On 25/06/2016 20:54, Jim Bennett wrote: > Now, what is my suggestion? How about if during the decode pass, that > a signal decoded within +- a few Hertz of the operator's current > receive frequency is the FIRST thing sent out to the decode display > window? That would give the operator of the RPi (or any other "slow" > processor) enough time to decide if his next transmission should be > aborted or proceed as normal. If the "decode code" would look to see > if (1) the Enable TX is turned on, and (2) there is a signal very > close to the receive frequency, dump it out to the display first. Now > THAT would be cool... :-)
Hi Jim, the way the decoding is done is not quite as you describe. We are very aware that not all CPUs are fast enough to decode signals on an HF band as fast as users would like. The first thing that is done is that the decoder concentrates on a small window around the DF of the Rx cursor and will always decode a detectable signal in that window first. This ensures that once you are in a QSO or simply have clicked on the waterfall or an interesting signal or double clicked a decode you are interested in; then an attempt to decode that signal is completed before anything else. When running JT9+JT65 mode there is another strategy employed as well if you CPU has multiple processing threads available, the two modes are decoded in parallel each utilizing a different CPU core. If you are seeing one mode *always* decoded before the other then either you do not have a multi threaded CPU or you have the settings such that one mode is taking much longer to process. If you are seeing *any* interleaved decodes (between the two modes) then you definitely have a multi threaded CPU. Things that can slow down decoding of a mode include: having the JT9 separator (blue line) set too low - this will slow down JT9 decodes because JT65 signals in the JT9 decoder passband slow down the JT9 decoder a lot. For JT65, with the latest decoder, you have control of the effort applied to try and decode detected signals - in "Settings->Advanced" you can adjust the "Random erasure patterns" spin box. This control can be set as high as 7 or 8 on fast CPUs but if you have a slow processor try lower levels for much snappier decoding. Lower values will reduce the effective sensitivity of the decoder but the difference in number of decodes is probably going to be very very small compared to the potential time savings. If nothing above helps with your situation then your best strategy is to use single mode but that is unlikely if you CPU is multi thread capable - which yours is. 73 Bill G4WJS. ------------------------------------------------------------------------------ Attend Shape: An AT&T Tech Expo July 15-16. Meet us at AT&T Park in San Francisco, CA to explore cutting-edge tech and listen to tech luminaries present their vision of the future. This family event has something for everyone, including kids. Get more information and register today. http://sdm.link/attshape _______________________________________________ wsjt-devel mailing list wsjt-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/wsjt-devel
