hessi

Using iTunes, Part 5: Isn't it Genius?

Wed, 15 Sep 2010 12:40:00 +0200

In the last episode, we talked about ways to get a good selection of songs that might be easier to rate than your main library. Most of this was fairly simple - last played songs, most often played songs, good albums. Now let’s introduce a little bit of crowdsourcing, shall we?

I’m sure you are familiar with Genius. If you have enabled it, you will for sure have seen the resource hugging ‘Genius: Gathering information about your library…’ task that iTunes is scheduling once a week. It seems you need at least an i7 iMac with an SSD RAID0 to let iTunes do its Genius preparation in a reasonable amount of time.

But what exactly does Genius do? Well, basically, it scrapes your library for every bit of information it can get: Which songs do you have in your library? Which of these are you listening to? How often do you listen to them? Which songs do you put in playlists together? Combine that with some good data mining on iTunes Store browsing and purchasing information and you can imagine that Apple has a fairly good understanding of songs - who listens to them, which songs are similar to other songs, what goes well together.

Thankfully, Apple decided to not only use this feature to sell more songs (“Customers who bought x bought y as well”), but they decided to give you a handy tool to create playlists automatically, as long as you give them a song as source. Mind you, we’re talking about songs in your own library - with (this part of) Genius, they don’t force you to buy more songs. It’s a shame that this has to be mentioned as something extraordinary now, isn’t it?

So, if you give Genius a song, it tries to find similar songs in your library based on the information Apple gathered from all the other Genius and iTunes Store users out there. And at least for me, quite often the choices are really good. Of course, this highly depends on the contents of your library - it is possible that iTunes can’t create any Genius playlists from your songs - please don’t blame me for that, blame iTunes users worldwide for not listening to the music you are listening to. Or blame yourself, for being so far off the mainstream…

Anyway, I think you can see where I’m heading with this. One folder in my library looks like this:

All of the playlists in Genius Source are, as you can see by their icon, Genius playlists which I simply created by selecting a song I know and like (e.g. any of the ones you can read here, from Sugarplum Fairy’s (And Please) Stay Young to Yello’s The Race) and choosing “Start Genius”:

iTunes creates a new Genius mix, which I simply save (top right corner) as a playlist and then move into my Genius Source Folder. Done.

Now, what do we do with this? As I said, this is all about crowdsourcing. What we have in this folder are playlists full of songs that other people (or at least Apple’s Genius algorithm) found to be similar to some good songs in our library. Some of them might already have 4 or 5 stars, but some of them might not be rated at all or just have 2 stars, either because we really didn’t like them or because we thought we didn’t like them when we half-heartedly listened to them a long time ago. Time to check that:

This is my rating playlist. Let’s go through the criteria one by one:

Playlist is ‘Genius Source’: take all songs from all Genius playlists
Rating is less than 4 stars: take only songs with 0-3(.5) stars
Playlist is not ‘duplicate’: a somewhat crude hack to make sure duplicate songs are not in there - more on that later
Playlist is not ‘Genius don’t bother’: take only songs that are not in this (static) playlist
Limit to 150MB selected by least recently played: start with songs I’ve never listened to, then go to old stuff

The procedure is simple - listen to songs in there, decide whether they should get a good rating or not. If the new rating is 4 stars or better, the song will be removed from the playlist - mission accomplished, we found a new good song. If the new rating is not so good or it stays the same, the song does not get removed from the playlist. Huh. What’s with that? [Update: Obviously, this is only true if the complete number/size of songs within your Genius Folder is below your selected limit. Since the limit is by ‘least recently played’, any newly played song would be replaced by an older song. Still, it would come back, probably soon…]

Well, for some time, I had criteria in there to make sure that a song gets removed from the playlist: Last Played and Last Skipped, both set to not within the last 1 month. The problem with that is that the song might return after that, because the crowd still thinks it’s a good song, although I wholeheartedly disagree - Newsflash, hessi: Not everything is about you. They simply don’t care, unfortunately. Now, if you set the threshold for these settings very high (somewhere along the lines of 5 years), the song does not show up anytime soon, that’s for sure. But unfortunately, neither do songs you did listen to within the past 5 years, whether you distinctly judged their rating at that point or not. Say it with me: Unacceptable.

My solution is crude, I know that, but it’s the best I could come up with: Every song that should not get a better rating than 3(.5) stars will be (manually) copied into the static playlist ‘Genius don’t bother’. From that moment on, I know for a fact that I’ve checked this song and it should not show up in my Genius Rating playlist anymore. hessi trumps crowd.

