Changdong Wang (Chad)
I created a total of 10 different plants, replicated randomly by some simple scaling, and rotating transformations. In Houdini, I first created some blocks to replace the city model and then I created some random points on the ground. Search through the point cloud and if a building block can be searched within 0.5m of the point, delete the point.
At the beginning of the part 3, I build a simple shape of objects and key a camera animation to fix the whole content of the video.
Then I start to building some building which are close to the camera.
For these closing builidings, I build real 3d models and for other buildings which are far from the camera I’d like to use normal or bump texture to show detail. Based on some codes, I only build some fundation model units to generate different buildings.
Then, I want to choose spesific areas to generate trees through attribute transfer and points cloud. because I don’t want put plants on the road and building, so if the the point too close to these models the program will delete them automaticly.
Then I import some trees from Speedtree. And use stamp to rebuild the size, transform and rotate.
for further work, I will add more different buildings and road detial on the scene to make them look more realistsic.
In week 7, I start a challenge video as part 3 of my personal project. I find a reference video below.
Obviously, The difficult point in part 3 is the huge scene with plants. Besides, I plan to increase details with some Roma buildings. So, there are some Roma style buildings and different trees in part 3. I will build these buildings automatically in Houdini and scatter trees of different shapes.
I have rendered part2 in week 6. Through trying the Houdini rendering pipeline I found that the render layer is easier to control in Houdini than in Maya. In Houdini, I can directly set force rendering objects and cancel rendering certain models or only keep the matte of some object in one render pannel. In Maya, if I want to separate the rendering layer I have to create different collections and add different objects or lights, in a word, it is not easy to use the render layer in Maya.
This week, I am starting part 2 of my personal project which focus on camera animation, lighting and material in Houdini. Normally, I am more familiar with the UI of Maya, however, I am learning FX in Houdini. It’s a terrible workflow to make the same material in two software and transfer file between Maya and Houdini frequently. For trying everything in Houdini, I have started to make the part 2 animation.
In the process of part 2, I split the environment fog. In my old workflow, I always render models and fog together. Although that way can save some processing time, it cost more time on rendering for denoising.
Once I separate them, I can control the quality of images independently. The cons are that I have to set different render layers, but the method will reduce rendering time.