was very important in making sure the work I was doing was clean and not filled with errors.
Fig 06. Boss Door Uneven Sculpt (Jones, 2017)
I then tried to use the lasso mask once again along with the clay buildup brush and smooth brush to apply an uneven sculpt to the flat section of the extruded panels. Although this ended up taking to much time, and didn't help the overall feel of the material I was aiming for, the piece ended up looking very jelly like instead of stone.
Thursday:
Fig 07. Corrugated Metal Texture Test (Jones, 2017)
With the Thursday I thought I would look into one of the texturing programs which I hadn't had the time or thought to look into before hand. The university labs had Substance Designer available to me, so I decided to look into some base level tutorials by the company that created the software, Allegorithmic. My thought process was that if I could create a base rock texture to apply to some of the assets, I would be able to get a better idea as to if they worked the way I wanted them too. I ended up following a three part tutorial on how to make a very simple PBR tileable material texture, one which was classed as corrugated metal. The tutorial was informative and easy to follow along to, explained in a good manner. This research allowed me to understand what uses the program has at a ground level, which better helped me to understand what options I have when it comes to applying textures to my scene.
Fig 08. Rough Stone Texture Test (Jones, 2017)
I decided to fiddle around in the program a little more to see what kind of results I could come up with by myself. I spent an hour or two working my way round the menus, knowing what I wanted to accomplish, but pondering over how to go about it, clipping back and forth between my tutorial file for reference. In the end I didn't come out with anything impressive, although I had created myself an uneven normal map pattern, which added the uneven rock shapes to my sculpt which I had tried to apply before hand. So with a few more tweaks I believe that I could make use of that to all my assets at a later date to add more variety into the flatter sections of the sculpts, therefore saving time and being more efficient.
Fig 09. Boss Door Finished Sculpt (Jones, 2017)
I finished the day by accomplishing the rest of the sculpt for the boss door asset. I made use of a variety of different brushes to accomplish the final look. These brushes included a set of Orb rubble brushes to apply the dents in the door, so that there looked like there was wear and tear available. A set of Orb slash brushes accomplished some more subtle additional wear and tear, although very hard to notice. Orb crack brushes also allowed me to add the other layer of detail visible, although the cracks could do with some more variation and work on them in my opinion. An alpha brush that I created in Photoshop was also used to create the image at the centre of the door, which after creation was just as simple as applying it to the mesh. The end result of the door was better than I had hoped originally, and I'm very pleased with how far I have come with the sculpting workflow, although I know there is plenty which I could improve on or do more efficiently, and hopefully I can correct some of these things in my next attempt.
Friday:
The Friday began with a small issue, I got started with my work, and I was prepared to high to low poly bake the sculpt which I had finished the day before, only to find the unwrap I had done was jumbled again for some reason, this didn't take long to unwrap again, although it still confused me as to how the mesh had jumbled up again. The baking process worked correctly without any issues, and I created myself a normal map for the asset. The asset looked good with the bake applied to it, although most of the details such as the cracks and dents needed to be much stronger, so I manually intensified the normal map in an effort to make the textures pop stronger. I then went into Photoshop and saved out the normal map with a technique I had researched online to intensify it. This was done by duplicating the normal map, and playing around with some of the duplicated channels and levels, along with an overlay setting to make the map pop further.
I also during this period of the day decided to play around with some of the colour values on the asset to see if any particular base colour or combination would help to make the asset view a little more towards stone. A dark grey/blue helped me visualise things a little more clearly.
Fig 10. Floor Panel Asset (Jones, 2017)
About midday, I felt fairly happy with the progress I had made with learning about how to apply new workflows, so I took a step back and decided to reconfigure the floor panel a bit so that it would fit better into the scene, along with fixing some prior small geometry issues. Asset smoothing groups were also manually applied to make sure everything viewed the way I wanted it too, along with finishing the unwrap for the floor panel.
Fig 11. Floor Panel UV (Jones, 2017)
Since the floor panel only needed some light tweaking, I still had the motivation to create another asset, the wall surrounding the Boss door that I had created seemed like a logical thing to create next. I already had an earlier design for this specific wall which has been mentioned in a prior weeks blog post, and I felt like the design would still work with all of the changes that I was making to the scene. I ended the day with getting the model to the stage where chamfers and final touches were probably the only things needed before the model could be unwrapped.
