Tbf Im tired with all this scaler and ray tracing talk. Wow you can tank a card with calculations it was not really made for. Impressive.
Ray / path / whatever you call it , will always be a simplified implementation of actual proper light tracing. Which is practically impossible because the formula runs to infinity. Ofc we dont need infinite bounces, but all this talk is just nvidia marketing , nobody really needs this - it hardly makes the games more pretty, just more easy to program - and spends a ton more ressources. I do not want my graphics card to deprioritize gaming graphics performance , to give way for lighting engine cores - So whats next, we need fluid simulation cores we need differential equation cores , we need blablabla. This was never the approach for a reason, it is insanely expensive and is really only a benefit to lazy developers.
Sure ray tracing is a cool technology - but its not new, its not groundbreaking, nvidia did not do the bulk of the work making this possible even and its just unfeasible anyway and immensely expensive to do. Usually what we do in gaming is make the best with what we have, we dont suggest people go and buy a full fledged real time physics simulator.... We can implement all kinds of cool physics into the games, it would just tank performance. Many old games have more than good enough lighting to be immersive , and most was just made with clever programmers and a curious mind. This idea of bruteforcing next gen graphics is outright stupid to me.
When you can demonstrate an algorithm that gets very close to the real results, without spending 150 watts on lighting, or implying that I need to buy a lightsimulator to get light in my game..... Then we can talk.
You can hardly claim to be anything but ignorant if you think that VSR spending 200watts upscaling a video on youtube is amazing .... wtf people.
In its essence it is a question of ingenuity vs brute force. Brute force is the expensive one, both in terms of dollars and watts. Im to the core not impressed whatsoever. It is a slippery slope for the entire industry imo. And is it really of benefit to the gamer when "all you need" is just an insanely expensive number cruncher , to give way for developers with zero clue about gaming engines and graphics. Its lazy, its stupid and its not to the benefit of the consumer.
Consider if in doubt about what to think, that actual real fluid dynamics simulations in practice is psycho difficult even with endless ressources. Look at formula1 as an example - again and again the numbers did not line up with practical results (these people have 100m$ computers for this). So we should just throw all ingenuity out the window and start doing fluid simulations the "physics" way? ... it does not make sense
I wonder , for the lulz, how jensens ray tracing would react to a double slit
😀That is to say, even this implementation is far from the "real" algorithm. It probably has no clue that light has wavelike properties, and that different wavelengths interact differently.
Facebook
Twitter
Instagram