After some late hours I have finally completed most of the Golem character's model. It isn't as precise to the reference as I would have liked but a very near likeness. This model's process is very different because I have no exact front and side orthographic views to work from. Just the one semi full frontal concept design. Also, obviously, sometimes there arise 3-dimensional impossibilities within a 2D design, if one does not strictly adhere to the rules of perspective. Here are some screen grabs of the current model:
One of the pros about this model is that all the pieces are rigid bodies and a therefore do not need to deform in any way (like the face of the Sheldon character). This allows me much more leeway in the modeling department and has thus been considerably more speedy.
Because I am being more free with the 3D interpretation of the design I have also decided to divert from the normal process of modeling, unwrapping and then texturing. I am planning to rather tweak, UV unwrap and texture pieces as I go along. This allows me to identify problem areas (such as the areas not visible in the original design that I have to freestyle), and adjust as needed as I go along.
I have also been watching some more tutorials related to rendering in Maya. Tonight I did some of the unwrapping and texturing of the three main 'henges' of the design and then had a go at some quick renders. I was extremely happy to see that some of my research has payed off. I believe I have found the perfect settings for rendering the model so it very nearly resembles the concept art. Here is a quick test render I managed to push out tonight:
Also, it seems that this setup is also relatively light on the render side, even at full HD. I still haven't cranked up all the settings for best quality and I am only rendering three textured models in the particular scene so it remains to be seen if the render times will stay low.
Here we have some videos I have been watching to help me decide how I want to interpret the 2D concept art in 3D regarding the use of lighting and skin shaders. While I like the 2D concept art very much I think the 3D application allows for some further exploration and pushing of the boundaries of what this character could end up looking like. Thus I will also be able to push the boundaries of my 2D illustration practice with the incorporation of 3D elements and skills within my process. There are many 3D illustrators out there that never venture into any kind of animation work, just using their skills for 2D illustration.
This 'Final Gather Lighting' tutorial has been most 'illuminating' regarding lighting my scene without the use of any lights whatsoever. This is great because it makes achieving the soft lighting from the original concept art much easier. There are other pros in regards to hard, crisp specular details in areas like the eyes, where that 'studio lighting' effect comes into play.
I was aware of ambient occlusion before watching this video but not aware that you could build it straight into your shader. This might be a way to go for me, we shall see. I would like the ability to change the colour of the ambient occlusion so might still have render it out as a separate pass and make adjustments as I see fit in Photoshop or after effects.
Sub surface scattering has always been something that I have wanted to master in my 3D work. It is great for creating more convincing skin. I think my character could greatly benefit from this effect. One of the biggest parts of the 'Pixar look' is their amazing skill in creating their iconic warm, beautiful looking skin in their movies. The biggest drawback for me here is that it will require splitting up my completed colour map into multiple parts so I am able to create the desired effect. This is a lot of extra work. Time allowing though, I think I might take a stab at it.
Showreel inspiration:
The first 50 seconds of this clip are the best for me. I like how the character has been to posed show off the character and hint at the context of the character and what kind of personality and genre you could expect to couple with him. The little addition of the rock he's standing on is GREAT as it suggests terrain and environment while not drawing attention from the character. This ties in with a conversation I had with Matt Johnson (my level 4 tutor). I mentioned to him my concerns regarding adding in some environmental elements. He brought up the example of an artist I love, Jamie Hewlett, creator of the art for Gorillaz. He mentioned how he solidifies much of his characters' context and personality by the use of interesting poses and suggesting little actions and quirks the characters might have. I was reminded that I touch on this very aspect of contextualising the character within my essay (attached to this blog).
Not sure I'm a fan of the choice of music in the clip above though.
This showreel has some better music. Upbeat while not overly so or distracting. He displays an array of skills including knowledge of key industry software, and with the Gambit character, he shows that he has knowledge of industry requirements in the form of a low polygon model for gaming, with high levels of detail through the use of texturing effects and Normal mapping.
This showreel is my favourite. I like the choice of music. He has achieved exactly the effect I had in mind when I started this project. His renders are crisps and clear, with very nice lighting. The background is part of the scene by making it an infinity curve as opposed to just sticking a textured wallpaper in the background. I love the smooth camera movements he has achieved. He shows the right details at the right times, e.g. the mesh. I don't like, however, that there is no suggestion off environment and the turntable spinning becomes repetitive as the reel goes on. This, to me, is not helped by a song that sounds really nice at the beginning but also starts to feel a little repetitive.
The book prop for my character is modelled and textured. As I am not doing any real animation at present I did not rig the book but rather just created open and closed versions for use in the showreel. Rigging would simply be a waste of time at this stage and isn't a skill high on my 'to do list' so to speak. Here are some screen grabs from maya of the basic, untextured geometry:
Here are some screen grabs of the open version of the book, all textured and pretty. Really loved doing the textures for the book. Adding in all the little scratches and ageing was a lot of fun.
