As mentioned, you can just upgrade 10 to 11. It works very well. We’ve done it several hundred times at work, I did it on my home system which is what I run my DAW on.
As for tweaking, I also agree that you shouldn’t really worry about it or spend time going overboard. The guide linked is a good one, but one to follow in order and only move on to the “consider” things if you are having issues. Don’t go nuts with optimizations if you don’t need to. You can make things worse.
The only optimizations I personally do are to disable all onboard hardware I don’t need. The WiFi, onboard audio, that kind of stuff., and to set it to max performance when using the DAW. I also keep my system and firmware up to date, but that really isn’t optimization that is just good computer maintenance.
You’ll find that as CPUs have gotten more and more powerful, there’s just less and less need to try and go crazy on optimizing things. Sure, in the Windows 98 days when everything was slow and multi-tasking was bad, you wanted to shut down EVERYTHING you could. Now? No, no need and most of those background tasks are doing something for your system.
The biggest two “optimizations” I’d recommend (which are on the list) really aren’t. They’d be:
-
Have a good audio interface. It makes SO much difference. You will solve way more issues with a good interface than any amount of messing with the registry. Driver and hardware quality matter. RME are pretty much the kings, but there are plenty of other good choices.
-
Turn up the ASIO buffer unless you have a real good reason not to. As he mentions: 1ms is about 1 ft in the air. So is the difference between 64 samples (1.33ms at 48kHz) and 256 samples (5.33ms at 48kHz) REALLY noticeable to what you are doing? If it is, could you instead just move the speakers closer, or wear headphones, and call it good? Ultra low latency is the thing that starts to be a big stress, so long as you can tolerate turning up the buffer a bit you often find there are no problems. My new system works fine at 64 samples, but the realtime gauge is pretty spikey and I wouldn’t want to try it with a big amount of plugins. 128 samples runs just fine though, very low and stable realtime load and it only gets lower with higher latencies.
Most of the time I see someone having a bad time and wanting to “optimize” it is either someone with a system that’s just too old for what they are doing, or even more likely, someone with a mediocre audio interface convinced they need to cut the ASIO buffer to the minimum. Even when you are playing live, you generally can tolerate some latency. I mean when you play electric guitar, how close do you stand to the amp? Any distance adds latency, yet it really isn’t an issue to be 10-15 feet away. However sometimes people over-focus on cutting down latency and it causes them headaches that just aren’t necessary.
Also also something that to note is that the actual latency you get can vary quite a bit at a given buffer size depending on the card, because that is only one of the buffers they have. Some implement more or less safety buffering and that can change things. So you can find a card that runs fine at “32 samples” because while the ASIO buffer is that low, it has a big 'ole buffer that you don’t see. It can vary even between OSes. Like RME sticks an additional safety buffer on their stuff of 32 samples on Windows, but 24 samples on MacOS. Sounds like MacOS is lower latency, but no, because on MacOS there’s also an additional Core Audio Safety Offset of 24 samples so 48 samples in total (you can shorten it to 12 samples, so 36 total, but at risk of clicks).