Audio Interfaces 101
Topics we’ll touch on:
- Why this is the most important purchase you can make for your studio
- What the specs mean
I am a hopeless nerd for computer recording. In fact, it’s kinda funny we use the term “computer recording” any more. As opposed to…what? I still remember the endless debates in the 90s about tape versus hard disk recording. I am happy to report that the purists lost. They usually do.
But, hey, maybe you are that guy and still believe that tape is the ideal medium for recording. That’s fine. You’d be right in many ways in terms of the sonic qualities of tape. However, I’ll tell you, essentially, what I told the countless freeks for piano samples standing on the retail sales floor. These folks continually pointed out how a workstation or stage piano doesn’t sound like the real thing. I’d tell them, “Then go buy the real thing.” Really, though, this argument on the merits of tape and hard disk recording are done now. Consider yourself lucky if you haven’t had to have this conversation.
Why this is the most important purchase you can make for your studio.
Let’s say you are a photographer. You would probably own a camera, lenses, cases, a copy of Photoshop or Aperture and a tripod stand. Now, out of all of these if you were challenged to say which is the most important piece of the puzzle you can purchase, that would be an easy answer. The camera is the most important piece in the process. We could first suppose you’d have to have talent and drive. I don’t mean the intangibles, but just for the sake of gear – the camera is the most important piece.
Your audio interface is the camera. If you are going to make or capture music to hard disk, your audio interface has to be as good as possible. Just like a computer. When people ask me which one I think they should get I always tell them to get the most pimped out, sick-featured behemoth they can afford at that moment. That way it can last you many years and you don’t buy a piece of software tomorrow that it chokes on. If you intend to make and record music and you don’t have to use your 2” Studer machine – get the best audio interface you can afford today.
The audio interface (or audio I/O) is crucial because it takes what happens in analog world, the real world and translates that into 1s and 0s – a digital signal. That’s big stuff. The same is true for the output stage of the interface – it’s crucial. It has to take 1s and 0s and turn that into voltage that accurately represents what is on the computer. We take so much of technology for granted that we don’t stop and think about what these gizmos do. It’s pretty incredible stuff; especially if you’re a geek like me. Then it’s endlessly fascinating. But, simply put, the interface is heart of the studio and the computer is the brain.
What the specs mean
Specs lie. Yeah, I said it. Specs lie. Wait, let me clarify; manufacturers lie. It’s not the specs’ fault. You need to read the specs and then consider that each part of an IO can influence or negate the other parts’ best efforts. For example: I saw one manufacturer that bragged that their interface had 106 dB signal to noise ratio. However, what they didn’t point out on the same web page was that their built-in mic pre had a signal to noise ratio of 98 db. We’ll get into to S/N ratios in a sec, but suffice it to say – your interface is as good as the weakest link in its chain.
Inputs: how many separate things (cables, channels, small dogs) can be connected into the computer at one time.
Outputs: how many separate output channels, cables and (yes) small dogs can be connected to control what is coming out of the computer. Discrete channels, if you will…
Side Note: The tricky thing here is to look at whether they mean analog or digital ins and outs. You could get excited to learn the Focusrite Saphire 56 has 28 ins and 28 outs. However, if you have no other devices in your studio that have digital connections, you’re probable only going to use the 8 analog ins and outs. Although a company isn’t lying to you about what it is, ins and outs can be tricky only insofar as you need to decide how many ins and out you really need. And when do you need them? When you hit the record button, at that very moment, will you have, say, 28 channels of audio to record? Or, are you going to overdub? Say, you only need 10 channels for an seriously mic’d drum kit. Tomorrow the guitar player and bassist will record together and only take up 4 channels on their takes; and so on and so forth. So, first things first – decide how many ins and outs you really need. Most of us, at the end of the day, need 2 to 4 outputs, by the way. I only need to hear stereo out of my recoding software after all.
