Following NVIDIA’s footsteps, AMD decided to launch their variant of RTX Voice dubbed as ‘AMD Noise Surpression‘.
RTX Voice comes as a package known as ‘NVIDIA Broadcast‘. It features support for filtering unwanted noise from your microphone, removing background without a greenscreen and even allowing your camera to adjust the picture based on your position. It is indeed a product of magic. This was released back in 2020.
However, there are some roadblocks. Officially, NVIDIA’s RTX Voice only works for RTX enabled GPUs due to its utilization of Tensor Cores. So sorry GTX users, NVIDIA seemingly forgot about us a while back. Although, fear not! The community never gave up and custom versions were made to support non-RTX GPUs such as Pascal, Turing and even lower GPUs. However, you do need CUDA support for this technology to even work.
So, was AMD left behind? Certainly. And as a counteractive step, AMD finally decided to release its version of ‘RTX Voice‘ in its last driver update…just two years later. Is AMD running on valve time? Do bear in mind, this is just the plain ‘Noise Filtering‘ solution offered by AMD so no greenscreen for you.
As is the tradition, AMD does want you to purchase its newer and fancier GPUs namely the RDNA 2 lineup (RX 6000 GPUs). In order to do so, they have kept ‘AMD Noise Surpression‘ limited to RX 6000 GPUs. A rather pleasant surprise isn’t it?
The community didn’t give in and decide to make their own drivers. Custom drivers named “NimeZ” have now been released for Terascale, GCN and Polaris–Vega–Navi GPUs, which include support for this new ‘AMD Noise Surpression’ feature. You just can’t beat the modders can you? Although, as is with everything un-official, this package does come with its own problems.
Reports of buzzing on using AMD NS (Noise Supression) with older GPUs came up. The icing on the cake is the fact that any custom driver is rather unsafe and may just end up voiding your warranty (But how will AMD know?). Fear not, custom drivers act as life support for GPUs currently at the end of their life. Not saying that your pretty RX 5700 XT on which you spend your yearly savings is old, its just marketing from these giants.
Buzz talk aside, this may have been done for particular reasons. The main reason being that RDNA 2 might have some feature on the hardware level allowing for flawless noise surpression (Maybe).
Is it comparable to NVIDIA, you may ask? Its like comparing FSR and DLSS. The latter uses hardware enabled upscaling. I’m not judging, decide for yourself because you may just like FSR better.