Resolume Forum

Hello All,

I'm capturing live camera (SDI, 2160p50, UHD not DCI). Some of the camera is masked using a luma mask on the layer above, with that layer set to mask mode (mask all below). The layers with the camera inputs and the mask layer are grouped together to avoid masking content on layers bellow. There are times when I use the camera inputs straight (no clip active in the mask layer), and times when I use the camera inputs masked (luma mask clip active in the mask layer). I have noticed significantly more delay (60-100ms extra) in the live camera feeds when I use a mask on them. I have used Arena for years and always been happy with the amount of delay, only in this specific setup have I ever experienced delay like this.

If anyone could shed any light on this that would be much appreciated. I'm not sure if this is a bug, or if this is expected behaviour.

System specs as follows. Threadripper Pro 5965WX, RTX4090, 128gb 3200mhz RAM, Decklink 8K Pro, 2 x 2tb PCIe 4.0 NVME SSD, Asus Pro WS WRX80-SAGE SE WIFI.

Kind Regards, Tom

Do you see the FPS drop too when masking is used, or only added latency?

Zoltán wrote: Thu Apr 20, 2023 17:01 Do you see the FPS drop too when masking is used, or only added latency?

Only added latency, the FPS remains solid at 50.

How are you measuring the latency compared to the capture playing without masking / grouping?

One instance of the input with masking, and one instance without masking, displayed side by side in the composition and output on the same feed to the same display, with both being in the same slice.

Admittedly I haven't accurately measured them, just going by eye and experience. When I get access to the servers tomorrow I will shoot the output with a slow motion camera to get a more accurate result. Shooting at 240fps should give us the the answer to the nearest 4.166ms.

Could you attach a few screenshots of how you have the clips set up?

Playing a capture source on layer1, and having the same on layer 2 in a group masked by a white solid color, group set to 50Difference blend I get no difference signal, so both the masked group and the straight capture clip are the same.
Luma mask sounds like you might be using an external effect?

Thank you for all your help. However there is no problem. I made a mistake based on bad information. Nobody should waste any more time investigating this.

So, what happened there, making your cameras delayed?

The source used with the mask was supposed to be an identical copy of the source used without a mask, and it was visually. However it was also routed through additional hardware upstream of the capture cards, the non masked source was not routed through this additional hardware.