On 8/29/21 12:30 PM, Gordon Weast via wsjt-devel wrote: > Ed, > > When I look at decodes on 20m today, I've seen multiple times when > signals offset by as little as 1 Hz are decoded just fine. > Basically, the protocol works fine with lots of overlap. There is NO > need to channelize transmission frequencies. I don't see that that necessarily follows. It is always desirable to try to make the channel utilization more uniform. The ultimate fix here is to frequency hop between every *symbol*, not every transmission, and to use FEC to fix the inevitable collisions. That's true CDMA. But it's not backward compatible with FT8, so it's a separate topic. > > Restricting transmissions to the channels you're proposing would > severely restrict the number of simultaneous users.
Why? The protocol would not necessarily inhibit transmission when it thinks a channel is busy. It could look at the signal levels and conclude that it wouldn't necessarily interfere with another transmission even if it transmitted at the same time. > > In any event, you can't avoid overlap since what looks like a clear > channel at your location will probably not look like a clear channel > at a different location. Again, that's the point of my original MACA paper -- that simply sensing whether or not the channel is busy doesn't really tell you what you want to know. Collisions occur at receivers, not transmitters. Phil _______________________________________________ wsjt-devel mailing list wsjt-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/wsjt-devel
