Is the shuffle feature faulty
 

Hi there!

I am playing a 350 song playlist in shuffle mode for some time now and it seems as if certain songs do appear significantly more often than others. So I'm beginning to suspect that the random process ist somewhat faulty.
Does anybody have the same impression?

Regards,
Malte

p.s.: By the way, the shuffle window size is set to 100 %, of course.

It's your imagination.

I've definitely noticed this too. If you want to just play all your songs in a random order, randomize the playlist then play them straight through. This will ensure that the songs you hear are random and are played only once.

i think its a pretty good random playlist thingie :). I like.

I think there is actually an option in the prefs (control+P) to set how randomly songs are chosen... but that might just be for when you randomize the list, Im not sure.

Since a couple versions back (2.76, or thereabouts), Winamp started to do true shuffle. That is, like a CD player, it will create a random virtual playlist based on the entries of your real playlist. This ensures that you shouldn't hear the same song twice, until all the songs in your playlist have been played at least once. The shuffle slider controls how much this virtual playlist is randomized. If you leave the slider to the left, everytime the virtual playlist is played through, it will remain the same. If you leave the slider to the right, a new virtual playlist is created everytime the previous one has finished playing.

So, as Ice said, it's all in your head.

I'm fully aware of that. This new method of shuffle worked really well until version 3.

But I've found they haven't included the new method in 3 and you have to use the randomize option on the actual playlist to get the same effect.

Yes!
 

First, thank you for the suggestion to manually shuffle the playlist. I had the same idea and insanely automated this with a little proggie before I encountered the Randomize option.

Then allow me a comment on the feature called "The shuffle slider" mentioned by rm'. If I'm guessing right you mean the "shuffle window size" as it is called in WinAmp 3.
Again, thank you for the information. I believed that it only tells WinAmp how far it is allowed to jump from one song to the other.
If your description of the process is valid the algorithm for this is not trivial since it must incorporate something like a floating shuffle window that controls for songs remaining in that window for too long.

And a last word on the technique of creating "virtual playlists" which get shuffled, then played thoroughly and finally getting shuffled again.
This would make the shuffling process an equivalent to drawing randomly from a box without replacement.
I truly doubt that WinAmp does this in Version 3, for I today heard Frank Zappa's Bobby Brown 3 times within approximately 5 hours - given my 350 song playlist.

Have a nice day,
Malte.

Hi again.

Just did a test with a playlist containing 6 songs labelled Track01 to Track06. All tracks were 2 seconds long.
I had WinAmp perform 126 draws at random, with the shuffle window size set to maximum (-> full shuffling).

The results perfectly support the objection made by InsOmniac. The first 10 draws produced the following sequence:

4 5 3 5 3 6 6 3 3 3

This sequence should not be possible if WinAmp really set up a complete virtual playlist to shuffle it afterwards because the latter would result in all of the songs in the playlist being played before a repetition of a certain song could occur.

Second, I visually checked the distribution of the draws. They came out as shown in the attached graphic.
I didn't do any statistics on that yet since I have no statistics package running on my notebook but I bet the distribution significantly differs from what would be expected to emerge out of a random process.

Regards,
Malte.

i have noticed this as well. while i have too many songs (and am too lazy) to try this method. however, from how often i run winamp, i have noticed that songs in the "middle" of my list tend to play more. it might be in my head, but i don't think so.