Now, I still owe you an explanation for the criteria: Playlist is not ‘duplicate’. If your library is bigger than a few dozen songs, you will be familiar with songs that you own twice. They are on the original album of the artist, they are on a Best Of, they are on a compilation. It’s not a live version, it’s the same damn recording. You can’t delete any of the songs, because iTunes can’t handle multiple albums per song (yes, I know, you can create a static playlist per album, but at the moment, I have 1703 albums - no, I do not want to to create a static playlist for each of these). Now, everything would be fine if the song isn’t particularly good and only got 2 stars: it will be played in the context of that album, but that’s about it. But what if the song is really good, you possible even want to give it five stars? You can’t give all instances of that song 5 stars, because that would mean that on average, you will listen to this song with your star-related shuffle playlists twice as often as to any other, since it’s in your library twice (or thrice, or…). Worse, the second instance of the song could sneak into your quarantine period and be played just a day after the first one. Unacceptable!

My somewhat awkward solution to this is to decide on a main instance of the song (usually on its original album, if available), give that one the real rating and give every other instance 2 stars. These instances are then copied into a static playlist called ‘duplicate”. As I said before, in my library 2 stars mean that the song can be played in the context of its album, but no shuffle playlist will look at them. Some smart playlists that don’t rely on stars alone (e.g. the Genius one introduced here) check their result against the duplicate playlist and filter out duplicates. The reason for that is that I might not remember that there is a main instance of the song in the library which already got a good rating. Problem solved, works for me…

Back to Genius - every once in a while, you should modify, add or delete the playlists. Either generate a new Genius playlist from the same songs (just click on Refresh in the top right of each Genius playlist) or find new songs in your library that you want some crowd opinion on. Just save the new playlist, move it to the Genius folder and your smart rating playlist will update automatically - as long as there are songs in the new Genius playlists that are a match to the criteria and limit of it, of course.

I hope I gave you another, new way of looking at your library - make sure to tune in to the next instance of Using iTunes later this week.

Using iTunes, Part 4: Let's get the Rating started

Mon, 13 Sep 2010 22:42:47 +0200

In the last part, I promised you some way to make the tedious process of rating more bearable. Let’s start up front by saying that I will not show you some magic way to get a library with 35.000 tracks rated in an hour - this is a lot of work, especially if you want to do it right. What I will show you, though, are some hopefully clever ways to make a selection of the songs to rate, based on other information in the library. I don’t know about you, but some songs are definitely easier to rate for me than others.

Let’s start with something simple. What can I easily rate? Something I’ve just listened to:

Type Music, Rating 0, Limit to 150 MB of most recently played. Setting the Play Count to >0 is just a way to make sure that the playlist really contains only songs you’ve already listened to. iTunes has the… interesting behaviour with Limit selections to go “beyond” the selection as soon as nothing else is available. If you only have 5 unrated songs which you’ve already listened to in your library, it will top up the list with songs you’ve never listened to (admittedly, they are definitely at the bottom of a “most recently played” list). These songs seem to be ordered by Artist, although I think I’ve seen them being ordered by Album at least once, as well. Unfortunately, I can’t replicate that anymore. I thought it depended on the sorting in the main library at the point of creating the smart playlist, but that logic seems to be wrong…

How you work with such a playlist is up to you. Sometimes, I go through the list and give ratings to songs without actually listen to them again - just out of memory. Sometimes, I choose this playlist to listen to in iTunes DJ - as soon as I remember the song, I rate it accordingly and either skip it or listen to it again. Obviously, 5 star songs have a higher probability to be listened to again than 2 star songs…

Now, you could argue that it doesn’t really help you to have songs in there you have listened to 2 years ago - who remembers them now? Feel free to add Last Played in the last x days/weeks/month as another criteria.

Which brings me to the next playlist - the more often I listen to a song, the more I might like it - or just remember whether I like it. For this case, I have a second playlist:

Similar criteria, but this time, the selection for the Limit is most often played. As you can see, I decided to exclude songs already in my most recently played playlist. It is up to you whether you want to do this and whether a song should preferably go into most often or most recently played. This is my choice and it makes sense to me. Once again, I use this playlist to either apply ratings directly or listen to the songs once again.

Note that iTunes automatically applies a sorting of your playlist depending on your Limit: Most often played: Play Count descending. Most recently played: Last Played descending. If you ask me, that’s a nice touch and I mostly keep it that way. If I’m very keen on seeing progress, I will change the sorting to the first column (item number) and see what vanishes and what appears. With the item number, you can be sure that everything new will be added at the bottom, even if it would be the first item to match the criteria. To see what I mean play an old song not rated: If the playlist “recently played” is sorted by Last Played, it will show up at the top - if it’s sorted by item number, it will show up at the bottom and remove the least recently played song in the playlist.

