Search results
Been searching and scouring the web for hours trying to find the best settings to utilize the M1 pro's amazing processor/GPU combo and it was a mixed back. I figured out what seems to be the best settings to utilize the optimized threads in the M1 to record as well as stream in 4k.
The main gpu has an easier time pushing your frames directly into it's encoder (the encoder is an ASIC device, and does not hurt gaming performance, maybe 1%) than having to send it across pcie into the frame buffer of the second gpu.
Oct 26, 2021 · Nvenc has the nice property of being able to directly encode from the frame buffer of the GPU, so on 1 GPU the raw picture data is not required to be copied even once through the pci-express bus, so 1 Nvidia GPU card with Nvenc (new) as encoder is the most efficient solution.
People have been seeing CPU usage of 200 and 300% when using OBS and I found a solution on the OBS open source page on github that members made to allow OBS to use hardware encoding which shifts the work to the GPU.
Jul 11, 2023 · Apple's M1 Macs, which integrate the CPU and GPU into one chip, perform best with the x264 software encoder. For 1080p streaming, we suggest a 'fast' CPU usage preset and a bitrate of 3500 to 4500 kbps.
Aug 27, 2021 · I'm trying to determine how effective an M1 Mac would be for transcoding video using FFmpeg (specifically resizing and adjusting bitrate). I can do single transcodes using a command like this: ffmpeg -I in.mp4 -nostdin -c:v h264_videotoolbox -c:a copy -vf scale=1280:720 -b:v 8000k out.mp4.
People also ask
Does a GPU encode a game if it's dedicated to encoding?
Should a 2nd GPU encode a second monitor?
Can I use a second encoding GPU?
Is Nvidia a good encoder for a PCI-Express card?
Can I run OBs on a different GPU?
Is Nvidia hardware encoding better than software encoder?
Feb 7, 2011 · For consistent scaling, specifically with systems with many CPU cores, you should divide the work by frames, because each portion of the encoding pipeline takes differing times to complete.
