Base Canvas resolution: dual stream - Elgato 4k capture card


New Member
Hey there!

Thanks in advance for any feedback or support!
I have a dual stream set-up Elgato 4k 60 Mk 2 Capture card in the stream PC. Im currently at 2160p resolution at 144 refresh rate on my gaming monitor.
The 4k capture software is taking in 2160p 59 fps.
In OBS I am working currently with a base canvas resolution (resolution of the PC monitor of the stream PC) of 1440p.
I am then downscaling to 1080p to put out to Twitch.

My question is: Should I have a base canvas of the original source from the 4k capture software (2160p) or keep the canvas at 1440 (stream PC base resolution).
Im not sure this even matters at this point after some testing, but wondered what the standard was if there was one.

Thanks in advance!


Active Member
If there's a standard, I haven't heard of it yet. Maybe someone else knows.

I've set up several rigs, and I've always had the canvas and output size the same. Any source that doesn't match gets resized as a source to fit the canvas.

I can also see an argument to minimize resizing, as a proxy for CPU load. That is, to match the canvas to as many sources as possible, and only resize once at the end. Of course, that involves guessing how much CPU it takes to resize vs. how much it takes to run your transitions and compositing at various resolutions, if it even matters on your rig.

If you really want to minimize resizing (if it even matters), you can set the game and its monitor to your final output size...

I still haven't seen the point of anything beyond 1080p60 for anything. That's already more than most people can see. (if you CAN see more, you might be too close - radiation from the TV was always a myth, but there's still something to be said about a screen filling your view) Most of my stuff is 1080p30, including cameras and final product. (screens are 1080p60)

Are more people than I realized producing for an IMAX dome or something (or gaming in one), where the primary focus is in the bottom-center of the frame, and so that part by itself is 1920x1080, but you have to have the same density everywhere else too?

If you're recording an intermediate file to crop, zoom, and pan later, I guess I can see that, but it also leads to a question of doing those things live, without compression artifacts or a ginormous file, instead of in post where you have to have one or the other.