GTA 6 is still a long time away, but meanwhile, the fans are busy dissecting the first trailer and getting excited about every little piece of information that comes their way. Recently, a patent by Rockstar Games’ parent company, Take-Two Interactive, suggests that GTA 6 will be hyper-realistic, particularly regarding character animations, especially the way characters’ limbs move. First discovered by Reddit user Fly-Low, the patent describes the use of machine learning for motion capture, wherein accurate hand details and movements are recorded.
Also Read: GTA 6 won’t launch on PC due to these 4 reasons: PC launch details, platforms, and more
What Does the Patent Claim?
The motion capture tech uses a machine learning model to track hand movements over time, using data from many sensors. This helps it predict future hand movements based on what it’s learned from past data.
“The systems and methods described herein overcome the imprecision conventional optical marker pipelines for the case of finger movement, where the slightest imprecision yields results that are not good enough for production,” the patent adds.
Also Read: GTA 6 may include gyms for fitness: Here’s how it could change your gameplay experience
What Could This Mean for GTA 6?
We have seen how detailed character movements have been in recent Rockstar Games like Red Dead Redemption 2—right from how Arthur swings the lasso to how he holds a gun. The addition of this motion capture technology through this patent could result in even better movement, making the game more realistic.
That said, you might not even notice the changes when GTA 6 releases. When something becomes hyper-realistic, one tends to not notice it all—this is especially true for character movement.
Also, it cannot be ascertained if GTA 6 will use the machine learning technology described in this patent, so until the game is released, this information should be taken with a grain of salt. B
Also Read: GTA 6 file size speculations: Fans predict between 200 GB and 750 GB ahead of 2025 release