So, we’ve covered songs which are possibly easy to rate, since we know them reasonably well - time to look for another source of unrated tracks - new songs:

The criteria is as expected - it should be Music, it should not have a rating and it should not have been played yet. All of this is limited to the most recently added songs. Now, once again, feel free to change this accordingly. It is my particular way that the actual Play Count of songs in this playlist should be 0. If it makes more sense to you to have some kind of “heavy rotation” playlist, where you put everything new and listen to it until you rate it, that’s up to you. I decided that songs I’ve listened to but didn’t rate should move into the “recently played” playlist - which they do automatically - and make room to older songs in this playlist. You might know this already: Your mileage may vary…

Next comes a playlist I don’t know whether I should really mention it - it’s damn ugly, but I haven’t found a better way yet and I have to admit I use it quite often:

Yes, I know - absolute values in the criteria. You’ve teased me about this in Part 1 already, but what can I say? It works…

So, what does it do? Imagine you’re in a line at the checkout and know you have about 2 minutes to spare. You don’t know how long it takes to read the next InstaPaper article and you don’t want to go for Twitter or your RSS feeds. Just fire up this playlist, listen to one song and rate it. Done. Depending on your library you might have a lot of spoken word and other crap in there, for example announcements in live concerts. These can be dealt with quite easy - I, for example, give them two stars: Never show up in any shuffle playlist but include in the album.

As I said, ugly but efficient…

OK, now for something more sophisticated. Imagine you have an album of which you have rated some songs quite highly, but there are still songs in your library from that album that you haven’t listened to or rated yet. There might be a good chance that you like these songs, as well. Up until iTunes 7.4, this was quite a hassle - sure, there were AppleScripts floating around to create static playlists with songs from albums/artists you seemed to like according to the rating, but it just wasn’t as nice to use as iTunes should be. Fortunately, in iTunes 7.4 Apple introduced Album Rating - without advertising it. This might be because it’s behaviour seems a little bit erratic if you just play with it, but actually, it’s not that difficult:

Album Rating will show you the average rating of an album based on the ratings given to tracks from the same album - excluding Rating 0. For example, if you have given one track of the library 5 stars and another 3 stars, and no other song of that album is rated yet, the Album Rating of this album is going to be 4 stars:

Now, these stars are a little bit different to the “official” stars: For starters, they are white instead of black. This shows you that they are applied automatically and not chosen by the user. But, and that’s an advantage as well as a disadvantage, they match criteria in smart playlists. And so we can create another playlist:

As you can see, I have to use an absolute value, once again: Album Rating is greater than 3 stars. You have to decide for yourself whether this value makes sense for you or whether it should be something like 4 or even 2 stars… This playlist will only be used if a song is not captured in recently added, recently played or most often played. See for yourself whether this works for you.

Another option I chose for this playlist is to use “least recently added” as the selector for the Limit. I decided that I already have enough playlists that look for new songs, there must be something that looks deeper, maybe even discovers old gems or - to be honest more likely - old sins of one’s youth, which in the end force me to change the rating of the songs responsibly for the average Album Rating…

I mentioned above that Album Rating’s behaviour can be erratic - we’ve already covered some of this in the comments, but I think it’s worthwhile mentioning it here, again: So far, we’ve used Album Rating as a read-only value. Note though that you can assign a rating to an album, which results in real (black) stars in Album Rating, but this has side effects: Not only does track rating effect the Album Rating, it’s the same the other way around. As soon as you give any song a real value for Album Track, all tracks of this album will get this rating and all songs that don’t have a specific rating yet will get this album rating (white stars, so at least you can see the difference):

Unfortunately, I have not yet found a way to keep these songs out of normal rating-based playlists, which means that tracks that are not rated yet could show up in your 5 stars playlists just because you rated the album with 5 stars. My advise would be to stick to one logic, either rating tracks or albums, but not both. Make use of the automatic process like I did with the playlist above, but don’t mess around with it too much…

And that’s it for today. I’ve shown you most of my playlists for rating, which so far helped me to rate close to 15.000 songs in my library. Hopefully, you will have success, as well. If you have other ideas on how to find songs that are easy to rate please share them in the comments.

I am still planning on writing about third party software to help organising your library - rest assured, it will be part of this series at some point…

Until then, have a nice day and thank you for your attention. I really enjoy all the feedback I get on these articles.

