[pskmail] Aw: Re: Fwd: Test Alpha Fldigi version for Non-USA HF links

From: "Rein Couperus" <rein@xxxxxxxxxxxx>
To: pskmail@xxxxxxxxxxxxx
Date: Fri, 13 Apr 2012 15:20:39 +0200 (CEST)

I think what you really need for this use case (1 -> n broadcast) is a proper multicast arq protocol.

I will give an example of what could be done:

1.: Use a mode with proper FEC, like PSK500R

2.: Use blocks of 64 or 128 characters with a checksum and number them 1 - 63

3.: For every 3 blocks add a 4th block which is the xor over the previous 3 blocks

4.: Broadcast this file

5.: Most stns will be able to rebuild the whole file providing 3 out of 4 blocks are received ok.

5.: At the end of the block ask all stations for a NAK (not received 100%)

6.: If no NAKs close the transaction

7.: If a NAK is received from somebody, ask for the list of bad blocks

8.: Broadcast the error blocks

9.: Etc... until no more NAKs are received.

This guarantees 100% copy from all stations on the list.

just an example, other protocols are possible.

Rein PA0R

Thanks Harry. I see what you mean now. The important point for me is that it is not a repetitive event but more likely single events randomly distributed in time.

So maybe we could have acceptable results with a simple fixed delay redundant stream with the existing FEC scheme on top. That should be less taxing on the net data rate while providing an acceptable level of droupout resistance.

Of not much value I suspect on HF, but it makes sense on VHF/UHF FM links.

Another option may be to widen/speed up MT63 for FM use.

Let me have a look.

73, John

On 13/04/2012 6:25 AM, "Harry Bloomberg" <hpb@xxxxxxxx> wrote:

It might be everything you mentioned. One possibility is somebody kerchunking the repeater during a transmission. Or perhaps on a weak signal into the repeater the signal drops below the squelch level. Or maybe an HT antenna is swishing about and the signal drops for a bit.

73,
Harry Bloomberg W3YJ
hpb@xxxxxxxx

On Fri, 13 Apr 2012, John Douyere wrote:

David and Harry,

Thanks for the clarifications, I can now look at some options.

A question though. What are the characteristics of the dropouts you mentioned? Are they the results of operator issues like un-keing the microphone/external noise while TXing using audio coupling, or propagation issues, or others?

[pskmail] Aw: Re: Fwd: Test Alpha Fldigi version for Non-USA HF links

Other related posts: