Note: This article is recent as of Windows 10 1909.

 

Table of Contents

Overview

If you’re just starting out with audio production on a Windows PC or are moving to Windows from planet Mac, you’re in for a world of… awesomeness!  Windows is an amazing operating system, especially within the realm of audio production. It offers many benefits over similar solutions from Apple with their Mac platform. It’s sad that the perception of contemporary audio production usually begins with “what Mac should I use?” because in Windows you’ll find a much better experience at a fraction of the price.  And personally, I find it just much more enjoyable on a general level.  The PC becomes less like a blackbox as Macs are and more of an open platform for you to alter and upgrade and build out to your heart’s content.

However, where as Mac is more or less a matter of installing your DAW and audio interface drivers and away you go, Windows users should take some time to configure their PC to be better optimized for music production.  This isn’t a strike against Windows so much as it displays the versatility of Windows and its history of supporting a larger array of form factors and use cases than Mac, which largely only exists on laptops and automobile-priced workstations.

This article goes over the steps I take after acquiring new machines or installing a fresh Windows installation.  I know a lot of people make double use of their laptops as their primary recording and sequencing device as well as more common productivity machines for email and web browsing, and honestly either usage is perfectly valid.  There are some considerations you want to make though because audio production involves near-real time audio so performance and optimization matter much more than a computer primarily used for simply Microsoft Office or Adobe CC.

To UAC Or Not To UAC

So you’re installing a program and a pop up from Windows blackens everything out and asks you “Do you want to allow this app to make changes to your PC?”.  Since the release of Windows Vista, Microsoft has included this protection system in Windows to let users and operators know when a program is attempting to install and make changes to the present installation.  The problem this solves is from the Windows XP-era of computing where so-called “drive by installs” were common where people would visit a website and through some obscure vulnerability in the browser and the OS, the website managed to secretly install software of all sorts without the user knowing, from backdoors, to adware, to malware and more.

With an audio workstation, problems can arise when some applications cannot write to a directory they feel they should.  This is very rare, however, and it’s completely possible to leave UAC settings as they are setup during installation.  If you are using your Windows laptop or desktop for both audio production *and* web browsing, email, gaming, etc, I would leave it on.  Personally, I don’t do anything on my audio production PC that’s not audio production so I turn mine off.   Why so cocky?   One, I don’t use pirated software, ever.   If you used pirated software, I would never log into your bank account or any other critical web service, because your PC has been compromised beyond repair with any numbers of backdoors installed.   Two, I keep all Windows Defender and firewall services active, offering protection in case any other PC on my network has been compromised.  Since my machine is entirely dedicated to music production, I don’t really have a need to be prompted to install the software I need installed.

If you are using your laptop or PC for audio production and general purpose use, my suggestion would be to leave UAC on and use the Microsoft Store as much as possible for any applications you want installed.  While it’s true that the Windows Store isn’t as populated as mobile app stores, Microsoft does provide a secure experience with a bunch of bells and whistles, some even included on a Windows 10 installation by default (such as Mail, Calendar, Netflix, and Spotify).  Additionally, you can find cool tools like Microsoft’s new Windows Terminal, Ubuntu images for WSL (Windows Subsystem for Linux, allowing you to run Linux in a local container with little to no performance penalty), Notepads (a kick ass Notepad replacement), OneNote (I personally use this heavily to keep track of projects), LastPass (for secure password management), and more.

UWP

So I brought up Windows Store, what makes it so secure?  In Windows 10, Microsoft has a type of application platform available that’s very secure called UWP (Universal Windows Platform).  UWP apps are often (though not always) installed from the Windows Store.  The benefits are both for security and housekeeping.  For security, these apps largely run inside sandboxes keeping UWP apps from willy nilly accessing and altering the filesystem Windows is running on.  For housekeeping, these apps make minimal or no entries to the system registry (which doing a lot can lead to negatively impacting performance).  I highly recommend for all non-music production apps to seek out UWP apps to use.

My Favorite UWP Apps

Here’s a list of my favorite UWP apps and other applications from the Windows Store:

  • Windows Terminal – A great command line tool that’s easier to customize and has solid copy and paste support
  • Ubuntu – For my website, I have a beta installation loaded in a local VM that let’s me test features and changes first.  I also keep my SSH keys in the linux VM so I can log into my cloud provider where my website is located.
  • Notepads – OMG, I wish Microsoft would by this guy out and replace Notepad with this app.  It’s so clean and easy to use.  Imagine having a modern looking notepad with tabbed window support!
  • Windows Central – I like to keep up with PC and Windows news, this app is great and includes support for comments.
  • Adobe Photoshop Elements 2020 – If you are also doing graphics work for album art or other types of promotional material, Photoshop Elements is really amazing.
  • Grover Pro – This is the best podcast player app in existence, of any platform.  Buy it now, thank me later.  Check out the Sonic TALK podcast from Sonicstate.com and the windowscentral.com podcast by searching for them in Grover while you’re at it.
  • LastPass – The best secure password manager out there, hands down.  There are also plugins available for web browsers such as Edge, Chrome, and Firefox.

 

Microsoft Store Apps For Music Production

Are you looking for DAWs or plugins for music production in the Microsoft Store and coming up short?  Yes, this is normal.  Your audio production tools and drivers aren’t going to be found in the Microsoft Store largely because the companies that develop these products already have software delivery channels of their own, so there’s currently little incentive to place things like Ableton Live or Access Virus Control Center or Avid Pro Tools in the store.  Additionally, some of the security requirements of applications distributed in the Microsoft Store are not compatible with audio production software as they tend to rely quite heavily on inter-software connectivity and software working together.  So, try to use the Microsoft Store for as much as you can with productivity apps and everything else installed is ideally only your audio production tools.  UWP is great, but we’re not quite at a point where audio production machines are going to make much use of it.  That doesn’t mean you can’t take advantage of what UWP offers now though.

