My personal definition for improvement is: Take a task, see how long it takes you to do it with a good quality level (not awesome, just good enough), then work out a way to reduce that time until you can do that same task with the same quality level from the beginning, but in a shorter span. Then, that remaining time you just saved, try using it to improve the quality. At the end, you’ll end up taking roughly the same time doing that task as when you first started, but it’s now a lot better. You have improved.
Efficiency is what we get taught at school and at our office jobs. It’s about reducing the time in a task, because time is money and so forth. But improvement, at least for me, is not about reducing the time, but increasing the quality. Time can be cut down later on. Once you can do the same task in the same time with a better quality, all you need to do is do that again. Find a way to do that new level of quality in a shorter time. Get efficiency out of your already improved quality. That’s when you get a win-win.
With that said, the first part of the texturing for the YF-19 was full of those little tests of trying to get quality but making small adjustments to the workflow so I can do that same thing faster every time. This was necessary because this is a beast of a map. It’s 6144×6144 pixels large, which is 6 times the 1024×1024 usual map you’ll find in a videogame. But since this is a transforming robot, that wasn’t enough, I needed a loooot of pieces put together in a single, huge square.
Once I finished the last bits of mapping the UV, I exported the huge PNG file to Photoshop and started assigning colours. Then I figured I’d rather spend extra time writing down what was what, so I wouldn’t have to keep going back to Blender and looking up at the map just to see if it was a leg or part of the chest. I wanted to know as much as possible from each piece without having to corroborate it in Blender. Things like the direction of the texture in relation to the model made sense, since I wanted to be able to tell where would things like the weathering and bits like oil would run to – away from the direction of the normal flight path. Things that are usually on the top would have a lighter shade because of the sunlight, while the things below the plane would be darker and muddier, since water and oil and gross stuff runs there with gravity. I’m trying to make a realistic texture, it’s what I’m trying to say.
So with these annotations I could be able to start cutting pieces out and naming them in layers. I wanted each individual piece to be easy to select, so I could just select it and apply a shade or a light or add panel lines without problems. I’m making a faster approach to the texturing, since I want to spend as much time as needed making quality out of each important piece.
After making sure the wireframe was following the mesh precisely to texture, I began the long and slow process of cutting each piece. Later on I’d find out that I didn’t need to be that precise, because I could always cut into smaller pieces when I needed to add details. The big important thing right now was to make a difference between the fuselage colour and the grey tid bits or the red and black rudders. In my mind I can change the full colour of the YF-19 from tan to white or black by just filling a single layer of colour. That way variations and different paint schemes will be easier to do later on. Again with the efficiency thinking.
The numbering system was eventually discarded, but I first thought about putting a number to each individual part. I had this idea of seeing how many parts would there be in total, but in the end I decided to not bother. The important thing was to name the part as simple and clear as possible, put the flight path direction if it was necessary, and then put things that would give me a clue of the colour, like “connection” or “metal ring”. I’m thinking this robot is built using similar materials for things like the hinges or connecting plates. So it makes sense they’re all using the same shade of grey and similar features like bolts or connection struts. Since I want to be able to texture them all in a fast way, I would spend more time naming these things.
The wing spoilers were a stupid thing. I spent a lot of time naming them, but as you’ll see soon, that didn’t make sense. They had to be mapped in their place with the upper part of the wing, since the paint job has a Macross logo that will include the wing and the spoilers. So it had to look like it was painted as a single piece.
I did things like separating with blue and pink things that were named from things that were not, then decided to just starting to assign a base colour to them to separate them even further.
Yeah, the nose logo. This is going to be a challenge. Those two parts are not even close in the UV map. the upper part of the logo is two different pieces, and the lower part is just one. Still not a clue how to pull that one off.
I also noticed the textures were not completely touching all the corners of each part. There had to be a small margin left in each part with the base colour, instead of filling each part precisely. I had to become, turns out, less precise to get a better result.
The moment I added those black wing stripes the whole thing started to really look like in the anime. I was happy!!
A small transformation shot to see if the textures were working well in the seams for the connection parts. Those untextured parts belong to the mechanisms, which will need their own UV map and are mostly mechanical grey things that I couldn’t add to the big image.
Ah yes, the stupid decision to map the spoilers separately from the wing body. I had to go back and put the UVs in position, reexport the UV map again and update the Photoshop file. But I’m glad it was not such a big deal.
Mapping the frontal part for the cockpit, this is getting closer and closer to the Anime look. Though I need to take a course on shaders. This material still looks like golden plastic. It’s supposed to be hi-tech planet re-entry resistant ultra-light alloy metal.
This one makes me so happy!!!
I’m gonna start going more and more precise on the paintjob till I have something that looks just like the Hasegawa model. Then, I’ll raise the stakes and go super detailed on each part. Weathering, panel lines, bolts, pins, handlebars, service doors, oil spills. Combined with the specular map and the bump map for the panel lines and bolts, this should end up looking a lot more hi-res.