Song repetition problem, with automatic playlist generation

The problem is that, from a library with more than 1000 songs to choose from, there are songs that have been played between 2 and 5 times and others that have never been played (at least 1/3 of the total).

There should be some system so that no song is repeated, while there are songs left unplayed.
 
It should be easier than all that. I have a library with 1538 songs. The system plays a song, based on the criteria (no artist repeat in 3 hours and no song repeat in 1 week). Then mark the song as played and on the next pass, choose a song that hasn't played yet. When all of them have been played once, the system starts again. That would be the ideal way. But it works differently.
In more than 2 months, from that library 371 songs have been played once, 1022 have been played between 2 and 7 times. And 145 have been reproduced 0 times. I don't quite understand what the logic of that is.
 
You can define one priority rule that will increase probability for tracks that have Playcount lower than average, it will solve the problem

In more than 2 months, from that library 371 songs have been played once, 1022 have been played between 2 and 7 times. And 145 have been reproduced 0 times. I don't quite understand what the logic of that is.
Random selection does not guarantee anything.
 
Those priority rules don't work. Furthermore, what and how determines what is the minimum or average of the number of tracks played?

Up here more than 7000 tracks, hundreds of which are never selected by playlist generator pro. Although I have set priority rules. And no, there are no songs disabled or date-bound in the track tool.
 
Those priority rules don't work. Furthermore, what and how determines what is the minimum or average of the number of tracks played?
They do work.
To calculate maximum and average it checks all tracks from category's source and does basic mathematical operations to determine those values.

Up here more than 7000 tracks, hundreds of which are never selected by playlist generator pro. Although I have set priority rules. And no, there are no songs disabled or date-bound in the track tool.
Random track selection does not guarantee a track will be selected. This is how random selection works. Out of 1000 tracks it can select 1 track 3 times in a row, that's totally possible. That's why there are repeat protection rules and priorities, so that you have some control and restrictions on the random number generator.
 
They do work.
To calculate maximum and average it checks all tracks from category's source and does basic mathematical operations to determine those values.


Random track selection does not guarantee a track will be selected. This is how random selection works. Out of 1000 tracks it can select 1 track 3 times in a row, that's totally possible. That's why there are repeat protection rules and priorities, so that you have some control and restrictions on the random number generator.
The random selection has to work in conjunction with the library database. If in the database, you mark the song as played and in the SELECT query you indicate that the song to be chosen has to be played= 0, it will limit the selection to songs that have not been played. This way you will prevent the random from repeating the songs more than once. At the end, when all of them have been reproduced, within the indicated parameters, you tell the database that the entire category will return to 0 and we start again. This way of doing this is much easier to operate than indicating priorities, which, honestly, you have to have a degree in engineering to understand.
 
The random selection has to work in conjunction with the library database. If in the database, you mark the song as played and in the SELECT query you indicate that the song to be chosen has to be played= 0, it will limit the selection to songs that have not been played.
It's not tasked to choose not played songs. It's suppose to randomly choose the songs that pass the filters, constraints and rules that you define.

This way of doing this is much easier to operate than indicating priorities, which, honestly, you have to have a degree in engineering to understand.
What's wrong with the priorities? You don't have to understand the math behind them in order to use them efficiently. You can simply define rules like "Increase priority by 100 [or whatever value you see fit*] for tracks where Playcount is less then AVG [average]" and that would be it.

* +100 priority increases track's probability to be selected by about 4-5 times.
 
Back
Top