When choosing an audio interface, most people look at the number of inputs and outputs, the number of microphone preamps, and (of course) the price… But what if I told you that you’ve been ignoring one of the most important factors?
In this blog, we’re exploring audio interface drivers… what they are, why they’re important, and the fact that they are not all created equal.
I’m a big fan of RME audio interfaces for several reasons, including the fact that they are known throughout the audio industry as the best in the game of audio driver development. But don’t worry, because the information in this blog will help you no matter what interface you choose to use.
Thanks to RME for supporting audio education.
What Is a Driver?
First off, what is a driver?
In the realm of computer science, a driver is a software component that lets the operating system and a device communicate (according to a definition on learn.microsoft.com).
When you connect a mouse, a keyboard, a printer, or an audio interface, there will be a driver that acts as a middleman between the hardware device and the application that you’re using.
The driver is often developed by the hardware manufacturer, but not always. If the hardware follows a published hardware standard, it’s possible that a standard driver can be used. But in the world of audio, things don’t necessarily stick to a rigid standard. For this reason, and for reasons I’ll share with you throughout this blog, it’s very likely that your audio interface has a custom driver that should be used for optimal performance.
You can usually find and download that driver by running a web search with your interface’s model number and the word “driver” or “downloads.”
Aside from the variations between hardware devices in the audio industry, there’s another important distinction that sets audio drivers apart from other types of drivers… and that’s the fact that audio is very time-sensitive.
To help you to understand the nature of a digital audio signal, let’s take a step back and think about a CD. Similar to a vinyl record, the information on a CD is organized in a spiral on a disk. However, being a digital audio format, a CD encodes the audio as PCM (or pulse code modulation).
The audio signal on a CD has 16-bit depth and a sample rate of 44.1 kHz. That means that every second there are 44,100 samples for both the left and right channel, with each sample consisting of 16 bits (and each bit representing a 0 or a 1).
In the case of a CD, the 0s and 1s are represented by either a pit or a land that will reflect light differently. What’s important to understand is that a CD is a streaming format, which means that a CD player reads the stream of PCM in a serial fashion.
If there’s a scratch on the CD, you may hear errors, which can result in annoying clicks or dropouts. It’s very important that digital audio streams maintain some minimum level of quality control (otherwise, the analog audio waveform won’t be properly reproduced when converted back to analog by the DA converter).
Getting back to the discussion on audio drivers, it’s not only important that an audio driver transports PCM data with minimal errors, but (especially in pro audio) it’s important that the driver does this quickly, because they are often used in time-sensitive situations.
Contrast this to a printer driver, for example. If your computer takes some extra time to send a document to your printer and if it takes a few attempts to send that data correctly, it won’t be the end of the world. But if your audio interface struggles to send high-quality digital audio quickly, the stakes are much higher (especially in professional situations).
How Audio Drivers Work
While the PCM on a CD is streamed to the converter somewhat linearly, the PCM data in a DAW or other computer application is not. That’s partly because the computer has other things to worry about, so it can’t simply dedicate all of its resources to the audio applications.
Instead, an audio interface driver will utilize a buffer to transport audio to and from the hardware device. Rather than a linear stream of bits, the driver will store a buffer of bits and then transport it, then store another buffer of bits and transport it, and so on.
This part is very important, so pay close attention here…
Imagine a scenario where you are recording guitar, playing on top of an existing track that contains bass, drums, etc.
Ideally, you will be able to hear both the sound of the backing track and the sound of your guitar. And beyond hearing them both clearly, it’s also important that those two signals are in sync.
This makes it much more comfortable for you as a performer. Imagine hearing your guitar in your headphones, delayed in time. You pick a note and a moment later you hear the sound of the note in your ears. As we all know by now from Zoom calls, it’s very difficult and distracting to hear yourself with a delay.
This is where the buffer size becomes very critical, because the buffer size determines the round trip latency. If your audio driver is saving up a buffer of 1,024 samples before sending it to the DAW and then it saves up 1,024 samples before sending that audio back to the audio interface so you can hear it, there will be 2,048 samples of time between the moment you strum your guitar and the moment you hear it in your headphones.
Assuming you’re recording at a 48 kHz sample rate, that’s about 43 milliseconds of latency (21.3 milliseconds each way), which is enough to distract just about anyone. And that’s for mono audio. If you’re monitoring stereo audio, you double that number!