Set Up Power Profiles (Do This If You Ignore Everything Else)

Power plans are profiles used by Windows to determine how much power it extends to hardware for varying degrees of performance.  The trade here is performance for power.  If you aren’t pushing your machine hard, it won’t expend as much power.  This is desirable if you’re wanting to use your laptop to browse the web in an airport and want your machine to last as long as possible.  However, with audio production this robs the machine of its true performance.  It is highly suggested to only do audio production when you’re powered in and doing no other tasks.  As such, it is wise to maintain at least two different power profiles based on your usage.  Windows 10 installations come preinstalled with three power profiles; Power saver, Balanced, and High performance.  While all three of these are perfectly fine for general desktop use, none of them are suitable for audio production.  But that’s fine, this is Windows and we can configure our machine to work like how we want!

To start, search for Control Panel by clicking Start, typing in Control Panel and clicking on its icon when it appears.

Next, we want to click on “Hardware and Sound” and then click on “Power Options”.  Next you want to click on “Create a power plan” on the left hand menu.   Doing so will take you to the following screen:

Here we’re going to choose “High Performance” and name your plan name something you will associate with audio production.  What we are doing is starting with an existing power plan as a template and naming it something else.  We will further modify this cloned power plan in the next step.  I have called my power plan “Island Dust Studio”.

Next you’ll want to click on the “Change plan settings” link next to your new plan.  In the next screen, click on “Change advanced power settings”.

So now we have a window labelled “Power Options”, and we have our newly create power profile loaded with all of its options listed below.

In this Power Options setting, assuming we have cloned the High Performance power profile that shipped with Windows, we want to make the following changes:

Power Profile: Hard Disk

Hard Disk > Turn off hard disk after > Setting: Never

If you are like me, you probably have a rather large sample library you’ve collected and compiled over time, not to mention the huge sample libraries you may house on additional drives for products such as Native Instruments Komplete and UVI Falcon.  In our music production power profile, we want to ensure that Windows isn’t periodically turning off disks to them or the OS disk after any length of time in a here-unwanted attempt to save power.

Power Profile: Sleep

Sleep > Sleep after > Setting: Never

Sleep > Hibernate after > Setting: Never

We don’t want our music production PC to go to sleep, because it should be plugged into a power source.  These settings will prevent Windows from automatically putting the computer in sleep mode.  Remember that if you use your PC for audio production and anything else (like gaming, work, productivity, binge watching Stargate SG1 like a boss, etc), you want to flip your power profile from your music production power profile back to one of the other normal ones afterwards.  If you use your PC exclusively for music production, feel free to leave the music production profile active all the time and reap the performance rewards.

Power Profile: USB settings

USB settings > USB selective suspend setting > Setting: Disabled

This is a very important option to toggle for music production.  I can’t think of a current workflow in music production that doesn’t utilize USB.  From audio interfaces, to MIDI controllers, to MIDI interfaces, etc, ensuring that Windows isn’t attempting to suspend USB ports intermittently in an attempt to save power is highly important to any producer on the Windows platform.  Setting this to Disabled makes sure that all USB ports are powered at all times.   If this isn’t set, you’ll have all sorts of glitches that most people contribute directly to Windows as opposed to it being a feature of Windows that is 100% togglable.

Power Profile: PCI Express

PCI Express > Link State Power Management > Setting: Off

This feature is much like the USB settings feature discussed above, however it pertains to your PCI Express slots.  And I believe (correct me if I’m wrong), your Thunderbolt devices (seeing as how Thunderbolt devices are simply PCI Express devices exposed through the Thunderbolt bus).  If you have an Avid HDX or HD Native PCIe card, an RME sound card, a Universal Audio UAD-2 PCIe card, or any other device utilizing PCIe in your audio production workflow, it’s imperative that you halt Windows from attempting to manage its power draw.

Power Profile: Processor power management

Processor power management > Minimum processor state > Setting: 100%

Processor power management > Maximum processor state > Setting: 100%

By now, I hope you’re seeing the trend in our changes to our custom audio production power profile:  keeping Windows from shorting power and reducing performance.  In actuality, this wouldn’t be such an interruption for audio production except for its intermittent effects presenting themselves as random quirks.  Much like the other settings, here we’re stopping Windows from trying to run the processor at a lower speed based on load. If we allowed a difference in these two settings for minimum and maximum processor states, the processor would vary in speed throughout our working in a DAW leading to all sorts of weird problems like threads that were fine all of the sudden overloading a core.   This setting controls the software portion of this feature, though do note that at the hardware level, your processor is going to be changing its speed as well, however audio applications tend to handle BIOS defined processor state features without problems, where as these Windows settings can cause issues if left enabled.

Flipping Between Power Profiles

So this part is relatively easy.  When you’re using your computer for web browsing, email, Netflix, Spotify and Skype, just remember to flip back to one of the original profiles.  And before you fire up your DAW, flip to this custom Power Profile that we just created and modified.  You can do this easily by going to Control Panel with the instructions before (click Start > type in “Control Panel” > click on “Hardware and Sound” > click on “Power Options”), and toggling between your power profile.  Laptop users can simply click on the Power system tray icon and choosing your plan from there.  The Power system tray option is not available on devices without battery power sources (such as desktop computers).

Screen Resolutions

Microsoft has supported a feature called HiDPI mode for some time.  HiDPI mode is using a much larger resolution needed for 1:1 pixel ratio for a display size.   What the heck does that mean?  In practice, think of a common 22″ 1920×1080 (1080p) monitor, and picture four times as many pixel in that display making the video feed generated by Windows for your desktop to appear much more crisp and clear, but with the size of desktop elements such as titlebars, widgets, and scrollbars appearing roughly the same size.   You’ve most certainly seen this effect used to some degree:  most modern smart phones use the feature extensively, Apple has used it in macOS since 2012 with the release of their Retina screens for Macbook Pros, and Windows has had some form of support for nearly just as long.   Unfortunately, one of Windows’ greatest strengths (its support of older software to continue to run without issues on modern Windows releases) can stop you from seeing applications take advantage of HiDPI mode because in many cases the technology wasn’t around or widespread when the developer of the software wrote it and they likely have not added support for it in any satisfactory manner since then.

