search

DSheirer / sdrtrunk

A cross-platform java application for decoding, monitoring, recording and streaming trunked mobile and related radio protocols using Software Defined Radios (SDR). Website:
GNU General Public License v3.0
1.6k stars 256 forks source link

Push Call Events to API #523

Open metaphoricalmarlinspike opened 5 years ago

metaphoricalmarlinspike commented 5 years ago

I've recently migrated my monitoring station from trunk-recorder to sdrtrunk, which has cut down significantly on the number of missed or garbled transmissions. However, for each transmission, I would like to be able to push a compressed (mp3/whatever) audio file with all the metadata from the call up to an Amazon S3 bucket or something like ScanOC. I have an existing project that processes the metadata from a call in trunk-recorder, re-formats the filename to fit my needs, and uploads it to Amazon S3. However, sdrtrunk does not seem to have an option to generate the JSON file with this metadata.

It seems like there's a better way to do this with sdrtrunk, given its event driven architecture. That said, I'm not too sure where to start--and Java isn't one of my strong points. Can someone point me in the right direction here?

davidmckenzie commented 5 years ago

This would be neat for #212 and #92 too.

My dream is to one day develop a fully web based UI that plugs into a remote sdrtrunk, but I'm also completely devoid of java skills :)

ImagoTrigger commented 5 years ago

set your aliases to Record then inspect the generated .wav files in /recordings

metaphoricalmarlinspike commented 5 years ago

I don't see anything in the audio metadata for the recordings that isn't in the filename already. Maybe VLC is missing something or I just don't know where to look? SDRTrunk outputs most of the information that I need in the file names, so I can probably just zero out the unknown fields and move on. But TrunkPlayer can also handle additional info such as emergency calls, multiple unit IDs in a call, and frequencies. At that point, the information would be getting too long to fit in the file name but could be embedded as metadata.

I have noticed that there are several files that don't have a talkgroup ID or any radio IDs attached. Most of them are shorter, but this isn't always the case. Is there any way to minimize the number of "orphan" calls like this? It's not a huge deal--in a sample of 5,381 calls there are only 4 with no talkgroup, 129 with no metadata at all, and 864 with no radio ID. Longer calls are more likely to have all the information. Is this an issue with the information not getting added to the file names, or is SDRTrunk not capturing this information in the first place? It's hard to tell from the logs because there is so much information in them. Trunk Recorder has a different process for capturing audio though; as far as I can tell, SDRTrunk will split transmissions from different radios across separate files so there wouldn't ever be a file with multiple radio IDs or frequencies.

DSheirer commented 5 years ago

Starting with release 0.4.0 alpha 7, each indiividual call segment is output as an individual recording, so ther should only be from and to radio identifiers. Frequency is included in the metadata. I'll have to add an emergency attribute.

When you see incomplete metadata for a recording, it's because the information wasn't decoded. Sometimes a call will be decoded where the beginning is missed. The from radio id is only treansmitted on every other voice frame, so there may be rare instances where that doesn't get decoded on a short call.