I’ve not tried this trunking software yet as I’m more into monitoring avation traffic, adsb, acars, and voice traffic. I’ve thought about the application of AI for this sort of thing but haven’t looked into it much yet, but you’re motivating me 😊.
Great idea, I want to try this with OpenAI Whisper on my own server(no GPU) and see if the performance is good enough to not have to use a cloud provider.
I was just wondering if you are using an API to OpenAI on the PI or using a custom model on AWS? Either way, it's a very cool project. Thanks for sharing.
It's actually just the native AWS transcribe model, I haven't custom tuned it yet. No OpenAI involved, check out the GitHub link for an example (you will need to set up AWS credentials)
Hi, I would love to help contribute to the development, and also to satisfy the needs that I have. I want to do this with the police scanner app, or scanning website. I like that you are using the RTL SDR dongles, but even with giant antennas, my basement apartment blocks almost all signals, even right next to a window. I pick up very staticy analog audio, and digital audio streams never pick up. If I can get an antenna outside, then I could make a setup similar to what you have, and decode additional talk groups to watch for things of interest, but since I rent, I really doubt that will happen. So my idea is, run the police scanner app / website in the background, that handles the police audio, then have the program constantly transcribing all the traffic, by saving 1 min audio files, of course there's going to be a small delay, but that's okay. I also want it to show on screen like you have, send me an alert for things that match my criteria, through something like MS Teams or Slack, which I know how to do. Then, additionally, play an aubile chime or message. That chime could just be a notification sound from that Slack/Teams channel or through Home Assistant which plays on my Google Homes. The reason why I want to do this is because I've been swatted quite a bit. It's not like I'm super popular or celebrity or anything, but I'm well known enough in my area and in my town, that I've been swatted several times, including a former employee who continues to try to call the police on me. I've been interested in this idea for quite some time, and I periodically search for police radio transcription options. There was another project I found that also used the Amazon library and one that used a Google library.
This is going to push departments to full-time encryption so fast. Really wish people would stop talking and showing this stuff. It's going to kill the hobby
I’ve not tried this trunking software yet as I’m more into monitoring avation traffic, adsb, acars, and voice traffic. I’ve thought about the application of AI for this sort of thing but haven’t looked into it much yet, but you’re motivating me 😊.
very impressive!
🔥
Great idea, I want to try this with OpenAI Whisper on my own server(no GPU) and see if the performance is good enough to not have to use a cloud provider.
Let me know!!
Totally interested in this and I'm hoping will be compatible with DSD+, I assume it would be as DSD+ has the same recording and output functions
I was just wondering if you are using an API to OpenAI on the PI or using a custom model on AWS? Either way, it's a very cool project. Thanks for sharing.
It's actually just the native AWS transcribe model, I haven't custom tuned it yet. No OpenAI involved, check out the GitHub link for an example (you will need to set up AWS credentials)
Hi, I would love to help contribute to the development, and also to satisfy the needs that I have.
I want to do this with the police scanner app, or scanning website. I like that you are using the RTL SDR dongles, but even with giant antennas, my basement apartment blocks almost all signals, even right next to a window. I pick up very staticy analog audio, and digital audio streams never pick up. If I can get an antenna outside, then I could make a setup similar to what you have, and decode additional talk groups to watch for things of interest, but since I rent, I really doubt that will happen.
So my idea is, run the police scanner app / website in the background, that handles the police audio, then have the program constantly transcribing all the traffic, by saving 1 min audio files, of course there's going to be a small delay, but that's okay. I also want it to show on screen like you have, send me an alert for things that match my criteria, through something like MS Teams or Slack, which I know how to do. Then, additionally, play an aubile chime or message. That chime could just be a notification sound from that Slack/Teams channel or through Home Assistant which plays on my Google Homes.
The reason why I want to do this is because I've been swatted quite a bit. It's not like I'm super popular or celebrity or anything, but I'm well known enough in my area and in my town, that I've been swatted several times, including a former employee who continues to try to call the police on me.
I've been interested in this idea for quite some time, and I periodically search for police radio transcription options. There was another project I found that also used the Amazon library and one that used a Google library.
This is great idea any updates? or a step by of how this was done?
Just found your videos. What are the rough costs for the AWS transcribe service? (I assume it's pay as you go/use?)
Hi! Yes you pay for however much audio you transcribe (time based). It's around 2 cents per minute!
Hey, this could be super useful for my job. Would it be alright to e-mail you?
Yeah just added my email to the channel
try that in the uk all emergency and government services are encryped and have been for 30years
Compile it to windows version and you can take my money...
This is going to push departments to full-time encryption so fast. Really wish people would stop talking and showing this stuff. It's going to kill the hobby