Using iTunes, Part 3: The magic of iTunes

Sat, 11 Sep 2010 10:19:22 +0200

So far, we’ve seen a few simple ways of dealing with iTunes libraries. They all have one thing in common - they depend on actually rating songs. Now, if you have a big library and have never actually rated a song… good luck. This is really tedious work, that might be fun for the first few ~~hundred~~ dozen songs, but after that…

Nonetheless, we have to start somewhere, so let’s give it a go. What are we looking for, actually? Simple: “iTunes, give me all songs that I haven’t rated yet”:

That was easy… but wait, this has a downside:

14 fuckin’ days! - And yes, this is my actual result for the playlist above, so don’t get your hopes up too high that everything will be fine as soon as you start using a “sophisticated” system…

Of course, if you like impossible challenges, take this playlist and be happy, but I am a person who likes to break stuff down, make it more accessible. Well, by now we all know how we can achieve this - the fabulous “Limit” system:

For the moment, let’s pretend we are starting from a big library, no ratings, no playlists. We don’t have any indication on what might actually be a good song. Consequently, it doesn’t really matter what we look for, we can just select the tracks to rate at random. 150 MB, that sounds good. And, bonus, you can sync it to your iP[[o|a]d|hone] and rate songs on the go. Yes, that can be fun. And more productive than playing another round of Doodle Jump or Angry Birds. If it’s Carcassonne, on the other hand…

Anyway, we have a playlist now, so let’s assign some stars! Go into the playlist and press play. Listen to the song for a while and decide which rating it should get. Ready? Apply the rating. See what happens? The song gets rated, the playlist updates itself and iTunes skips to the next song. Which is actually a surprise to me - until fairly recently (definitely iTunes 8, maybe even 9) it simply stopped after that. Good move, Apple, but - of course - still not good enough. Granted, in some cases this might be exactly what you are looking for, but let’s imagine you want to listen to the song until the end…? Or you’re so far beyond the border to insanity that you can’t take it that you’ve listened to 85% of the song until you could finally decide what the rating should be like and then iTunes removes the song and your library will never reflect the fact that you’ve actually listened to nearly the whole song! Your whole quarantine system might be in jeopardy! (How sad is it that this is actually not much of an exaggeration on how I feel about this issue? - yeah, I thought so…)

Which begs the question, what shall we do about this? Ah, I see your proposal - clearly, “Live updating” is the culprit, so let’s get rid of that. That is a possible solution, yes. Now be honest - are you satisfied with that solution? Sure, it lets you finish the song even after you’ve rated it. But to actually get new songs, you will have to manually trigger the update. Unacceptable.

Luckily, there is a solution. Have you ever looked at the top of your list of playlists? There’s one special playlist, called “iTunes DJ”. The core functionality of iTunes DJ is easy to explain: choose a playlist as source and play songs out of that playlist on shuffle. Sure, in the last few versions of iTunes it grew into some kind of monster on its own, with guest remote controlling and even voting capabilities, but let’s go back to the basics for the moment: it takes a playlist as source. I wonder what happens if we give iTunes DJ our “random not rated” playlist as source, play a song in there and rate it. Go ahead, give it a try. See what happens? It rates the song, the song gets removed from the smart playlist, but iTunes DJ continues to play it.

And this, ladies and gentlemen, is the magic of iTunes.

99% of all the music on the my computer is being played through iTunes DJ. While I play songs, I can work on playlists, I can change sorting or the filter on the main library, I can rate and edit songs as much as I like - iTunes DJ just continues to play. And I can always return to the list of songs being played at this moment. If I want to see what I listened to or skipped in the past half hour, I look at iTunes DJ, it will tell me. iTunes DJ is a simple, elegant, no-frills music player. Everything around it is a somewhat bloated, but very capable music management system.

Now, I know that you can’t listen to everything in shuffle. Unfortunately, iTunes DJ doesn’t have a non-shuffle mode. Or does it? Imagine you want to listen to a full album in the correct order. Look for the album in iTunes (either in a playlist or in the main library). Mark all songs of the album. Right-click:

Look at the first two lines of the context menu. “Play Next in iTunes DJ” and “Add to iTunes DJ”. Both will put all marked songs in your iTunes DJ queue, exactly in the order they are at this moment. The difference? “Play Next” does what it says, it will put the songs in between the current and the next song. If no song is played right now, it will start with the first marked one. “Add to”, on the other hand, simply appends the marked songs at the end of your iTunes DJ queue. So how do we make sure that iTunes DJ doesn’t return to shuffle mode after the album ends? Easy, its source has to be empty. And yes, that works as you expect it to, just give it an empty playlist as source. Small drawback - iTunes doesn’t believe that you really want to do that:

Yes, as you can see I have a special playlist exactly for this purpose. And no, it is not a smart playlist. Although this might be a nice exercise: Define the most complicated criteria that will deliver an empty smart playlist in every iTunes library.

While we are at the topic of playing a full album, there’s a neat little trick I’d like to mention here. Have you ever wondered what the small arrows next to Song Title, Artist and Album are for? Click one. If you have an iTunes Store account, it will open the store and show you the song/album or the artist in the store. Very weird behaviour, since you already have that song. Why would you want to buy it again?

