With a fixed key frame interval, key frames cannot be distributed as dynamically as required, so either quality suffers or disk space increases. For example, key frames are often reasonable on scene changes when image changes completely. For recording, no transmission errors can happen, so the encoder should be allowed choose on its own when to put in keyframes. For streaming, a forced interval of 2 is required, so that if packets get lost during transmission, you don't have to wait for more than 2 seconds to get the next full image.

Giving the encoder the freedom to choose its own keyframe interval is optimizing the output. This will not increase the system load, only the file size of your recording (and the quality, of course). If the current quality isn't good enough for you, switch to 'Indistinguishable Quality, large file size'. This is a variable bitrate constant quality rate control that will produce videos with as much bitrate as necessary to ensure the requested quality. These say 'high quality, medium file size', which is internally the CQP rate control.

If you start recording, on its own or at the same time as streaming, OBS creates a second encoder and encodes the recording with the settings from the 'recording' part of your screenshot. With a RTX 2080, this is producing very good stream quality - only to increase if you are able to increase the bitrate, which is dependent on your internet connection and on your viewers. it's streaming with 6000 bitrate, which is internally the CBR rate control. If you start streaming, OBS uses the settings from the 'streaming' part of your screenshot, i. Actually, your settings are the optimized settings.

Best stream settings for streamlabs