<html><body><div>Kevin,<br>  This sounds really hopeful.  Are you still working on the idea that latency is not permitted?  Or are you allowing latency?  <br>  I would really like to see this working across G8BPQ nodes.  If the system used G8BPQ on both ends, we could use SLIP or twisted pair over-audio and take the radios out of the test.  I would come up with test radios and digital hardware if this would help.  I could do desktop PCs or Raspberry PIs.  I can loan DRSI/Kantronics 9600 baud data radios (2 ends)  as well as TNC-PI 1200 baud devices.   <br>  You and I are almost in network range with the TARPN system.  I'm putting up a 3-port node in the next two weeks weather permitting at NC86 x i-40.  <br></div><div>    Tadd / KA2DEW - Raleigh NC<br><br><a href="http://tarpn.net">http://tarpn.net</a>      Terrestrial Amateur Radio Packet Network<br>tadd@mac.com<br><br>On Dec 30, 2014, at 12:44 PM, Kevin Otte <nivex@nivex.net> wrote:<br><br><div><blockquote type="cite"><div class="msg-quote"><div class="_stretch"><span class="body-text-content"><span class="body-text-content">I migrated the code from my personal SVN over to GitHub to make <br>interaction easier. I also made some headway trying to do text encoding <br>of codec2. Full rundown with github links: <br><a href="https://groups.google.com/forum/#!topic/digitalvoice/vqJ-HnER2lU" data-mce-href="https://groups.google.com/forum/#!topic/digitalvoice/vqJ-HnER2lU">https://groups.google.com/forum/#!topic/digitalvoice/vqJ-HnER2lU</a><br><br>On 08/24/2014 02:38 PM, Kevin Otte wrote:<br></span></span><blockquote class="quoted-plain-text" type="cite">tl;dr:</blockquote><blockquote class="quoted-plain-text" type="cite">- Code is at <a href="http://www.nivex.net/svn/dvopus/" data-mce-href="http://www.nivex.net/svn/dvopus/">http://www.nivex.net/svn/dvopus/</a><div style="width:0px; height:0px;"> </div></blockquote><blockquote class="quoted-plain-text" type="cite">- I think I've got some buffering problems.</blockquote><blockquote class="quoted-plain-text" type="cite"></blockquote><blockquote class="quoted-plain-text" type="cite">So, I had a crazy idea: send low bitrate (I chose 7.2Kbps based on what</blockquote><blockquote class="quoted-plain-text" type="cite">other protocols are using for "high bitrate" to leave room for headers,</blockquote><blockquote class="quoted-plain-text" type="cite">etc.) Opus data using 9600 baud KISS TNCs. I have a Kenwood TH-D7Ag and</blockquote><blockquote class="quoted-plain-text" type="cite">a TM-D700A, both of which have those in it. Piece of cake, right?</blockquote><blockquote class="quoted-plain-text" type="cite"></blockquote><blockquote class="quoted-plain-text" type="cite">Well, getting the encoding and the KISS framing coded up took some time,</blockquote><blockquote class="quoted-plain-text" type="cite">but it was all pretty doable. Piping the output of the transmit program</blockquote><blockquote class="quoted-plain-text" type="cite">to the receive program via an mkfifo works just fine (with headphones!</blockquote><blockquote class="quoted-plain-text" type="cite">The delay is low enough to create feedback). Hooking it up to the</blockquote><blockquote class="quoted-plain-text" type="cite">radios, notsomuch.</blockquote><blockquote class="quoted-plain-text" type="cite"></blockquote><blockquote class="quoted-plain-text" type="cite">First I tried sending from the D7 to the D700. My ears heard data, but</blockquote><blockquote class="quoted-plain-text" type="cite">the D700 never returned any frames. I flipped things around and data</blockquote><blockquote class="quoted-plain-text" type="cite">started flowing, but it didn't stay that way for long.</blockquote><blockquote class="quoted-plain-text" type="cite"></blockquote><blockquote class="quoted-plain-text" type="cite">The transmitter would burst a few frames, then unkey, and start again.</blockquote><blockquote class="quoted-plain-text" type="cite">The txdelay is 200ms, so that was killing any real throughput. I imagine</blockquote><blockquote class="quoted-plain-text" type="cite">the computer would be more than capable of keeping the buffer full so it</blockquote><blockquote class="quoted-plain-text" type="cite">wouldn't have to unkey, so I wonder if I have the other problem. That</blockquote><blockquote class="quoted-plain-text" type="cite">is, I'm swamping the buffer in the radio. These things were only ever</blockquote><blockquote class="quoted-plain-text" type="cite">designed to send APRS frames, and KISS mode isn't even in the manuals.</blockquote><blockquote class="quoted-plain-text" type="cite">Eventually I overran a receive buffer somewhere (probably missed a frame</blockquote><blockquote class="quoted-plain-text" type="cite">separator) and tripped a failsafe in the code.</blockquote><blockquote class="quoted-plain-text" type="cite"></blockquote><blockquote class="quoted-plain-text" type="cite">So, I have a prototype, but it doesn't really work. A partially blinking</blockquote><blockquote class="quoted-plain-text" type="cite">"Hello World" LED is still an improvement over a doodle on the back of a</blockquote><blockquote class="quoted-plain-text" type="cite">napkin, but this still might not be viable.</blockquote><blockquote class="quoted-plain-text" type="cite"></blockquote><blockquote class="quoted-plain-text" type="cite">Some ideas for improving matters:</blockquote><blockquote class="quoted-plain-text" type="cite">- Increase the bitrate, maybe 8Kbps</blockquote><blockquote class="quoted-plain-text" type="cite">- Adjust the frame size. not sure whether to go up or down. Up (60ms)</blockquote><blockquote class="quoted-plain-text" type="cite">would reduce the per-frame overhead but increase the latency. This might</blockquote><blockquote class="quoted-plain-text" type="cite">help the buffering issue too.</blockquote><blockquote class="quoted-plain-text" type="cite"></blockquote><blockquote class="quoted-plain-text" type="cite">Anyway, that's what I've been messing around with lately.</blockquote><blockquote class="quoted-plain-text" type="cite"></blockquote><blockquote class="quoted-plain-text" type="cite">73 de Kevin N8VNR</blockquote><blockquote class="quoted-plain-text" type="cite">_______________________________________________</blockquote><blockquote class="quoted-plain-text" type="cite">Linux-ham mailing list</blockquote><blockquote class="quoted-plain-text" type="cite"><a href="mailto:Linux-ham@trilug.org" data-mce-href="mailto:Linux-ham@trilug.org">Linux-ham@trilug.org</a><div style="width:0px; height:0px;"> </div></blockquote><blockquote class="quoted-plain-text" type="cite"><a href="http://www.trilug.org/mailman/listinfo/linux-ham" data-mce-href="http://www.trilug.org/mailman/listinfo/linux-ham">http://www.trilug.org/mailman/listinfo/linux-ham</a><div style="width:0px; height:0px;"> </div></blockquote><blockquote class="quoted-plain-text" type="cite"></blockquote><span class="body-text-content">_______________________________________________<br>Linux-ham mailing list<br><a href="mailto:Linux-ham@trilug.org" data-mce-href="mailto:Linux-ham@trilug.org">Linux-ham@trilug.org</a><br><a href="http://www.trilug.org/mailman/listinfo/linux-ham" data-mce-href="http://www.trilug.org/mailman/listinfo/linux-ham">http://www.trilug.org/mailman/listinfo/linux-ham</a><br></span></div></div></blockquote></div></div></body></html>