Getting started in nature sound recording presents a lot of confusing and overwhelming choices. In the following pages, I cover not only the equipment I am using, but also the decisions I made in choosing the equipment, some equipment comparisons, and general guidelines to consider.
- My current kit
- Choosing a sound recorder for nature recording: the essentials
- Microphones for nature recording I.: types and arrays
- Microphones for nature recording II.: different types for different situations
- Options for recording ultrasounds
- Protecting microphones from the wind
- Audio recording with a smartphone
Pingback: Bare Bones Recording Equipment for the Amazon – Hard Light Equator
I needed to connect a large distance with ethernet at a zoo so they could video record the birth of a lion. instead of using the specialised POE repeater devices I simply used one POE injector with a switch in between. The other end did have power though. The total distance we needed to cover was 120m. This worked fine with the longest section being about 80m.
Again this is extending not a USB connection but a small cheap computer. If you setup a jack daemon on the raspberry pi then I think you can connect audacity over the network to the pi mic.
I’m new to the whole outdoor recording thing, but very familiar with software, video and networking. Hoping to learn a lot more about audio so I can build some network mic recording software. Interesting to read all your blogs it helps me build a requirements list for recording software.
Another approach is to use power over ethernet to power a raspberry pi computer and then extend your microphone from there. There are devices that can daisy chain POE injectors to extend 1GB ethernet to between 500m-1km in 3x hops. You could record digital sound from 1km away 🙂
I’m interested in connecting up a digital mems chip to the CM108 chip in one of the cheap usb microphones. invensense have ones with -46dB sensitivity and 65dB SNR. In principle such a mic should perform pretty well based on the paper specs.The cm108 has a max sample rate of 48000 though.
Has anyone else done anything like this ? I was going to buy a USB mic and try ripping out the analog electret mic and replacing it with a high end digital mems mic but it would be nice to know if anyone else has tried this.
Kim
Kim, thank you for your comments and questions. I hope someone else can chime I on the USB-over-Ethernet chain, as I haven’t tried anything like that. The mems chip you mention sounds pretty noisy (generally for nature recording you want >75dB SNR), but should be sufficient for testing. I’ve been interested in learning more about raspberry pi applications for recording – although so far my interest has been using one as a controller for a recorder, to program start and stop recording times. But setting one up to do remote recordings sounds very interesting!
Thanks for the tip regarding what to expect from SNR values. I’m new to interpreting the ratings so that was very helpful. I’ve looked further and I think that the MEMs microphones are not appropriate as they seem never to have better SNR than 65.
I’ve order some EM172 and EM173’s. I want to use this with the cm108 chipset as another soundcard I tried based on another chipset seemed weak in the amplification, the level was very low.
How are you getting on with using a raspberry pi controlling applications, is this working as you would like it? Automation is an area where I have a lot of experience in as I developed software for a home automation product (And still work on it) that I tried to make commercial once. I recently ported my software to a raspberry pi, but I should point out I do not make use of the GPIO pins on the pi but rather use an external I/O controller from Phidgets. Specifically the Phidget 8/8/8 controller. This isolates it from the pi board and importantly for me it works out of the box with my software. It adds to the cost however, as such a controller costs even more than the raspberry pi. If this is something you I can help you with let me know.
Thanks for the comment, Kim. I’m interested in hearing about how the Primo capsules work for you in the long-range set-up.
I haven’t started with the raspberry project at all, just something I’m interested in trying at some point.
I finally got some primo capsules to try out. I ordered both an EM172 and an EM173. I’m not quite sure about how to correctly wire up the EM173 so I haven’t tried that yet. For the EM172 I’ve tried out several cheap usb sound cards and usb microphones. Some of the more common ones had a poor result, i.e. the amplication with low so it would need pre-amplication to hear quiet sounds. One device worked better than the others, it was a USB microphone. I determined that it was based on the C media CM108 chip. I ordered another sound card to test that was also based on this chip. Strangely, the sound card that also had an output on it would disconnect with errors within a few minutes. Even though it doesn’t make much sense it seemed that the microphone based on the same chip amplified more.
So I’ve settled on the usb microphone. I replaced the electret that was in the mic with the EM172. Worked wonderfully. Initially I extended the connection using another cable that came connected with a cheap lavalier microphone, replacing the mic in that device, but the extra cable length resulted in picking up A/C hum. So I went back to replacing the microphone that came with the USB mic with the EM172, connection length from the mic to where it was digitised was then only about 5mm, no A/C hum.
