One very intriguing feature of InvenSense's TDM mics is that currently up to 16 elements can be daisy-chained in an array and sampled simultaneously at 48kHz. Data is then clocked out at ~3MHz on a single data line, 24 two's complement bits per element. Not good for ultrasonics, but for an audible-audio steered beam it'd be great. I suspect it would take a lot of array processing power.
See pg 12: https://www.invensense.com/wp-content/uploads/2016/10/DS-000121-ICS-52000-data-sheet-v1.1.pdf
Tom
Reply by Steve Pope●October 22, 20162016-10-22
angrydude <simfidude@gmail.com> wrote:
>I also would say that (sufficiently large) array of omni-directional
>microphones is much better than few directional mics - you can point it
>in any direction at any point in time, e.g. after recording is done
Excellent idea.
I know form folks in the National Park Service that they deploy
many many audio sensors for wildlife detection. If they are not
using the above beamforming approach presently, they certainly should be.
Such a technique might also make it possible to detect bat sounds and
other animal sounds with an overhead system, such as a blimp.
Steve
Reply by angrydude●October 22, 20162016-10-22
On Saturday, October 22, 2016 at 6:27:14 PM UTC-4, Tom Becker wrote:
> > Have you tried making your own elements?
>
> No, but maybe someday I'll tinker with a 30mm Polaroid 200V electrostatic but I am, for now, put off by the patterns that large-diaphragm sensors exhibit. High voltage is surely also more problematic.
>
> I don't want narrow lobes. I enjoy recording in stereo to hear the bat flight dynamics; omnidirectional, figure-eight and cardioid patterns, sometimes recorded in Mid-Side which allows manipulating the pattern later, suit that well.
>
> Most government, commercial and academic bat freaks are seeking species identification and population counts best achieved with more directional mics (for which more than one company gets $1200 each: http://www.ultrasoundgate.com, for example).
>
> The equipment is often exposed to dew that requires conformal coating and heaters to combat. My mics have eight resistors that warm the assembly to prevent condensation and consequential corrosion - an instant serious noise generator - and that's at 12-48V. I expect 200V must be a PITA.
>
>
> Tom
"Most government, commercial and academic bat freaks are seeking species identification and population counts best achieved with more directional mics (for which more than one company gets $1200 each: http://www.ultrasoundgate.com, for example). "
Wow !
Those "bat freaks" certainly have some bucks to spare cause this route is more expensive than the one I suggested (e.g. National Instruments DAQ hardware)
I also would say that (sufficiently large) array of omni-directional microphones is much better than few directional mics - you can point it in any direction at any point in time, e.g. after recording is done
Reply by Tom Becker●October 22, 20162016-10-22
> Have you tried making your own elements?
No, but maybe someday I'll tinker with a 30mm Polaroid 200V electrostatic but I am, for now, put off by the patterns that large-diaphragm sensors exhibit. High voltage is surely also more problematic.
I don't want narrow lobes. I enjoy recording in stereo to hear the bat flight dynamics; omnidirectional, figure-eight and cardioid patterns, sometimes recorded in Mid-Side which allows manipulating the pattern later, suit that well.
Most government, commercial and academic bat freaks are seeking species identification and population counts best achieved with more directional mics (for which more than one company gets $1200 each: http://www.ultrasoundgate.com, for example).
The equipment is often exposed to dew that requires conformal coating and heaters to combat. My mics have eight resistors that warm the assembly to prevent condensation and consequential corrosion - an instant serious noise generator - and that's at 12-48V. I expect 200V must be a PITA.
Tom
Reply by Tim Wescott●October 22, 20162016-10-22
On Sat, 22 Oct 2016 11:18:39 -0700, Tom Becker wrote:
> Thanks for the link.
>
> I've been recording bat calls for years and some of the data has made it
> into academic literature. I choose to make my own mics; instead of
> spending major money buying commercial bat sensor systems and software,
> I choose to enjoy wasting large amounts of time developing my own.
>
> In the field I currently use a pair of 14-element phantom-powered
> electret pressure arrays feeding good preamps and an RPi; another pair
> of them is permanently mounted above water in Cape Coral and yields
> audio - shifted -22kHz at the moment - that's streamed at
> http://78.46.38.217/amb9u;http://78.46.38.217/amb9 is unshifted. Windy
> today.
>
>> Let us know about the path you take and how it goes...
>
> Initial AVR counter results are forthcoming.
>
>
> Tom
Well, so much for suggesting that you buy electret elements and roll your
own preamp...
Have you tried making your own elements?
--
www.wescottdesign.com
Reply by Tom Becker●October 22, 20162016-10-22
Thanks for the link.
I've been recording bat calls for years and some of the data has made it into academic literature. I choose to make my own mics; instead of spending major money buying commercial bat sensor systems and software, I choose to enjoy wasting large amounts of time developing my own.
In the field I currently use a pair of 14-element phantom-powered electret pressure arrays feeding good preamps and an RPi; another pair of them is permanently mounted above water in Cape Coral and yields audio - shifted -22kHz at the moment - that's streamed at http://78.46.38.217/amb9u;http://78.46.38.217/amb9 is unshifted. Windy today.
> Let us know about the path you take and how it goes...
Initial AVR counter results are forthcoming.
Tom
Reply by angrydude●October 22, 20162016-10-22
On Saturday, October 22, 2016 at 5:40:12 AM UTC-4, Tom Becker wrote:
>=20
> It looks like a good 192kHz analog sound card for the RPi, though.
>=20
>=20
> Tom
People have used Wolfson card for bat recordings with good success.
Here is just one quick google search result:
http://www.afraidofsunlight.co.uk/weather/public/Pi%20Bat%20Call%20Recorder=
.pdf (not sure why he uses external electrets - the onboard mics are high-q=
uality MEMS mics)
Card is good for recording, maybe for streaming too, but if you need to do =
something else then you are on your own:)=20
For university-type of research I would use some high-quality data acquisit=
ion card or USB DAQ device e.g from National Instruments, with high-quality=
analog microphones, preamps and PC or laptop running controlling software =
program (e.g in Matlab) - be prepared to shell out some $$$$
=20
Let us know about the path you take and how it goes...
On Friday, October 21, 2016 at 11:50:27 PM UTC-4, Tom Becker wrote:
> > ... solder those tiny mics yourself...
>
> No, thankfully Knowles can provide samples on passive "coupon" boards that have a five-conductor standard 4mm-centers edge connector. Tiny, indeed. Two-pin 0.1" for reference: https://www.dropbox.com/s/rndx30dda7w3ztw/KnowlesCoupon.jpg?dl=0
>
> > ... Wolfson Audio card for Raspberry Pi...
>
> A natural - if it took PDM in.
>
>
> Tom
>>> No, thankfully Knowles can provide samples on passive "coupon" boards
How many samples can they provide and how much do they charge per board ?
I'be seen some manufacturers charging as much as 80 bucks for a simple soldered on the board mic "sample" (no preamp, no nothing, just mic)
thanks, but no thanks (but still a better deal than trying to solder it yourself which is pretty much impossible)
> > ... Wolfson Audio card for Raspberry Pi...
>
> A natural - if it took PDM in.
As far as I remember Wolfson has 2 high quality digital MEMS microphones soldered to the main board interfaced to their DSP with all of this PDM crap being taken care for you :-)
If you just want to make some high quality bat recordings then save yourself from all the trouble and go Wolfson (or similar) route