In this article, we have focused on a topic that we believe will be of great importance to our industry in the coming years: artificial intelligence (AI). AI is also becoming increasingly important in the creation of 3D configurators.
The reason is simple: the computing power we have in devices goes beyond the human brain in several areas. This means that the machines are able to absorb information and process it faster than humans. Let`s think, for example, of modern cars that can drive autonomously without a driver. This is only due to the intensive use of AI.
In the medical field, thanks to the intersection of an incredible amount of date, we are able to accurately predict whether patients will be exposed to certain diseases and, on this basis, prepare treatments in time that can save their lives.
It`s no coincidence that companies like Google, Facebook, Tesla, Nvidia etc. are investing massively in this type of technology to be more effective with their products and to provide us with a virtual experience that is more similar to our tastes. After all, by studying our behaviour, the computer can offer us what we are more interested in. The market leader in this technology is the one who can theoretically ensure almost infinite computing power with GPU computing: Nvidia.
One of the applications of AI is directly related to what we do every das: rendering.
It is very impressive what the new Denoinsing alhorithms can do on the basis of AI. Last year at the Siggraph, an amazing application was presented that could render a scene almost in one sample per pixel and allowed the AI to reconstruct the image in the missing pixels.
Nowadays, denoinsing is limited to blurring the pixels to ensure that there is not enough detail in the final image. Instead, thanks to AI, denoinsing can create pixels so that the final product looks as if we rendered the image with high quality settings and higher resolution.
In any case, Nvidia is not the only one that goes in this direction, as Disney also moves in this direction and the application of this technology was already brought on the screen with the film Finding Dory.
What does this mean for us as users? It`s easy to understand.
In the future there will be much more real time on the internet. In the future, rendering will change dramatically and no longer be a process that won`t take many minutes, hours or even days to calculate. On the contrary, it will be a queation of moments, because the AI will perform so much in the future, what the GPU and/or CPU could not reproduce so far.
We do something in the viewport and will see the result immadiately on the canvas. If so, we don`t even need to worry about the settings in the rendering process.
There will be rendering engine interfaces where there is only a single “Render” button that you can click to create render elements for compositing. If we think about it carefully, all the parameters we have today in the various renderers are derived from the approximation we have to apply each time to limit the computation time of our images.
Do you think this scenario could be frightening? Do you think that this will somehow compensate for the quality of our work?
In general, I think it will certainly be easier to produce images and animations based on computing power and that users will be able to focus on the creative aspect that we think is more important in production.
The real question is: “Will machines one day be able to create images based on fewer human inputs alone because of their impressive learning speed?”.