Update - Week 03 - 2022
A last minute contract for the weekend, lot's of optimization and assets for the painterly and all round incredible labour hours this week
February 25, 2022
Weekly Sprint Stats:
Points Velocity per day
This week had a last minute contract that could be done over the weekend, so I did it. I still managed to get some me and love time - but it was still a 7 day work week. The other project also advanced at a good pace, but not as fast as I had like, but fortunately the client has clarified timing which give the painterly some extra breathing room to work. I still have to keep the pace up though. This last minute project I could do over the weekend for some much needed timely cash in hand. I did do the job, plus revisions on both Saturday and Sunday, quite proud of myself - for a prototype, it was great.
This week had me slaving away 7 days a week, though in the weekend resting time was still included. I still got in a few extra days of hours in though - which did put a dent into the rest of the week. You can tell my points and daily were highly on point and productive. Probably the most productive week in a while! I usually get quite a happy high working like this - but I know my body and real-life relationships take a toll. So I usually have to learn to balance myself and stick to the health routines of eating well, exercising, and taking breaks. Sometimes workng less is doing more.
The Painterly Project
The painterly project during the initial part of the week got a lot of love, with new assets, new workflows to paint the assets quickly, a new map, new particle effects and also optimization (dropping viewport preview stutter from 3-5FPS to 30fps) with enough power to animate and work and do layout without waiting for the computer to respond. Now I could paint and colour how I would like.
Other aspects was the particle system. Here were my conclusions:
- Using a grid was expensive, because stacking planes to instance planes to instance instances was very slow. Here I learned using the process time on the nodes helped optimize that. I found randomizing the position of a single line was much more efficient to get a volume.
- I don't need to have the particles face the camera if the particles are constantly round enough and the texture based on Fresnel, also more optimal.
- I get a kick if I cull the particles.
- I can link the cloud and use it how I want with library overrides without too much of a performance hit
So with that I could build my coveted official DOF system, include some controls in the null object so that I could also override colour, saturation, luminance and other attributes of the point cloud strokes. This is the final development for the "Sunny Rain" technique for spacially coherent painterly post-processing.
After I worked on that, I then continued to refine my animation systems and the canvas edges and it's animation. I also started to work the map of the world, adapting a height map and turning it into the valley of the city. In the end it would be a good 225 square Kilometers, so.. I would have to come up with a way to get good backgrounds and layout based on camera and story. I will have to be particular here, because making an entire city is not enough.
So with it all, i started to put a scene together, started to find the issue with scale incoherence (or just massive things to small things) and started the ground work to start working the shots.
The weekend project was about taking huge LED screens on a building and creating an illusion ofr depth for keen tik-tokers and instagramers on the street. So I took this cablecar technique, quickly photobashed some buildings, found some CC0 assets and adapted what I needed to a very fast animation to a mapped building here in the city. The project had some revisions - and it was a big success - rendering in no time with Eevee. I even used some free LCD effects on the screen, and mapepd the actual building to 3D geometry to help me line up the illusion. These screens are seen around the world with animals, water and other effects to give large "wow" factor for publicity. And now I know how they did it. I didn't worry too much about how the rendering to LCD screen and mapping would be, but I managed to get the prototype well enough.
So all in all, thank you for supporting me.