In music production, this problem presents itself in a number of different ways.   In HiDPI mode, your DAW might appear fuzzy as Windows essentially takes the original size of the application canvas and stretches it out (leading to a pixelated view).  Plugins might also appear impossibly tiny even if the host DAW supports HiDPI mode.  In most all cases, the increased screen resolution requiring four times as much bandwidth for the video signal puts unnecessary strain on the video card, which can affect real time audio applications.  In the UWP applications I mentioned before, you’ll notice that virtually none of them suffer from these issues because they were all built using technology specifically taking advantage of some of the newer features in Windows.

If there were a way to fix this problem and maintain HiDPI mode, I would do it in a heart beat because when it works it looks beautiful.  Unfortunately, that’s not the case.  It’s not really Microsoft’s fault, because they have already provided resolutions for these bugs for both the newer UWP apps as well as the older legacy win32 apps.  And it’s completely understandable that many DAWs don’t support this mode well yet or at all because developers of DAWs generally have to build their entire UI widget and form stack from scratch as a DAW will contain many options that simply can’t be represented by standard UI toolkit widgets.  Rethinking  and retooling their entire UI would impact more important updates to their software for performance and workflow improvements so they’re likely not heavily prioritized.  Cubase, for instance, just recently received HiDPI support and it’s far from perfect.

I recommend to work in a 1:1 ratio mode, at least when you’re working on music.  For all the reasons listed above, you’ll put a lighter strain on your computer resources that you can then devote to audio production work, and you won’t have any issues in the middle of a project requiring you to close everything, fiddle around with application properties toggles for HiDPI settings and troubleshoot.   1:1 mode essentially means running your desktop in a manner where 1 pixel represents 1 pixel, where as in HiDPI mode 1 pixel might be represented by any number of pixels depending on the scaling factor.

Practical Example #1 (my Alienware Aurora R8 with dual 4K displays)

For my main tower, I have two 28″ Samsung displays with a native resolution of 3840×2160 (4k).  However, I run at a resolution of 2560×1440 with the setting “Change the size of text, apps, and other items” set to 100% (1:1).  I have attempted to use the native 3840×2160 resolution before with scaling set to 200%, but ran into many issues, specifically with plugins that don’t yet take advantage of any resolution scaling technologies.  Functionally if you were to look at my monitors with either setting, everything would appear to be the same size, but application developers just haven’t caught up to taking full advantage of Windows 10’s HiDPI technologies yet, so I go with the settings that work. Because my scaling settings are by an easily divisible factor (100% to 200%), running at this non-native resolution actually looks perfectly fine because on my displays I essentially have a reverse HiDPI effect happening where 1 pixel is being stretched out evenly to 4 pixels on the monitor.  This is suitable on desktop displays (and I got these dirt cheap on Woot.com), but it wouldn’t look as pretty if you’re attempting to use a non-native resolution on a 4K TV as their pixels displayed don’t always correlate to physical pixels on the TV.

Practical Example #2 (my Microsoft Surface Book with 3000×2000 display)

I have found edge cases where HiDPI mode has no issues.  In my workflow, I have a sampling station set up in one area of my studio where I have my turn tables and tape deck set up to go into my Surface Book where I record and further process the audio before storing it on my network drive that I later access from my Alienware tower.  I use Ableton Live 10 Suite for this task, probably one of the best adapted DAWs to utilize HiDPI settings.  On this machine, I use no plugins except for the native ones inside Ableton Live 10 Suite that are fully supported by its HiDPI implementation.  So on this machine, I don’t go into 1:1 mode.

Experiment, But Be Mindful Of The Issues

For your own setup, I’d say experiment.  If you find yourself with a HiDPI display but all of your plugins and your DAW works fine in HiDPI mode, then feel free to ignore me and carry on.  If you have intermittent issues with the odd plugin showing issues or your DAW looks fuzzy when ran, then perhaps heed my advice and just run at 1:1 mode.  It depends entirely on the software you’re using.  If your laptop resolution is in HiDPI mode and you have say a 1080p external monitor attached as a second display and you try dragging most audio production software across, you’re going to have a bad time.

The good news is that Microsoft seems to have spent more time and focus than they have on legacy win32 programs more in the past year than they did the preceding seven years.  Microsoft Office is still largely deployed as a win32 program and supports HiDPI settings quite fine, so hopefully audio production software developers start taking this into account, too.

DPC Latency

One area that you might have come across if you have investigated Windows as a suitable platform to do audio production on is DPC latency.  It can be quite the rabbit hole which I’m hoping I can sufficiently summarize for you here without you needling to Bing or Google your way into a week long obsession and overnight forum archeology.  I don’t want to gloss over it because if you’re of that attitude, you might has well just buy a Macbook Pro and deal with your slower speeds and inevitable burnt out components, but I also don’t want to give the impression that it’s something you have to completely familiarize yourself with in order to optimize your rig.

What is DPC?  DPC stands for Deferred Procedural Call.  Microsoft has created this functionality to let higher priority tasks have higher priority over lower priority tasks.  This is obviously favorable for audio applications because in ideal settings we want to treat audio functionality as though they are real time.  Windows is not a real time operating system though.   Thankfully, Microsoft have engineered Windows to handle these high priority tasks called by applications as close as it can get to real time… as long as your have the right hardware.

Quite often, most DPC settings can be resolved on a lot of hardware by putting into the place the power plan profile that we created above.  However, many other things can affect audio related DPC tasks. However there are also other things that can cause other tasks that have been set as high priority to processed even before the audio tasks.  When these non-audio DPC tasks consistently get in the way of audio DPC tasks called by audio programs, we call this DPC latency.  They manifest in a number of ways, commonly skips and stutters in the audio and drop outs entirely as audio is being processed or played back.

Tools Of The Trade: LatencyMon

A company called Resplendence has a tool called LatencyMon that lets you parse and view some special features of Windows helping one to identify sources of DPC events that might interfere with audio production, as well as analyze your hardware and software to determine if it’s suitable for audio production.

