Sample Rate 101 — Which Should I Choose

Transcript:

Alright. Now, let’s talk sample rates.

Here’s the thing about this debate, and what you should choose. In terms of what actually makes a fidelity difference in terms of the overall product of whatever you’re mixing or recording, encoding data is the biggest difference. Bit-depth is going to be well below that in terms of how noticeable it is, and sample rate is even below that. We’re really splitting hairs when it comes to sample rate.

So know that if you do everything in 44.1, you can still get top of the line, top quality sounds without losing anything. You’re not making compromises.

The reason being is that sample rate is not about taking slices of a sound like a film shutter. That’s not how it works. We’re taking slices of data. That data can accurately reproduce a continuous wave. We’re not losing anything. The only limitation to that is how fast that wave can actually be.

If the wave is moving faster than the time shutter allows, then suddenly we start having problems, because our time based data doesn’t know where to put things. It becomes a problem, and it starts saying, “Okay, this super fast wave is actually this slower wave.”

That’s called aliasing, and it creates a form of distortion that’s annoying. It thins the record, it causes weird sound. So, we don’t like it.

That said, if we just chop off everything that’s over a certain frequency range, we don’t have to worry about those super fast waves, and that’s what sample rate is determining. It’s determining where we’re chopping off those points.

So, the sample rate for 44.1 is chopping everything off at about 22kHz. Well, human hearing basically goes from what they call 20 to 20. 20Hz to 20kHz.

Now, I’ve tested my hearing many times. My hearing goes to about 18.5kHz. That’s really, really, really good. It is considered exceptional for a male in his 30s to have hearing that goes up to 18.5kHz, and I can even squeak by and discern a little bit going on at 19.

Even if you were a child with perfect hearing, you still would unlikely be able to hear 22kHz. That’s about where dog whistles are starting.

So, do we need that stuff? I would say no. Personally, I don’t even want to hear stuff that’s that high, it’s annoying. So, I don’t feel like I’m losing anything.

That said, okay, we do have to use a low-pass filter. Yes, that low-pass filter is going to have certain effects at the upper resolution of our sound. Alright, so we might want something higher.

There’s nothing wrong with that either. Your CPU, if it can take it, sure, work 88.2. I like to record 88.2, I don’t think it’s a bad thing. Then I’ll down-sample and mix at 44.1. Most of my plug-ins have oversampling anyway. They’re going to read as if there’s enough information for 88.2 to work, and then down-sample again so that it doesn’t need to.

Alright, what about super high sample rates like 192? Well, again, you are getting technically more resolution. We can’t hear the vast majority of that resolution, but the bats will appreciate your record, I’m sure.

The problem is that we’re still working in analog gear when it comes to the going in and coming out of the computer. We have converters, and they are hardware. The fact is that 192 thousand samples per second, that’s getting into nanosecond world. You know, there are certain physical limitations that come up, and you actually get distortion from the step gates not being able to fire accurately at that speed.

So for accurate representation, 192 right now in 2014 is actually worse. Then again, inaccuracies can be pleasing, so you might like it. There’s nothing wrong with that either.

Anyway, the bottom line is no matter what sample rate you’re choosing, the actual end effect on your audio is so subtle, it doesn’t really matter, so if your CPU is going to start interfering with your ability to mix properly or record properly, just go with 44.1. There’s nothing wrong with it.

If you happen to have a super computer that can handle higher sample rates, go for the higher sample rates.

Again, the difference is going to be barely perceivable, if that, but why not? You grab the highest sample rate that you can effectively as long as it’s not interfering with the rest of what you’re doing.

The bottom line is even if you’re recording 16, 44.1, you can still get commercially viable in 2014 sound. There’s nothing wrong with it. I usually work 24-bit, 44.1 for mixing. I usually work 24-bit 88.2 for recording, just because I’m working with systems that can handle that fairly easily. Sometimes I’ll even mix 88.2, but that’s only because I’m preserving the optimal information for longevity’s sake and because there is the slightest pinch of difference.
However, that 0.1% difference is not what’s making or breaking my record. Trust me.

And if the higher sample rates and bit depth were interfering with my ability to work, that would have so much of a greater impact on the actual outcome of the record. I feel that the detriment vastly outweighs whatever fidelity I would be getting.

Alright, guys. Until next time.

Matthew Weiss

Matthew Weiss

Matthew Weiss is a Grammy nominated and Spellemann Award winning audio engineer from Philadelphia. Matthew has mixed songs for Snoop, Sonny Digital, Gorilla Zoe, Uri Caine, Dizzee Rascal, Arrested Development, 9th Wonder, !llmind & more. Get in touch: Weiss-Sound.com.
Smiley face
Recommended