Now go back to your library. Click the arrow again, but this time, hold down the alt key while doing so. Instead of sending you to the store, it will show you the main library, but with a filter: Only all tracks of this album or all songs by this artist. See what I’m getting at? A few lines above, I suggested you can simply add a full album to iTunes DJ by marking it and use “Play Next in iTunes DJ”. With this functionality, it takes you not more than two steps from every song in any playlist to having it played in the context of its album in iTunes DJ. Alt-click the arrow, press Command-A (mark all), right click and choose “Play Next”. OK, these are more than two steps, but they feel like two.

To make life even more easier - you don’t actually have to alt-click to remain in your library. Just open a Terminal and write:

defaults write com.apple.iTunes invertStoreLinks -bool YES

Restart iTunes and from now on a simple click will activate the filter in the main Library, while alt-click will bring you to the iTunes Store - whatever that is for…

And with this glorious example of Apple’s strategy to keep stuff simple on the surface but very powerful beyond, this ~~advertisement~~ ~~iTunes Help article~~ episode of iTunes for the professional OCD person comes to an end. Be sure to tune in next time, when I will open up my heart a little bit further and give you an idea of the the ways I try to make rating less tedious and more… no, who am I kidding? It’s still tedious.

Using iTunes, Part 2: Quarantine periods

Tue, 07 Sep 2010 19:45:00 +0200

In the first part of this series, we’ve created smart playlists for songs rated with 5 stars (and 4.5 stars, and 4 stars, and…). As long as you will meticulously maintain each song’s rating (we’ll come to that later…), you can now simply chose the playlist 5 stars and know for sure that you won’t get Miley Cyrus or Tokio Hotel (and if your mileage varies on that one I want you to leave this blog now. And never come back again).

Anyway, having good songs at your disposal at any given moment is nice, but how about something more sophisticated? Not like, I don’t know, music radio stations? You know, “The best songs from the 60’s, 70’s, 80’s, 90’s and today”? And in the end, it’s just the same Top 10 playlist on shuffle… way to ruin “Männer sind Schweine”, thank you Antenne Bayern.

So, what we actually want is a way to make sure songs are played with a reasonable repeat rate - ideally based on their rating. Guess what: iTunes can do that.

Let’s have a look at my main playlists:

Obviously, the higher a song is rated, the more often I’d like to listen to it. The description should make clear how I do that - “5 stars 1y”: 5 stars songs not listened to in one year. 4.5 stars: 2 years. And so on. Whether you will copy my (let’s face it) ridiculously long quarantine periods, as I like to call them - one year for the highest rated songs, 5 (sic!) years for 3-star songs - or go for something more reasonable, like 14 days for 5 stars and 2 months for 3 stars is entirely up to you - it highly depends on your daily music throughput and the number of songs you have. For me, these figures work - and by work I mean that all songs are reasonably often repeated but the playlists are never empty or close to empty.

Once again, note that I use folder to combine playlists with OR. Both, the “4 & 5 days” as well as the “3, 4, 5 days” folder are actually playlists I use quite often.

In detail, all of these playlists are similar:

Despite the obvious Media Kind, rating and period for “Last Played”, I decided to put two more time-based criteria in there - “Last Skipped” and “Date Modified”. Sometimes, I’m definitely not in the mood for a particular song and therefore skip it. Usually, this alters my opinion of this song for some time, therefore I will give this song a timeout without actually changing its rating. I decided on half the period of the actual repetition rate per rating, but once again, feel free to change that.

Now, speaking of skip. Apple introduced the skip counter in iTunes 7 and it has some interesting logic to it: If you skip a song after it is played for at least 2 seconds but less than 20 seconds, it increments the skip counter. Otherwise, it doesn’t do anything. Huh…

Actually, the 2 second rule, I can get behind. If I’m familiar with an album or a playlist, I sometimes quickly skip a few songs until I come to the one I want to listen to. Having these skips count in the library would result in “unfair” treatment to these songs. 20 seconds, on the other hand - I wonder how they came up with that… I could imagine their thinking went along the way of: Skipping a song within 20 seconds means you don’t like the song. Skipping a song after that period is more a result of: “oh, I found something else, let’s switch.” - Well, could be, anyway.

So, whatever their reasoning, these are the rules behind skipping, therefore this criteria is of limited value to you if you are a slow thinker and need at least 30 seconds to realise that this is not the song you’d like to listen to right now. Better hurry up!

This leaves us with “Date Modified”. And me with a dilemma. It’s not that I don’t want to explain what I’m doing here, but actually I don’t want to do it right now. It is quite peculiar and requires a lot of additional information that I’d like to keep for a later post. For the moment, either ignore it or think of this criteria as something similar to the skip counter. Yes, I know, that’s unsatisfying… bear with me, will you?

