Google Stadia May Eventually Become More Responsive than Local Machines Thanks to ‘Negative Latency’
Google Stadia is about to launch. Though we don't know the exact date yet, the cloud-based streaming platform for games is set to debut at some point in November.
There's anticipation to see how latency will fare in a non-controlled environment, with many still doubtful that Stadia can deliver a responsive experience, particularly for fast-paced games.
However, speaking to EDGE Magazine (issue 338), Majd Bakar (VP and Head of Engineering for Google Stadia) boldly stated the goal is to actually use something called 'negative latency' in order to eventually make games feel more responsive than possible on local machines such as PCs and game consoles.
Ultimately, we think in a year or two we’ll have games that are running faster and feel more responsive in the cloud than they do locally, regardless of how powerful the local machine is.
Bakar was also recently interviewed by Wired and provided a little bit of extra info there.
Our platform and infrastructure allows for techniques that create additional time buffers. We can generate frames in less time than it takes consoles or PCs, and with our machine learning experience, we have built models to help with the prediction and generation of content faster. This counteracts the impact of network distribution time.
Machine learning is involved, then, with AI algorithms likely doing some predictions on what will happen next in the game in order to reduce latency. On paper, the concept is not too dissimilar from GGPO (which, by the way, just went open source and is now available through MIT License). It remains to be seen whether
When it comes to decoding, some had been wondering whether Google Stadia would use the AV1 codec. As it turns out, it's still based on Google's own VP9, though it has been significantly improved for the occasion as revealed by John Justice (Head of Product Management).
The VP9 codec is open, but in Stadia's case, we do special work in the encoder that makes it super fast. That encoder still follows the VP9 standard format, so every VP9 decoder out there can read the streams. But our encoder is specially optimised, making it much faster. We put a lot of special sauce in there.
In related Google Stadia news, the launch lineup lost DOOM Eternal due to the game's delay into next year but gained Red Dead Redemption 2.