I also added in the middle pages as extra geometry to add to the realism of the book a little and break up the solid shapes. The grey visible in the image at the edges of the middle pages won't render in the final images. This is due to a 'transparency map' that I created that will mask off those grey areas. This saves me from modelling all the edge tears in with geometry. This way it will render quicker but still have all those great little ageing effects. This in line with my understanding of requirements within a professional production pipeline.
Here are the textures I painted that are applied to the model in the images above:
Getting the cover design to work took me a little more time a finesse than expected but I believe the end result is up to the standard I have been working at up until now.
Here we have the texture for the inside pages of the book. This one texture map will be shared between 2 pieces of geometry.
This is the transparency map for the middle pages. A transparency map works exactly like a mask in Photoshop or After Effects. 100% white is 100% visible and 100% black is 100% transparent. This has been another case of how previously acquired knowledge has worked to my advantage, as understanding how to make this map work was relatively simple (it took some finesse getting it to render properly within Maya).
Now that the texturing is done it is time the rig the Sheldon character. This will allow me to pose the character for the final renders. This is a step I cannot skip and yet it feels like a waste of time as I would prefer to move on to the Golem model so I can move things along to final renders and compositing. Another example of the work that happens behind the scenes that few people appreciate. Here is a clip I watched a day or two ago that might be an alternative to full Maya rigging:
The advantage is that it's a MUCH quicker process (Maya rig will take at least a day for a basic rig), but it has some major drawbacks regarding the accuracy of the rig and the ease of use when it comes time to pose the character. Time is running out and this might be something better explored within the summer holidays. Either way way at least I am now aware of this method of rigging.
In the spare moments I have between working on the main assets for the showreel I have been making very loose little storyboards so I can start to plan the flow of the reel and what kind of music and mood I would like to shoot for. In a following post I'll post up some of the tutorials I have been watching to fill in some of the gaps in my training and also some of the other showreels I have been watching to be more aware of industry standards and which parts of the creation process would be most important to highlight within the final showreel.
It's been a while since my last post so there is much to show. Texturing for the Sheldon character is all but finished. Any further tweaks would only happen before final render time, after I have set up the shaders in maya with the appropriate attributes for shadows, reflectivity, ambient occlusion etc, etc. Here are two clips from zBrush of the textured head (where the most detail is). Keep in mind these are hardware (not software) renders and don't yet reflect the final look. Also I noticed that the colours appear a little washed out. An issue that will disappear at final render time from Maya as these are hardware renders. Make sure to watch them in HD.
Head textures, just flat colour, no shading. Might try exploring this look as it suits the style of the concept art quite well. It is something I haven't tried before and I feel could yield some nice results.
Same head and textures. The only difference is that this one has some hardware shading in the material. The hair colour has also been fixed since. And the eyes have been textured. :D
Texturing took longer than expected but I am very happy with the result. The Face took the longest time by far, as I had many issues with matching UV seams up so they aren't visible in the texture maps. All in all it was very enjoyable and it has been very gratifying to see the character imbued with more life. I was luckily able to draw on previously gained experience in advanced Photoshop methods and specifically texturing skills for this phase. These skills are also evolving as I learn about new texturing methods (such as the ability within zBrush to paint directly onto the model).
All the colours I used I sampled straight from my concept art which has been a massive time saver. Here is the colour palette I made for myself to take the guess work out of colour picking:
Pants is the South African way of saying trousers ;D
There are some changes in the colour due to the necessary conversion from CMYK colour space to RGB. This was obviously needed as the final textures will be used in a screen based outcome. Here are the 2D colour maps I painted that are to be applied to the character for the final renders:
The head colour map
The arms colour map (there actually two versions to introduce variance to the skin details, none of us are totally symmetrical after all :D).
The shirt colour map. A deceptively simple looking map. Took me ages to get some of these colours in the right place without having massive seams visible on the model. Gradients are a nightmare to match up.
The belt and buckle colour map. I also did a bump map for this map.
The trousers colour map.
The shoes colour map. This map also has an accompanying bump map.
What are bump and normal maps? Bump maps are grey scale images that one plugs into the bump attribute of a shader. Values above 50% grey appear to push the surface of the mesh out, while values below 50% grey create the appearance of cavities in the mesh. In the example below all the darker areas will simulate the indents in the soles of the shoes:
Normal maps are 'advanced' versions of bump maps. Normal maps use RGB values to map the direction of mesh 'Normals' to the XYZ directions in 3D space. This also creates the illusion of detail on a lower polygon 3D mesh. This allows for shorter render times while still supporting high levels of detail at render time. Here is the Normal Map for the face of the character exported from zBrush:
Normal maps are new to me and I am excited to master their possibilities in my process. This is of course a skill highly valued in game design production where low poly models are important in relation to processing power. Game characters need to be rendered very quickly on screen by need to appear to have high detail levels. The drawback of Bump and Normal maps is that they do not affect the silhouette of the model, meaning that some unwanted hard edges can be visible at render time. Displacement maps are a workaround here but are much harder to setup and cost much more in processing time at render time. I am hoping to get away with Normal maps as a happy medium between Bump and Displacement maps in this project.
