Getting Audio Signals Into The Digital Audio Workstation

A Bit Of Audio Theory

More......

All sounds that the vocal chords project and ear receives are analog; these days, all audio processing inside your computer is digital (bit streams of digital 1’s and 0’s) encoded and decoded in such a way as to represent its analog counterpart. The purpose of an audio interface is to faithfully convert the analog input into a digital form. The computer you are using has one of these. It is almost certainly a single channel device that converts the input into a 16 bit word sampled at 44.1kHz, exactly the same format as every CD you have ever played. So what wrong with using this to record? Actually, not much, except of course that you can only record one channel at a time. Oh, and 16 bits isn’t really anywhere near enough!

More Bits

Consider this; 16 bit analog to digital conversion means that the amplitude of the recorded signal is split into 65,536 levels. Each of the 65,536 levels is represented by a 16 bit word. This might seem like a lot, and if every sound you recorded was at or near the loudest sound you wanted to record, you’d be OK. But what if you had a softer sound in the mix, your recording had a high dynamic range? If you had a 40dB, (100:1) dynamic range, you’d only have 655 bits to represent the quietest sound. These days, the generally accepted sampling depth is 24 bits – a whopping 16,777,216 bits, a 256 improvement in dynamic range. Now your quietest sound has 16,777 bits at its disposal; a huge improvement.

Now let’s consider the sampling rate. 44.1kHz is used in CD format; 65,536 levels sliced in time at 44,100 times a second. If you do the math, a one minute recording will need about 10MB of disk space, and a 3 minute song would require about 30MB. So you can see that this is very convenient for CD with a capacity of 700MB being able to hold more than 20 songs. Now, if you record at 24 bit, 44.1 kHz, you would only get 15 45MB songs on a CD. If you recorded at a higher sampling rate, such as 88.2 or 96 kHz, this would reduce it to around 7 or 8 90MB songs.

More Hertz

So why record at anything better than the industry standard CD quality? I would suggest that for most of us, recording at 44.1 kHz will be quite adequate. I personally cannot tell when my classic rock band is recording at 44.1kHz or 88.2kHz, and doubt if you will either, and quite frankly, you will probably get a much bigger improvement in sound quality by upgrading the quality of your A/D converters or pre-amp or compressors etc, than you will by increasing the sampling rate. If you have reasonable quality gear, or are seriously considering upgrading in the future, use 88.2kHz. This is what I use. As for recording at 192kHz, a waste of disk space, you’ll never notice the difference.

However, far as bit depth is concerned, I highly recommend that you record at 24 bits (or greater) because:

► You will hear a noticeable improvement in the dynamic range. Quiet passages will sound as clear as the loud ones; your noise floor will be lower.

► You won’t need to operate your audio interface on the edge of its maximum input all the time, you will be able to drop it by 3dB or so and give yourself more headroom. You won’t need to use as much compression.

►You will generally hear a more clarity and presence from the individual instruments in the mix.

So, how does the digitized sound get out to your speakers. Well, just as the audio interface used an Analog to Digial Converter to get the sound into the computer, it uses a Digital to Analog Converter to get it out, Easy huh?

Back To Top

Buy.com is a Great Place to find Computer Hardware and Software