Frequently Asked Questions
microMARS Acoustic Recorder
What type of SD Card should I use?
It is MANDATORY to use SanDisk Extreme Pro Cards with your microMARS recorder.
Why is the sample frequency not staying the same on my microMARS?
microMars is automatically optimizing the actual sample frequency for the closest match to the user specified frequency. microMARS uses a digitally controlled oscillator (DCO) to generate the sample clock, but the DCO changes frequency in steps or intervals. So, you will generally see some discrepancy. This discrepancy however will not impact the accuracy of a signal analysis such as a FFT, because the actual sample frequency is specified in the metadata of the .wav file and will be used by common acoustic analysis software such as Adobe Audition, Cool Edit, Audacity and Triton to compute accurate results.
Why is the sample frequency sometimes changing from one recording segment to the next?
In the current version of the recorder firmware (V1.30), the recorder will re-tune the DCO before recording each segment, to hit the closest match available at that time. You may see a very slight change (a few Hz) because for example the temperature has changed. You may also see a step function change (a few hundred Hz) because a different frequency is now the closest available match to the user specified target. Note: We are aware that this re-tuning between sample can cause problems with some forms of data analysis. In the next firmware version, segment-by-segment best match hunting will be disabled in favor of keeping the same DCO setting throughout a mission, but measuring and specifying the actual sample frequency in the meta data of each recording segment.
Can I obtain diagnostic information about my recordings?
Yes. Once you have configured your recorder using the microMARS Dock software, select View→View Raw Coms. In the filed at the bottom, type this 'Exit Dock Request' command and then hit the Send button: $B,0,EDR
Now watch the blue pane on the right. Various diagnostic messages will appear.
A good quality assurance method for your recorder configuration is to use the $B,0,EDR command to watch some of the initial diagnostics of your new recording, making sure all is OK. Then, connect the battery, disconnect the USB cable (in this order!), and button up the recorder.
In some cases you may wish to record diagnostics for a long period, such as overnight. For such purposes, use a terminal software such as Teraterm. Once the recording is running, exit microMARS Dock while leaving the USB cable connected. Launch Teraterm, and select the recorder's COM port, using a 9600 baud rate. Enable the Teraterm log function to generate a text log of all diagnostics. Let run for as long as your test requires.
Here is a link to the Teraterm Wiki page. You can find a pointer to a download page within.
Why does microMARS experience data losses?
If you watch the microMARS diagnostics, you will notice that some data losses may be reported after each recording segment is completed. The diagnostics specify how many seconds into a recording each loss occurred, and how long it was. These losses are caused by occasional slow SD card writes. microMARS uses pairs of SD cards in a ping-pong fashion, writing a 'page' of data (256 samples) to the Bank A card, then Bank B, Bank A again and so on. If one card isn't done committing its page of sample to FLASH memory by the time the page buffer of the other card is filled up, a data loss occurs.
In general, the losses are very short, on the order of a few milliseconds (about the same length as a geo-location click). The should only occur once every tens of seconds or minutes, but generally more with higher sample rates and less with lower sample rates.
Specific SD cards are selected for use with microMARS, and it is important to use the card models that Desert Star or your recorder distributor has selected and tested. Recorders shipped directly from Desert Star Systems are equipped with a starter set of 32 GB SanDisk Extreme PRO cards in each slot. We have also tested and qualified the higher capacity versions of the SanDisk Extreme PRO.
Which SD Cards are qualified for use with microMARS?
Desert Star Systems has qualified the SanDisk Extreme PRO card series for use with microMARS. This is the only card series we have qualified. We have found these cards to have a good combination of low data losses, moderate power consumption, and no noticeable noise coupling into the recorded acoustic data.
If you bought your recorder from a distributor such as Cetacean Research Technology, they may have qualified different cards. Please contact your distributor for card compatibility and test results.
Why are recording segment lengths generally a few seconds shorter or longer than specified?
In continuous recording mode, microMARS starts a new segment as soon as the previous recording segment is completed. For reasons of reliability, SD cards are shut down and then booted up between recording segments. The boot process duration however is changes in duration, on the order of three seconds, introducing some variability (with Firmware V1.30).
Note: We are aware that walking time stamps for the recording starts can be problematic for data analysis. It's better to have segments start on a rigid timeline. The next firmware revision will fix this problem by ending each recording segment slightly early, enforcing a sufficient gap that allow worst-case SD card boot-up time and keeping the segment time stamps on a precise schedule.
What's the difference between a sloped-frequency response curve and a flat-frequency response curve?
Flat-frequency response curve
Underwater recorders are traditionally optimized to provide a flat frequency response, i.e. uniform sensitivity across the operating frequency range. The principal advantage of a flat frequency response is that the resulting .WAV files can be used directly to generate dB re. 1 µPa (or equivalent) referenced spectra, perform various sound measurements or play the files with a 'natural' sound quality. Hydrophones with a flat frequency response are designated by a number directly after the dash, such as MH33-1 or MH33-2. The -1 or -2 defines the gain programming of the hydrophone, which determines both the maximum signal the hydrophone can receive without clipping, and the noise floor.
