Recording @96Khz

This section is now closed.

Audacity 1.3.x is now obsolete. Please use the current Audacity 2.0.x version.

The final version of Audacity for Windows 98/ME is the legacy 2.0.0 version.

Recording @96Khz

Permanent link to this post Posted by Interloper » Fri Jul 25, 2008 10:10 pm

I have an M-Audio Fasttrack Pro. Generally I use this in Linux, but need to use windows to set the device to its maximum sample rate of 96Khz. However, I can't record in Audacity at this rate. I've tried setting the default project sample rate in Audacity to 96Khz, which it accepts. I have set the M-Audio control panel so that only inputs are enabled, and 24 bit depth is selected, which then reports a maximum sample rate of 96Khz. When I open Audacity, it describes a project sample rate of 96Khz, and an "Actual rate" of 96Khz. However, as soon as I hit record, the "actual rate" changes to 44.1Khz. I can't seem to find any options anywhere to set the recording rate so that Audacity does actually record at 96Khz?

I'm expecting that Pro-Tools will allow recording at the faster rate, but even though I have my laptop with me at home, my dongle is at work, so I can't use it. :x

I'm attempting to record bats. I don't know which frequency the local species uses, but after reading some bat recording pages, it seems to be worth a try.

Any advice?
Interloper
 
Posts: 5
Joined: Fri Jul 25, 2008 9:58 pm

Re: Recording @96Khz

Permanent link to this post Posted by steve » Sat Jul 26, 2008 3:18 pm

Which version of Audacity are you using?
Try both 1.3.4 and 1.3.5

Interloper wrote:I'm attempting to record bats. I don't know which frequency the local species uses, but after reading some bat recording pages, it seems to be worth a try.

That explains why you want the high sample rate, but do you have a microphone that will go high enough?
Most dynamic mics roll off their frequency response around 15 to 16 kHz, and condenser microphones around 18 to 20 kHz.
9/10 questions are answered in the FREQUENTLY ASKED QUESTIONS (FAQ)
steve
Senior Forum Staff
 
Posts: 34850
Joined: Sat Dec 01, 2007 11:43 am

Re: Recording @96Khz

Permanent link to this post Posted by Interloper » Sat Jul 26, 2008 4:42 pm

I was using 1.3.4, I'll try 1.3.5. TBH I was hoping it was my unfamiliarity with Windows that was causing the problem, and that I'd missed something obvious.

I'm using an electret condenser microphone. A Sony ECM-77B. Some of the bat pages I read said that for some species of bats, an electret condenser microphone will pick up frequencies higher than the claimed 20Khz or so, but that signal to noise may be poor. Given that they were recommending cheap condensers, I thought using an expensive one had some chance of success. If it doesn't work, which may well depend on my local species of bat more than anything. If that doesn't work, I might build a frequency divider circuit, to be driven by a transducer.
Interloper
 
Posts: 5
Joined: Fri Jul 25, 2008 9:58 pm

Re: Recording @96Khz

Permanent link to this post Posted by steve » Sat Jul 26, 2008 5:45 pm

I'm not sure why the sample rate should change when you start recording, but the correct step by step procedure for setting up Audacity for recording at 96 kHz is as follows:

1) Open Audacity
2) "Edit > Preferences > Audio I/O" and check that the correct Recording and playback devices are selected
3) Go onto the "Quality" tab and select the sample rate and bit depth (make sure these are valid for your hardware.
4) CLOSE Audacity and RESTART (this is not always necessary, but in some cases Audacity requires the restart for the changes to take effect

Note that if you import a file that is a different format, Audacity will use that format (it will override the default that you have set).

Yes some electret and condenser microphone will go over 20kHz, but an expensive one may not go any higher than a cheap one. Cheap electret microphones have very little active circuitry in them - sometimes no more than a single FET (a type of transistor). On the other hand, expensive microphones may have quite complex circuitry. For the microphone to have a response above 20 kHz it is necessary that the built in circuit also has a response beyond 20 kHz

Also, it will be necessary to keep the microphone lead quite short, preferably using low capacitance shielded cable. It is also possible that your sound card may limit the frequency response. Good quality sound cards will generally employ a low pass filter to reduce interference, so it will also depend on the frequency response of the sound card.

