We were interested in finding out whether 2 Audiomoths can replace 1 Song Meter. Financially and logistically, it would be very nice.
For the Audiomoths, we used protective plastic cases with a hole drilled into them to let sound pass through an acoustic vent (GAW112 by Gore) and made a test outdoors in autumn of 2018. We strapped an Audiomoth to each side of our SM2Bat+, which used custom microphones (SPU0410LR5H-QB) that are at least more performant than SMX-U1 mics (in review). We recorded for 1 hour and 20 minutes 30 min after sunset near to a pond. This test is more fair than the one mentioned in the other forum because all microphones have equal protection levels (GAW 112 vents).
The cases handled morning dew (there was plenty) well. They felt a bit damp inside tough, and it won't be easier to handle in tropical regions, so maybe little bags of silica gel could help (the kind you can put in an oven to recharge/dry out).
The Audiomoths recorded sound, and every main bat pass (there were ~half a dozen, 2 morphocalls) that was detected by the corresponding Song Meter microphone was also recorded by the Audiomoth. These are good news.
The Audiomoths record distorted bat calls when they are strong (but before distortion):
This confirms what we found in the post https://www.openacousticdevices.info/support/device-support/sound-transmission-with-and-without-cases-comparison-with-sm2bat). You see lots of additional harmonics, most above 100 kHz. I think Joe Chun-Chia Huang also highlighted that somewhere on the forum. We recommend the SPU0410LR5H-QB mic on the Audiomoth, it has been proven to be good and could be compatible.
Signal-to-noise ratio looks better with the SPU0410LR5H-QB, and detection distances should correspondingly be better too: really faint calls would be missed by the Audiomoth, check screenshot.
There are vertical bands of noise in the Audiomoths.
UPDATE: according to Alex Rogers, these are due to the SD card writing and can be avoided by using a certain model.
To conclude, 2 Audiomoths can replace a single higher-end dual channel recorder well, if you can live with small inconvenients (distorted ultrasound, avoidable noise ticks). It might be worthwhile to find out how to synchronize the recordings from 2 units accurately (also for triangulation and time-of-arrival measurements that might be useful to some).