I want to create a picture book for my 2 year old son. He will have to undergo a heart surgery in a couple months. If everything goes well he will have to stay in hospital for couple weeks until he recovered enough to go home. My goal is to create a children’s picture book for him so he has a nice present from dad because I won’t be able to stay with him the whole time - f*** COVID .
→ I didn’t get even remotely to the point where I would have been able to render multiple images of the same level. I was once again overwhelmed by the scale of the environment and keeping FPS somewhat in check.
I’m very proud of the end result though because in the process I got to a point where I was very frustrated because I still had no artistic direction. I then ordered the book “The Visual Story” and was already able to incorporate some key things from reading the first few chapters. Especially regarding depth, I feel like I improved a lot. I added a couple depth cues to make the scene more deep (longitudinal planes, textural difference (a bit), movement, aerial diffusion, overlap).
My son likes all kinds of construction machines (e.g. excavator, trucks, front loader, etc.). So I think it’d be a great idea to create some small construction site. I’ll have to make sure that it doesn’t get too complex. So for starters I could just add an excavator that digs a hole and a truck which takes care of the soil. Maybe also a bulldozer.
The environment will be something close to where we live, with many hills but no mountains. Season should be spring or summer.
So far I only have a very hazy image of what the story could look like. Maybe we the truck could drive to the construction site. Then we have various images from the construction site, showing the various machines at work. Then maybe how the truck brings the excess soil to some other place.
Something I’m also not sure of is to what extent I should include people. I could use Metahumans, but they aren’t really fit for a construction site (wrong cloths). Also it’s a lot of work for a beginner like me to create all the individual poses and make each look natural (especially also facial expressions).
This project isn’t going great. I have about 3 weeks left and am still at the beginning with my landscape. I’m once again facing the issue that the landscape feels far too big for me to make something that looks realistic. To be honest, I should have restricted the level to a very small corner that is limited by cliff walls or something. That would’ve been much easier.
The reason why I took so long and it still looks so bad and empty is because I spent about 2 weeks studying Unreal Sensei’s Masterclass content. Furthermore, I made a terrible mistake in the beginning and wasted a lot of time.
The mistake I made was to not adjust the directional light strength but adjust the brightness of each and every material asset instead. That meant I had to tweak every single material manually and the same again once I realized it’s better to just adjust the directional light strength instead. The reason I hesitated to change the directional light strength is because the sky got so dark when I reduced the strength. However, thanks to Epic’s Stack-o-bot example project, I learned that you can adjust the sky color separately and still get a bright blue sky even with lower directional light strength.
The mannequin’s in the scene are also a really cool trick I learned from one of Epic’s videos. You can add the mannequin mesh to the procedural grass spawner of your auto landscape material. That means you’ll always have a reference human ready wherever your camera goes and even at different distances. Very handy!
.