Audio circuits will typically have much more noise at high frequencies, so this is another thing that you will be fighting against.

If your sound card will support the necessary high frequencies, it may be possible to interface an ultrasonic sensor directly to the microphone input and record that way - alternatively, trawl through the spec. sheets of electret capsules, some of the very cheap capsules have surprisingly good specifications.

The Panasonic WM-61A has a good high frequency response and low noise. It is available from Digi-key.
9/10 questions are answered in the FREQUENTLY ASKED QUESTIONS (FAQ)
steve
Senior Forum Staff
 
Posts: 34850
Joined: Sat Dec 01, 2007 11:43 am

Re: Recording @96Khz

Permanent link to this post Posted by kozikowski » Sat Jul 26, 2008 6:13 pm

<<<Sony ECM-77B>>>

I can't find the actual curves for this microphone, but sometimes you can extrapolate beyond the ones they give you. Unless they don't give you one.

I can tell you Sony isn't stupid and they only give you a 3dB specification of 20 KHz is because it looks grand on paper and lets them sell a lot of mics. I'm betting the curve drops as a very large rock after that. No recording benefits from microphone response above and below the human hearing points. Virtually all vocal microphones have a roll-off switch for the low end and I can imagine how happy you would be to have your Boston Pops recording trashed because of ultrasonic sounds on the stage.

So I wouldn't be shocked to learn that the response is aggressively limited after that.

Any way to test it? Put a signal generator into a very small speaker and see how your microphone does. My General Radio "Audio Oscillator" doesn't run out of steam until 1 MHz. My hearing won't go that high.

You can buy dog whistles that will generate significant energy at ultra-sonic frequencies.

Koz
kozikowski
Forum Staff
 
Posts: 26300
Joined: Thu Aug 02, 2007 5:57 pm
Location: Los Angeles

Re: Recording @96Khz

Permanent link to this post Posted by Interloper » Sat Jul 26, 2008 6:32 pm

Thanks Steve. I can't change the length of the cable on the Sony microphone as it isn't my microphone. I can for other microphones.

The sound card I'm using is a M-Audio Fasttrack Pro. I set it up in 96Khz/24bit sampling mode, which I can only do in Windows. I'd presume that the circuitry would not lowpass filter @ significantly less than 48Khz, as it would seem to defeat the purpose of having 96Khz sampling. Lowpass filtering at ~48Khz would be necessary to prevent aliasing anyhow, and would cut out bat calls higher than that. But that's always going to be the problem with this approach.

I've seen the Panasonic WM-61A recommended on "bat sites" (for want of a better description). But while it is extremely inexpensive, I'd need to get around to buying it and building at least some support hardware. I'd like to try that after trying easier things first. There is always the option of building a frequency divider circuit. That has considerable limitations, mainly that any amplitude envelopes of the bat calls are lost. But at least I could be reasonably sure that it would work no matter what frequencies my local bats are using.

Do you know if large diaphragm condensers would be a good shot? Something like an AKG414. I thought, perhaps hurridly, that the large diaphragm would be insensitive to high frequencies solely because of its size. Is this likely to be wrong. I could keep the cable short on the 414 (or other mics) due to their not having a permanently attached cable.

Interfacing an ultrasonic sensor directly to my microphone input sounds a very interesting idea. Particularly if I could cannibalise the sensor for a frequency divider circuit should the recording @96Khz option not work.

Thanks Koz. I have access to a large range of microphones as well as signal generators. And should be able to find a small speaker somewhere.
Interloper
 
Posts: 5
Joined: Fri Jul 25, 2008 9:58 pm

Re: Recording @96Khz

Permanent link to this post Posted by steve » Sat Jul 26, 2008 7:47 pm

Interloper wrote:I'd presume that the circuitry would not lowpass filter @ significantly less than 48Khz, as it would seem to defeat the purpose of having 96Khz sampling.