Before I use the playlist criteria as a nice way to bridge to the next playlists, give me a few sentences to explain the limit. Of course, you could use this playlist without any limitations (and actually, I do have the same playlist without limitations somewhere in my list, to see how many tracks would actually meet all criteria), but there are two reasons not to do that:

First, if you own any kind of portable music player and have more than a few songs in your library, you will be familiar with insufficient space. Limiting the songs you sync to your device is a must, and of course, for iTunes aficionados like us an automatic solution is the only one acceptable. Furthermore, to be safe from unpleasant surprises due to high bit rates or long songs, neither a limit on the number of items nor on the length of the playlist is acceptable. It’s always a good idea to calculate with fixed sizes per playlist to have an absolute maximum of space allocated to this playlist on your device(s).

Second, and in my opinion even more important, is the idea of weight. I mentioned above that the enclosing folder are my main source of music. As you know, these just combine all their playlists in them. Now, having songs from 3 to 5 stars in a playlist, further filtered by the last time I listened to them is one thing, but I’d like to make sure that 5 star songs are still more common in my mix than 4 star songs, which should be more common than 3 star songs. How to do that? Easy, just make sure you have more 5 than 4 or 3 star songs in your playlist. Therefore I gradually reduce the size per playlist:

5 stars: 300 MB
4.5 stars: 250 MB
4 stars: 200 MB
3.5 stars: 150 MB
3 stars: 100 MB

As long as your bit rate is similar over most of your songs and you don’t rip your most loved songs with thrice the bit rate of everything else, this should result in three times the amount of 5 star songs as 3 star songs. Sounds like a nice weight criteria.

One more thing on limits - I mentioned above that the long quarantine period works for me - together with the size of your playlist there is one more thing to think about: Number of songs meeting the criteria without limit at any given point. If you are in any way as particular as I am, you want to make sure that you not only don’t run out of songs in the playlist, but even that the amount of “spare” songs not in the playlist due to the limit doesn’t run below a certain threshold at any given time. With some bad luck, this would mean that all songs in the playlist have been played around the same time in the last period. And that would make things more boring. Which, of course, is unacceptable. And yes, I agree, that’s quite anal. I’ve been called worse.

Now, finally, let’s have a look at some more playlists, shall we? The last criteria in these playlists is: Playlist is not 5s 6m 8x”. I think you can imagine what that means…

As you can see, within “3, 4, 5 days”, there is another folder called “4, 5 new”. And how do these playlists look like? Take “5s 6m 8x” as an example:

Very similar to the playlists above, as you can see, with one addition: “Plays is less than 8”. The result of that is that songs in here are usually either neglected or new. Both good reasons to give them better attention. Therefore these playlists shorten the quarantine period per song significantly - as long as the song is played less than 8 times. The actual values can easily be seen in each playlists description:

5 stars, 6 months, 8 times played
4.5 stars, 9 months, 6 times played
4 stars, 1 year, 5 times played

I know, that’s not linear and doesn’t fully line up with the normal playlists. It bothers me, as well, but I’ve tried it with 4.5 stars, 1y, 6x and 4 stars, 1.5y, 4x - there are just not enough songs left most of the time… consider it work in progress for the time being.

Once again, there is a nice, linear limit in place:

5 stars: 300 MB
4.5 stars: 200 MB
4 stars: 100 MB

Note that although these playlists are all about “new” songs, I didn’t get tempted to use a limit based on “Most recently added” or similar. Although that would certainly help getting the new songs into the playlist, it would also put too much emphasis on them, so that old songs would never make it into this playlist until I’ve listened to all newer songs within the quarantine period and would not have bought new ones since then. Unfortunately for my bank statement, that usually happens quite a lot… Furthermore, and for me even more severe, it will result in quite a static playlist, since the same songs will be added in the same order every time their quarantine period runs out. Once again, unacceptable.

I’m not going to mention this time that there’s another folder which can be used as a playlist, too… oh well, guess I just did. Well, in case you haven’t noticed yet, there’s a folder. Hurray!

And with this more detailed look into my quite OCD ruled ~~brain~~ iTunes, it’s time to call it a day. Once again, I hope you’ve learned something or at least considered this worth reading. Without repeating myself, let me reassure you that more is going to come, hopefully soon. We still have to find out how we’re actually going to rate songs…

Using iTunes, Part 1: Introduction to stars

Mon, 06 Sep 2010 08:10:00 +0200

I consider myself an avid iTunes user. In fact, I would call myself an iTunes fan, although I know, accept and often join the typical rants about iTunes. Yes, it is bloated beyond belief and so far, Apple doesn’t seem to care about treating us with something more modern - 64 bit? Maybe at least Cocoa for a change? No way Jose.

Still, below its overwhelming interface with such gems as iOS App Management, Apple TV storage hub or eBook support and besides the fact that it is more or less now the gateway to the iTunes Store, iTunes is quite a good music management tool and still a music player - a good one at that.