Sample Rate: how many “samples” are taken in one second’s time. If you have a 96 kHz recording, you’ve just recorded 96 thousand individual samples of one second. It would be like 96 thousand (or whatever sample rate you choose) pictures taken of one second of audio. This is important because more information is better in computer world. Having said that, I am done with being surprised at how many high-end studios are doing 24-bit, 48K recordings. I would venture a guess that most are using 48K (K, as in “thousand) sample rates, followed in second place with 96K and a distant third for those recording in 192K. Big sample rates are important if you are doing a lot of micing. If you do a lot of Reason or line-level recording from synths you won’t get much benefit from 96K or higher. If you are using mics and hard disk space is of no concern – rock the 96K recording.
Bit Rate: the number of bits (units of information) that are processed per second. Back to the photography comparisons. Think of bit rate as megapixels. The more you have, the better “depth” of field you can get. Dynamics get better, the audio picture gets clearer and suddenly life gets better. Oh, and people start to look at you as someone to be reckoned with. But, only if you record at 24-bit. Seriously. Record at 24-bit, not 16-bit. Let me illustrate for you just why in one sense. There is twice as much information in a 17-bit recording than there is a 16-bit recording and twice again as much info in an 18-bit recording as a 17-bit recording. So, a 24-bit recording isn’t just 8 thingies better than 16-bit. It’s, like, a gazillion thingies better.
Another Side Note: This is the part of the discussion where you interject, “ah, yes, but aren’t CDs 16-bit 44.1K recordings?” To which I say, “your Run DMC cassette tape back in the day wasn’t recorded to cassette tape – it was recorded to a fat 2” reel to reel machine first.” You start with the highest resolutions possible so that by the time this music gets down to a stereo .WAV or MP3 format – it has everything going for it and nothing working against it.
Frequency Response: this is a measurement of how well your interface can represent frequencies from 20Hz (way low, bassy bass) to 20kHz (way high tweety tweetness). The “±” symbol means plus or minus. If you see that an interface does ± 0.4dB from 20Hz to 22kHz this means that the resulting signal it captures and reproduces is give or take 0.4 decibels accurate. Accurate to what really happened in the real world at that moment/frequency. The smaller this number (± whatever) is, the better.
Dynamic Range: the difference between low-level thermal noise in the electronic circuitry and high-level saturation resulting in distortion. People use this and “signal to noise” as one and the same. From clean to nasty you might say. In between these opposites is where your pretty music can sit. So, your audio interface may have “106 dB of dynamic range” for instance. The bigger this number, the better.
S/N (Signal to Noise Ratio): If something says it is “-106 db S/N”, they mean you have to go down 106 decibels from the loudest a signal can be (before distorting/clipping) to where you hear the inherent noise in a piece of electronics. The bigger this number, the better.
THD+N (Total Harmonic Distortion plus Noise): This one can get really heady if you want to dig in online. Suffice it to say a company is testing to see how much distortion is in the sum total of their electronics – a distortion creating unwanted harmonic content plus the inherent noise in the electronics they are using. The smaller this number is, the better.
Crosstalk: a signal on one circuit or channel is transmitted to another circuit or channel. You might read “-118 dB @ 1 kHz.” This means that you’d have to go down a full 118 dB at 1 Hz (this is prime midrange area we are talking, a lot of vocal content is around 1kHz for example) before you would here this unwanted crosstalk content. The bigger this number, the better.
Now these last few categories of specs are where you can really fudge numbers. Besides, we are talking about total harmonic distortion… Let’s pause here for now, because I am starting to nod off. I mean, “crosstalk?” Yeesh. In part 2, we’ll discuss such fun topics as latency, usb vs. firewire vs PCIe and sound quality differences amongst the interfaces.
This was my very first blog ever. I can’t believe the day is really here. I have a lot of people I want to thank. First, I’d like to thank my valedictorian, U of Michigan English major mother. We did it, Momma!
Secondly, I like to thank my lord and savior Bob Newhart, who’s comedy albums kept me up at night as a child.
Lastly, I’d like to thank the pimp ass bitches at Unqiue Squared cause they pay me in beanie babies. Hell yeeyah!
Thanks for reading, and don’t forget to comment. I’m here every day.
~The Unique Geek