Sampling Rate/Playback Rate.

Hi there -
Recently, I’ve come to Audacity due to a need to work with high sample rate audio (192kz and higher)…

Having said that, I do not always work with a high sample-rate interface and frequently audition what I’m listening to with the 44.1 or 44.8khz samplerate built in to my soundcard.

I am curious what happens to the higher-frequency audio when I listen to a 192khz file at 44.1. My guess would’ve been that there would be aliasing, but it seems to play “normally”.

Is there a way to know what goes on behind the hood with this?

Thanks.

44.8khz samplerate built in to my soundcard.

Can we assume that’s really 48000 video rate?

need to work with

What does “work with” mean?

there would be aliasing, but it seems to play “normally”.

How old are you? If you’re anything past about 40 with a history of club attendance, your hearing doesn’t go up much beyond about 14KHz. None of these gymnastics are going to be audible to you at all.

Yes, you can certainly send test signals through the system and see what happens to them, but I don’t know that you can inspect the sound and be able to tell.

Audacity can have some messy surprises for the scientists and the experimenter. It’s not a WAV editor. It’s a sound production editor and if there is ever a split between having a product be scientifically accurate or sounding good, Audacity will almost always take the good sound route.

That and none of this is going to make it past the analog filters in your soundcard.

Why are we up at 192kHz sample rates? That will pass some navigation radio transmitters. Not the content, the actual radio.

Koz

Hi Koz -
Sorry about the typo: 48khz is the sample rate I am working with, not 44.1.

I understand that human beings cannot hear about 20khz. I know that, personally, I can barely hear above 15khz.
But there are many things in the world that make sounds that humans can’t hear - and technology helps us to hear and identify them.

My question is about sampling rate, though.

In a program like Max/MSP, when you extend an oscillator beyond 20khz (on a 44.1 or 48khz card), it starts to alias. You begin to hear the frequencies starting over from 0hz.

So, I am curious if Audacity has a similar aliasing effect. If not, I would like to know what “happens” to the audio information that falls outside of the playback sample rate.

I can add that, ultimately, the goal will be to down-sample and pitchshift the audio into human hearing… but before that, it is important to me to understand how it is working with the software at hand.

But Audacity does seem to support these higher sample rates, where other audio softwares do not. If there is no interesting having people use a 192khz sample rate, why is it included in the program?

Yes, sorry I meant 48khz.

I am working with some ultrasonic recordings of animals. I know that I can stretch+pitch it all down, if necessarily.

I’m curious what is happening in the meantime, though. In a program like Max/MSP, when you exceed 24khz (through a 48khz samplerate), you start to get aliasing. The frequencies start again from 0hz.

In Audacity, I’m not sure the same thing happens. If it did, wouldn’t that mean some of these ultrasonic sounds would be back in the human hearing range?

So, I am curious if Audacity has a similar aliasing effect. If not, I would like to know what “happens” to the audio information that falls outside of the playback sample rate.

As Koz says, Audacity usually does the “right thing” (which means low-pass filter to prevent aliasing when downsampling). If you go into Preferences there are “quality settings” and I don’t know how “bad” the low-quality setting is…

So, I am curious if Audacity has a similar aliasing effect. If not, I would like to know what “happens” to the audio information that falls outside of the playback sample rate.

Audacity will (usually) filter-out “invalid” frequencies.

There should also be hardware filtering in your soundcard so you can’t record beyond the Nyquist limit. And again as Koz says, we don’t know what your soundcard is doing with ultrasonic frequencies. It may be limited to audio frequencies, even at high sample rates. There are soundcards specified up to 192 or even 384kHz, but you rarely see frequency response specs up to the Nyquist limit (if you can find frequency response specs at all).

Any soundcard can play (or record) a “high resolution” file but often the actual hardware is running at 48kHz and the drivers are downsampling (playback) or upsampling (recording) without telling you. Any-old soundcard can “play” a 192kHz audio file, but you may only get the audio range (up to around 20kHz).

The same thing happens if you view a high-resolution image on a low-resolution monitor or print on a low-resolution printer. The Windows drivers "secretly’ downsample/downscale and you may never know what’s happening…

I’m curious what is happening in the meantime, though. In a program like Max/MSP, when you exceed 24khz (through a 48khz samplerate), you start to get aliasing. The frequencies start again from 0hz.

I don’t know what these programs are doing. A proper downsampling algorithm will always low-pass filter as necessary as will a proper analog-to-digital converter (or soundcard).*

A good digital signal generating algorithm (such as the one built-into Audacity) will refuse to do something impossible. i.e. Audacity won’t allow you to (attempt to) generate a 30kz waveform at a 48kHz sample rate.

But Audacity does seem to support these higher sample rates, where other audio softwares do not. If there is no interesting having people use a 192khz sample rate, why is it included in the program?

Well… It is one of the “standard” audio sample rates. The “pro studio standard” is 24-bits/96kHz.


\

  • Filters are imperfect so you can get some low-level aliasing. With “normal audio” it’s usually filtered well-enough to be to be inaudible.