I have also modelled and textured the book prop model, previously highlighted on this blog. Those wii images and textures will be uploaded in the next post.
This is just a quick update to mention that I have now added my recent essay on the relationship between character design and the viewer as a page to this blog. The page is creatively named 'Essay'
This essay encapsulates much of my research and thoughts regarding the convergence of narratives and their audiences, as mentioned within my learning agreement. Much of the research I did on the essay topic has informed my decisions at various points throughout this project and thus I felt that it's presence here on this blog would be most appropriate. Happy reading :D
Here are two WIP turntables of my Sheldon character. The character is now ready for export back to maya and then hopefully some texturing will follow soon.
A day spent in zBrush scratching around and watching and reading tutorials for just about every button. The character detailing is coming along quickly though and I'm learning a lot. Hopefully if all goes well this character will be done well before week's end. Fingers crossed.
It is such a great experience adding in all the little nuances that make a character come to life. :D
Fears are mounting that I might not have the time left before deadline to finish all that I had initially intended to do for the project. I lay awake at night fretting about this even though I am currently putting in 10-12 hour days in front of the pc. I set out with the intent of responding to three of Carl Jung's Archetypes but trying do do anymore than the Persona will inevitably result in ruining this project.
It is such a bitter sweet as I am so happy with the concepts I have made and how they are materialising into 3D assets, but I just wish I could do more without sacrificing quality. I am taking solace in the fact that this is a learning opportunity after all and if I knew exactly how everything was going to pan out then I obviously didn't learn anything. Experimenting and learning new things seems to be taking longer than anticipated.
As this is to be a showreel of sorts quality must inevitably take president over quantity. One of the top showreel tips I have come to be aware of is "less is more". there is no point in rushing a second character just for the sake of a longer reel.
Looking back on my progress, from the inspiration gained from Carl Jung's Archetypes to where I am now with the characters and props I have and am creating, I am satisfied with the trajectory of this project and excited to reach it's conclusion with a final product I am proud of.
Just a quick update. The Sheldon Parsona character is almost through the modeling stage and will soon be ready for some textures. That's where the fun starts. Here are some quick renders of the progress. He still needs some tweaks in the modeling department, but I'm happy with the result of the translation from 2D to 3D. Things never work out exactly like the concept art but I'm pleased with how little I had to 'jippo' the model in this case.
I can't wait to get him out of that boring 'T-pose' and into some more interesting custom poses.
Not many people get how long it takes to do many of the things that go on behind the scenes of a 3D character. Things I don't have to bother myself with when working on a 2D piece. Take for example the topology of the lines on his face. These lines have to be carefully planned out in such a way that they form circular edge loops that compliment the shape of the face and underlying 'musculature'. This is so the character's mesh deforms well during animation. Bad topology often leads to artifacting during the rendering process.
Another hold back on the speedy process is UV mapping. This process makes it possible for a human to interpret the mesh on a 2D plane and therefore use painting programs (like Photoshop) to apply painted custom textures to the character model. Bad UV mapping can lead to resolution and other texturing problems down the pipline.
As the fun texturing part is drawing near I am also turning my attention to new texturing techniques and shader setups I might imploy. In this regard I have been watching some more tutorials online to expand my knowledge to draw from when the time comes. Here is a clip I found particularly interesting and helpful.
It's been a while since my last update. Just before the Easter holidays I had a critique with two lecturers and a selection of piers from my course. All in all it went well. The main critique being that I needed to hurry things along as the deadline is approaching.
Over the two week holiday I completed all the final character and prop concept sketches, including the final colour. This was a challenge for me as I do not feel colour is one of my strong suites. Luckily, through much much trial and error I feel mostly pleased with all the results. Some changes my still occur during the 3D phase. During the second half of the holiday my progress was stunted due to the arrival of my sister visiting from Cape Town for a week and an 3000 word essay that was due today(12 April).
My sister's visit was not a total loss of project time as I finally had an opportunity (and the funds) to visit Stonehenge, which obviously luckily played very well into the fact that my Golem design is based loosely on the stones from that wonderful site. The photos I took were helpful in relation to scale and to be better aware of how weather erosion and other natural forces has affected the look of the stones. I am happy with my design and feel it is a happy middle ground between the reality and my personal interpretation via the character. Here is a small selection of some of the photos I took on my visit to Stonehenge:
Now that the holiday and essay are behind me and there is only four weeks left of this project it is time to blast full steam ahead. In this spirit I have already started blocking out my character in 3D (in Maya). My experience was a little lacking and I felt out of practice so I made use of a set of tutorial DVDs to help, and hopefully speed up, my progress. This specific set of four DVDs was procured via post from Digital Tutors, a leading online resource for digital artists, and is called: Creating Stylized Females in Maya. Here follows some images and a short clip from the beginning of the training series:
These tutorials have been massively helpful in rounding out my knowledge of some 3D concepts and bettering my workflow in relation to animation production flow for less problems down the line when it comes time to rig and pose my character. With any luck the next post will be filled with progress shots of my lovely characters.
