Audio interfaces

Above a modest price, interfaces all sound the same#

In my opinion, past a certain price point, interface sound quality is a non-issue. Under normal listening conditions almost everybody would fail a properly level-matched ABX test between a mid-tier interface and a mastering-grade converter costing thousands. I ran 1,500 loopbacks through a premium converter and it barely changed. It is a wonderful product. But who is going to do 1,500 loopbacks? Nobody. Mid-tier gear is already transparent, so nobody hears the upgrade.

Measured differences do exist. One cheap interface I tested had more than twice the distortion of a mid-priced one, and both were still low enough to be transparent for a single pass. That matters mainly when a signal goes round trip after round trip. The exception is the cheapest plastic junk. For why conversion is otherwise a solved problem, see DACs.

There is a golden mean per channel#

Work out the price per channel of analogue conversion. Around €30 to €60 per channel is the golden mean: expensive enough not to be crap, cheap enough not to be a rip-off. The ultra-cheap end doesn't meet my standards. The ultra-premium end is excellent but gives no audible benefit for the price, and some of it charges you for plugin stuff you don't need. Large high-channel-count units and older ones bought second-hand can be very good value.

If you want something beautifully made with relays that click nicely, buy it. I've got a Swiss watch. Just don't claim it sounds better.

Count the analogue channels#

Because competent interfaces sound the same, the I/O decides which is best for you. Most manufacturers add up analogue, digital and headphone channels and name the product after the big number. In most cases, only the analogue channels count, because using the digital ones means buying more converters anyway.

If you record musicians, sooner or later you will want to record a drum kit, and to do heavy modern drums properly you need about ten mic preamps; see drum recording.

Judge the mic preamps, not the line inputs#

Most people record microphones, so test the preamps. A low-sensitivity dynamic mic like an SM7B at maximum gain is where bad preamps struggle. On decent gear the mic preamp is isolated from the line and instrument circuitry, and when that is not the case, it can introduce more noise and crosstalk. More under preamps.

A pot on the headphone output is bad design#

Professional interfaces should not have potentiometers on the headphone outputs. Every touch changes your level and you can never get back to it, and consistency depends on recalling your exact working level (see monitoring level). Headphone monitoring is not an afterthought. It is essential to the modern engineer. Whether the recall uses digital control or relays doesn't matter.

The headphone amp also needs low output impedance. Amp and headphones form a voltage divider, so a high output impedance skews the frequency response of dynamic headphones and loosens the bass. Many interfaces on the market have high output impedance, and you can measure it yourself with an oscilloscope and a resistor. See headphones.

Digital outputs#

Many speakers now have DSP crossovers built in. If you send them an analogue signal, you go from digital to analogue in your audio interface, back to digital inside the speaker, then back to analogue again, for no reason. Isn't it neater to use a digital output (S/PDIF, TOSLINK, AES or ADAT) and plug it straight into the speakers? Even if you don't currently own digital speakers, you might do in the future. The same goes for a monitor controller: it should have digital in and digital out, and almost nobody makes one.

Class compliance makes the driver argument moot#

"You need that brand for its drivers." A class-compliant interface needs no drivers at all. I recorded a full drum session into Linux on a 2015 laptop with no drivers and no problems. What does still matter is whether the control software for the interface's mixer runs on your operating system; see operating systems.

The one I chose, because it ticks every box on this page, is the MOTU 10pre. Disclosure: I bought it with my own money, then approached MOTU about a full-length review, and they refunded the price in exchange for making the video. I would have bought it anyway, and I've been a fan of their gear for at least a decade.