Right now I’m using this in an urban environment because my primary usecase is for security. To this end I’m using it with AGC turned on. The result is very sensitive very clear sound but of course by default it’s on top amplification and I can hear the motorway from a couple of km away all the time. Local sound overwelms the motorway noise though.
At the moment I have the audio connected through darkice software into icecaster2 to provide a webbased live broadcast of the mic that can be played simply with a webbrowser in a smart phone. As the sound is digitised locally at this point the rest is just transport details.
The microphone comes with short piece of cable surrounded with coiled metal so that you can bend it into shape. It’s not that convenient so I intend to desolder off the USB from the sound card part, which is in the cylinder right next to the microphone and replace this with a piece of a normal 1.8m USB cable where I have cut the end off. As it’s digitalized within mm from the mic this extra cable length is of no consequence anymore. It physically resembles the longer cable connected examples I’ve also seen on this blog except that there’s no change of A/C hum, which should not be a problem in the wild anyway.
All in all a good start, I’m going to order a bunch more mics to test witth as I want to develop some flexible recording software for use with it.
The USB mics I ordered from aliexpress for this purpose had the part number SPC-3204.
Kim
Sounds like a great start! Glad the Em-172 capsule is working. It seems to be able to record high frequencies pretty well (I’ve picked up quite a few bats with my recorder set to sample at 96 kHz). Not sure what the maximum is. You’re running the mic into a Raspberry Pi, right? What is the maximum sample rate you can set that to? Probably not your primary interest, but might be interesting if this is for a zoo setting.
The raspberry pi can sample high rates no problem. But the particular USB microphone I’m using uses the C media CM108 chip which only supports 44100 and 48000 sample rates. I would need to use something like a startech USB sound card to support the higher rates. One of the cheaper startech USB sound cards supports 96KHz 24 bit audio.
I was able to get the USB sound card working on my desktop PC. Had to repeatedly “remind” the Sound Recorder aplet and Audacity to use the USB card instead of what’s built into the motherboard. Once past that, I was able to record (and play back) through the USB device with mics or patch cords plugged in, however I got feedback if I left the mic plugged in while playing back. Otherwise sound quality seemed fine, but I hardly have golden ears.
Then I tried adding the USB to cat5 adapters with 25 feet of cable in the middle. That still worked!
The next step will be building say a 200 foot cat5 and if it doesn’t work, chop off ten feet at a time until it does. Then verifying everything on a laptop which might not want to supply as much current to the USB.
What version of Windows are you running on your PC?
Windows XP. CMedia says their chip inside will work with everything from Windows 98 SE on.
Wow. Didn’t know anyone still used XP. The reason I asked is because Windows changed their USB address (at least regarding audio) either before or after XP and again with 8. I have a USB MEMS ultrasonic mic that I used to use on a Win 7 tablet – worked great. I upgraded the tablet to 8, and even though the software finds the mic, if you hit record, nothing happens (same on a laptop with 8.1). Only one software maker I’ve found has a way around this, and they are way too expensive. I’m hoping it’s fixed with 10. I can use the mic on an Android phone via USB host and USB OTG, but there are only a couple of apps to record that way, and they are bare bones. Much more fun if you can see a spectrogram while you are recording. So what is the chain on your USB-cat5 system? BTW, I’ll be traveling a lot over the next two weeks, so I won’t be able to check messages daily.
I was able to verify that the same USB chain worked on another PC with Win7, but I have no Win8 or tablet experience. I’d guess what you are experiencing is either driver settings that you aren’t accessing or issues that need attention from the software gods.
The chain is: “computer” mic with 3.5mm TRS plugging into a USB sound card (or a USB mic that combines these two), then the USB to cat5 converter dongle, cat5 cable with RJ-45 on both ends, then the opposite cat5 to USB dongle plugged into the PC.
The USB sound card is powered from the PC through all that wire. I’ve only gone out to 25′ of cat5 but at that distance everything looks the same as with the short leash without dongles and cat5.
It turns out that the generic Windows driver for USB sound cards properly handles gain and provides sufficient user control! The “AGC” checkbox apparently toggles the 20dB boost without trying any funny stuff. Then the gain slider has about five discreet steps, making settings repeatable. Measuring dBs, the total slider effect minimum to maximum gain pretty well matches the 22.5 dB spec in the USB chip datasheet. It looks like gain and “AGC” are indeed analog amplification ahead of the ADC digitizer, instead of manipulating — and sometimes throwing away — bits in software.
