Hi everyone,
I was tasked by a client to design podcast studio capable of 4k and green screen.
I will be using 4 pieces of Canon CR-N700 PTZ cameras.
At this point full budget for professional solution isn't available so for a interim purposes we would like to use OBS.
I have approximately $ 10000 US for the server and the SDI capture card and I'd like to pick your brains regarding OBS requirements strengths and weaknesses.
I already noticed there is a limitation with CUDA acceleration I for some reason cannot assign more then 3 sources to it.
Does it mean I need to have 2 separate cards?
When I tried to run preview when the cameras were set to 4k/60Hz I couldn't get a stable preview, I'm not sure if I have something wrong with my hardware configuration.
Some basic questions first
1) for this project CPU or GPU FFMPEG encoding / decoding?
2) if GPU, can I use two cards?
3) if I can use two cards which is the minimum required for this task (real-time 4k-60Hz) I was considering two A5500
I haven't really played much with background removal in OBS, is this CPU intensive task or is it accelerated?
4) ideal scenario would be to record individual sources while the live stream is going out to allow for post processing at full resolution
5) audio integration
From what I understand the USB device has some bug, anyone used Dante? It would be easiest solution for me to have DVS on the server and assign individual microphones to respective video streams.
I will have 3 people with miniature microphones being recorded by 4 PTZ cameras and I plan to use one beam tracking microphone as a backup solution where I can split it to 3 separate Dante channels based on the direction the voice is picked up from. Can this be achieved? Each source would have 3 Dante channels assigned, body microphone, beam tracking microphone and mix. I want to be able to record it ideally as a video with 3 separate audio tracks for each of the sources and output with mix track.
I was tasked by a client to design podcast studio capable of 4k and green screen.
I will be using 4 pieces of Canon CR-N700 PTZ cameras.
At this point full budget for professional solution isn't available so for a interim purposes we would like to use OBS.
I have approximately $ 10000 US for the server and the SDI capture card and I'd like to pick your brains regarding OBS requirements strengths and weaknesses.
I already noticed there is a limitation with CUDA acceleration I for some reason cannot assign more then 3 sources to it.
Does it mean I need to have 2 separate cards?
When I tried to run preview when the cameras were set to 4k/60Hz I couldn't get a stable preview, I'm not sure if I have something wrong with my hardware configuration.
Some basic questions first
1) for this project CPU or GPU FFMPEG encoding / decoding?
2) if GPU, can I use two cards?
3) if I can use two cards which is the minimum required for this task (real-time 4k-60Hz) I was considering two A5500
I haven't really played much with background removal in OBS, is this CPU intensive task or is it accelerated?
4) ideal scenario would be to record individual sources while the live stream is going out to allow for post processing at full resolution
5) audio integration
From what I understand the USB device has some bug, anyone used Dante? It would be easiest solution for me to have DVS on the server and assign individual microphones to respective video streams.
I will have 3 people with miniature microphones being recorded by 4 PTZ cameras and I plan to use one beam tracking microphone as a backup solution where I can split it to 3 separate Dante channels based on the direction the voice is picked up from. Can this be achieved? Each source would have 3 Dante channels assigned, body microphone, beam tracking microphone and mix. I want to be able to record it ideally as a video with 3 separate audio tracks for each of the sources and output with mix track.
