Facebook
Google
GitHub
Linkedin
So I figured in order to localize sounds, it may be simplest to use a
volume meter with a microphone array.
Each microphone may be an FM transmitter with an amplifier to amplify
the volume, or some other microphone connected device to save samples
of sounds. Filter out anything that isn't human speech to make it
easier and,
Each mic represents a different direction, north, south, east, and
west, along with a time stamp associated with it that also goes along
with a snapshot from a camera with the same time stamp as a reference.
Each sample, from each mic, is compared via the volume meter. Whatever
sample has the highest volume, is going to show the direction of a
specific word or phrase.
An application like Jasper can be used on a raspberry pi to save
samples of sounds/voices and covert them to text. The text will have
the digitized version of the analog signal from the sound also
associated with it, and that can then be played through a speaker to
the volume meter. The pi will cycle through each mic's sample, and
play it through one speaker for the volume meter.
This can be done quite quickly, and a buzzer can vibrate the
microphone/device that showed the highest volume, indicating the
direction of a particular word or phrase.
If there's an easier or cheaper way to do this I don't know what it
is. If you're only interested in a few different words, this might
make things easier. Those words are filtered out and saved, whereas
everything else is either saved elsewhere or discarded.
Reason I wanted to use an FM transmitter is because I already have the
components for at least a few of them and a raspberry PI can only use
a USB microphone, so with a RTL SDR I can record at least 5 channels
simultaneously, in other words the audio from 5 transmitters
simultaneously. It just has to be played through the speaker so the
volume meter can determine which one produced the loudest sample.
This could of course be used for other sounds like footsteps from
animals or humans.
another thing is, the transmitters can be used for other purposes once
detached, as with the SDR having multiple purposes. If actual
coordinates could be matched with the direction of sounds, then a
compass can be used to locate the source of sounds, rather than simply
a reference point from a mic or snapshot.
A smartphone app may replace the volume meter, also, but the
touchscreens can be a pain in the ass to use, so I'd probably rather
go with the volume meter as it's only for that specific purpose.
Anyway I obviously have minimal knowledge of this stuff so I'm just
looking for what a lay person can put together cheaply. The LEDs in
the volume meter aren't really necessary if a microcontroller or the
pi itself is used to simply measure the power level from each
sample/mic. And I realize I'll have to learn some programming to do this more efficiently. Just wondering if anyone has any input/tips
volume meter with a microphone array.
Each microphone may be an FM transmitter with an amplifier to amplify
the volume, or some other microphone connected device to save samples
of sounds. Filter out anything that isn't human speech to make it
easier and,
Each mic represents a different direction, north, south, east, and
west, along with a time stamp associated with it that also goes along
with a snapshot from a camera with the same time stamp as a reference.
Each sample, from each mic, is compared via the volume meter. Whatever
sample has the highest volume, is going to show the direction of a
specific word or phrase.
An application like Jasper can be used on a raspberry pi to save
samples of sounds/voices and covert them to text. The text will have
the digitized version of the analog signal from the sound also
associated with it, and that can then be played through a speaker to
the volume meter. The pi will cycle through each mic's sample, and
play it through one speaker for the volume meter.
This can be done quite quickly, and a buzzer can vibrate the
microphone/device that showed the highest volume, indicating the
direction of a particular word or phrase.
If there's an easier or cheaper way to do this I don't know what it
is. If you're only interested in a few different words, this might
make things easier. Those words are filtered out and saved, whereas
everything else is either saved elsewhere or discarded.
Reason I wanted to use an FM transmitter is because I already have the
components for at least a few of them and a raspberry PI can only use
a USB microphone, so with a RTL SDR I can record at least 5 channels
simultaneously, in other words the audio from 5 transmitters
simultaneously. It just has to be played through the speaker so the
volume meter can determine which one produced the loudest sample.
This could of course be used for other sounds like footsteps from
animals or humans.
another thing is, the transmitters can be used for other purposes once
detached, as with the SDR having multiple purposes. If actual
coordinates could be matched with the direction of sounds, then a
compass can be used to locate the source of sounds, rather than simply
a reference point from a mic or snapshot.
A smartphone app may replace the volume meter, also, but the
touchscreens can be a pain in the ass to use, so I'd probably rather
go with the volume meter as it's only for that specific purpose.
Anyway I obviously have minimal knowledge of this stuff so I'm just
looking for what a lay person can put together cheaply. The LEDs in
the volume meter aren't really necessary if a microcontroller or the
pi itself is used to simply measure the power level from each
sample/mic. And I realize I'll have to learn some programming to do this more efficiently. Just wondering if anyone has any input/tips
