Perhaps a nice addition to the scheduler would be ability to flush out the play counters to zero by group or scheduler code, this would be a very useful feature for currents, something you could do like once a week or when adding new tunes. On Oct 10, 2011 4:45 PM, "Robert Jeffares" <jeffares.rob...@gmail.com> wrote:
> The Scheduler makes a selection for [in this case MUSIC] by looking > through the database for a track that matches the GROUP you have > specified [we have SINGLE OLDIE ALBUM PLAYLIST] and then the SCHEDULER > CODE you may have specified as well. > > At some point it looks at the Scheduler Rules to see if you have allowed > this kind of track to play next. We don't play 'slow' after 'fast' and > we favour 3 in a row 'reggae' 'country' 'blues'.. > > It works for us > > Of the tracks that meet this criteria it picks the one with the LEAST > NUMBER OF PLAYS. > > If you have just added a track to a GROUP in which the majority of > tracks have been in the system for some time and have been played a > number of times your new track is more likely to be scheduled tomorrow > until it racks up a few points. > > This is/was known as a bubble rotate back in the day, and is not a bad > thing, IMHO, however this may not suit you. > > It's easy enough to go into mysql and set all the play counters at zero, > which solves the problem. > > Over time all tracks get even airplay unless you get carried away with > the scheduler rules. > > > > Now this opinion is not based on an intimate knowledge of the scheduler > code, but on a long term observation of it's behavior, and some analysis > of the play counters to settle an argument with a constant listener that > we play a song he hates far too often. [It airs just under once a week > on average] In fact we found that there were many songs that played > slightly more often, and reviewing the list found he liked all the > others, it's just one that irritated him. > > If you are going to add tracks, add a bunch if you can, it softens the > impact, if that's an issue > > regards > > Robert Jeffares > Big Valley Radio > Thames > New Zealand > > > On 09/10/2011 22:27, Hoggins! wrote: > > Hello Scott, > > > > Could you point me where in the code the scheduler's "brain" is ? A few > > months ago, I briefly looked into it without any findings, and I would > > be very interested in the way it "thinks". > > > > As for our problem, it looks like we are quite alone on this one. But > > I'm pretty sure everything is well configured, as we followed the ROG to > > set up our events, clocks, etc. > > > > Maybe should we nuke it all and start from scratch ? Of course, no harm > > would be done to the music database. But how could we know if those > > tables themselves are not the cause of our problems ? > > > > I must say it is a very strange behavior. > > > > Hoggins! > > > > > > > > Le 08/10/2011 21:10, sedwa...@xmission.com a écrit : > >> Hi Hoggins > >> > >> We have not seen any problems with artist separation, we have it set > >> in the groups and it seems to work (we are using 2.0.2 on Ubuntu 11.04). > >> > >> We had some issues with the same cart playing several times in one day > >> and the same carts playing two days in a row at about the same time. > >> We couldn't figure out how to prevent that so I went in and hacked the > >> scheduler so that it tries not to play any carts that have aired in > >> the last two days until we can figure it out (isn't Free Open Source > >> Software wonderful!). > >> > >> -Scott > > > > > > _______________________________________________ > > Rivendell-dev mailing list > > Rivendell-dev@lists.rivendellaudio.org > > http://lists.rivendellaudio.org/mailman/listinfo/rivendell-dev > > -- > Robert Jeffares > Big Valley Radio > Thames New Zealand > _______________________________________________ > Rivendell-dev mailing list > Rivendell-dev@lists.rivendellaudio.org > http://lists.rivendellaudio.org/mailman/listinfo/rivendell-dev >
_______________________________________________ Rivendell-dev mailing list Rivendell-dev@lists.rivendellaudio.org http://lists.rivendellaudio.org/mailman/listinfo/rivendell-dev