Fig 12. Boss Wall Asset (Jones, 2017)
Saturday:
The Saturday morning I planned to begin by making things a little neater in my master scene in 3DS Max, I went through each of the assets and made sure they had correct naming conventions, were placed correctly, and also had a specific material applied to them depending on their current stage, for example if they were unwrapped or not.
Fig 13. Average Wall Asset (Jones, 2017)
After this was done, I got to work on some of the assets which still needed some adjustments, the wall panel that will surround the other type of door only needed some minor tweaking, although extra details were added as well as taken away, so this asset took a little more time than the floor panel had the previous day. The geometry was changed and optimised along with having smoothing groups applied to the asset manually again.
Fig 14. Archway Extension Asset (Jones, 2017)
I swiftly moved on to create an extension piece for the wall section I had created, since the wall panel needed extra length to make sure it connected with the other sections of the scene. I decided to use a kind of archway wall panel which was set back into the wall enough to lay candles or statues into the alcove. This was created to the stage of smoothing groups being applied manually also.
Fig 15. Progress Image 2A (Jones, 2017)
After creating the extra wall extension, I'm still worried that the scene may be too complicated or busy once I apply other assets or even add the doorway pieces into the scene. I feel as though the pieces in the image fit together well and seem to fit the theme, but I think the scene is on the verge of being overpowering again.
To finish off the day, I open up Unreal Engine 4 and started setting myself up a folder structure and uploading a few of the pieces into the relevant folders ready for whenever I start using the engine. I also took a little time to look at how substances are imported and used in the engine for the sake of general knowledge, that way I knew the pieces I bring in if any can be used appropriately.
Sunday:
Usually Sunday is the day I take off to make sure that I have some spare time to myself, although considering throughout the week I had done dotted pieces of things such as going to a hypnotist show etc, I felt as though I needed to balance it out a little bit and at least go in to do some work for a handful of hours.
Fig 16. Slab Extension Asset (Jones, 2017)
During the time I decided to create another wall extension piece in an attempt to see how an even simpler wall piece would work in the scene, hopefully adding some negative space to give leeway for some more of the detailed assets. After creating this piece and setting up the environment to test it in for comparison, I found that the wall that the extensions were attached too, was having some heavy smoothing group issues, even though they registered as correct. I heard of a different approach by another student to applying smoothing groups by just simply adding a smoothing modifier to set a base group for everything. After this, they would then make an edit poly modifier group for each smoothing group and then just stack them. I attempted this to see the result, and it fixed the issue I was having in the viewport. I also checked this issue in Substance Designer to make sure it wasn't just the viewport in 3DS Max, and it still persisted.
Fig 17. Progress Image 2B (Jones, 2017)
The feedback I attained from fellow students, were along the lines of that the stone slab pieces help bring it out of the sci-fi zone, although it makes the floor and other pieces possibly not match, and makes the scene lose consistency with shape language etc. Although, even though the archways work with the environment, it seems too much like something else compared to a stone tomb or temple, an example given was a library. This gave me a lot to think about, especially with how I wanted to advance with the project.
Overview:
This week has felt much more productive than prior weeks, although i'm also starting to debate if i'm wasting a lot of valuable time. I'm very pleased with the research and new skills I have gathered, and that I have proven to myself that I can accomplish the workflow I want to up to the sculpting stage. I have also created a set of fairly nice assets, even if not all of them get used, I haven't been slacking, I have been working moderately hard on trying to advance with the project. The worry is that I'm creating these pieces and doing things which will get scrapped and have to be redone again because I'm doubting my direction or ability, I feel as though the Monday session with the tutors will need to be another harsh one, as I feel as though my skills aren't the issue here, but my reference and planning. I know that I can advance more comfortably if I take a step back and review what I'm doing by getting some proper reference, and drawing a line in the sand with what I will do direction wise, to then stick with it and see it through to the end.
- Ryan Harlee Jones
Comments