And I think that Rick, KN6KB, was being modest about the 80% detection. I did not find that the software would ever transmit on what I, as a human, would have considered a busy frequency. However, there were times that it did not want to transmit because of what it perceived as a busy frequency, but I would have.
The one thing that may have to be improved is the ability for the software to ignore a continuous carrier caused by a local internal or external birdie as it is extremely sensitive to the slightest carrier, even ones you can barely see on the screen. Even fleeting carriers. It seems to me that if you can detect SSB, then you can pretty much detect most modulation. The software had the ability to be adjusted by the operator for the level of detected signal by "x" dB using an on screen slider. I hope others will suggest to Paul Rinaldo, when they submit their recommendations for an HF digital mode, (or maybe an addendum?), that this software is already invented and ideally should be used, rather than having to reinvent it. Rick seems like a very reasonable person to me and not quite as much into the politics of the Winlink 2000 systems as the main owner/administrator. And remember that that ARRL may be able to provide some input into this considering that they so strongly supported Winlink 2000 with a "stacked" committee that insured a particular outcome in the decision making to support this kind of activity. Based upon the overall attitudes one gets from the Winlink 2000 administrator and his supporters, I would expect that the last thing they would want is to have a competitive system, that builds upon other systems, such as PSKmail, and incorporates some of the SCAMP software components, run at a moderate to high speed, and do it on the MS OS (Microsoft Operating System) platform. And yet, that has to be what will eventually evolve if we are able to set up truly robust and decentralized systems wherever we want them and need them and not be under the control of a central group. 73, Rick, KV9U Dave Bernstein wrote: > As is often the case in engineering, Jose, perfect is the enemy of > good. What Rick KN6KB discovered while developing SCAMP's busy > detector was that he could detect CW, PSK, Pactor, and SSB at an ~80% > confidence level without enormous difficulty. SCAMP beta testers were > amazed by the effectiveness of this first iteration. > > Pushing the confidence level from 80% to 100%, however, would take > years -- if its even possible. But a busy detector that works 80% of > the time would cut QRM from unattended automated stations (like > WinLink PMBOs) by a factor of 5! > > Your comment that "many think it is simpler than it really is to do > it WELL" is frankly moot; Rick demonstrated two years ago that useful > busy frequency detection was implementable on a PC and soundcard. > > 73, > > Dave, AA6YQ >
