Separate names with a comma.
Discussion in 'General Production Forum' started by kroguk, Dec 29, 2016.
Anyway you are all talking crap. Linux OS where it is at. The end.
How can I be wrong about personally not seeing any performance difference between Win 7 and 10?
I know it is not the greatest of tests, but using the Live 9 performance test as a rough barometer, my results on both Win 7 and 10 were 11% CPU usage for the base test and up to 102 tracks before experiencing audio dropouts.
What exact performance improvements did you experience when you upgraded?
Testing by a couple of reputable DAW builders reveal no discernible difference between Win 7 and 10 using pre Redstone build of 10. I tested with Maschine and came with the same by doing a load test with same heavy project until crackles occur. You have to compare on same rig and even then there could be some software or missing tweak on your Win 7 install that is causing difference.
One thing to note is I think Cake is testing with Sonar which always had an issue with a spiking core for many years and why it was excluded from the DAWbench testing. It could be that there is something about Win 10 that improves this but that issue was specific to Sonar.
There have been reports across the board about multithreading getting better .. and there are discussions everywhere about it. If you are looking at overall CPU usage (you quote 11%) you are barking at the wrong tree. You need to look at all the cores and see how flat the overall core usage is. ie do all cores have similar percentages, do they vary wildly or not. You will note core 1 in windows 7 is always dramatically larger for instance the rest of the cores. In win10 it's still higher but at a far lower ratio, it's quite dramatic. The core usage graph has been getting much flatter. Of course if you are running a load of apps that aren't handling cores very well (ie Maschine) that's another problem.
Not just Sonar. It's notable that Sonar has certainly improved core balancing and has recently introduced a new feature for it when you are in overloading scenarios. And yup very much aware of what you say. Sonar def have some good Windows 10 optimisations in there.
i tried windows 8 but i couldn't find my programs so i went back to windows 7.
Should have tried 8.1 and enabled the start menu. I liked 8.1 but it's true that 8.0 was a stinking piece of dog turd in regards to the UI when it came out. However I actually didn't enable the start menu in 8.1 as I quite liked the start screen, I had to spend quite lot of time customizing it though to get it into a decent state (remove MS metro dogturd apps completely via a script, or just delete the icons, then add the apps you use). Once done fine, but there's no doubt the default state was pathetic and most people wouldn't have figured this out or couldn't be arsed to do so (understandable).
Interestingly there is a long running Win 10 thread on the Gearslutz forum. A major contributor to this thread is an MS employee who works on the Windows audio team. The audio improvements to Win 10 have been outlined (improved latency for WDM/WASAPI, new MIDI APIs, Audio Stack and Core optimisation etc.). However, there has been no suggestion that there are any discernible performance improvements for current DAW users using ASIO ( A major interest for Win 10 is MS support for TB3).
Certainly there will be people who were experiencing problems in Win 7 that disappeared when they upgraded to Win 10. Equally there will be people who encountered problems on upgrading. In both cases the problems are likely due to drivers or having different software running (e.g. a different AV).
Also the test I quoted is using an audio application to playback audio so it is recreating the scenario of a typical DAW user; looking at MS performance graphs does not.
Fully aware of that thread, and from memory the guy said something like ASIO still gives the best audio driver experience. So that's why agreed with you earlier on when you first mentioned ASIO here.
Yes, the performance improvements noted are the same that comes up every time. Mostly reports from those who did not properly benchmark one against the other using same DAW projects or just looking at graphs. Some people will get better performance when doing a new upgrade but that same performance boost would have been there with a clean install of the same OS. Other than Sonar, I have not seen any substantiated claims of improvements from Win 7 to now for DAW's. The WASAPI improvements are really not relevant here.
Since Win 7, Windows beats out OSX pound for pound for audio but the benefits of aggregate audio devices are a nice feature of OSX for those that need it. The midi implementation is also better as noted but the benefits there are not necessarily relevant in all use cases. Beyond that, you pay more for less on OSX for audio. Even the stability is now better across most host on Windows. Less OS specific bugs in host and plugins under Windows by comparison. Protools is the exception as usual.
Just went from a MacBook pro to a Windows 10 desktop. Both experiences are good. I only switched because I wanted the power of a desktop and Mac desktops aren't offering much bang for the buck.
On another note, I came across a guide to tweaking Windows for music from Native Instruments in the past. I can seem to find it. anyone have it bookmarked?
Here you go: https://support.native-instruments....1729-Windows-Tuning-Tips-for-Audio-Processing
Since im posting here i might as well give my 2 cents about the OP:
I used windows for about 12 years and Macs for about 7 years, anyone who says either one is 100% better and perfect lives in fairy tail land. Each have advantages and disadvantages, none is perfect or better.
Mac's run overall smother and are a bit more reliable. There are almost no virus and 0 maintenance is required (defrag, scan disk, etc). Its expensive as hell (but it retains a lot of value, you can sell a 4 year old mac for a good price, its a lot harder to sell an old PC). Macs have a lot less software application options (and less cracked apps for those who care...)
PCs have a lot more horse power for the price (mostly), especially desktops, are super customizable and upgradable (again.. desktops). Require more care, maintenance and IMO tend to get a bit slower over time as junk builds up in the OS. PCs allow for more uses, like 3D applications and games, we all know how bad macs are at games.
The perfect solution? Both.
I have a hackintosh desktop. Pretty much a PC that has dual boot with both Windows and OSX, in my personal case i use OS X for anything work/music related and Windows to play some games and for some applications that don't exist for OSX.
NI does a great guide there. That's pretty much all what I do and I've never read the guide until now, I notice looking at task scheduler is missing, disabling some startup apps and services (in detail) as well as deferring windows update and exluding antivirus paths. The only thing I disagree is enabling background services in windows. It sounds logical and people have been doing it for ages, but it's a myth.
Edit to add the MS guide as well:
And if you have windows 10 you should note there are various powershell scripts available to remove all the crappy metro apps you don't need, and a whole load of optimization can be done in the settings areas when it comes to sync'ing, background tasks, privacy. I also disable IP 6 and network scanning.
Thanks @D-One ne and @CakeAlexS !!
I'm interested in knowing why NI supports enabling background services.
It's not a question of NI supporting it, it's just that NI recommends it for better performance in the KB article. However I'm pretty certain the mighty NI are incorrect here with this advice and if you click on the links I posted it should be pretty clear why, I think NI should update their article.
The reason why they recommend it is I suspect that it has always been a default recommendation with DAWs and Windows since forever... and it seems to make sense until you actually look at what this actually does in detail (again see the links).. I certainly did it for a while. It appears however we have all been doing the wrong thing after all these years.
Did a post about Win10 creators update coming tomorrow. I hope everybody is at least backing up tonight...
Not a myth. You should not make such a claim without testing it in a controlled environment as theory is often debunked by real world test. Some of us testers had the same question and one of the developers stated background is better on with Maschine for ASIO performance. We load tested same project on same system with only that as the variable and it proved to be a significant performance benefit with it on background. I tested and verified this on Win 7 but have not on Win 10 though.
I've also found that it can be system-dependent with performance typically being better when set to background (in my experience), but performance can be better for some systems when set to programs.
Good to know. I wonder if the difference is in how the system is tuned? I have not tried on a system that has not been tuned for audio with minimal extras running in the background.
I'd guess it's to do with how its tuned (i.e., different interactions) and perhaps just how the components/drivers work e.g., out-of-the-box budget and office systems and out-of-the-box gamer systems will have different components/drivers with different outlooks on what is important and how they work.
The recommendation is really just to spend ten minutes and try your system on programs and then on background, each with the same heavy project, and see which seems better.