no subject

[firecat]

Does your iTunes randomizer have favorite songs? I have over 3000 songs in mine, and this one has come up at least once a day over the past week.

Comments

[pixiecrinkle.livejournal.com]

Mine gets stuck in certain artists or themes it seems. I had it going one night and some friends were over and talking about keeping parity between gay men and lesbians in a political group they were involved in and I swear to you every song that came on was by Ani Difranco or Melissa Ferrick. It was odd.

[ckd]

I have a trick to solve that; I use a smart playlist with a setting of "Last played not in the past X days". That way, no song will be played too often.

(I actually have a bunch of playlists that combine based on rating/time pairs, so 5 star songs not in the past 14 days, 4 star songs not in the past 70 days, etc.)

[firecat]

Oh neat! I hadn't explored it deeply enough to know it could do that.

I don't know that I want to solve it, but it's good to know I can...

[epi-lj.livejournal.com]

Do you know if you can do that with shared iTunes libraries? I have a fileserver which shares out all of my mp3s via rendezvous and daap. I connect to that from my Mac to listen to music.

[ckd]

I don't think you can do it with a daap setup, unfortunately; that doesn't update the last played info, in particular.

[blackfyr.livejournal.com]

How do you do that?

[ckd]

In iTunes, choose New Smart Playlist... from the File menu. You'll get a dialog box with "Match the following condition:" and a few other choices. To add additional conditions, press the "(+)" button at the right. This changes the box to let you require "any" or "all" of the listed conditions; most of my playlists use "all".

In my case, my "5 @14" playlist has two conditions:
- rating is *****
- last played is not within the last 14 days

I also have "live updating" and "only checked songs" turned on. (I uncheck songs to remove duplicates of the same cut of the same song.)

I then have other playlists "4 @70", "3 @1yr", etc. The final step is a smart playlist which uses "Match any of the following" and simply lists "Playlist is 5@14", "Playlist is 4 @70", and so on. This puts them all together into one mega playlist.

One trick that works well with this and "live updating" is to turn on crossfade playback with a one or two second delay. This keeps iTunes from removing the song from the playlist before starting the next song, which would interrupt your listening....

[lysana.livejournal.com]

Oh, yeah. I've posted my iTunes top 20 before. It's amusing.

[ckd]

My Top 20 probably says too much about my moods to post.

[epi-lj.livejournal.com]

Mine has about a hundred songs that it plays orders of magnitude more often than any others, and about a thousand songs it has never played even once.

[abostick59.livejournal.com]

Assume: average song length of 3 minutes; 3000 songs; the song selector is truly random.

What is the expected number of songs which will have been played ten or more times over the course of a week?

What is the expected number of songs that won't have been played at all in a week?

These are calculable numbers.

A quick-and-dirty handwaving kluge would be to figure that there are 7*24*60/3 = 3360 timeslots to fill with 3000 songs, so that the expected number of times a single song was played is 1.12, and plug it into the Poisson distribution.

Results left as an exercise to the student, who might turn out to be surprised.

I just did the math

[abostick59.livejournal.com]

For the assumptions I listed, here is what I get:

Expected number of songs never played: 978.84
Expected number of songs played once: 1096.3
Played twice: 613.93
Played thrice: 229.2
Four times: 64.18
Five times: 14.38
Six times: 2.68
Seven times: 0.43
Eight times: 0.06
Nine times: 0.01
Ten times : 0.00

We'd expect to hear between either two or three songs played six times over the course of the week. There is a 43% chance that there is one song played (on average) every day.

This has been an obsessive-compulsive moment.

Re: I just did the math

[firecat]

Thanks! You rule!

(I would have said so sooner, but there's no net access on the Mendocino County coast.)