See the source image

 

The causes for DPC latency affecting audio generally fall in one of four categories:

Common Cause #1:  Network Devices

In Windows, network devices are given high priority because networking by design has to be.  Without getting to much into it, TCP/IP networking revolves around a concept of transmitting and receiving encapsulated packets with data formatted in various ways depending on the payload.  Modern network devices (1Gbps is common) tear through this workload pretty effortlessly, however at the penalty of its workload being treated as higher priority than most.  An especially problematic element comes into play with regard to wireless networking.  In addition to maintaining the overhead of constantly receiving and sending packets across a network, it has to use a radio to maintain a link with a remote access point and encrypting all traffic to and from.  All of this penalizes the computer in that it takes up DPC resources that could otherwise be attributed to audio production work.

Solution: If you are on a desktop, I wouldn’t use a wireless access point at all if I could help it.  I use ethernet cable to get my network connection to my tower and I don’t do network intensive things while trying to produce music.  If you are on a laptop, like my Surface Book, I have a hub that provides an ethernet port.  Some motherboards on desktops and laptops even allow the wireless network device to be disabled entirely.  If you see NDIS.SYS in LatencyMon consistently pushing you into the red, investigate this further.

Common Cause #2:  Power efficiency profiles

If Windows is trying to govern power draw of different devices constantly, eventually you will run into a scenario where a task flagged as DPC is set to be processed in the queue and suddenly Windows will depower some piece of hardware like the CPU to prevent the tasks from being completed on time, resulting in drop outs and clicks and pops in the audio.

Solution: If you followed my directions above for creating a custom power plan profile, then you’re set!

Common Cause #3:  Video Devices

Simple analysis –  buy AMD GPU over NVIDIA or Intel GPUs, all day every day.  For a gaming PC, I would give the opposite device as NVIDIA tends to have stronger performance unmatched by AMD.  However NVIDIA seems to consistently saturate a PC with DPC tasks, even when the GPU is not heavily utilized and only driving desktop applications and not 3d games.  In my desktop, I use a Radeon RX Vega 64 and it’s glorious.  Not one stutter and my audio playback has never been less than perfect.  Intel GPUs also seem to affect DPC latency more than Radeon in that Intel GPUs are actually part of the processor and using the Intel GPU to drive your displays can make the Intel CPU run hotter and be more prone to throttling.  What is throttling?  Throttling kicks in when the CPU is detected to be operating above a safe temperature threshold and the CPU speed is decreased.  While we disabled Windows from doing this on its own with the power profile, a mainboard will throttle a CPU if it runs too hot.  Keeping on all non-essential tasks off the Intel CPU (like its built in GPU), is best practice.   It’s not in the same league as the hit you take on an NVIDIA GPU however, so if it’s all you have to use, it’s nothing big to worry about, its impact is just measurable in non-direct ways.

Solution: Buy AMD GPU

Common Cause #4:  Audio interface

For near real time performance of audio DPC tasks, you want to make sure your audio interface can process the bandwidth and latency required.  Ultimately this means completely and thoroughly rejecting the idea of ever using onboard audio for any serious audio production.  Onboard audio interfaces are limited in functionality and quality and even using 3rd party ASIO drivers, you’ll never get the same performance as even just a $100-$200 dedicated audio interface.  To other producers that are also PC gaming enthusiasts, I try to use an example of what you get out of an Intel GPU vs an NVIDIA GPU, one is just measurably better in every way compared to gaming and the same is true for the audio interface within the realm of audio production.  The lower latency the better.  The less transport abstraction you have to go from audio interface to the OS is even better.

Solution: Avid HDX PCIe + HD Omni > Presonus Quantum Thunderbolt interfaces > USB-C audio interfaces > USB2 audio interfaces.  You’ll need $10,000k to even get started with an Avid Pro Tools Ultimate system for hardware and software licenses, so a reasonable place I’d start is either a Presonus Quantum if you have a Thunderbolt port or a Presonus Studio Series USB-C audio interface if you have USB-C.  If you only have USB like I do on my Surface Book, I can vouch for the Behringer UMC1820 being an excellent and cheap solution.  Always make sure you use the ASIO driver that is available for the audio interface vendor you’re using and not a generic ASIO driver like ASIO4ALL.

Using LatencyMon

When you fire up LatencyMon, remember that you’ll always see DPC tasks.  The goal isn’t to get rid of them, but to make sure they don’t interfere with the audio process that LatencyMon is simulating in place of a DAW and push it to the red.  Be sure to read their documentation.

 

Audio Devices

I previously touched upon my views of audio interfaces and their effects on DPC latency.  But beyond that, there’s a world of reason why you should pay such special attention to the audio device you’ll be using for audio production.  For one, it defines your IO count, or the number of input and output channels that you have to work with.  If you’re like me, you’re constantly pushing audio in and out of the computer to be processed by outboard gear like compressors, limiters, guitar pedals, and effects processors.  Since that’s my workflow, a 4×4 (4 in, 4 out) interface wouldn’t suit my needs.  Another area of attention you’ll want to spend time on is the type of audio interface.   These are all questions you’d be asking yourself if you were on a Mac as well, but here we need to take into account hardware trends that are different from Apple.  While Apple has pretty much put all their eggs into the Thunderbolt 3 basket, devices designed for Windows are thankfully more user friendly and capable of hosting a wide array of different ports and different devices.

Connection Type (Thunderbolt 2/3)

Again, based on my understanding of the Thunderbolt spec, it’s essentially a way to expose additional PCIe lanes to external devices.  Commonly on Mac, you’ll see people wasting this potential on goofy ass things like a single Thunderbolt 3 to USB dongle made by Apple for $30.  It’s crazy people are ok with this and is primarily one of the reasons I stopped using Apple for audio production.  Using devices like that is essentially like buying a device with four external PCIe slots and using one of those slots for a single USB2 port.

