In general:
- the higher the resolution, the larger the amount of data that has to be processed. Double the resolution means 4 times the data.
- the higher the framerate, the larger amount of data has to be processed. Double the resolution means double the data.
This tells about the raw picture data and the relation between resolution, fps and amount of data.
With compression (encoding), this raw data is converted to the actual data you store or stream. The relation between resolution, framerate and amount of data is roughly the same as with the raw data. Exception/rule of thumb: the larger the amount of data, the more efficient the encoding, so it's not exactly 4 times / 2 times any more but a bit less.
The higher the compression, the smaller the encoded data.
The lower the original image detail and complexity of the game, the smaller the encoded data (fast shooters often have less image complexity, thus compress better than highly complex sophisticated graphics of some RPGs)
The smaller the encoded data, the better the image quality (because you can put more information into the available space)
With recording, the amount of data is mostly unimportant, because usually you have unlimited disk space.
With streaming, the amount of data is crucial, because you have limited bandwidth.
So you need to differ between recording and streaming with your settings.
With recording, you want highest quality and don't care about disk space. So you are looking for settings that produce encoded material that looks indistinguishable from the original game, and still doesn't waste disk space.
With streaming, you are constrained by available bitrate. Not by your upload bitrate alone, also by the bitrate your viewers are able to download.
In the end, what settings do you need?
For recording, you use a quality based rate control with your encoder such as CQP (Nvenc), CRF (x264) or ICQ (Quicksync). The crucial parameter is the quality parameter, that you vary between 18-25. Lower values mean higher quality, so this parameter controls how much detail is removed. The quality of the resulting video is guaranteed and doesn't care about high or low motion scenes. This is what you need for archiving or postprocessing.
For recording with the purpose of archiving or postprocessing, you never use a bitrate orientated rate control (CBR, VBR), because these modes lower quality for high motion scenes to fulfill the bandwidth constraint, and they bloat the file for low motion scenes for no reason, because they hold up consuming the given bandwidth, even if absolutely nothing is moving on the screen and you only record a still image.
For streaming, some rules of thumb apply. There are recommendations from streaming providers for bandwidth:
https://stream.twitch.tv/encoding/
https://support.google.com/youtube/answer/2853702
tl;dr:
If you want to record with OBS, use simple mode, Nvenc as encoder if you have a Nvidia card, and the quality setting of "Indistinguishable Quality". This is the correct setting regardless of resolution and fps. Internally, this uses the CQP rate control with a customized appropriate CQ value according to resolution.
If you want to stream with OBS, see above streaming provider recommendations.
It may be necessary to reduce resolution or game fps, if the computer isn't able to keep up with processing the data, resulting in laggy or choppy video.
