David_Harry
Second Officer
United Kingdom
Offline
|
That’s actually a very unrealistic user case scenario for testing any encoder.
The problem with what you consider to be a ‘typical user case’ is that it’s not the same for the next person and it’s not testing the encoder but the CPU and GPU.
For instance, someone loads up a crap load of noise reduction, that’s just gonna hammer the CPU and GPU and bring the whole system to a halt. This will then mean the encoder itself is waiting for the buffer and doesn’t show the speed the of buffer, rather the strain on the CPU and GPU.
Next person doesn’t use noise reduction but loads of colouring and masking, again same problem but this time the timeline and main video buffer are acting different to user case one but once again doesn’t show the encoder speed but rather the buffer fill rate.
This is why when people ask for ‘real world’ examples, it’s only real world to whoever is doing the exact same thing and not real world for, in this instance, the encoder.
Don’t forget, the encoder here on M1 Max and Pro is separate tech to the CPU and GPU and does have a measurable performance level, which can be measured when it’s not waiting for the video/frame buffer to fill up and feed it.
|
|