That said, Thunderbolt itself is quite interesting with regard to audio production in that is exposes PCIe lanes to external devices.  That’s kick ass.  I went with the Prosonus Quantum, which is a Thunderbolt 2 device I have connected to my Thunderbolt 3 port on my Alienware Aurora R8.  Presonus has built out their Quantum line for the ability to daisy chain IO as well, so I could hook a 2nd Quantum or a Quantum 2 to my Quantum and have even more IO if I wanted.  While it provides the same speed and functionality if I was getting something like an RME HDSP PCIe card, it provides the ability to host it outside of my tower and effortlessly offer expandability that I wouldn’t be getting off the physical constraint of a physical PCIe slot.  Because of its use of Thunderbolt, it gets amazing audio latency results often around 2-4ms, though I tend to only use that sample size when tracking stuff in from my preamps.

USB4 is coming out soon which as far as I’ve read seems to indicate it’s mostly based on the Thunderbolt 3 spec.  Hopefully they are compatible and more and more devices start targeting it, because up until now it’s been quite rare to find a Windows PC with Thunderbolt support.  Some laptops like the Alienware 17R series natively support it, while desktop computers depends on the motherboard used and whether or not the vendor added Thunderbolt headers to the board and offers a Thunderbolt card for it.  And because Thunderbolt cables themselves have firmware, it’s not always true that a Thunderbolt converter for 2 to 3 will work on Windows as easily as it does on a Mac, because most cables are still made by Apple.

Connection Type (PCIe)

As I mentioned earlier, PCIe offers the same speed and functionality of Thunderbolt 3 albeit without its expandability by way of daisy chaining devices together, especially ones that handle it so elegantly on a driver level like Presonus managed to pull off.  In addition to not having this same level of expandability, a PCIe audio interface burns one of your PCIe slots that you could use for things like DSP cards like Universal Audio’s UAD-2 devices that power amazing low latency plugins.  I think we’ll see this format become more rare as time goes on, and its advantages might be beaten when USB4 becomes standard and offers combined PCIe lanes theoretically providing better performance than PCIe.

Connection Type (USB)

USB 2 and 3 are very common on most devices today, and audio interfaces using them are quite good.  Even though the audio is send through a USB bus abstraction, most computers will handle it without issue.  It’s pretty much a no frills experience but that lack of wonder is often offset by these devices being much cheaper than their faster port alternatives.

 

Fresh Install of Windows 10

I would highly, highly consider doing a fresh installation of Windows 10 before embarking on setting up a Windows 10 PC for audio production.  If you’re building a PC from scratch, you’re doing this anyway, but if you just bought a PC desktop tower or laptop, immediately pull up Bing and search for “Download Windows 10” and you’ll see this page or some variation of it if you’re not in a US English region.  All you need is a USB key and enough time to go make yourself a coffee while it downloads and installs the Windows 10 installation media to the USB key.

Why do a fresh installation of Windows?  It’s certainly not a jab at Windows 10 and you want pure Windows 10.  It’s because often OEMs that sell you Windows 10 PCs also include bits of software to facilitate software installation and driver support, something that Microsoft already largely provides with Windows Update, but usually the OEMs feel the need to feign value add and litter your Windows 10 installation with their own software right from the factory.  It’s a pain in the ass to deal with, it usually always eventually breaks down and causes issues down the line once your warranty runs out, and more importantly it can and will affect your ability to consistency have good audio performance.

I am a subscriber to Office 365 (soon to be called Microsoft 365 according to WindowsCentral.com), so I have tons and tons of cloud storage via Microsoft’s OneDrive service.  I throw all of my personal files in there, along with artwork, sample collection backups, tax returns, business receipts, etc.  I suggest getting in the habit of doing something similar, because I’ve found it aids me an getting back up and running in case of replacing a machine when upgrading to a new one.  I wouldn’t use it primarily for storing your audio projects, but it would be suitable for storing backups of your audio projects.

What About Drivers?

Thankfully, Windows 10 is developing in such a way that Microsoft releases new versions twice a year.  With each release, the available drivers also increases to compensate for the newly available hardware devices.  I have found that on reinstalls Microsoft consistently nails getting the right drivers for everything I need.  The only manual driver installations I’ve had to handle in the past several years have been for my Audio Interfaces and their ASIO drivers, and the drivers needed for my pair of MOTU Midi Express 128s.  I bring up my experience here because if you used Windows XP or Vista back in the day and switched to Mac, you might still be thinking you have to go searching for all of your device drivers when attempting a reinstall and I can thankfully tell you that this process is now much easier and that’s largely not required any longer.  It’s still good practice when doing a reinstall however to make sure you have drivers on a usb key for your network interface.  If you have that, everything else can be found pretty quickly by Windows itself.  Maybe this isn’t best practice, but it’s my experience.

What about Microsoft Account to log in?  Should I use it?

A LOT of people get really upset when they reinstall Windows and find that Microsoft really tries to advertise their logging into Windows with your Outlook account.  It’s really not a big deal.  To keep my main tower installation as simple as possible, I use a local account (just like Windows 7 had), but I could just as well use a Microsoft Account.  I simply don’t use my main audio production PC for productivity though, so apps like Mail and Calendar aren’t present to provide benefit where I wouldn’t have to log into each one rather than letting my Microsoft Account handle populating all of that for me.  If you are using your Windows PC for both production and productivity though, rest easy using the Microsoft Account to log in.  Outlook kicks ass, Skype is great, OneDrive integration with Files is the bomb, and it syncs wallpaper and bookmarks between your Windows devices.  If you use a local account it works perfectly too, but if you want to use something like OneDrive you’ll need to manually log in using the application.

ZOMG! Telemetry!  Microsoft is spying on me!

Yeah, no.  Windows 10 is deployed on more than a billion devices.  Microsoft collects telemetry data to determine when features aren’t being utilized, what things people are having trouble with using, and more.  Remember that useless People Bar that appeared a couple of Windows 10 releases back?  Remember how they quickly started hiding it and are preparing to remove it soon from new installations?  That was from Microsoft collecting telemetry on the service.  Don’t go running crazy 3rd party tools to disable telemetry and remove stuff because it’s likely to break something critical down the line and reinstalling production machines is a pain in the ass.