Even with the booster off and gain minimized, what looked like self-generated mic noise was showing up, corresponding to about 50 dB or 40 dBA ambient. That’s not very quiet! It’s OK for me because I’m studying urban noise pollution, but not good enough for wildlife or desert soundscapes. Anyhow I wish I had a quieter mic just to see whether mic noise or noise inside the USB device is limiting. If you have a quiet mic with a 3.5mm TRS plug, you could buy a USB soundcard on eBay for a couple bucks and find out. Or just bring it over. I’m in Culver City not far from LAX.
But for the noise floor issue, hardware and drivers look good. Software for PCs also looks good. I was recording short test clip with the Sound Recorder applet, looking at waveforms in Audacity and getting LEQ and FFT specs with VSLM, so I spent $0 on software.
What is the advantage of cat5 over regular mic cable? If you’re not opposed to soldering, I have directions on my site for DIY mics using inexpensive and quiet Primo EM-172 capsules. Self-noise of 14dB (company specs).
Apparently Win8 works ok with regular USB mics, so it would likely work ok with your chain. Something about this MEMS mic…the company that makes them is moving in other directions, so pushing it toward smartphone and tablet apps and worrying less about Windows. Does your soundcard have two inputs for stereo, and can the software interpret it as stereo (not dual mono)?
The issue isn’t regular mic cable vs. cat5, it’s long analog cable runs inevitably adding noise and subtracting high frequencies vs digitally being able to send forward enough juice to power the soundcard while sending back the digital signal in good enough shape so host can decode it. This either works or it doesn’t. As long as it works there is no degradation with length.
Building mics from the Primo capsules is a great idea and would almost certainly work plugged into a USB soundcard, provided pinout on the 3.5mm plug matches what the soundcard jack expects and the capsule gets a bias voltage within its acceptable range.
Cheap USB soundcards have mono mic input. There should be pricier ones with stereo mic inputs, using a single USB connection on the digital side. A deluxe model might also have stereo line input, or you might be able to persuade a mic input to handle line by disabling boost and turning down gain.
I like MEMS mics because they have less sample variation in sensitivity and frequency response. Their characteristics are more stable with variations in temperature and humidity, and they change less over time. I should think they would be suited to the intermountain environment subject to wide day/night temperature swings and generally low humidity but occasional showers plus summer monsoon in the Colorado basin.
Most smartphones have 2.5mm or 3.5mm TRRS jacks for headsets with four contacts: left earphone, right earphone, ground and microphone (or another permutation depending on brand and model). Windows mono mic inputs are (practically?) all 3.5mm TRS (three wire): mic signal, mic bias and ground. There are adapters mediating between the two worlds but they don’t span all smartphone variations.
Ah, I think I’m starting to understand now. So something like this might work for either stereo PIP input or line in: StarTech.com 7.1 USB Audio Adapter External Sound Card with SPDIF Digital Audio Sound Cards ICUSBAUDIO7D? The power output of smartphones with USB host seems a bit low; do you think they could power something like this over a long cable run? I’d be very interested in seeing the actual noise figures for a phone or tablet plus cat5 (and adapters) and soundcard and mics, versus a small recorder with cables and mics. Of course, there’s always the issue with battery life, too.
My MEMS mic is pretty noisy, and not very sensitive. Works well with bats and other ultrasonic callers which tend to be loud. Way too noisy for ambient recordings. It does seem impervious to temp and humidity swings.
Almost all smartphones have TRRS jacks now, including most Windows phones. The big exception are the Sony xperia phones and tablets, which went with a TRRRS jack, allowing stereo mic input. But the only mic that matches is their own stereo mic.
No question USB current delivery is limited. Probably desktops>laptops>smartphones, if you can get any out of a smartphone at all and it doesn’t clobber battery life. You might have to resort to cutting the +5v line and splicing in a second power source for soundcard and mic. Zeno Sloim touches on this: https://forum.xda-developers.com/showthread.php?t=480899
Stereo input USB soundcards might want more current than the cheap mono-only versions. eBay has some miniature $6 meters that plug into the USB chain to display voltage and current. Possibly useful debugging tools for power supply issues!
I believe microphones will be the main noise sources, with the possible exception of low-noise mics like the homebrew you mentioned, with the 14dB noise floor. In that case the main chip’s preamp sub-circuits between input jack and digitizer might also add significant noise and upgrading to something with higher grade chips than C-Media’s could be in order. Once the signal is digitized, it’s presumably fixed no matter how long the cable or how much electrical noise there is further downstream.