At the time of writing, my iTunes Library consists of 17307 items of music (not counting movies/TV shows, apps, Podcasts, Audiobooks etc.). Luckily, iTunes takes care of the gruesome task of organising these files, leaving me with a database-like interface to my music based on metadata in the files (and the library itself, but let’s forget that for the moment).

Now, I decided to write about my usage of iTunes because I’ve seen so many people being fascinated by the most basic smart playlists that I decided it’s time to chime in and show my… advanced? anal? way of “working” with iTunes.

I’d like to see myself as seriously interested in music, always hunting for new bands and at times very keen on trying to find the right song for the moment, but on the other hand, most of the time I’m listening to music I don’t want to think about what to listen to, I just want to make sure I will like it and it’s not something I’ve heard five minutes ago.

One more thing you should know before we start is that I’m not an album listener. Yes, every once in a while I decide to listen to an album from the first to the last track, and neither iTunes nor my system will hinder me in doing so, but nearly everything I do in iTunes is counter-intuitive to album listeners. For me, music is about individual songs.

Enough with the warnings already - time to dive in:

I’m sure by now you have realised that I heavily rely on stars - iTunes’ way of rating songs from 1 (or even 0) to 5 stars helps you quickly finding a bunch of songs you will usually like. For most people, these songs will be marked with 5 stars, therefore it’s always a good idea to have a smart playlist will all your 5 stars songs. Since I am not only interested in my best rated songs, but also in the distribution of stars among my library, I have a smart playlist for each star category:

Straightforward, yes, but a few things might still be worthwhile discussing. As you can see, I like to use Playlist Folder - for two reasons: First of all, they help to tidy up the list of playlists - since I have about 250 at any given time this is quite useful. More important though, each folder is a playlist, as well. It contains all songs in all playlists below it, in other words it’s a simple OR of all its containing playlists. Very helpful indeed. Since my threshold for songs like ‘ok, well…’ and ‘not bad, actually’ is between 2 and 3 stars, I decided to group my star playlists as 1/2 and 3/4/5. Of course, your mileage may vary…

Furthermore, I consider these playlists one of my most important ones, therefore I’d like to have them at the beginning of my list of playlists. iTunes’ way of sorting playlists is a little bit particular (and interestingly enough different to the way the Finder sorts files): It sorts Folder before smart playlists before normal playlists. This repeats itself within Folder - Subfolder up front, then smart playlists, then normal playlists. Within each type, the sorting is alphabetically, with numbers before letters. Symbols are mostly in front of numbers, some are after letters (e.g. Ω). Confusing, yes, but with the help of Folder you can basically have your playlists in any order you wish. Want some very cool mix playlists (normal, not smart) directly on top of the list? Make a Folder “! Cool Mix Playlists” and put all your playlists in there. They will be on top of the list, even before my “[1] Ratings” playlist. Helpful.

Last, but definitely not least, I’m sure you’ve discovered that I have actually seven playlists - for 5 stars. So, what’s up with the ‘all 3.5 stars’ and ‘all 4.5 stars’ playlists? Well, it turns out internally iTunes doesn’t count to 5 when assigning stars to tracks, it counts to 100. Logically, 0 is 0 stars and 100 is 5 stars - which leaves 20 = 1, 40 = 2, 60 = 3, 80 = 4. But what is 70? Yes, you guessed it, 70 is 3.5 stars. Surprisingly, although iTunes doesn’t normally support half-stars, it can display them:

Now, whether you will need such fine granularity in rating your songs depends entirely on you, of course, but if you need it - now you know iTunes is there to support you.

So, how are these smart playlists actually defined, you ask - well, pretty easy. This is what all full-star playlists look like:

A simple AND operation (Match all) on the actual Rating and the Media Kind ‘Music’. Of course, if you want, you can leave the Media Kind Filter out of it, but I’d like to make sure that no music videos, Podcasts or Audiobooks are in my 5 star music playlist. Once again, your mileage may vary.

A limit is not necessary, obviously, since we wanted to have a playlist will all songs matching this rating. Myself, I like to check the ‘Match only checked items’ box, since I use the check-system in iTunes to mark tracks that don’t exist on the harddrive but should remain in the library - these, obviously, should not be used in playlists. Sounds weird, but it happens…

Live updating is a must for me, but I do pay the price. Every time I change the rating of a song, iTunes updates all smart playlists in the background and checks whether the song now belongs in this playlist. On my 2.4Ghz MacBook, this doesn’t really help iTunes’ sluggish performance, believe me. Nonetheless, the alternative, having smart playlists that aren’t so smart and display old, invalid information, is simply not acceptable. And in case you wonder what happens if you do uncheck the box: The playlist gets created as soon as you press OK and doesn’t change afterwards, even if less or more songs meet the criteria - until you go back in the playlist definition and press OK again, that is.

