iCAD Dispatch - Tone Detection

[xicarusx]

Hello,
I have created a PWA that does Tone Detection and shoots out alerts to different delivery methods.

This runs in Docker and designed to work on Linux. Windows is UNTESTED currently, but it may work. Using WSL2 on Windows is probably the best way to utilize a windows install.

Supported Tones:
Two Tone
Long Tone
HI/Low (Warble)
Pulsed (On/Off)
DTMF -> WIP semi useable results because of different speeds/lengths

Delivery Methods:
Email
Discord
Telegram
Pushover
Make.com
N8N

Helpers:
Transcribe -> Transcribe audio using an OpenAI compatible transcript service
Address Extraction -> Extract Address from Transcript using OpenAI API and Google Maps API
Incident Extraction -> Extract Incident category from Transcript using OpenAI API to classify.

Ingesting Audio:
Accepts audio in the same way RDIO does. So it should be compatible with most software that can upload audio to RDIO. It will even accept a downstream connection from RDIO.

Tone Finder:
Find tones in uploaded audio. This should help you decode tones so you can create triggers that will alert when these tones are heard.

Multi-Radio System Support:
Can upload from more than one system to one instance of the application. SO no need to run multiple instance of a tone detect software.

Triggers:
Robust Triggering system, you can trigger on more than one tone set or require more than one type of tones to match.

AND Mode -> Requres ALL tone rules to match to trigger the alert.
OR Mode -> Requires any one tone rule to match the trigger to trigger the alert

Per trigger Pushover: setup delivery groups in Pushover Per Trigger or Globally for all triggers
Enable/Disable a Trigger from appearing in Discord/Make/Telegram unless accompanied by other triggers that are enabled for those services. Example -> EMS agency you don't want to trigger alerts to Discord or Make unless it is dispatched WITH a different trigger like a Fire Department you have created a trigger for.

Text Placeholders for alert messages:
We expand variables to text to show different values.
Examples:
{trigger_list} -> Shows a list of trigger names in your text.
{transcript} -> shows transcript
{audio_url} -> expands to the URL of the audio file.

Installation and docs

GitHub - TheGreatCodeholio/icad_dispatch

Contribute to TheGreatCodeholio/icad_dispatch development by creating an account on GitHub.

github.com

 

[webstar22]

Hello,
I have created a PWA that does Tone Detection and shoots out alerts to different delivery methods.

This runs in Docker and designed to work on Linux. Windows is UNTESTED currently, but it may work. Using WSL2 on Windows is probably the best way to utilize a windows install.

Supported Tones:
Two Tone
Long Tone
HI/Low (Warble)
Pulsed (On/Off)
DTMF -> WIP semi useable results because of different speeds/lengths

Delivery Methods:
Email
Discord
Telegram
Pushover
Make.com
N8N

Helpers:
Transcribe -> Transcribe audio using an OpenAI compatible transcript service
Address Extraction -> Extract Address from Transcript using OpenAI API and Google Maps API
Incident Extraction -> Extract Incident category from Transcript using OpenAI API to classify.

Ingesting Audio:
Accepts audio in the same way RDIO does. So it should be compatible with most software that can upload audio to RDIO. It will even accept a downstream connection from RDIO.

Tone Finder:
Find tones in uploaded audio. This should help you decode tones so you can create triggers that will alert when these tones are heard.

Multi-Radio System Support:
Can upload from more than one system to one instance of the application. SO no need to run multiple instance of a tone detect software.

Triggers:
Robust Triggering system, you can trigger on more than one tone set or require more than one type of tones to match.

AND Mode -> Requres ALL tone rules to match to trigger the alert.
OR Mode -> Requires any one tone rule to match the trigger to trigger the alert

Per trigger Pushover: setup delivery groups in Pushover Per Trigger or Globally for all triggers
Enable/Disable a Trigger from appearing in Discord/Make/Telegram unless accompanied by other triggers that are enabled for those services. Example -> EMS agency you don't want to trigger alerts to Discord or Make unless it is dispatched WITH a different trigger like a Fire Department you have created a trigger for.

Text Placeholders for alert messages:
We expand variables to text to show different values.
Examples:
{trigger_list} -> Shows a list of trigger names in your text.
{transcript} -> shows transcript
{audio_url} -> expands to the URL of the audio file.