BIOS/UEFI Updates

Before started to install your DAW software and plugins, I would highly suggest taking the time to locate and install any available BIOS updates.  If you aren’t familiar with the process, it usually involves running some executable from an OEM or motherboard manufacturer that restarts Windows and boots into a special BIOS update environment and then reboots back into Windows.  It can brick your computer if the power is suddenly lost so don’t do it during a storm where the power might go out, and even better do it when your computer is plugged up to either a UPS battery or has a built in battery like laptops.

Start Thinking About VST Installation Location Before Installing VST plugins

One of the things that I really don’t like about Avid Pro Tools is that it uses AAX plugins and they all get installed to ONE location on your OS.  If you won’t be using Pro Tools and are using something like Ableton Live or Steinberg Cubase, you’ll want to go ahead and start planning where your VST installations will be set to because you’re going to need to pay special attention to this after you freshly install Windows 10.  What I do is let everything install to their default location from the installer and make a note of these locations and then create shortcuts to them inside C:\Program Files\Common Files\VST2 and C:\Program Files\Common Files\VST3 manually.  If any plugin installers hard code a location to something like C:\vstplugins\ you can also just create shortcuts back to their associated folder in C:\Program Files\Common Files\.  VST3 plugins usually have a *.vst3 file extension while VST2 plugins usually have a *.dll file extension.

It’s a matter of housekeeping and there’s nothing more frustrating than installing a new plugin and not being able to find it available in your DAW afterwards.  In managing your own group of shortcuts, you can also place them in folders named for their Vendors or by their usage (like Delay, Reverb, Compressor, Limiter, etc).  Ableton Live handles this quite well and keeps it organized.  If you use Cubase, it will read that entire directory and you can create your own custom plugin list layout in Cubase’s Plug-In Manager.

Alternatively you can set each plugin installation location during install, but this might be quite time consuming if you own a lot of plugins.

Storage Needs

I can’t think of a producer I know that isn’t checking out storage prices at least once a week.  Thankfully, Windows 10 is amazing in its support of various storage needs and requirements as well as easy integration with its OneDrive cloud storage solution.  It’s very possible that all you need in a single drive to record audio in and bounce the stems off to be mixed elsewhere.  Or maybe you are a sample hoarder and have older PCs all over your home network acting as various storage hubs.  Maybe you want RAID 1+0/10 to ensure you never lose data due to another crappy drive.  Windows can facilitate any of your needs. It scales from rinkydink $200 el cheapo laptops from Walmart to the biggest datacenters in the world.

SSD or HDD?

In 2020, I think it’s safe to say most storage scenarios should revolve around some implementation of SSD at the bare minimum.  The prices have fallen substantially in the 8 years I’ve used them and the storage sizes are ever increasing.  Definitely use an SSD for your Windows 10 installation.  You don’t want to have to stop and wait for some part of Windows to load from the slow spinning disk into memory and be processed.

HDDs do have their place though as they are more cheaply available for larger storage sizes in the 4TB+ range, however one should be aware that they can easily be broken due to their moving parts.

Gotta Keep Em Seperated

For audio production, the general consensus seems to be it’s a good idea segregating project data and OS data.  This means a minimum of two storage disks, one for Windows 10, your DAW, and plugins and one for your projects and sample collections.  I go a little overboard and have a third production drive to exclusively put sample collections on, keeping it separate from my projects drive.

Please Plan Proper Backups To Prevent Piss Poor Data Loss

Macrium Reflect is an amazing tool allowing you to make images of your drives as well as snapshots at completely customizable intervals.  When set up, I store these backups on two separate HDD drives here at my home studio as well as OneDrive.  Once a month I inspect the weekly snapshots to make sure they can be restored if needed and keep an eye on the space used to determine if I need to seek additional storage.

You really can’t be careful enough with your data, and that goes double for music production.  Even if you had the projects backed up on a drive, could you load those project files up on the same computer after you reinstalled everything from scratch?  Do you have all of your custom homebuilt presets collections, SYSEX dumps, and plugin installations the same afterward? If not, you might have trouble opening those project files successfully.  Just having the project files backed up on another drive is false sense of security, at best.

Utilize different types of storage, too.  If my home studio burned down, the latest snapshot of my whole studio is on my OneDrive account ready to be reimaged on new drives in the event of a catastrophe.  If my computer was struck by lighting, my HDDs are in the other room and only attached when doing new snapshots.

I have felt first hand the loss of a project due to poor data management and I wouldn’t wish that on anyone.

Should I RAID?

Obviously this is somewhat difficult to do on a laptop and might not be possible, but some desktop users might say “well, I have 4 SATA ports, shouldn’t I be raiding? And if so, do I even need backups?”.   RAID is definitely good.  It stands for “Redundant Array of Inexpensive Disks” and in practice is means that instead of storing data to just one drive, you have #1 one or more drives extending the available storage by combining storage available on multiple drives, #2 one or more drives mirroring each other so data written to one drive is also written to the other making a constant mirror or #3 some combination of these.

You might have heard of RAID striping where you take two drives and turning them effectively into 1 drive visible to Windows where their storage is combined (RAID 0).  This is the worst idea ever.  Do not do it.  Ever.  What happens is eventually one of the drive fails (with its half of data that’s on it) and the other drive is inaccessible because part of its filesystem is on the other drive.

Striping however is very attractive when combined with mirroring like RAID 5, 6 or 10.  Mirroring means that one of the drives can fail and the other drive can take its place when reporting to Windows, and an alert will tell you a drive needs to be replaced.  RAID 10 is fairly standard in datacenters.  It’s also very, very expensive and physically requires a lot of space in your computer or storage computer.  If you are comfortable with managing this, knock yourself out, because unless both mirror drives fail you won’t have to go through the reimaging process from your backups.

However, its cost can be pretty prohibitive when everything is factored in (in addition to needed the storage space to keep the snapshots of the disk images on other media).  As a result, I don’t personally employ it.  I have unopened standbys of each type of my drives made by the same manufacturer, the same size and type ready to go in case I need to reimage.  The strength of my backup plan rests not in not needing to reimage (which is very easy anyway with Macrium), but in having backups in different geographic locations and machines.  Aside from that, there’s the whole topic of hardware raid and software raid and performance penalties of using either.