Sloped-frequency response curve
Sloped frequency response hydrophone end-caps increase in sensitivity with frequency. At low frequencies, they are less sensitive and can thus accommodate louder signals without clipping. At high frequencies, sensitivity is greater and the noise floor is lower, meaning that fainter signals can be detected. This sloped frequency response matches the typical noise profile in the ocean, with lower frequencies experiencing louder noise. Thus, by selecting a sloped frequency response end-cap, you may be better able to measure loud industrial noises without clipping while at the same time detecting faint biological signals such as from distant marine mammals vocalizing at higher frequencies. In other words, a sloped frequency response end-cap provides a greater dynamic range (great ratio of loudest to weakest discernible signal) than is available with a flat frequency response end-cap.
Will I be able to record whales (low-frequency sounds) AND dolphins (high-frequency sounds)?
Yes, IF you use the microMARS, then you will by default be using a hydrophone endcaps will a sloped-frequency response curve, which will enable you to hear both sounds without clipping.
Are the microMARS clocks calibrated to each other for time syncing?
No, They are not synced in standard microMARS. Each has its own independent clock and they will drift relative to each other.
Actual sample rates are never exactly the same as what I sent them (can be off by as much as 5%; either higher or lower)
That is correct. microMARS uses a 'digitally controlled oscillator' (DCO) to synthesize the sampling clock. But, it cannot hit all frequencies; the frequency selected will be the closest available to the target. Also something to be aware of, is that the selected frequency will be somewhat recorder-specific and that the DCO re-tunes after each recording segment to compensate for things like changing operating temperature. So, the frequency may vary slightly. However, the measurement of the actual DCO frequency is very accurate and is stated in the .WAV header. Therefore frequency analysis is not impacted by this.
The .WAV file length doesn't exactly match what I expected based on the length set; for example, when I set file length to 1 min, it can be anywhere between 0:59 to 1:02; the length seems related to how much over/under the actual sample rate is from the one programmed.
Correct. This is because the recorder powers down and re-initializes the SD card after each recording segment. This process takes about three seconds and is somewhat variable. This is done to assure that any potential recording problems related to the SD card is confined to a single recording segment and would auto-correct on the next segment. While we haven't observed such recording errors like a bit slip, it is at least theoretically possible and this method provides a fresh start every segment.
We do however realize that having slightly walking time stamps can be problematic, and so the next firmware version will adhere to a rigid schedule but enforce about a 5 second break between segments to assure the next segment will start on schedule.
Start times of files are always a few seconds off from the start time that is set through the microMARS Dock Program
This is due to the same as the previous question, and will be corrected in the next firmware update.
There appears to be ~2-4 seconds missing between consecutive files
Two questions above addresses this, as well. However, these breaks will remain and indeed lengthen by about 2 seconds as we move to the rigid schedule in the next firmware version. If the breaks are problematic, a recommend fix is to use longer segment lengths such as 1 hour so that the gaps are less frequent. Here it is very important that only the SanDisk Extreme PRO cards are used, as these cards produce the lowest number of sample losses due to slow card response. The sample loss locations are logged by the recorder (they are generally < 10ms and generally only once every minute of few minutes). The data loss log memory is limited to 80 entries per segment per memory card. So, if a slower card is used, this log may be filled and a new segment start is then forced.
Which battery should I use?
With the four battery option, you will get about two months of continuous recording. For longer deployments, you can specify the duty cycle. Also, you can have two or three recorders, and program each to start recording when the previous one has stopped. This can be done in a linear fashion; so first two months for the first recorder - second two months for the next, and so on.
It’s not clear from the manual, whether one has to follow a specific launch procedure: unplugging, re-plugging battery packs etc. I.e How long can the electronics pack be without external power for the Mission details to be retained?
Specific launch procedure would be as below.
Plug in microMARS to software and connect device
Input your desired settings
Format the memory cards that are present in the microMARS, after you have received a successful format window you can then write your settings to the microMARS
Before disconnecting form the computer, attach the battery
Feed the battery through the housing and then attach the hydrophone element.
Close the device and there should be a red LED blink roughly ever 20 seconds.
Is there any way to resize the Raw Comms window since I can’t always see all the text that is displayed?
Regrettably there is no way to resize the window with the current software. We are working on a new revision which will bundle both the card reader and the microMARS dock software together. In this new revision resizing the window will be addressed.
We have recently started work on a networked microMARS capability for the Office of Naval Research. This project will provide a capability of microsecond level time synchronization between recorders to localize sound sources. It's a one year project, so results are still a little out, but this is just a heads up of the microMARS growth path in case your research needs go in that direction.