Join Dave Schroeder for an in-depth discussion in this video Sample rate, part of Digital Audio Principles.
- View Offline
The process of Converting Analog to Digital is one of the most important processes when it comes to working with digital audio. It's important to understand what goes into it, because you'll be making decisions about how analog is converted to digital in a lot of different situations. So I want to talk about how we do that, and what it actually is. To be able to take sound and use it in the computer or in the digital domain, we need to come up with numeric representations of those analog signals, or those electric signals that are moving through our cables and wires.
So there are two sets of values that determine how accurate these representations are. Sample rate and bit depth. Let's take a look at sample rate first. Sample Rate is actually just what it sounds like. It's the rate of taking samples. It's the amount of samples we take per second, or there are a number of times each second, we look at a sound and take a measurement of that sound. So the number of times we're analyzing, or looking, and scrutinizing a sound to see what's actually there. Now a Sample is an individual piece of information.
So each time we look, we say, hey, what's there, how louder, what frequencies are there. Then we write it down in a little notebook, which becomes a digital piece of information. It's basically the process of coming up with the numeric representations of what exists. Sample rates comes into play, because how many times you look has a big effect on how accurately you see what you're looking at. Needless to say with audio, we need to take a lot of samples to get a pretty accurate picture. There are a lot of things going on there, a lot of changes.
So sample rates are expressed as a frequency of samples per second. So they're expressed in hertz. We take a lot of them. A sample rate of 44.1 kilohertz in one single second, we're taking a look and sampling that sound 44,100 times. So that's a lot of looks. So every time we take a look, we record a sample, which is more or less numerical data. That is digital information that our computers can read and understand. The higher that sample rate is the better the sound quality we'll have.
Let's take a look at a few examples. So Higher sample rates = greater accuracy. I've a couple of visuals here. This is a analog signal coming in. These little gray bars represent taking a sample. So this is our sample rate. So these red dots kind of represent what we're seeing. When we take that look or take a sample, this is what we find. We know that this wave is at this point, at this time. This wave is at this point, at this time. So you can see here we have a lower sample rate, and so you get fewer samples.
It's pretty obvious. Higher sample rate, more samples. So we're taking more looks, and getting more little bits of information, more little measurements, and more little numeric representations of what's happening here. So in the next slide, we'll connect those dots. You can see that if we connect the dots at the lower rate, we miss out on a fair amount of information involved in the slide. In this one, we really miss this whole big spike. We missed some of the peaks for sure. We get a lot of that kind of middle information, but we missed out on some of the real extreme things happening here.
With a Higher sample rate, we get a much more accurate representation of that analog sound. This is why higher sample rates are better. So as you would expect, the higher the sample rate, the more accurate the reproduction of the sound, which in a way of saying the better sound. So the higher the sample rate is the more accurate the analog to digital conversion will be. Here are a lot of the sample rates you'll come into contact with digital audio. From the lowest being 8 and right now the highest being 192. If you go and buy an audio CD at the store, it's going to come in at 44.1 kilohertz.
Now 44.1 is really considered the minimum sample rate required to achieve high-quality digital audio. We come to this rate as a result of the Nyquist theorem, which concludes that a sample rate should be at least twice as high as the highest frequency you're trying to record or sample. Since humans can hear up to 20 kilohertz, a sample rate of twice that or 40K is required. Now the reason we got up a little bit further to 44.1, gets a little bit technical. I'm not going to go into it in this title. But certainly if you want to find out more about it, you can look up the Nyquist Theorem.
Any sample rate lower than 44.1, you can usually detect degraded sound quality. We'll listen to a few examples. Rates higher than 44.1K such as 48, or 96, and 192 are used quite a bit now in digital audio. They provide some really excellent results. We'll talk a little bit more about kind of some of the trade-offs of using different rates. It's not as simple as just saying, well, let's use the best rate. There are a few things you need to take into account. We'll talk about that in just a little bit. So let's go ahead and check out a few examples of things at different sample rates.
We'll start with something at 44.1. (music playing) So you can hear that's pretty good quality. It's pretty even. And this is the standard, if you buy a CD in the store or something like that, this is what you hear. It's not the highest quality available. It's generally accepted as very, very suitable quality. Now let's check out something at a lower rate at 22. (music playing) So at 22, we're actually taking half as many samples. So we're getting half as much information.
And obviously, we're not going to have as accurate representation of the sound. Finally, let's check out one at 8K. (music playing) So hopefully, it's easy to hear that there is a big difference in the quality between 8K and 22, and even more of a difference between 8K and 44.1 kilohertz. So at 8K, we're taking more than five times fewer the number of samples than we are at 44.1 kilohertz. The sound is noticeably not as good.
So that's a quick look at sample rate. Next, we'll look at bit depth. Then we'll think about how when it combined with sample rate will affect the overall quality of our digital audio.