Meer Power
Boring synth stuff

32-bit Audio is Cool, Sometimes

I haven’t posted anything in a while, so to start things off again, let’s talk about something that’s pretty dull, but has huge implications. The importance of 32-bit audio. To make it better, there will be probably be a sheep metaphor. This won’t be an incredibly in-depth article, but I’ll let you know what you need to know to make good stuff. This article totally sucks, don’t say I didn’t warn you.

Sooo… Digital audio is basically just a huge stream of numbers that represent the amplitude of an analog signal. The bit depth refers to how many bits each number is made of, since this is all binary stuff. With a greater bit-depth, more numbers can be represented. 2-bit audio only has two bits, so there can only be 00, 01, 10 and 11, if you know what I mean. That’s sort of important, but this article will sort of be about the practical implications, and not a bunch of math. 32-bit audio, then, has a crapload more zeroes and ones and can represent a huge amount of different numbers.

When audio is recorded digitally, usually it begins as an analog signal like electric current from a microphone or something. The analog signal is turned into a digital signal at the.. analog-digital convertor. The convertor chops up the analog signal into little pieces, at the sampling frequency, and then turns the pretty continuous analog signal into a quantized digital signal. If the convertor has a sampling rate or frequency of 44.1khz, and a bit depth of 24 bits, it will chop the signal up 44.1 thousand times a second, and assign a number consisting of 24 zeroes and ones to each ‘chop’ or sample. This paragraph was supposed to be about how you can’t record at 32 bits, A/D and D/A convertors can’t do it, everything these days is recorded at 24 bits. But, inside the computer or wherever it’s recorded to, using 32 bits is super useful, sometimes.

Since there is a limited number of zeroes and ones available to represent amplitudes, there is a limited dynamic range, and there is a point where the amplitude is high or loud, the bits can’t represent it. There’s a point that you can’t go over, no matter how hard you try, and this is called clipping. But 32-bit audio usually uses a different kind of math, so you can clip if you want to. If you want to know the difference between fixed-integer and floating-point numbers, you can try wikipedia. In FL Studio, everything inside of it runs at 32 bits, floating point, even if it was recorded at 24 bits, fixed-integer. BUT, the digital audio needs to be 24-bit before it gets converted into an analog signal that you can hear. If you want to have stuff clipping all over the place inside of FL, or it happens accidentally, don’t worry, as long as it isn’t clipping before it’s converted into 24 bits, you’ll be okay.

Let’s talk about sheep. You have a sheep farm, on a mountain. Each side of the sheep farm is a sheer cliff and 10 000 feet down there are pointy rocks. With 24-bit audio, if your sheep walks off the sheep farm, it falls to it’s death, and you cannot get it back. 32-bit floating point audio puts an invisible glass floor out from the edges of the sheep farm, so they can wander wherever they want, without dying. As the shepherd, you must keep in mind that this convenient glass floor will disappear at sundown, so you have to get all your sheep off the glass and onto the farm before then, or they die.

In conclusion, clipping is bad.

Advertisements

3 Responses to “32-bit Audio is Cool, Sometimes”

  1. this post is both amazing and informative.

  2. Man, the sheep analogy is GOLD! I have never ever heard it expressed so well!
    lol… oh and yes, 32bit audio is very important. You want to preserve your audio signal at its highest quality before dithering. 32bit is the answer. Nice.

    dauntless

  3. I’ve been thinking recently about the loudness war and how we should all just have automatically adjusting speakers/media players that make everything the same volume. But then people would get lazy and gradually we’d just start hearing more white noise. I don’t know…


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: