MADRIX Forum • NDI bug
Page 1 of 1

NDI bug

Posted: Wed Apr 20, 2022 11:41 am
by CARLOSLUCES512
Hi there we are immersive in a huge new install ( around 250 universes) We have all the system ok just We observed a problem with the NDI ,we are receiving this signal from another device (mac) with MILLUMIN software .
When we play effect from Madrix it is ok ( CPU AT 80/100 ) but when we uses SCE capture and play de NDI the CPU to put at 100/ 100 and the GPU don ´t have any process or just a bit.
We are using a 11th GEN INTEL (R) CORE (TM) I9 11900F @2500 GHZ ,NVIDIA GEOFORCE RTX 3080 TI AND 3G RAM AT 2133 MHZ.
THANK YOU SO MUCH IN ADVANCE.

Re: NDI bug

Posted: Thu Apr 21, 2022 1:43 pm
by Schulze
Hi CARLOSLUCES512,

Which resolution does the NDI stream have and which Matrix Size do you use in MADRIX 5? The down or upscaling of the NDI stream requires additional performance. If your CPU is already utilized higher than 80%, this could of course lead to a utilization of 100%.

THe GPU is used to display the Previews in MADRIX 5, the rendering of the effects is done on the CPU.

Re: NDI bug

Posted: Fri Apr 22, 2022 7:37 am
by CARLOSLUCES512
Hi We have a resolution in Madrix of 2560 x 1080 x 1 and the signal incoming of NDI is the same...So what can we do? thank so much for your help.

Re: NDI bug

Posted: Fri Apr 22, 2022 1:49 pm
by Schulze
Hi CARLOSLUCES512,

The huge resolution is the main issue. Given that your CPU is already nearly completely utilized, it seems to be not sufficient to handle this huge matrix size. NDI is using a compression in order to transmit the content, therefore a decompression on the receiver side is also needed. This of course also requires CPU performance and it also explains the additional CPU utilization. I also need to note that 2560 x 1080 x 1 = 2.764.800 pixel is way above a Full HD resolution, which would be 1920 x 1080 = 2.073.600 pixel, this of course requires a high performance CPU in order to render the effects.

Re: NDI bug

Posted: Sat Apr 23, 2022 10:49 am
by CARLOSLUCES512
Thank so much we are trying to reduced the size of our matrix.