HD EAS fill in missing fields

[TheDaChicken]

The implementation of NRSC5's EAS decoding has some unknown data that is skipped. I came across this video where it shows time of the EAS alert which may or may not be apart of the message. https://youtu.be/5ZpY9mv9Es8

I wonder if somewhere in NRSC5 EAS contains the time and duration/end time of the EAS alert??
In America, SAME contains the start and duration time: https://en.wikipedia.org/wiki/Specific_Area_Message_Encoding#Header_format

[pclov3r]

It looks like yes?

Testing my resampling tool from a remote server in ND this happened so i guess so.

THo, Not sure if there is actual data that sets the start and stop time or if it's just text.

[argilo]

I expect @TheDaChicken is referring the unknown bits in the control data, rather than the text of the emergency alert:

nrsc5/src/pids.c

Lines 182 to 190 in 310fa37

	off += 8; // unknown
	off += 12; // CNT CRC
	off += 8; // unknown
	*category1 = decode_int_reverse(bits, &off, 5);
	*category2 = decode_int_reverse(bits, &off, 5);
	off += 9; // unknown
	*location_format = decode_int_reverse(bits, &off, 3);
	*num_locations = decode_int_reverse(bits, &off, 5);
	off += 1; // unknown

When I get a chance, I'll take a closer look at those bits in the sample recordings I have. But my recollection is that they don't vary much, if at all.

[argilo]

The corresponding transmit code is here: https://github.com/argilo/gr-nrsc5/blob/7bb202d594f345934d3fd5a4f99f546f376f9aff/apps/send_alert.py#L60-L68

I expect the "unknown" bit values I put in there were the same as what I saw in all the alerts I had received.

Some clues about the control data can be found in U.S. patent 8,138,915: https://patents.google.com/patent/US8138915B2/en

Things may of course differ between what was patented, and what was actually implemented.

It seems the RX_IDD_2206 document may describe the control data, but sadly it's not public.

[TheDaChicken]

I expect @TheDaChicken is referring the unknown bits in the control data, rather than the text of the emergency alert:

nrsc5/src/pids.c

Lines 182 to 190 in 310fa37

	off += 8; // unknown
	off += 12; // CNT CRC
	off += 8; // unknown
	*category1 = decode_int_reverse(bits, &off, 5);
	*category2 = decode_int_reverse(bits, &off, 5);
	off += 9; // unknown
	*location_format = decode_int_reverse(bits, &off, 3);
	*num_locations = decode_int_reverse(bits, &off, 5);
	off += 1; // unknown

Yes, that is what I meant. The message in the control data could contain the time. My HD EAS recording didn't include any time mentioned in the message.

I wonder if the data of the unknown bits is so similar due to using the EAS "duration" instead of a timestamp. A timestamp would hugely vary. I could be 100% wrong. Control data may not include this information.

[argilo]

I had a look over my collection of recordings, and observed the following:

• The first group of 8 unknown bits is always 00000000.
• The second group of 8 unknown bits is always 10010010.
• The group of 9 unknown bits is almost always 101101101, but in one instance it was 100001100 (seen on KQMV).
• The final unknown bit is always 0.

U.S. patent 8,138,915 shows the following fields in the data structure:

• location information (4-64 bits)
• category type (12 bits)
• scope (3 bits)
• message kind (3 bits)
• status (3 bits)
• reserved (1 bit)
• CRC (7 bits)

The location information, category type (which actually occupies 10 bits) and CRC (which actually occupies 12 bits) are all decoded by nrsc5 today. My best guess would be that the group of 9 bits represents the scope, message kind, and status, and that the final unknown bit is reserved. I have no idea what the two unknown groups of 8 bits represent, but I doubt that they are a time or duration.

[argilo]

My best guess would be that the group of 9 bits represents the scope, message kind, and status

I'm not so certain about this guess, because the most common "101101101" pattern would suggest that the scope, kind, and status are all 5, but the tables in the patent only show valid values going up to 3 or 4.

[TheDaChicken]

Thanks for checking that out!

I may have another HD EAS recording of KQMV. I haven't looked through it. I am not sure if looking through it would be helpful. I do have another HD EAS station, KPNW-FM or KRWM which may also have the different group bits?

That said, those don't sound super useful? Could be interesting to decode if figured out. I'd assume that the bit count is incorrect for scope, message kind etc.

Scope may only be two bits. In the patent, the first bit in scopes are always zero. The group of unknown 9 bits starts with 10. That would be two

[argilo]

Yes, it's certainly possible that the bit lengths could be different than what's shown in the patent.

Unfortunately we have too little data to draw any conclusions at this point. More recordings would definitely be helpful. Also, I'd be interested to know whether any receivers use the unknown bits. My Sangean receivers only seem to care about the category and location codes.

The group of unknown 9 bits starts with 10. That would be two

Actually one, since integers are stored with the least-significant bit first.