Latency in home recording

A question I often see asked by people new to home recording is:

When I record, there’s a horrible echo in my headphones, what am I doing wrong?

The short answer is you’ve just come up against something called “latency”, but the long answer is, as usual, a bit more complicated.

To start to understand this, let’s take a look at a basic home recording setup. Generally you’ll start with a computer running Digital Audio Workstation (DAW) software. This can be Pro Tools, Reaper, Cubase, Logic Pro, or any of a growing number of other options. The DAW allows you to record, edit, playback, and otherwise process audio to make it sound how you want.

But first you have to get the audio into the DAW. Your voice is analog, the DAW is digital, so you’ll need something to convert analog signals to digital. Now most computers these days already have one of these A/D converters built in – the built in microphone uses it. But that converter is usually not professional quality, and neither is the microphone, so the next step up is generally to get an external audio interface.

Audio Interface Connections

These come in even more flavours than DAW software so be prepared to do some research before you buy one, but basically what they do is convert an analog signal to digital, like your built-in microphone, but better. And they generally don’t come with a built-in microphone, so you’ll need to invest in one of those as well (I think you’re starting to see why not everyone does this).

So there you are, you’ve got your mic, interface, and DAW all connected, you fire it up to record your dulcet tones and … well … you’ve got this horrible echoey sound in your headphones (Oh, didn’t I mention those yet? You’ll need a decent pair to get good results). You try everything, different mic, different cables, different room, but nothing fixes it! So that is when you go online and ask that question.

And you’ll get many suggestions, most of which you’ve already tried, but the real answer is that you’ve just learned about latency – the hard way.

Latency Diagram

Latency, to put it in simple terms, is the time delay introduced between your audio interface and your computer. That A/D conversion process takes a tiny amount of time, transferring the digital signal to your computer adds a tiny bit more, and the DAW tacks on a bit more time to process it. Then it has to go back to the interface, and be converted back to analog to go to your headphones. All that time adds up. It generally only adds up to a few milliseconds, but your brain is pretty sensitive, and will easily hear that delay. And because you’re probably also monitoring the direct analog signal from your interface (ie. before it does that huge round trip), you will hear that as a short echo, and that can be very off-putting.

Decision, decisions

So what can you do about it? Well there’s basically only 2 things you can do, and as usual there’s compromises involved. But generally the solution is to switch off one of the signals – if you’re only listening to one of them, then you won’t get an echo in your headphones, right? Well sort of. Let’s look at both scenarios:

  1. Your first option is to switch off input monitoring in your DAW software. This means that digital signal won’t get sent back to your interface, and everything will be great! But hold on, what about that cool EQ and reverb you’ve setup on the track in your DAW? BOOM! It’s gone! Unless you have hardware effects or an interface that can process plugin effects on-board, all you’ll be able to hear in your cans will be your raw, unprocessed, vocals. Yuck! But at least there will be no delay or echo…
  2. Your second option is (not surprisingly) to switch off direct monitoring on your audio interface. This may be a switch or knob on the box itself, or may be done via the software driver on your computer. Either way, it means you’re now only getting the return signal from the DAW in your cans. And it’ll have all those neat effects you added in your DAW. Great right? Well sort of. There won’t be an echo in the cans any more, but there will still be a delay due to the round trip latency, so what you hear in the cans will be slightly behind what you sing…

Something else that’s pertinent to mention here – there’s generally more options you can use to reduce the latency if you choose option 2. Your DAW will most likely have a setting for “audio buffer size” or similar, which allows you to change the size of the buffer it uses to smooth out the processing it does. Setting this value small will reduce the latency, but below a certain point (depending how much processing it needs to do) will start introducing pops & clicks into your playback, which is arguably worse than echo. So when recording, it’s best to switch off as many plugins as possible, and reduce the buffer size to the lowest setting you can get without pops & clicks. Your audio interface might also have a buffer size setting, you’ll generally want to set that as low as possible too.

So that’s latency in a nutshell. If you’ve made it this far, well done! And have fun recording!