Hardware for high sample rate audio processing in bioacoustics

[phocoena]

Hello, I'm not really sure that this is the forum for a question like this, but I figured I'd ask here just to be sure.

I am interested in learning to record, manipulate, and play back audio for animal species that hear and vocalize in ultrasonic ranges. It is my present understanding that electronic devices I would use for recording and playback need to be capable of a 384kHz sample rate to capture and play sounds above 96Khz (i.e sounds at ~120kHz). I know effectively nothing about the requirements on the analog hardware side of things regarding recording and playback (both in air and water) at these frequencies, and I would love to be pointed to resources on that topic!

However, my main question is regarding digital audio manipulation: Does my computer require some sort of sound card or audio interface to manipulate sound files in ultrasonic ranges? Or is that already possible with any standard PC CPU, and only the recording and playback hardware really matter? I would like to perform pitch shifting and scaling between ultrasonic and human-audible frequencies, as well as low- and high-pass filtering in ultrasonic ranges, for example.

 

[DVDdoug]

The Nyquist sampling theory says the sample-rate must be at-least twice the signal frequency. Simplified, that means at-least one sample for the top-half of the wave and one sample for the bottom-half.

You can get audio interfaces that go to 768kHz, but they tend to be rare and expensive. Also, audio interfaces are made for audio so they may have filtering and may not go to the theoretical maximum. Audio interfaces and soundcards include an analog-to-digital converter (recording) a digital-to-analog converter (playback) and a microphone preamp.

Also you can be "fooled" because, for example, you can configure Audacity to record at a higher sample rate than your hardware supports and the drivers will up-sample without telling you. (Similar to printing a high-resolution photo on a low-resolution printer.)

Regular microphones (and speakers) are also generally limited to the audio range. Audio interfaces work with "pro" stage/studio microphones which have balanced XLR connections. Regular computer mics use an unbalanced connection. They are not interchangeable, but you'll probably be using some sort of special ultrasonic sensor anyway. High frequencies are attenuated in the air, but I assume the distance will be short. I don't really know about water but whales communicate over long distances using low frequencies.

Once you have a digital "audio" file, it's no problem for the software. It's just a series of sample values. "Processing" doesn't involve the soundcard and it's not necessarily done in real-time so processing speed isn't an issue. I just tried a quick experiment in Audacity... I opened an audio file and exported it at 768kHz. I then I re-opened the "high resolution" file and ran a 200kHz low-pass filter which of course did nothing. Then a 200kHz high-pass filter which gave me silence (a flat line). I didn't try any "crazy" pitch shifting.

Note that Audacity is intended as an audio program and people doing scientific work sometimes run into limitations. People doing serious scientific work often use MATLAB (or a MATLAB clone).

 

[nerdstrike]

For transforming audio there's almost certainly some copy-pastable python code lurking on the internet, if you should find the regular audio editing tools aren't doing what you want.

Wav files are endearingly easy to analyse/manipulate. I would be happy to have a play some time if you're not a coder. Only necessary if you the up with hundreds of recordings to process the same way.

Once upon a time I had a Creative Audigy card that had enough onboard grunt to re-pitch audio in real-time, but we haven't needed bespoke hardware for a long time to do most things to audio.

 

[Blumlein 88]

The main bottleneck will be microphones to record at those frequencies.

Here are some:

UltraSoundGate - Avisoft Bioacoustics

Avisoft Bioacoustics Hardware and software for investigating animal sound communication: ultrasound microphones, integrated acoustic data acquisition systems, recording and sound analysis software optimized for applications in bioacoustics.

avisoft.com

These people also have the other gear and software for this purpose. Maybe they have competitors I'm not sure in this particular field.

Home - Avisoft Bioacoustics

Hardware and software for investigating animal sound communication: ultrasound microphones, integrated acoustic data acquisition systems, recording and sound analysis software optimized for applications in bioacoustics.

avisoft.com

You'll notice some gear uses 200 volt phantom power so you'll need specialized recording interfaces to work with that.

Worth noting that at 100 khz air will absorb about 300 db+ at a distance of 100 meters. So these elevated frequencies won't travel far and still be picked up by the microphone. Perhaps you already knew that.

Another place with gear you might need.

Wildlife Survey & Monitoring | NHBS

Here is one claiming to work at 250 khz.

Ultramic UM250K | Dodotronic

The first ultrasonic USB microphone. You only need a PC or Android smartphone and a USB cable to listen and record Ultrasound.

www.dodotronic.com

Software they suggest. Nothing too exotic. Audacity and Reaper, two free bits of software are listed.

Bioacoustic Software | Dodotronic

Windows software SEA Sound Emission Analyzer SeaPro is a software developed by CIBRA and AEST to make easy the realtime analysis and recording of bioacoustic signals in a wide range of frequencies and in different applications. Multi-Instrument Multi-Instrument is a powerful multi-function...

www.dodotronic.com

 

[staticV3]

Does my computer require some sort of sound card or audio interface to manipulate sound files in ultrasonic ranges?

