Video Device Library support for NVIDIA Rivermax

What would it take to add the NVIDIA Rivermax support in the Video Device Library?

Rivermax is a fully-virtualized streaming solution that complies with the stringent timing and traffic flow requirement of the SMPTE ST 2110-21 specification. Rivermax enables the future of cloud-based software defined broadcast.

Rivermax runs on Nvidia Mellanox ConnectX-5 or higher network adapters, enabling the use of common off-the-shelf (COTS) servers for HD to Ultra HD flows.

Single CPU core receives or sends 8 UHD 2160p60 or 32 FHD 1080p60 streams

2 Likes

Hello, I would also like to ask about support for SMPTE 2110 (using Rivermax). Are there any plans to add this in the future?
SMPTE 2110 looks like a great standard and Rivermax is very interesting in terms of performance.

2 Likes

+1 for this, compression of the NDI signal with alpha in high resolutions is really noticeable on LED screens (TD to d3). Disguise are also planning to implement this protocol as input so it would be really useful combination. Renderstream is nice but not always needed for simple things.…

Bump again. device input output on this library over normal 10G(mellanox?) chipsets for 2110 would be super useful.

Bump. 2110 is only becoming more prominent, support in TD would be killer.

Yeah, this is going to be hugely helpful for uncompressed data streams over networks

+1 and bump, Rivermax will solve a lot of issues related to “traditional” signal routing, surclassing NDI compression and artifacts. (European Song Contest 2024 relied on ST2110 for video management and routing with Disguise -former d3- servers…)
So please take a look on it!

1 Like

+1 bump. I did a thorough hobby research on that topic and the hardware accessibility of low latency, multi stream, SFP and Ethernet network cards to run with TD and nvidia rivermax would tick all the bells.
Native integration would be awesome!
A GeForce nic with an affordable gaming rtx card can be a quite powerful multi application streaming/broadcasting/ai platform, which is still affordable!
I don’t know if nvidia GPU direct is still a thing, but that would be also an awesome feature within TD too or can shared memory also be used between pcie lanes?
Thank you very much in advance!

I guess something like this, but for Touch-Designer would be helpful, if that’s an option:

+1 for sure.