My game plan is getting the USB soundcard/cat5 working on a laptop, then attempting repurposing smartphones as handheld sound meters and recorders. I won’t be eager to take on the hurdles of usb/cat5 extension, hosting and additional software development with smartphones unless the laptop approach doesn’t work out. If I were schlepping equipment for miles on foot I’d probably feel differently.
I appreciate your input on noisy MEMS mics. I’ll have to see if they will still work for me on urban noise, which probably is even noiser!
I had no idea there was TRRRS! I do know there is a lack of standards for TRRS. I bought a straight-through TRRS 2.5mm male to 3.5mm female adapter and plugged in mics that work perfectly my iPhone and WM6.1 Palm Treo, then plugged the adapter into a Motorola Q9M. No dice: Motorola used a different pinout. I am narrowing my search down to smartphones with 3.5mm jacks and the same pinouts as iPhones. Unless I can find software that flips the display, I’ll also want the jack to be on top. It looks like Samsung is a good bet.
The newer samsungs (s3 and later) all use the ctia (Apple) standard. I think motorola is a holdout on changing over, and sony is creating a new one. I cover this in detail on my “Audio recording with a smartphone” page. So I think almost anything will work except older phones and Sony’s. Unfortunately, I know of no place where those specs are listed. Not that it would save anything in portability, but would it be possible to add some extra power to the usb sound card? Or maybe a battery-powered usb-hub?
There’s this on pinouts https://pinouts.ru/headsetsheadphones/ — quite a lot to wade through. I’m glad to hear things are trending toward standardization!
A battery powered hub would be a good thing. It probably needs a “host” port it doesn’t power, lest the hosting smartphone decides it’s being charged and puts its USB into charging mode, disabling ordinary communication.
My eBay search was ambiguous at best, but such a device may exist. C-media’s sound card chips only draw tens of milliamps so a battery could probably last up to a week of extended work days.
Thanks for the link on the pinouts; I added to the smartphone page.
My plate is already pretty full with desktops plugged into A/C or laptops plugged into lighter sockets, both with USB over cat5 extension for unattended monitoring/recording. I’m also committed to working on analog mics plugged into smartphones, for work on foot. Delegating USB hosting from smartphones feels like an appropriate strategy. I can certainly pass on research I’ve done.
Oops, should be 150′ (feet), not 150″ (inches)!
There’s a whole ‘nother possible approach to quasi-outdoor recording, provided you are within 100’ or so of AC power for a desktop computer or a vehicle with a good battery to power a laptop, or your laptop has a strong battery and you don’t mind lugging it around.
There are USB dongles in pairs that supposedly let you plug (RJ-45) up to 150″ of cat5 Ethernet cable in between. These are pretty cheap on eBay. Then there are inexpensive USB mics as well, or you can get an even cheaper USB soundcard and plug in a mic conforming to SoundBlaster specs, i.e. TRS plug with tip=mic, ring=bias voltage, sleeve=ground. You probably already have some lying around.
Cheap USB mics and USB “sound cards” almost invariably house C-Media chips such as the CM108 or CM119. These chips are decent if not spectacular performers with 90dBA signal-to-noise claimed, and their power draw is low enough so the resistance of a pretty long cat5 cable won’t reduce USB voltage supply below the 4.5 volt minimum. Then you have everything digital and pretty much impervious to electrical noise and high frequency attenuation after the short route from mic to chip. The chip only consumes milliwatts compared to tens of watts in a laptop and a hundred or so for a desktop, so exposure to electrical noise is inherently limited. AFAIK from reading datasheets, these chips don’t have built-in AGC or noise gates.
Unlike the paucity of Windows Mobile recording software, there’s plenty of excellent desktop/laptop recording software to choose from; much of it free. Audacity is only one example.
The cheap eBay stuff should be sufficient for proof-of-concept; then you can upgrade microphones and USB soundcards at will. I’ll be working on the proof-of-concept part and will endeavor to keep you posted.
Interesting approach. I think the iPad crowd does something similar, and probably has many of the bugs worked out. I’m not aware of any USB mics that would be particularly well-suited for outdoor recording – most are designed for close studio work, like podcasting or maybe DJ’ing. But as you mention, it is possible to adapt a decent mic to USB, although I’ve never heard of using cat5 for that. Some of the noise gate issues are software related, so it’s just a matter of checking for them. I’m looking forward to seeing how you progress on this, and if any can be adapted down the road to phones that have a USB host.