To reduce the round trip latency, you can decrease the buffer size. Changing the buffer size from 1,024 samples to 64 samples will decrease the latency for mono audio down to only about 2.6 milliseconds, which is much more tolerable and completely imperceptible for most people.
With quick and rough math, we can equate 2.6 milliseconds to the time it takes for sound to travel about 3 feet from a guitar amplifier to your ears.
As a quick side note, increasing the sample rate will decrease the latency calculation for a given buffer size. The formula is:
Latency = (Buffer Size / Sample Rate) * 1000
Also, notice that the bit depth is not a part of this formula. That’s because the buffer size is measured in samples, not bits. However, that is not to say that bit depth has no relevance as you’ll see in a moment.
Not All Audio Drivers Are Equal
This is all good in theory. Just turn down your buffer size and you should be good, right?
The problem is that reducing the buffer size gives your computer’s processor and your audio interface driver less time to process and transport the data. So if you’re running a large session with lots of I/O channels or lots of plugins running in the DAW, you will probably reach a point where you’ll start to hear errors in the audio (as clicks and dropouts). Or, your session will simply crash.
That means the buffer size that is possible will be determined by the processing power of the computer and the efficiency of the audio driver. Plus, certain operating systems and interfaces will have safety buffers which could result in a deviation between the theoretical latency you calculate based on the buffer size and the real latency you’ll experience.
There are a few types of drivers that you may encounter when setting up an audio device.
Mac users will only see Core Audio drivers, thanks to the nature of Mac OS. However, Windows users may see options for MME, WDM, WASAPI, and ASIO.
There’s a particularly big difference between the driver capabilities on Mac OS versus Windows that I often encounter as a content creator and livestreamer. This is the fact that most applications on Mac OS will not provide the option to select specific input and output channels, while the Windows driver offers much more flexibility.
For example, the RME Fireface UCX II interface connected to a Mac only offers a stereo in and out within certain applications, but that same interface connected to a Windows computer gives me the option to create submixes much more freely and easily. This isn’t really a limitation of the interface, but of the operating system and driver capabilities within the operating system.
The RME Fireface UCX II is a powerful tool in setups that require flexibility in routing, making it a great option for many professional audio applications.
In general, pro audio applications on Windows computers will utilize an ASIO driver. The main difference that sets ASIO drivers apart is that they facilitate a direct communication between the application and the hardware device, as opposed to the other types which pass the audio through the operating system.
So, at the very least, you’ll want to ensure you’ve installed the ASIO driver from your interface manufacturer. This will optimize the performance of your setup and allow you to achieve the lowest possible latency.
But it’s not all about latency. There are other reasons why RME is known as one of the best when it comes to drivers, and you should consider these additional factors no matter which interface you end up going with.
For one, you want to choose an interface that is efficient and stable. A driver that can theoretically provide extremely low-latency performance is only valuable if it can ACTUALLY provide that in a real-world application. You don’t want a driver to crash your session, especially if this is your job.
An important element to driver stability is operating system compliance and compatibility. And that requires a development team that keeps a close eye on operating system updates. You can theoretically dial in your system and then disconnect your computer from the Internet and never update your computer again, and many do this.
But when you need to update your computer in order to maintain functionality or obtain new features, you better hope that your audio interface has a driver that will be compatible after updating your operating system.
This isn’t usually an issue within the first few years of owning an interface, but one would hope that high-end interfaces (like those from RME) will continue to work for years or decades to come. And they have a great reputation for providing prompt, reliable updates even for older interfaces. For example, the HDSP 9652 PCI card, released in 2001, just had a driver update in February this year (2024).
Finally, the software package offered for an audio interface is too often overlooked. Many users rely on routing tools that come with interfaces.
Choosing the right audio interface is about more than just the number of inputs, outputs, or preamps—it’s about understanding how the interface’s drivers work. High-quality drivers, like those found in RME interfaces, ensure reliable performance, low latency, and long-term stability. Whether you’re a musician, producer, or live sound engineer, the efficiency and stability of your interface’s drivers can make or break your audio experience.
So, when you’re selecting an interface, don’t overlook the importance of drivers. By investing in an interface with reliable, well-maintained drivers, like the RME Fireface UCX II, you’re setting yourself up for years of high-quality, uninterrupted audio performance.