Another imagineware post – in which a stripped-down version of radio station programming software is proposed to integrate with iTunes and similar database music players.

One of the things I like most about using iTunes as my main music player is that it does a couple of things very well. It’s a database of music, so it enables me to organise and search my collection easily. I can build playlists and compilations. I can randomise those playlists or my entire collection, and I can even specify queries that trawl my collection looking for tracks that meet a certain range of criteria (year of release, genre, etc.), and bulid those into automated ‘smart’ playlists.

It seems that there’s only one more step to go in order to build the bridge between what iTunes can do, and what a piece of professional radio programming software like ‘Selector‘ is capable of.

Selector is, in its simplest form, based on the way music used to be chosen at radio stations 50 years ago or more. The DJ would have a box of records in the studio, and would select the record at the front of the pile, play it, then put that record to the back of the pile and select the next one.

Of course, DJs are not to be trusted, so radio stations inserted a method of control. Instead of piles of records, a playlist would be generated by a music programmer.

The programmer would have a pile of cards that represent the box of records. Each card had the name of a record on it. The programmer would take the front card, write the song on the playlist, and then put that card to the back and select the next one. What the DJ actually played on air could be checked against the handwritten playlist.

However, radio requires a greater degree of complexity than that – so the idea of more than one box of cards arose – an ‘A’ list with just a few cards (perhaps a dozen) representing the real ‘hits’, a ‘B’ list with the larger collection of current popular songs and perhaps a ‘C’ list to drop in the odd perennial favourite.

The programme director of the radio station would then select a song from each of the boxes in turn, drawing more frequently from the B list to fill up the schedules, but of course with fewer songs represented in the A list, those songs would appear more often.

Of course, with separate boxes, there had to be rules. It may not be desirable to play two songs by the same artist next to each other – or indeed, within the same hour. If you play a song by the Beatles out of the A box, and then there’s a Beatles song in the B box, there needs to be a rule that says ‘put that card about 10 cards further back in the pile and select the next one’. That’s called an Artist Separation rule.

And then there’s the problem of solo artists. You don’t want to play a Beatles track and then a Paul McCartney track. There needs to be a rule for that too.

Now we’re getting a bit sophisticated, let’s have a rule that says you can’t have a really upbeat song right next to a really slow song. That just sucks the energy out of the listening experience – or vice-versa, where it’s an alarming transition. Far better to ramp up or down with a medium tempo song in between.

Also, there are some songs you don’t want to hear in the daytime, but work perfectly late at night. Let’s have a daypart rule. It’d be great if I didn’t hear the same song at noon today that I heard at noon yesterday, so let’s make that a rule too. It’d be good if say 30% of the music was from New Zealand – but to make it a bit flexible, let’s say 30% within any 3 hour block. Insert content rules.

The Selector software, and its few competitors in the radio station software game have got this database management of music nailed. It works really well. You might not agree with the selections your local radio station ends up with, but it’s hard to fault the software that makes it all go – it wraps the element of surprise up in the satisfaction of predictability. You can’t be sure what you’re going to get next, but you know it’s not going to get it wrong in the ways that you’ve planned for.

Of course, at the professional level, it’s all a bit more complicated than this. There are some black arts and tricks that radio programmers employ to keep their stations tight, polished and successful – according to their own criteria of success.

Now, of course, all this control over the parameters of the music playlist enables the safe programming that makes so much of music radio so incredibly dull. But the software doesn’t cause the safe programming. It’s the economics of commercial radio that has blanded out the airwaves. But personal music collections aren’t subject to the same economic pressures. I like to be surprised, but I don’t want to be disappointed or irritated.

One of the joys of being a music fan is in the organisation of the music – not simply, as many people suspect, the hoarding. I don’t want to bother with a complete professional radio programming software package just to build a few rules into my smart playlists – but I do think I should be able to make some rules.

For instance, I’d like to build a playlist made up of 30% jazz music from 1955-1965, 20% tracks from recent albums that I’ve given 4 or 5 stars to, and the other 50% from the broad collection. I have a handful of current favourite songs that I want to hear more often than everything else, but I don’t want to hear two Talking Heads songs in the same 2-hour block.

And then I’d like to make another playlist that is my own personal nostalgia station, with songs drawn from the period 1979-1987. Ian Dury, the Members and XTC should come up far more often than Men Without Hats or early Depeche Mode, but there has to be some cheese in there to give it that laugh out loud cringe factor that makes good nostalgia playlists so worthwhile. This time, however, I’d be quite happy to get a couple of two-in-a-rows in there. One good Smiths song deserves another.

This need not be terribly complex or sophisticated – and I think it would be a fairly simple addition to build in to iTunes version 7 – or make a fabulous third-party software plugin. Something drag-and-droppy and entirely intuitive. The metadata that would enable it already exists within the songs on your hard drive, and iTunes is already a music management database with built-in player. It can also already stream across local area networks.

Opinion leaders (read: celebrities) are already posting their personal playlists on iTunes as ersatz compilations that you can buy from the store. Why don’t radio programmers and smart enthusiasts post their playlist parameters online? You may not have the same songs on your computer as they do – but their rules might suit your collection just fine. For instance:

Playlist rules for My Hit Radio:
15 songs from the A-rotate folder
80 songs from the Current Hits folder
30 songs from the Gold tracks folder
1-hour artist separation
2-hour track separation
Any 3 songs in a row must have a total of at least 10 stars between them
There must be at least one 5-star song in any group of 8 songs
Songs must be between 3 and 5 minutes duration.

That’d do you for a good few days of listening. And when you want to change the station, you can just use a different set of parameters. It’s still your music, it’s still going to come up in an unexpected order – but it’s just that much smarter than simply clicking the ‘random’ button.

Trouble is, I’m not a coder and have no burning desire to become one. So again, this idea’s up for grabs. Call it Shuffle 2.0 – or perhaps MyStation. Something like that.

Again, if you make a pile of cash out of the idea, 10% seems fair.

Links:
RCS Radio Software
iTunes
A Guide to Creative Radio Programming

Tags: radio, music, playlist

Previously: Real iGroovy