True but as it has been pointed out several times, that sort of comparison is a task that an UNO is poorly equipped to handle. Try some of the other suggestions and see if they are reliable. Use a very directional microphone and attach it to the camera in such a way as to prevent pickup of other sounds.
+1
I think that with a properly set threshold and insulation this will work totally fine. No arduino needed..
Yes, I think it probably will.Certainly easy to try.
The idea is good, luckily the camera is not mounted directly to the glass, instead it’s mounted on a small swivel bracket on the edge of the frame. The window is very large and part of a sliding patio glass door so any slight noise from in or outside could be a problem. I will check it out.
If you mount the sensor directly to the camera you will be able to set its sensitivity very low and since the camera siren is orders of magnitude louder than any extraneous sources they will pose no problems. If something other than the siren produces a comparable magnitude sound it probably should be investigated anyway.
Would you reckon that feeding those peak gaps into a counter and logging the number of counts within a set time period, then reset the counter and repeat the process to produce x number of total counts against a predetermined time period. Does this make sense? Is this what you have in mind, taken directly from the mic.
That is the challenging part, but I think it should be doable. But it's not just a simple matter of feeding it to an analog pin!
Get stuck in and try! The payoff would be the versatility of being then able to use a DFR Player.
Not if you could tape the microphone directly onto the speaker or the camera - then the threshold can likely be set very high
ok I’ll see what I can come up with
OK, I'll try to help in parallel with you.
To make sure we're in sync, I suggest two stages for your project.
- Obtaining a signal that say an UNO/Nano can safely use from a camera speaker which cannot be accessed directly. You gave me a sample. Can I assume you propose to use the same method for your finished project?
- A sketch that will reliably deliver a 5V high from that.
Hi, yes , I can use a microphone near to the camera ( as others suggested ) it doesn’t need to be shielded from external sounds as we are only interested in harvesting the siren. This signal will be amplified and diode rectified so we are left positive peaks only, we may not need to tailor this signal any further other than to have sufficient amplitude to work on. If we set up a counter to count the peaks within a set period ( haven’t analysed the signal myself yet) based on your test image, and then repeat the count a few times to see what we come up with, the percentages may be quite close.
Once we have a count that proves reliable, then we’re good to go. BTW we can take a number of counts, because the time delay between this siren sounding and the external sounder being activated can be a few seconds, it’s not critical. Jevray
I intend to use the microphone, its placement is not critical now, as I have an alternative idea for processing the signals it picks up. We are only interested in the sirens signal. Alexa hears sounds continually but only responds to your voice. Thanks for posting Jevray
Ok, understood, sounds good.
Thanks for jumping in
Just seen your post as I opened the thread to post mine! Here it is:
I played your signal on my shed workshop's ancient Windows XP PC, as that's where I have PicoScope installed. Good news and not so good news. As we already saw in my GoldWave screenshot it's 'clean', so won't need noise filtering. But at around 250 mV peak-to-peak it's going to need some amplification before we can expect an Arduino to be happy with it. And, of course, rectification and probably some DC offset.
Anyway, I've had it for today!
Very similar.
Goodnight, back on the case tomorrow or Sunday.
Ah good old Pico, I worked for a company that built the pico scopes, once they came off assembly they were inspected, tested and calibrated, one of my jobs. They were finally certified by Pico. They sent me one of there top feature packed models for my personel use.
Anyway back to this project, the signal has to be lifted to logic level as agreed so that’s not a problem, also taking into account the half way rectifying diodes forward voltage drop of 700mV, post amplification. Hopefully it won’t need any filtering as we must keep those amplitude spikes clean, retaining the rise time. I have put some code together to play with, at the moment just using the iPads waveform I saved to experiment with.
The UK weather at the moment is causing havoc due to drought in our very large garden, so having to spend some time away from the bench.
Wish I still got my Pico, my old pc tower died several years ago, and the Pico was sold on eBay. 
(post deleted by author)
Yes it’s an interesting project, just ordered a MAX9814 mic preamp board, I’m also experimenting with an Lm393.