Posted on 17th April 2018 by Callum
When using a computer for recording, producing or even listening to music, you’re sure to have come across audio interfaces. Rich Tone attempt to cut through the confusion and give you the lowdown on what these nifty devices do. We’ll look at why they are important and what to consider when buying!
These are also known as soundcards or AD/DA converters, which helps to explain their function a little better. This stands for Analogue – Digital / Digital – Analogue. These boxes convert the digital information on a computer into analogue audio, and then back the other way around as well. Your computer will have an audio interface built in, although these are usually very basic. As soon as you start getting serious about producing or even listening to music you’re going to want to upgrade your soundcard! Audio interfaces not only provide you with professional-quality audio, but will often feature multiple inputs and outputs, as well as connections for MIDI devices to further expand your production capabilities.
What is the difference between these types of audio? You’ll likely have seen discussions (or arguments) online about which is better. We know that media like vinyl and cassette tape are analogue, whilst CDs and MP3s are digital audio, but what really is the difference?
Essentially the difference is that analogue is continuous information, whereas digital is discrete. This means that analogue information can take any value, and digital is fixed to set values. Analogue information is what we experience in the real world, and digital data is how our computers can approximate it.
A good metaphor to help explain continuous and discrete data would be a person’s height. Even though we would estimate to 180cm for example, we would need an almost infinitely accurate enough tape measures to pinpoint the exact height of a person down to the individual atoms. The actual height may be somewhere in between 180 – 181cm, but for normal requirements then 180cm is a sufficient approximation. This is exactly how digital audio works – it is a near enough estimation of an exact analogue audio signal.
We could go into masses of detail arguing about the merits of digital or analogue audio, but essentially these days we need both. Analogue will always be more accurate, but you can’t exactly pop a vinyl record into your band’s dropbox to share a song demo can you? And this is the benefit of great audio interfaces – the perceptible difference between analogue and digital audio is always shrinking as technology moves forward!
It’s quite a straightforward process once you’ve got your head around it. For converting analogue to digital, the computer reads, or ‘samples’, the analogue audio at a particular moment in time. Doing this repeatedly traces an outline of the analogue signal. This data is then saved as binary digital information in the computer as a stream of 1s and 0s. And to go from digital to analogue it does the opposite; uses the binary data of sampled points and joins them together to create a “continuous” audio signal. In essence that’s all there is to it!
As you can see form the diagram above, the digital signal is a collection of point samples taken from the analogue signal over time. These point samples then join together to re-create a close approximation of the analogue signal. The distance between the samples is known as the sample rate – see below. For a more detailed explanation of this, have a look at this blog post on l2pnet.com.
This is the speed that the computer takes a reading of the analogue audio signal. CD quality is sampling at 44.1 kHz, meaning that it reads the audio 44100 times a second! This figure is important as it determines the accuracy of the interface, and what frequencies can be detected. The Nyquist-Shannon theorem relates to sample rate, stating that the maximum frequency that can be detected is half of the sampling rate (great link from University of Indiana here for a more detailed explanation).
So a CD can reproduce frequencies up to 22 kHz, above the upper limit of human hearing. This was long considered sufficient, until it was discovered that humans can perceive frequencies above this. Nowadays we see sample rates as high as 96 kHz in professional and home studios, which will be detecting frequencies up to 48 kHz. Essentially, the higher the sample rate the more accurate the audio approximation. There’s also more data, so your audio files will end up larger at higher sample rates.
The graph above shows the first half with a low sample rate, and the second half with a higher sample rate. See how the samples become closer together, and are able to more accurately approximate the shape of the curve. The bars represent the sample data itself. This contains both an amplitude and a time code, so that the computer can reconstruct a sound wave from them in the right order. See how many more samples you would need to get a completely smooth curve – this is why high quality audio files are so large!
This is the resolution of the sample, and determines the number of possible values that the sample can take. A good way to understand this is back to our example of taking someone’s height. This would be measuring in say metres, centimetres or millimetres. Millimetres is the most accurate as it gives us the smallest margin of error.
A ‘bit’ in the digital world is a slot available for a 1 or a 0, which is the binary information that computers use. Below are the first few examples to help explain what this information actually means:
1-Bit: 0, 1 – 2 possible values
2-Bit: 00, 01, 10, 11 – 4 possible values
3-Bit: 000, 001, 010, 011, 100, 101, 110, 111 – 8 possible values
4-Bit: 0000, 0001, 0010, 0011, 0100, 0101, 0110, 0111, 1000, 1001, 1010, 1011, 1100, 1101, 1110, 1111 – 16 possible values
See how adding just one bit at the front doubles the amount of possible values each time. CD Quality is 16-bit, meaning there are 65,536 possible values (hence why we stopped at 4-Bit!). As the resolution massively increases with each bit increase, today we see 24-bit audio interfaces which have 16,777,215 possible values! Again, the higher the bit depth the more faithful the audio reproduction, and the larger the digital file size.
