When reducing the BW, the signal doesn't change, but the amount of noise does, and that changes the S/N. In a system that didn't do optimum digital filtering, reducing the BW to match that of the signal provides a S/N improvement. In SSB, for example, increasing the BW to more than the signal BW decreases the S/N.
Here's a paper that discusses it in the general case: https://www.wiley.com/legacy/wileychi/hbmsd/pdfs/mm395.pdf <https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=18&ved=2ahUKEwjBxOqM6JLeAhUOJt8KHd4DBlsQFjARegQICBAC&url=https%3A%2F%2Fwww.wiley.com%2Flegacy%2Fwileychi%2Fhbmsd%2Fpdfs%2Fmm395.pdf&usg=AOvVaw2AAgDJSbeQLpAJ5g4OSo-4> However, as Joe pointed out, it doesn't provide a decoding advantage, because the digital filtering (I assume using FFT bins) already filters at the optimum bandwidth. So unless there is AGC pumping from a very strong signal or something like that, the pre-filtering I suggested doesn't help. Oh, well... 73, Frank KF6E On Fri, Oct 19, 2018 at 11:18 AM Black Michael via wsjt-devel < [email protected]> wrote: > A decrease of bandwidth by a factor of 6 will increase reported SNR by > approximately 16dB. > But that's just a matter of what noise reference you use...not any real > change in the signal level. > If you used signal peak instead of RMS the SNR reported would be really > low....it's all relative to the value chosen for N. > > You could just as well divide the signal level by some really small number > and have 100's of dBs shown....the signal never changed though. > > de Mike W9MDB > > > > On Friday, October 19, 2018, 9:58:10 AM CDT, DG2YCB, Uwe <[email protected]> > wrote: > > > Frank, seems indeed to work! Just tested during RX of station KG4HF with > bandwidth of either 3 kHz or 500 Hz. See the following S/N comparison. > Astonishing! > > > > > > 73 de Uwe, DG2YCB > > > > *Von:* Frank Kirschner [mailto:[email protected]] > *Gesendet:* Freitag, 19. Oktober 2018 15:48 > *An:* WSJT software development > *Betreff:* [wsjt-devel] Improving S/N > > > > One thing I haven't seen discussed on this reflector is improving the S/N > by narrowing the receiver bandwidth. It is no surprise that decreasing the > bandwidth received increases the S/N, by 10 to 15 dB, sometimes more. When > I see a station I want calling at -24 or so, I can narrow the BW and get > solid communications. This is very easy with the graphical presentation and > digital filtering of the Flex 6600, but with a little practice, could be > done on any modern receiver. > > > > This means that, if we knew where the stations were calling at -30 or so, > we could focus on them and bring them up to the point of making contact. > Since you can't decode stations at that S/N, it would have to be done "out > of band." Is there any thought being given to setting up an FT8-only DX > cluster with exact frequencies? > > > > What a fascinating hobby! > > > > 73, > > Frank > > KF6E > _______________________________________________ > wsjt-devel mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/wsjt-devel > _______________________________________________ > wsjt-devel mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/wsjt-devel >
_______________________________________________ wsjt-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/wsjt-devel