It’s something to consider if your storage requirements aren’t as big as mine and you have the funds and machine to pursue it.

 

My Studio’s Storage Layout

I’ll go over my needs and how I’ve used Windows to make that possible and also describe some other solutions that might be what you’re looking for.  Keep in mind that though my drive sizes might be large, the disk image snapshot created is not necessarily equally as large because I don’t use 100% of my disk’s storage space.  This means I am effectively able to oversubscribe my backup storage needs for now, but will need to increase my backup storage needs after my production storage reaches a certain threshold.

Alienware Aurora R8 (main production machine)

  • C:\ 1TB M.2 SSD for OS, DAWs, and plugins
  • D:\ 2TB HDD for one set of C:\, P:\, and S:\ backups and snapshots
  • P:\ 2TB SSD for projects
  • S:\ 2TB SSD for samples (shared on the network so I can interact with it from my Surface Book on a mapped drive)

Surface Book (sampling machine, easily replaced because no data really stored to it)

  • C:\ 256GB SSD for OS, Ableton Live.  All project work is exported to the Samples drive on the Aurora R8.

External HDD #1 (Weekly Snapshot of Aurora R8, not constantly connected)

  • E:\ 4TB HDD

External HDD #2 (Daily Snapshot of Aurora R8, not constantly connected)

  • O:\ 4TB HDD

OneDrive (Weekly Snapshot of Aurora R8, geographically remote)

  • 1TB cloud storage

So, assuming I keep up with my obsessive backup habits, at least two meteors would have to hit earth before the Arcade Daydream project files and masters would be lost.  Because I archive older projects I can keep my production environment size requirements for backups quite low, it’s not like I need 5GB (1TB+2TB+2TB) to make a successful backup, but I am coming close to running into the boundary of the non-business OneDrive account with my backups of C:\, P:\, and S:\ which means I’ll probably need to move to the OneDrive For Business solutions before long.

When using OneDrive, I suggest adding the backup to a folder within a backup folder, and then after it finishes syncs going into the OneDrive settings tool and go into “Manage storage” and uncheck that subfolder to be locally synced.  Effectively this means you place it in your local OneDrive folder and remove it from your local storage after it finishes uploading, which means you won’t have backups of your back in your backup yo dawg I heard you like backups.  No, but really, you don’t want a backup of your C:\ taking up space in your newer backup of your C:\.

Does it require a bit more interaction than just me copying the backup images to external drives?  Sure, but OneDrive is not in my home studio so I feel much safer about it.   Plus OneDrive is dirt cheap, there’s no reason not to have at least the 1TB plan.

The Crazy “I Have Too Much Money” Solution

Believe me, I’ve thought about doing this, but then I remember that I would have to spend less money on equipment so I stop myself.  Another approach you might take to storage solutions is to purchase a dedicated storage server from a vendor such as ABMX who sells SuperMicro servers in an easy to use store interface.  Some server form factors let you store massive quantities of storage disks and provide hot swappable drive bays letting you pull and replace drives if they go bad without shutting down the server.  When you start investigating such products, you’re going to have to ask yourself whether or not you want to go the route of hardware based RAID or software based RAID because you can’t switch back and forth (or easily return equipment like PCIe RAID cards) once you commit.

It’s more common to rely on hardware RAID for turn key solutions and they do provide some advantages such as super fast caching for the most used and accessed data stored which would be a boon if you’re storing projects over the network to one of these devices.   Similar servers sans a hardware RAID card can be purchased which rely on your OS such as Windows Server or Linux managing RAID and despite what some people might think, MDADM (the RAID management utility on Linux) is very mature and makes managing hot-swapping faulty disks for new ones fairly easy.  I have never used these kinds of solutions for my home studio because aside from it being an incredibly expensive solution, it’s also quite loud as these devices are meant to be stored inside a datacenter rather than a studio closet.  I have however used software RAID extensively at previous jobs I’ve worked at where I managed huge data storage pools.   If you decide to go this route, you’re going to want to seek better information and companies like ABMX can gear you towards what you need based on your stated requirements and goals, so I just wanted to put the bug in your ear making you aware of them.

See the source image

It’s definitely possible to do something like this on the cheap end if you have a spare PC laying around with a lot of SATA ports and enough space to mount the drives, but at that point you’re going to not get a lot of return of your investment compared to my single drives + OneDrive cloud approach to production and backup data tiers.  And more often that not, I think the people that cheap out on the storage approach find themselves in situations where they are ill equipped to manage their own storage server once they get it up and running.

System Tweaks

Are they needed?

Not really.  Windows 10 today is pretty lean and mean by default.  I don’t think you’re going to push your machine much more beyond our custom power profile we made earlier by pulling two decade old levers in the OS written for entirely different processors and ancient versions of Windows.  I’ll go ahead and list the common ones, but know I don’t suggest using them on any machines that shipped with Windows 10.  In my experience on Windows 10, most issues stem from USB suspend settings interfering with audio interfaces and MIDI interfaces running over USB and getting quirky behavior.

Scheduling

Historically, it’s always considered to be best practice to adjust the Windows scheduler on DAW machines to “Background services” instead of the default setting of “Programs”.  However some DAW developers have said this is no longer necessary with some saying it’s still absolutely necessary.  I keep it set it as default, but I wanted to bring up that this is still present in many best practices Windows configuration guides.  I’m 99% sure it’s no longer necessary based on my experience.  If you are a DAW developer (specifically at Ableton or Steinberg), I’d love to know your take.

Speedstep

Much like the ancient instruction to set the scheduler to Background services mode, Speedstep was another boogeyman of sub-optimization in the realm of music production machines.  Now, there was very much a reason for this a decade and a half ago.  I very clearly remember Cubase performing much better on a friend’s rig when we turned Speedstep off.  It probably still is if you don’t have adequate cooling.  Speedstep was an Intel CPU feature that allowed the CPU to temporarily ramp up clock speed of the CPU to process more data with the expectation that the speed would come back to normal speed later.  Like the setting in our customer power profile we made earlier, we want the CPU speed to pretty much be at 100% constantly because variable speeds can introduce glitches when we have a steady stream of DPC tasks constantly being triggered.  These days, in practice and after having used old Pentium 4 as well as my current Core i9, I would leave it on and be more watchful of the CPU clock speed *dipping* below the processor’s advertised base speed.  That produces much more interruption to real time audio workflows than the processor temporarily becoming faster than it usually is.  My theory is that before the problem manifested in Speedstep activating, then the processor becoming faster but also hotter, then the system tried to actively cool the processor by, but to a lower clock speed and not its base clock speed, with that slower throttling always being the real issue.

