Many thanks Alex - that would be very useful.

Thanks Alex - I will try with the Extreme card.

Ps I have just remembered that the recording schedule in the 2 instances was different. In the first instance I was recording hour long files. In the second the device cycled recording for 55 seconds then slept for 5 seconds. Maybe this explains the difference?

That's a brilliant tool - thanks Alex. I only had chance to have a quick look, but it seemed like it was only possible to save a wav file of up to 1 min long. Is that right ? Jim

Many thanks Alex Here is one that doesn't work: https://1drv.ms/u/s!AuwFtHzw2ExNgakW3hDTZ6dtLu_cew?e=ZpWMeZ High pass filter applied in Audition. Error message: -- Downsample -- 20220723_003035.WAV - Could not find ICMT chunk ID. Is there a way around this problem? I am recording without any filter applied as I want to montitor for bats and birds / mammals at night. I plan to play these back at a public event and want to minimise noise. So for the bats I want to high pass filter around 20 kHz and then downsample. I will then apply a low pass filter around then same frequency to listen for the birds / mammals. Jim

Thanks Alex, Yes migrating bird calls are around around 1-10 kHz range (maybe 0.5-8). So I will test out what you suggest. Right now there seems to be very little migratory movement, so I will try in a few weeks. Jim

Thanks Alex, That's very helpful. So very briefly it looks to me very flat in audible range than a big peak at around 24 kHz before falling at higher frequencies. I have been trying out for 'nocmig' recording and using 32 kHz (to save space / power) but I get of lot of 'bat noise' in the audible range - that is aliased from the higher frequencies and that big peak around 24 kHz helps to explain that. So maybe I should sample higher and filter out. I didnt really want to go much higher than 48 kHz just for space and power saving. So I will try that next and see how it compares. I have previoulsy used a telinga Stereo mic with parabola and it is interesting to see how different it looks - I think the telinga amplifies low frequencies (complicated by the spatial aspects) and has no ultrasonic response. I only get a tiny bit of bat signal ( restricted to around 18-20 kHz - using 48 kHz sampling) that doesn't interfere with the birds. I plan to do some side by side comparisons - but comparing these two barn owls (on different occasions) I think gives a bit of an idea of the frequency characteristics. https://xeno-canto.org/742903 (AudioMoth) https://xeno-canto.org/742905 (Telinga) Jim

Thanks David I have a subscription to Adobe so have the lastest version. I have the options of 'Interpret Sample Rate' - or 'Convert Sample Type' (both under the Edit menu). The former allows you to hear the sound at a lower rate, whereas the later make a permanent change with more options (including filter). Jim

Thanks David i can do it in Audition which is what I normally use. The issue is that that antialiasing filter will only work if you have sampled at a high enough rate in the first place (2x highest f in the environment) - otherwise the aliasing is ‘baked in’ and cannot be removed. So if as I my case there are lots of bats around but I am only interested in birds I have to sample a lot higher than I need and then downsample - (which takes more power and storage) or put up with interference. This is why the proposed analog filter is such a good idea. Jim

Alex just a quick addition to my earlier answer. I think it would be a useful addition to the downsampling tool to add the low pass filter set at the nyquist limit of the new rate - as a User option. I understand what you say though as to why you don’t have that at the moment. Jim

Thanks Alex In practice I suspect the number of occasions where a bird and bat call will overlap will be pretty small so for long overnight recordings I will use 32 kHz I think and use the higher frequencies for shorter recordings around dusk where both bats and birds will be of interest. Thanks for the earlier demonstrations of noise reduction with downsampling. Would you get the same effect just by using a low pass filter? (of course you would have a file size bigger than it needs to be in this case) Jim

Hi Alex, That is really helpful again. I can see the pros and cons of filtering first before downsampling. I was just wondering if it would be technically possible to apply some kind of analog filtering that could be switched in or out. Possibly for a future development. I had very limited knowledge of electronics so don't know if that is possible. I am just looking at some tests I did overnight last night at 32kHz (listening for migrating birds - rather early in the season I know - and usually I would use 48 kHz, but just trying this out as I don't usually see any bird calls above 10 kHz). There was a great deal of bat calls coming down into the lower frequencies - very loud at times. I could see some benefits of somehow removing above say 15 kHz prior to sampling for this type of bird call application. I know there is a filter option in the configuration, but as this is digital after sampling I don't think it would work in this case (we discussed in another post). Just seen this post which describes the issue nicely https://www.openacousticdevices.info/support/configuration-support/bat-calls-at-low-frequencies Jim

Thanks for clarifying Alex, i think I was getting confused because if when there was higher frequency bat calls and I downsampled I could hear them in the audible range so I thought all frequencies had been shifted - but it now dawns on me that this is due to aliasing. i had also noted this in some earlier tests - for example when recording at 48000 I was hearing pipistrelle bats at around 3 kHz very much sounding like a hetrodyne detector tuned to 45 kHz - this is due to the undersampling giving an Artefactual signal. A sort of happy, but not very reliable, accident ! Thank you for cleaning that up - it’s beginning to make sense to me ! 😀

Thanks Alex I think I probably am ok at 192000, and the noise was less when I tested at that. jim

Thanks Alex

Thanks Alex I don’t think that does what I want ? So I recorded at 250000 to see if there were high frequency bats in the area. In fact there was not much of interest at high frequency but there was some bird calls at a few kHz. So I want to just discard the high frequencies and keep the lower frequencies- just as if I had sampled at 48000 to start with. So making a smaller more manageable file, without any pitch shift. is that possible? sorry if I have misunderstood jim

Just reading some other similar posts and I see it might be to do with SD card writing or batteries? - my card is a SanDisk Ultra 32 GByte and batteries are Energizer rechargeable. Device ID 249BC30461CBBA03 Firmware: 1.8.0

Thanks Alex, I guess I can set up the sleep cycle mode to record on for 59 mins - and sleep 1 min. which would have a similar effect, although with 1 min gap between one file and the next. Jim

Thanks Alex

Sorry should read 'all frequencies' not 'and frequencies'