Character Evaluation

  26/02/21 - Week 21

Here is my finished character along with a lit scene in unreal engine.

This project has been interesting, difficult at times but overall rather fun and has introduced me to character sculpting which I enjoyed a lot.

The sculpt of the character itself, in my opinion, is probably the most successful part of this project, applicable to the brief and I really like the design I have come up with for my character. However, it is very tailored to my own interests and aesthetics. The design didn't change too much from the initial version though I didn't end up having enough time to figure out how to create a good looking mask for my character unfortunately.

There is still a lot for me to learn with sculpting characters and this definitely shows in some areas, such as the hair which I struggled for a long time to try to get right and looking the same amount of curly as I wanted it in my design. The wrists, interestingly, were also hard for me to get right and I still think from some angles they are too thin but from others they look fine. Having this depth aspect to creating characters is new to me and I still have a lot to gain and learn from it.

Moving onto retopping of the character, though I have resulted with topology that has facial and body animation loops -  this was something that I found really difficult and it ended up being a really time consuming and tiring part of this project. I didn't struggle with the face retop but having to have to find my own resources to try to retop the body and clothing, I found it immensely difficult. In hindsight and from feedback, it would have been easier to create a very basic version of the body and then add loops to it. The way I decided to do it was through looking at retop reference and trying to approach it the same way as the face, filling it in and trying to create good animation loops.

The worst retop section is definitely the hands, overall the rest of the model doesn't have too much faulty topology, but the hands were very difficult to me and I couldn't figure them out and therefore they aren't up to scratch at all.

Another thing I have ended up with is wasteful use of topology, initially I wanted the shirt to be visible through the top but I ended up finding that this looked unappealing. So, all the topology that would have been visible now isn't and is a waste.

As for the textures of the model, I also struggled here, but mainly on the clothing more than anything else.

The skin texturing was really fun and I enjoyed putting nuances and story to the characters body. Though I do really like the face and I think its one of the strongest parts of the texturing, I do think some of the bruising and cuts don't read instantly which is a bit of a shame but I hope that to other people it looks better than what I currently see. Another one of the best parts of the texturing in my opinion is the arms and hands. This part was also really fun to texture and I think the results look realistic and to the quality I initially wanted. 

As for the clothing, there is definitely an issue here which is quite disappointing. I found it hard to get enough wear and tear without making it look really distracting, I don't think I managed to find this balance. The reason why I tried to put more damage than I would have initially thought is that last term I got heavily marked down for apparently not including enough damage into my textures where wear would be present - so I tried to amp it up. The combination of confusing feedback and not really knowing how to apply such damage to clothing has sort of resulted in a weird result.

Here is how I used the texture budget and some shots of the areas of texturing that I like the most:

After finishing the textures I moved onto skinning and rigging. Though, initially I didn't have too many issues here, I ran into a major one at the end. This was that part of the skeleton had somehow merged with my actual mesh. I don't how how I had managed this but the only way to fix is to merge the skin modifier which would include getting rid of the actual skeleton, losing any actual animation and resulting in just a pose. This is why I have been including both the initial pose and the posed character as I understand that a pose without a skeletal mesh is not as useful. I do want to make it very clear that the character was posed with proper rigging and skinning in Max, including weight painting. You can view the evidence of this in the previous post - due to time constraints I was not able to re-rig my model.

The resulting pose is not particularly exciting but I think it fits with the character and I think it looks alright in terms of awkwardness, though there's still things I could improve with the actual structure of the pose to improve on the shape language anatomical logic.

Finally, the presentation of the character in unreal. This was also tricky for me as I do struggle with presentation in unreal. I tried my best to create an interestingly lit character with a nice light rim on one side to separate the character from the background.

In some ways, I feel I achieved this and it is definitely the best I could do with unreal engine. I still feel that my character just looks so much better in marmoset and I haven't got the same sort of quality on the roughness and flatness in the lighting that portrays my textures the best and without such high contrast. For comparison I will show my character in marmoset and in unreal.

If my character looked like it did in marmoset I think I would be much more pleased with the outcome of this project as I just feel like unreal has shown a lot of faults in texturing and made my work look a lot less professional. Another reason why I just feel as though the marmoset renders look a lot better is because of the background, as recommended by my tutor I didn't leave the model in a black void but I felt as though this looked a lot better as the shadows and the light on the floor just is distracting. Marmoset has a really nice compromise here which includes a blurred background which I couldn't figure out in unreal.

'

Overall, this project has been fun yet frustrating in some places. As mentioned, I sort of feel like the final renders in unreal aren't something that I am particularly proud of and it makes me notice a lot of the issues I had with this project, the texturing specifically. However, when thinking about this project in alignment to the brief, I think there are successes and I feel as though I have met the baseline requirements, portraying a good level of sculpting, design and pbr usage. The texturing itself, technicality , workflow and portrayal in engine is of adequate quality in my opinion. 

As I have described in this evaluation, there is a lot that I could and wanted to improve on and I think a lot of this is purely down to being inexperienced in the field of 3D character creation, I know that if I were to redo this project I would be able to do it quicker and of a higher quality. I could also include some of the things I weren't able to in this project, one of those being shaders as I did not have time to look into this and apply them to my textures. 

In my unreal file, I have included the model in the default pose as this is what I would wish to be marked on, though I also have in the files the posed variant, though obviously lacking any skeleton. This part of the project was a disappointment and I really wish that this had ended up correctly though there is no harm in having practice here.

There were some positives and a lot of negatives in this evaluation, though in the end this was my first attempt at high to low poly character modelling, having to practically learn zbrush and lighting in unreal on my own - I think I have created something that has just made me want to fix my mistakes and misunderstandings and do a lot better next time.