Installation and docs

GitHub - TheGreatCodeholio/icad_dispatch

Contribute to TheGreatCodeholio/icad_dispatch development by creating an account on GitHub.

github.com

View attachment 194247
View attachment 194245View attachment 194244View attachment 194243

I'm running it, blows anything currently out there out of the water. This screenshot is from my discord room where the pages are currently going, yes that is a map from ICAD.

 

[Muxlow]

I just started using it today and so far i like it. Dont have the map thing yet as i dont have the needed stuff to do that, the voice to text has been hit n miss. We dont do automated pages and the dispatchers talk super fast or away from the mic so the AI kinda gets half the page wrong.. But so far its gotten the address correct 99.9% of the time so thats a plus

Im just having everything sent to pushover (other then audio since the fpt server i use dont have the correct protocol thats needed for icad) so more or less 1 pushover alert is the audio and second is the page with all the fancy details

 

[shajoe44]

Will this work on a DMR Cap Max system?

 

[Muxlow]

Is it a fire dept on that system? How do they get paged? 2 tone or a warble tone or kinda like the p25 trunking stuff now a days when they just send an alert call to a specific radio id (base station radio) and that sets off the station alerts

 

[shajoe44]

EMS is on the system. It is a warble type sound then the dispatcher gives out the call. Fire dept has not switched over from VHF yet.

 

[xicarusx]

EMS is on the system. It is a warble type sound then the dispatcher gives out the call. Fire dept has not switched over from VHF yet.

Yes it will pick up those alert tones, it may take some tweaking of the Hi/Low Tone settings to dial it in, but it shouldn't have an issue. If they use the same frequency each time. Then you may not know which station is being dispatched, but you can send an alert based on the dispatch.

 

[shajoe44]

Thanks, I will have to dig deeper and learn how to install and use. I love learning new things here.

 

[Muxlow]

Few ways it could work

#1 Dispatch pages them over analog with their own 2 tone or long tone set. Then you can say this is station whatever. And another tone set is for another station. Kinda how things mostly work now

#2 they page out over DMR but its patched into the analog side. So its still analog paging but its coming over the DMR side. SO would work just the same since its still a 2 tone being sent

#3 they page over DMR using the warble tone but each station/department has their own specific alerting/paging talkgroup. So everytime talkgroup ...400 comes up that is station 4's pager "channel"

#4 they page over DMR with warble but they send it directly to a specific radio ID #. Station 1 gets a page it goes out on the regular dispatch or paging talkgroup for example talkgroup 100 is "paging" so you hear it on that but only radio ID 1000 hears the page

Kinda sorta have to "see" how the system is doing the paging to move on from that i think. Either way i think iCad will work its just finding out all the background stuff to make it work

If you run DSD+ on that channel it should give you some useful info as to talkgroups being used, radio id numbers and will even show if the dispatcher has sent a private call or call alert to a specific radio. Lots of fun figuring stuff out

 

[shajoe44]

I don't think it will work on this system, since it's encrypted. I am able to hear it because I have a radio issued to me by EMS service. I see everything on DSD that you mentioned. Thanks for the information.

 

[Muxlow]

Ah yeah thats a game stopper. Well... sorta
Could take the radio and put a 3.5mm cable on that from the speaker mic and connect that to a computer so it can get that audio directly from the radio itself so iCad can still get the audio it needs to work

 

[webstar22]

Ah yeah thats a game stopper. Well... sorta
Could take the radio and put a 3.5mm cable on that from the speaker mic and connect that to a computer so it can get that audio directly from the radio itself so iCad can still get the audio it needs to work

There is an app called VoxCall, you can use the RDIO section of that app and feed any audio source right into ICAD.

GitHub - aaknitt/voxcall: A sound-activated recorder with support for uploading audio to OpenMHz

A sound-activated recorder with support for uploading audio to OpenMHz - aaknitt/voxcall

github.com

Might be another app coming, but too soon for details.

 

[Muxlow]

Might be another app coming, but too soon for details.

Such a tease! But i think i understand what/why

 

[sc2142]

I would really love to use this. Do you have plans on a Windows application or no? Ive tried to do this with the instructions you have provided, however I can not for the life of me figure out Docker.