I was looking through the source code and couldn't find the answer to this question. As far as I am aware: the scene is grabbed, composited and scaled via GPU, ran through the encoder and rescaled before being streamed or saved to disk. I am curious about this last scaling step as it seems to be using my GPU despite others saying that the encoder scaling is done on the CPU and to avoid it.
I am using NVENC on my GPU. Does the encoder (NVENC in my case) perform the Rescale Output stage itself or does it simply encode, pass back to OBS then OBS rescales it on the CPU with no hardware acceleration? It seems as though the GPU is doing extra work when Rescale Output is enabled thus I believe that the encoder performs the scaling at the same time itself on the GPU, but want to clarify from someone.
I understand that Output Resolution and Rescale Output are two different stages and this question does not concern the Output Resolution. I am trying to optimize my recording setup to reduce stutter.
Thanks!
I am using NVENC on my GPU. Does the encoder (NVENC in my case) perform the Rescale Output stage itself or does it simply encode, pass back to OBS then OBS rescales it on the CPU with no hardware acceleration? It seems as though the GPU is doing extra work when Rescale Output is enabled thus I believe that the encoder performs the scaling at the same time itself on the GPU, but want to clarify from someone.
I understand that Output Resolution and Rescale Output are two different stages and this question does not concern the Output Resolution. I am trying to optimize my recording setup to reduce stutter.
Thanks!