You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
[FEATURE] Support for GC570D (Live Gamer Duo)? #10
Is your feature request related to a problem? Please describe.
The AVerMedia Live Gamer Duo (GC570D) does not have Linux support, and I believe it is somewhat of a sister card to the GC573.
Describe the solution you'd like
My hope is for the Live Gamer Duo to be added to this project if feasible. This card has 2 HDMI inputs as well as 1 HDMI output for pass-through for one of the HDMI inputs. Ideally it would have feature parity with the Windows driver.
Hardware / Environment Context
I have a Live Gamer Duo and I can provide some context for the hardware such as lscpi info if needed. As mentioned, the card features 2 HDMI inputs and 1 HDMI output. Both of the inputs send a 1080p60 signal to PC, but the HDMI passthrough for one of them supports 4k60 HDR/1440p144/1080p240. Additionally the passthrough supports 7.1 / 5.1 Ch. LPCM surround sound. As for my environment, I'm running CachyOS, a RTX 4080, Ryzen 9 7900X and 32GB of ram in case any of that matters.
Additional context
I have attempted some reverse engineering of the Windows driver for the Duo myself and can provide support and context to potentially speed up any progress. I however do not have hardly any experience in reverse engineering or coding in C. I do have the Duo so I can actually provide testing support.
Thank you for opening this feature request and using our template! Having dual-input support on Linux would be fantastic for content creators and streamers.
Before we dive into anything, I want to be completely transparent and set realistic expectations:
No Physical Hardware: I do not own a GC570D myself. My work on the GC573 was heavily anchored around having the physical card on my desk for daily testing, probing registers, and validating with real consoles.
Purely Experimental & Theoretical: Without physical hardware access, any investigation here is purely theoretical research and blind testing. Reverse-engineering capture hardware remotely is very challenging, so please don't get your hopes up too high just yet.
That being said, I am happy to take a look at the architecture to determine how much (if anything) the GC570D actually shares with the GC573 / GC555 family (e.g. whether it uses a similar Artix-7 FPGA DMA ring or an entirely different chipset/bridge).
To do an initial sanity check, could you share a few diagnostics from your machine?
PCI Identification:
lspci -nnk -d 1461:
PCI Capabilities & BAR Memory Mapping:
sudo lspci -vvv -d 1461:
Windows Driver Context:
What are the names of the Windows driver files (.sys, .inf)?
In your reverse-engineering attempts, did you spot any references to receiver chips (e.g. IT6805, IT6802, Sil9x, etc.) or FPGA logic?
Once we have the PCI IDs and BAR layout, we can see if our driver even recognizes the board layout or if it's a completely different beast.
Repository owner locked and limited conversation to collaborators on Oct 11, 2026
Activity
Everlite commentedon Oct 11, 2026
Hi @PolygonPrincess,
Thank you for opening this feature request and using our template! Having dual-input support on Linux would be fantastic for content creators and streamers.
Before we dive into anything, I want to be completely transparent and set realistic expectations:
That being said, I am happy to take a look at the architecture to determine how much (if anything) the GC570D actually shares with the GC573 / GC555 family (e.g. whether it uses a similar Artix-7 FPGA DMA ring or an entirely different chipset/bridge).
To do an initial sanity check, could you share a few diagnostics from your machine?
.sys,.inf)?IT6805,IT6802,Sil9x, etc.) or FPGA logic?Once we have the PCI IDs and BAR layout, we can see if our driver even recognizes the board layout or if it's a completely different beast.