For the half-star playlists, there are actually two ways of creating them. If you have enabled this trick, you can simply chose a half-star rating in the smart playlist definition - [UPDATE: This does not work as expected. The resulting playlist is correct, but if you ever go back into the smart playlist definition it changes the stars back to the lower full value, i.e. the half dot gets lost. Very inconvenient. I therefore don’t recommend this way of creating half-star lists. Thank you Chris for pointing it out.]:

Alternatively you can work with full-star steps and still get any half-star playlist you like:

And on this bombshell, it’s time to end ~~the show~~ this post. If you have learned something, great, I hope you will use it for good and come back for more. If none of this was new to you please come back anyway, there will be a lot more exciting things in upcoming parts - at least I hope so…

Thank you for sticking out to the end. And if you just skipped here: Shame On You!

Internetsperren: Mail an Unicef

Tue, 12 May 2009 10:25:53 +0200

Folgende Mail habe ich am Wochenende an Unicef geschrieben, als Antwort auf ihre schon etwas ältere Pressemitteilung zum Thema Internetsperren:

Sehr geehrte Damen und Herren,

auch wenn ich bisher kein besonders großer Spender war, sah ich Ihre
Arbeit und Ihre Ziele als unterstützenswert an und habe dies nach
meinen Möglichkeiten auch unregelmäßig gezeigt.

Ihre bedingungslose Unterstützung der vom Familienministerium
getriebenen Internetsperren lässt mich allerdings fragen, ob Sie auch
in anderen Bereichen so wenig Sachverstand besitzen, dass Sie
kontraproduktive, Wahlkampf/PR-getriebene Aktionen nicht hinterfragen
und Spendengelder somit nicht bestimmungsgemäß einsetzen.

Ich möchte Sie daher bitten, meine Kontaktdaten aus Ihren Datenbanken
zu löschen - von mir werden Sie sicher kein Geld mehr bekommen.

Vielen Dank.

Mit besten Grüssen,

Christian Heßmann

Gerade erreichte mich die Antwort:

Sehr geehrter Herr Heßmann,

vielen Dank für Ihre Mail. UNICEF tritt bei dem Problem der Kinderpornografie für eine differenzierte Herangehensweise ein. Wir sind der Meinung, dass der Schutz von Kindern vor schweren Verbrechen auf der einen Seite und das hohe Gut der Informationsfreiheit auf der anderen Seite nicht gegeneinander ausgespielt werden dürfen. Wir haben in unserer Stellungnahme vom vergangenen Freitag ausdrücklich darauf hingewiesen, dass mit der Einführung von Zugangssperren die Verbrechen an Kindern nicht aufhören und diese allenfalls ein Schritt zur Störung des Massengeschäfts sein können.

UNICEF tritt seit Beginn der Debatte für eine klare gesetzliche Regelung ein, um ein rechtsstaatlich kontrolliertes Verfahren sicher zu stellen. Zugangssperren sind sicher nicht die Lösung des Problems. Deshalb tritt UNICEF für ein ganzes Bündel von Maßnahmen ein, um systematisch den Kinderschutz zu verbessern, die Strafverfolgung zu verstärken und den Opfern zu helfen.

UNICEF setzt sich weltweit für internationale rechtliche Standards im Kinderschutz ein, wie zum Beispiel auf dem von UNICEF mit initiierten Weltkongress gegen die sexuelle Ausbeutung von Kindern und Jugendlichen in Rio. Wir haben die Bundesregierung aufgefordert, die internationale Strafverfolgung zu verbessern und dazu beizutragen, dass Täter, die kinderpornografische Bilder auf ausländischen Servern einstellen, endlich verfolgt und entsprechende Seiten abgeschaltet werden.

Um Kinderpornografie wirksam zu bekämpfen muss auch in Forschung investiert werden, um die Nutzungshäufigkeit und die Nutzungsstrategien der Konsumenten sowie die Organisation solcher Netzwerke besser zu verstehen. In seiner Programmarbeit unterstützt UNICEF in vielen Ländern Projekte zur Prävention von sexueller Ausbeutung und Hilfen für die Opfer dieser Verbrechen.

Ich hoffe, dass wir Ihnen hiermit unsere differenzierte Position ein wenig deutlicher machen konnten.Über Ihre Unterstützung für diese wichtige Arbeit gegen die Ausbeutung von Kindern würden wir uns sehr freuen.

Mit freundlichen Grüßen

Die Bahn ist dem Genitiv sein Tod on TwitPic (via @zufall - @toco)

Thu, 26 Feb 2009 11:41:03 +0100

Die Bahn ist dem Genitiv sein Tod on TwitPic (via @zufall - @toco)

tumblr might become my favourite Twitter client, just because of...

Thu, 26 Feb 2009 08:44:23 +0100

tumblr might become my favourite Twitter client, just because of its thread feature. :-)