DV Dongle Downer

A DV Dongle arrived in the post today - somewhat of a surprise as the online store of Waters and Stanton were showing it out of stock when I ordered it a few weeks back, and is still showing "Back Order" as I write this. I must have been lucky for once!

Anyway, the DV Dongle is a device that simply provides the hardware codec for AMBE, the digital audio compression standard currently used in the D-STAR DV mode. It is somewhat controversial, being a proprietary component of the otherwise open D-STAR spec. The DV Dongle isn't cheap, but it's the only product that I'm aware of that enables computer software to act as a D-STAR DV gateway to the network. Obviously, this has the upshot of allowing you to participate in D-STAR calls even when having a radio isn't too convenient (such as when I'm travelling).

A least that's the plan.

The DV Dongle fully supports the Mac, and indeed the supplied software to drive it "DV Tool" is a Java app, employing the Java audio framework, along with some native components to shuffle audio streams to and from the DV Dongle, connected to a USB port. In the installation instructions you are requested to download the latest serial drivers for the FTDI UDB-Serial bridge chipset. I had done this several weeks ago as I have other devices that use this chipset (including the Black Cat USB to CI-V interface). In any case, the software starts up and seemed to connect to the dongle hardware OK.

My problems began when trying the simple loopback audio tests with my Logitech headset (a recommended configuration). The software offers two loopback tests - one without needing the dongle (presumably just the Java audio stack), and one via the AMBE encoding/decoding hardware. My experience with either of these was not good - choppy sound with a distinct periodic transient (less than 0.5s I think), and a nasty lag. The lag with the dongle in the circuit is getting on for being between 2 and 3 seconds! In any case, the results are hardly broadcastable, and I believe I have a problem somewhere that is introducing enough latency that the audio streams are starved with the 0.5s period. I could also be sequential/blocking behaviour somewhere, where asynchronous behaviour is required. Anyway, at the moment I can only guess.

I've made a posting on the DV Dongle Yahoo discussion list, and so far I have been entreated to check that I am using the 1.6 Java stack (rather than 1.5 that apparently has issues). This has certainly been the case, though I have switched between 1.6 and 1.5 in a quest to find if that makes a difference - it did not.

So, hopefully some further ideas will come forth and I can experiment some more. I'm assuming that normal operation results in smooth/clear audio, and it looks like people are successfully using this on Macs - so I remain optimistic that this will turn out to be some irritating little piece of configuration that will eventually be a simple fix. Fingers crossed.