It does not.

For recording, check out the Cosmos ADC which supports 768kHz sample rates and has flat frequency response to 300+ kHz.

 

[kemmler3D]

Does my computer require some sort of sound card or audio interface to manipulate sound files in ultrasonic ranges? Or is that already possible with any standard PC CPU, and only the recording and playback hardware really matter?

You'll want a decent, recent CPU to work with 384 or 784KHz audio since it's several times more data than normal audio, but you don't need anything special. Even the free audio editor Audacity supports up to 384Khz.

The other comments about needing special mics are correct. Most audio gear, especially microphones, is designed for human hearing and so won't work properly at higher frequency. You also need to be careful about mic preamps if you use an analog mic, as they won't be designed for ultrasound either.

 

[BillH]

A data acquisition system by national instruments will offer DACs that easily go into the hundreds of kilohertz at 16 bits.
USB-6423. 16-bit

• 250 kS/s (Max. rate: 1 channel)
• 32/16 AI (SE/DI)
• 4 AO
• 16 DIO

Or
USB-6451

• 20-bit
• 1 MS/s/ch1 (Max. rate: 8 channels)
• 16/8 AI (SE/DI)
• 2 AO
• 16 DIO

We use things like this to build arbitrary digital control and acquisition science experiments

 

[Philbo King]

Hello, I'm not really sure that this is the forum for a question like this, but I figured I'd ask here just to be sure.

I am interested in learning to record, manipulate, and play back audio for animal species that hear and vocalize in ultrasonic ranges. It is my present understanding that electronic devices I would use for recording and playback need to be capable of a 384kHz sample rate to capture and play sounds above 96Khz (i.e sounds at ~120kHz). I know effectively nothing about the requirements on the analog hardware side of things regarding recording and playback (both in air and water) at these frequencies, and I would love to be pointed to resources on that topic!

However, my main question is regarding digital audio manipulation: Does my computer require some sort of sound card or audio interface to manipulate sound files in ultrasonic ranges? Or is that already possible with any standard PC CPU, and only the recording and playback hardware really matter? I would like to perform pitch shifting and scaling between ultrasonic and human-audible frequencies, as well as low- and high-pass filtering in ultrasonic ranges, for example.

It sounds like you would be better off looking for data acquision (a/d) and d/a gear designed for ultrasonics. Tektronix, National Instruments Keysight (or whatever they're called this week, formerly Hewlett-Packard) and several other companies serve this market. But it definitely isn't going to be cheap; for a rough guess, 20-50k$ to get started, unless you can score it on the surplus market.

 

[Lambda]

what about the e1da cosmos adc?
it supports high Sampling rates

 

[AnalogSteph]

You'll want a decent, recent CPU to work with 384 or 784KHz audio since it's several times more data than normal audio, but you don't need anything special. Even the free audio editor Audacity supports up to 384Khz.

The limit in Audacity is actually higher than that. You can enter almost arbitrary numbers as a project rate, although I found at least 2.1.3 can be quirky... it wouldn't accept some combinations entered via the top row numbers but then the same via the numpad worked. (I could generate tones just fine at 12,288,000 Hz, for example, and an absurd 98,304,000 Hz still doesn't seem to pose a problem. I could even export a WAV file, though file sizes go over 4 gigs easily and reimporting those doesn't work. Newer versions support WAVEX.) Whether it can record like that is anyone's guess. Might need a custom build with ASIO support compiled in.

384 or 768 kHz shouldn't be a challenge to any modern-ish CPU if latency isn't too much of a concern. When 96 kHz first went mainstream, people had 1 GHz class processors, and back in the day I did plenty of 16/44 kHz recordings on a Pentium 90. An e-waste grade system would probably be fine assuming it was decently powerful in its day (not some wimpy AMD E-series APU or something, but a C2D E8400 or Sandy Bridge i3 ought to do it).

 

[phocoena]

Thank you for all the links and brand mentions. Certainly gives me a bit more leads and options to look at breaking into bioacoustics, which like everything I seem to take interest in these days, is very expensive and technically overwhelming. Are there any text resources anyone could suggest I read to get a decent start in understanding the more technical language in the field of audio science and hardware? About the closest I've gotten to it involves eurorack modular synthesizers...

 

[kemmler3D]

Thank you for all the links and brand mentions. Certainly gives me a bit more leads and options to look at breaking into bioacoustics, which like everything I seem to take interest in these days, is very expensive and technically overwhelming. Are there any text resources anyone could suggest I read to get a decent start in understanding the more technical language in the field of audio science and hardware? About the closest I've gotten to it involves eurorack modular synthesizers...

The Audio Expert by Ethan Winer is a good general purpose book for audio knowledge.