The diagram above shows some examples of using different bit depths on a single frequency. See how the higher bit depths result in a smoother line. This is because the resolution is larger, meaning smaller jumps to the next possible value, and resulting in a smoother and more accurate sample.
This is the time that the computer takes to do all the processing required, and why you may hear a delay. You can adjust this using the buffer size, which is how long you’re giving the computer to do its processing. If the buffer size is lower, the computer works harder to process, but if it’s too low you will hear drop outs and audio errors. Finding the balance that your computer can manage is the trick here! Check out Focusrite’s support page here for more information on latency issues.
These are two useful settings you may see on audio interfaces and can help to reduce errors. Aliasing is the negative effect of sampling at too low a rate, and creates incorrect ‘ghost frequencies’ in the output audio. This goes back to the Nyquist theorem mentioned above. Anti-Aliasing is placing a filter that removes all frequencies above half the sample rate on the way into the converter.
The diagram above shows aliasing in action. The blue line represents the original signal, and the blue dots indicate when the interface takes a sample. The red line is the resultant signal from the samples. As you can see, there aren’t enough samples to show all the peaks and troughs of the original sample, and this generates a frequency in error. To remedy this, the sample rate ought to be increased to capture more detail, however this isn’t always possible. Anti-aliasing simply removes all the audio frequencies where this would become a problem.
Dither is another tool which may be useful to minimise errors. When the audio is sampled, it may fall in between two possible values. Remember our height example, they may be somewhere between 180 and 181cm. We could increase the bit depth, (ie measure in millimetres), but this may result in too large a file, or be above the capabilities of our interface. Dither is a small amount of random noise added to the signal which helps to nudge it closer to one of the values, reducing the possibility of an error.
Have a look at this site from Maarten de Jager which features some great diagrams explaining this in more detail.
Audio interfaces use a load of possible connections to go from the unit to your actual computer. The most common these days is USB, although audio interfaces can come equipped with FireWire and ThunderBolt connections. The consideration here is simply the speed that these connections work at, and what plugs you have available on your computer!
FireWire was faster than USB 2.0, but USB 3.0 trumps this and is on a similar level to ThunderBolt. Macs tend to use ThunderBolt and Firewire, although will have USB ports too. The rest of the PCs out there are more likely to use USB ports. Just check this before you get your interface because different versions are compatible with different computers.
Now away with all the complicated techy stuff, this is a more straightforward consideration. Audio interfaces will have a number of inputs and outputs, and what you need depends on what you want to do! Often the minimum is 2 inputs and 2 outputs. This means you can plug two things in, and send two channels to the computer. If you want to track multiple channels at once, for example a drum kit or a full band, you’re going to need more than 2 inputs and outputs!
Above as an example we have three products from the Focusrite Scarlett range, the 2i2, 18i20 and the 2i4.
Often interfaces for home use will come with an onboard pre-amp, which takes an XLR input and boosts it up to a usable level. The quality of your preamps will do a lot for your recordings, so just be aware that the cheaper audio interfaces will have lower-quality preamps!
A great place to start (after reading this blog of course!) is over on the Rich Tone Music website. The Scarlett range from Focusrite are some of the most popular interfaces and for good reason. These are super-high quality (192kHz 24-Bit) audio interfaces and at a very reasonable price. There’s options ranging from the 2-channel Solo right up to the Scarlett 18i20, so you will find one that suits.
Other brands to look at include PreSonus, whose Studio range comprises some great-quality pre-amps for the price. There are entry-level home-use interfaces available from M-Audio and Mackie, whose new Onyx range looks stylish and operates at a level of high performance.
There’s also some great audio interface packages available, that will also provide software, microphone, and cables. This is brilliant if you’re getting started as it knocks a few things off your shopping list in one swoop!
If you’re looking for something more serious, then check out the Focusrite Clarett interfaces. You will find these in professional recording studios all over the world!
So there you have it, an in-depth look at audio interfaces. You can’t get away from them, you’re going to need one if you want to get recording at home. Fortunately you now have plenty of information informing your decision on which to buy, and a good understanding of what they actually do! Again, check out our related products below to pick one up today and get your home studio ball rolling.
Receive weekly news of special offers, discounts, new products and promotions.
You can unsubscribe at any time.
Rich Tone Music Ltd is a company registered in England with company number 05285423 and VAT Number 870 3855 09