LOG #001
Small Flat where Merithèque’s Derivative Intelligence (M’) will spend some time learning the entire knowledge of the Non-verse concentrated in the city neural core.
---
When I was studying Visual Effects, I always wondered why artist would work on a shot basis and only make what’s important for the frame. Most notorious example was the use of Matte Painting where the artist would only draw the part of the mountain that needs to be within the frame. Everything surrounding it wouldn’t be drawn. I argued with my teachers that the future was to go full CGI and build sets before actually shooting anything, that way you have absolute freedom to navigate the camera and choose the axis you want. The problem at that time was the technology, rendering would take a huge amount of time and it was only logical to work only what’s within the frame even if that means re-doing all the work if the camera axis is shifted because this work time would still be less than the render time. It’s only recently that I think 0MN1 truly became possible (after 10 years of work) with the rise of GPU rendering (and further improved with OptiX) and denoising algorithm such as Intel Open AI Denoise.
We approach cinema to a more traditional way: Build a set first and shoot whatever you want within it without being constrain to re-work your set for every shot. CGI feels more “real” in their craft once the set is dressed up, allowing us to move the camera however we want within a second and instantly render the piece. That kind of feedback tends to blur the limit between the digital world and the physical onset action of moving a camera.