Not really. As you get close to the nyquist frequency, representation of waveforms gets very iffy. Consider a sine wave at around 20kHz and a sample rate of 44.1 kHz, that gives you around 2 samples per cycle - not a lot of detailed information in that. Push your sample frequency up to 96kHz and for that same sine wave you will have 5 samples per cycle - still not a great deal, but a lot better. So 96kHz can do "20Hz to 20kHz" audio a lot better than 44.1kHz sample rate can. But what's the point of M-audio providing a frequency response beyond 20kHz (unless they have identified a niche in the market for people that want to record bats and dog whistles).

That's not to say that it won't work, just that it may not.

Interloper wrote:Do you know if large diaphragm condensers would be a good shot? Something like an AKG414.

I would expect that an AKG414 would roll off above 20kHz. Large diaphragm condenser microphones are capable of responding to extremely high frequencies (the moving mass is extremely low), but I would think that there would be design considerations such as rolling off the response well below the resonant frequency of the diaphragm, and avoiding picking up ultrasonic interference (such as from radio gear and lighting dimmers).

I think this is going to call for a fair bit of experimentation on your part. I'd go for trying to record a dog whistle or similar before venturing out with the bats.
9/10 questions are answered in the FREQUENTLY ASKED QUESTIONS (FAQ)
steve
Senior Forum Staff
 
Posts: 34850
Joined: Sat Dec 01, 2007 11:43 am

Re: Recording @96Khz

Permanent link to this post Posted by kozikowski » Sun Jul 27, 2008 1:26 am

A word on the Field Effect Transistor used inside Electret Condenser microphones. It's not a microphone amplifier. It's in there to transform the signal at the condenser which is reasonable level but very delicate with no horsepower behind it, to the same signal with enough authority and force to make it down the cable. Not only isn't the signal any louder when you do that, you actually lose about 10% in the process.

Used that way most FETs will manage frequencies through audio and well up into the radio bands. This cries out for limiting so turning on a radio (cell phone, BlueTooth Adapter, WiFi Computer, ) next to the microphone doesn't create problems.

And yes, the 96000 thing isn't to extend the upper end. It's to make what you have a lot more accurate and with much lower distortion. 44100 will not manage frequencies up to 20 KHz without some interesting tricks like dithering, modulation, and circuit ringing. It depends on most people's inability to hear all that working. Some people can and they were very vocal about CD quality in the early years.

96000 will manage 20 KHz audio straight with no tricks.

Koz
kozikowski
Forum Staff
 
Posts: 26300
Joined: Thu Aug 02, 2007 5:57 pm
Location: Los Angeles

Re: Recording @96Khz

Permanent link to this post Posted by Interloper » Sun Jul 27, 2008 6:47 am

I'm starting to think that using analogue electronics to convert the bat sounds to the audible spectrum is likely to be the only real possibility. Frequency divider circuits are simple. And I've emailed a work colleague with an electronics background about others.

Though, I'd still like to know how to record @96Khz for future reference.
Interloper
 
Posts: 5
Joined: Fri Jul 25, 2008 9:58 pm

Re: Recording @96Khz

Permanent link to this post Posted by steve » Sun Jul 27, 2008 8:53 am

I am so hoping that this works. It would seem a shame to have to use a frequency divider - somehow moving it away from the actual bat sounds. Let's hope that you have big butch bats with (relatively) deep voices.

Some posts ago we were talking about recording at 96kHz - any joy there?
9/10 questions are answered in the FREQUENTLY ASKED QUESTIONS (FAQ)
steve
Senior Forum Staff
 
Posts: 34850
Joined: Sat Dec 01, 2007 11:43 am

Next

Return to Windows



Who is online

Users browsing this forum: No registered users and 9 guests