My suggestion is to leave it on if you’re on a K series processor which is specifically meant to be overclocked by default.  If not, then turn it off in the BIOS if you’re unable.  And then count yourself lucky you’re using Windows because if you were on a Mac, the firmware would let you run as fast and hot as it could well past the highest suggested safe temperatures before turning on the airplane engine fans, leading to Macs quickly burning out both their CPUs and GPUs.  Apple *still* won’t use NVIDIA GPUs again because they presumably threw them under the bus trying to explain away their own terrible approach to thermal design.

USB Hubs and Port Types

The USB spec is a funny thing.  While its form has survived nearly two decades thanks to its backwards compatibility and cable type, there are some things you’ll want to pay close attention to both when sourcing devices and hooking them up.  First is to be more mindful of which USB types are which on your machine, much more so compared to if you were simply using the machine to play video games, use Microsoft Office, or watch Netflix.   I mentioned previously that audio production software is acutely sensitive to anything that might interfere with near real time events being processed by your computer.  To add to this, I would advise keeping USB2 devices on USB2 ports, USB3 devices on USB3 ports, and not to mix and match randomly.  Especially with audio interfaces running over USB, they seem to be very sensitive to some limitations of backwards compatibility resulting in strange and intermittent device behaviors such as pops, crackles, distorted, or late audio.  Additionally, devices that use MIDI in some way I would advise to hook directly into the computer as opposed to a USB hub exposed by a USB port expander device or a Thunderbolt or USB-C hub.  I’ve had much trouble attempting to run my MOTU MIDI Express 128 devices through a hub both on Windows and Mac.  On Mac, the problem never went away because a hub is required due to Apple only providing Thunderbolt 3 ports on their newest devices.  On Windows, and especially with my Alienware desktop. I have plenty of USB ports of all different types and the problem went away.  The same holds true for devices such as MIDI keyboards or devices like the Ableton Push 2.

I can only speculate as to what the root problem might be but I have a feeling that like networking, there’s a penalty cost to encapsulating and bussing USB2 traffic through a USB3 hub.  I say this because through hubs I’ve experienced MIDI events coming in way off time at consistent intervals, so it’s definitely some sort of delay problem.  However, devices such as PC keyboards, mice, and webcams seem to be unaffected by such limitations so save your hub for these kinds of devices.

Airgapping : Should I Keep My Computer Offline?

If you’ve searched and searched on Bing or Google for best practices on maintaining a music production computer, Mac or PC, you might have come across a number of different posts or forums suggesting that it’s a good idea to keep you computer offline and never upgrade.  I think there might have been a time where this would have been reasonable advice to some degree.  However, today we live in an era where much of music production software available isn’t consistently using offline friendly approaches to product authentication.   Back in the day, a service known as iLok was very prevalent and so as long as you managed your USB keys and put your software licenses on an iLok key, you never had to worry about whether or not your product would be authenticated when you tried to use it.  A lot of software is still available to authenticate via iLok such as Avid’s Pro Tools.  Companies like Steinberg have implemented their own similar approach with eLicensers.   However with increasing use of the internet, many companies have implemented their own authentication procedures and many of them are simply incompatible with the practice of airgapping your production computer.

While it’s still possible to airgap if you plan to keep your music production PC installation spartan and rely mostly on outboard gear and iLok plugins only, I’d advise not to try iceskating uphill.  Instead I implore you to use Microsoft Office’s Excel to make a master spreadsheet of a list of all of your music production software, licenses and where licenses are currently activated. And store that spreadsheet on OneDrive so you still have access to it on other computers and phones.  Some software requires their own installation and authentication managers to make this process easier but believe me, my ass has been saved by being vigilant about keeping track of my licensing.   If you haven’t already registered as an LLC, I also encourage you to investigate doing so and keeping track of your purchases for tax reasons.

Aside from software growing away from facilitating offline use, the truth is modern computers running Windows 10 are very well placed and equipped to handle updates easily and keep your computer secure and running smoothly.  Airgapping afterall was meant to solve the problems of update breakages and security.   If you keep your computer offline without updates, it will be insecure if you need to plug it up to a network to reauthenticate a plugin or some other piece of software.

Conclusion

And that’s really all there is.  Windows 10 is an incredible operating system. It simply requires less esoteric tweaking than its former iterations did.  It’s now thankfully much more important to focus on the quality of your components than worrying if Windows is treating multithreaded audio processers efficiently.  And I really can’t stress enough, do not use ASIO4ALL with the audio interface built in the motherboard and then complain that Windows is buggy.  Get a dedicated interface.  And get a somewhat fast processor.  If you’re buying a Celeron or Pentium processor today and plan to run a lot of effects and virtual instruments, you’re gonna have a bad time.   That’s not to say that ultra portable and underpowered devices don’t have their place though.  The Surface Book has worked wonders for me in what I use it for (sampling and preparing the samples for use) and I’ve heard of many others using even the Surface Go for similar reasons.

If you’re moving from macOS to Windows and you haven’t used Windows for a long time, don’t be inclined to simply blame Windows for any woes or confusion.  Windows 10 is on a billion devices, chances are it’s just you.

And really, buy good components for audio production.  It’s so damn important.  Don’t get an Intel Core i5 if you’re going to be using plugins, don’t get less than 16GB RAM, don’t not have a backup plan.

I suppose this really should have just been a tutorial on simply how to customize your power plan, not use HiDPI scaling modes, and to turn off your wireless network card, so apologies if it was long winded, so I hope you enjoyed my commentary.