Monday, December 10, 2007

Assessing the Development of a Station: ASUD

Nowadays, the inputs that are used by a player for a particular station are almost entirely public. As long as, the person who created the station has not set their profile to private, you can read the lists of artists, song seeds, thumbed up tracks and thumbed down tracks.

Back in the early days, a few of us gathered at (now defunct and cybersquated). We came up with the idea of measuring the development of a station as a list of the number of Artist seeds, Song seeds, Ups and Downs (ASUD). Tim, IIRC, when he created the first version of proposed a measure D which was equal to the sum of the numbers of Ups and Downs, but the rest of us had already grown used to ASUD. And I think it's still a useful place to start.

Artists It's good to keep the number of artists in the station separate from the other numbers because the number of tracks called upon by an artist seed is the one thing about a station which can change without the user's input. Pandora analyzes and makes available for play new tracks by an artist on an on-going basis. The number of tracks available for popular artist who is still recording will always tend to increase over time. The number is more static for more obscure artists. (Interestingly, the founder of Pandora, Tim Westergren, was in the band YellowWood Junction which has exactly three songs available for play and that number has not changed in the past year at least.)

Songs Each song which is added to the list of song seeds for a station becomes available to form the basis for a set of four songs to play on the player. As similar songs are added to the list the station generally becomes more and more focused. Thus, because the number of tracks represented by an artist seed is fairly random, dynamic and difficult to assess, the number of songs seeds is probably the second most important measure of a station's development (after the number of thumbs down).

Ups The entire list of thumbed up songs serves as a single possible basis for a set. I used to believe that all the thumbed up tracks could individually form the basis for a set, but the Pandora FAQ now definitively states that that is not the case. Thus, the number of Ups gives a good idea of how much the creator has listened to a station, but it's the least important measure of the development of a station.

Downs On the other hand, the number of thumbed down tracks is an absolutely crucial indication of how well developed a station is. No thumbed down track will ever play on the station again, and, as long as no track by an artist appears in any of the other three lists (artist seeds, song seeds, or thumbed up tracks), two Downs ban all the tracks by that artist from the station. Thus, unless you are consciously moving tracks into the list of song seeds, cutting out songs one by one via the thumbs-down is the most usual way to incrementally improve a station.

The best measure of a station's development would be a count of all the songs which can be the basis for a set and a count of all the tracks that cannot play on that station. Unfortunately, those two numbers are difficult to calculate and until they become readily available ASUD is the best surrogate we have.

Friday, November 30, 2007

Dear Santa Pandora

After nearly two years of listening to Pandora I think I'm finally ready to state with some clarity my wish list for improvements to Pandora. Here're are my Christmas wishes for the coming year:

Filters: I'd like to be able to filter out the songs which play on a station on at least a subset of the genome characteristics. Yes, I know there's difference between the focus traits and those which comprise the underlying genomes. And I know that the genomes do not necessarily have many or even any traits in common. But creating stations which only play songs with humorous lyrics or only female vocalists should not be impossible as it has been to date.

Moving Ups To Seeds: It’s now fairly clear that once the number of thumbs-up given to a station reaches a certain point (maybe 20 or 50) that additional thumbs-up have virtually no impact to a station. And so to continue to shape a developed station you must operate on the list of seeds. I’d like to have an additional button under the thumb-up list on the station page which says “MOVE CHECKED TO SONG SEEDS”.

Banned Artists: I’d like a button on the station page that enables the user to calculate and display the artists which are currently banned on a station.

Identify Unencoded Song Seeds: This problem is becoming smaller over time since fewer unencoded tracks are being added to the db. Still, it is unconscionable that there is no way to tell when a song seed is, in fact, unencoded and therefore its presence in the list of song seeds is bringing in all the material of an artist. We should be able to push a button under the list of song seeds on the station page to show any such cases in the list of seeds.

Mobile Quickmix Play: Please display the originating station for the track being played.

Fix Searching for Classical Tracks: I’m so glad Classical is now available, and I can accept the various compromises like only one recording per opus, but Classical is truly crippled right now by the fact you can’t add individual tracks to stations.

Net Song Seeds: I’d like to have a button on the station page that would calculate the net number of seed tracks and net number of banned tracks for the station. The called program would generate the list of all encoded tracks available as seeds (both from the artist and song seed lists) remove all the tracks from the thumbed down tracks and banned artists for the station and return the count as the net seeds. Similarly, the program would return a count of the tracks from thumbed down tracks and banned artists. These two numbers would be the true measure of the current development of a station.

Oh, and a pony, I suppose. I love Pandora, and it would continue to be my primary source of music even with no improvements, but it could be made even more fun for me with these improvements.

Wednesday, October 24, 2007

Station Story: Prog Rog

If you created a QuickMix with my Prog Rog and Pioneers of Electronica stations, you'd basically have everything I was listening to in high school. I lived in Hermosa Beach in Southern California at the time, and a few blocks away was a Music Plus store which was part of a regional chain of record stores. There was a small bin for electronic music which single handedly weaned me from the easy listening radio stations that my parents listened to in the car.

The first record I ever bought was Walter (later Wendy) Carlos' Switched On Bach when I was in third grade. The synthesizer captured my imagination even at that young age, and so when we moved to So. Cal. and I started high school it was a natural progression for me to start listening to the electronic music of the era and rock groups which featured synthesizers which, at the time, meant Progressive Rock.

The holy triumvirate at the time were ELP, Yes and Genesis. But, thanks to some staffer at Music Plus who'd handwrite little reviews on address stickers and put them on selected albums in the import bins, I also discovered Novalis and Goblin during that era. I was so pleased when these latter two groups were covered by Pandora this year. Oddly, though I did purchase a copy of "Leftoverture" at the time, I did not consider Kansas to be Prog Rock at all then though I do now. I also was skeptical about Queen because their albums of the time boasted that "no synthesizers were used on this record". But I eventually accepted them into the fold as well.

Pandora has helped me flesh out the station a bit. The most out-there band in the set has to Ruins which came long after the others on the station, but provide some truly interesting and difficult (in many senses) tracks. On the other hand, Starcastle, PFM, Van Der Graaf Generator and Egg are all much more central to the genre.

The station occasionally strays into more mainstream rock one side and Fusion on another, and, in general, the station is below average on the monthly listening tests. Nevertheless, some of my favorite music from my high school days is played by this station.

Station Review: The Darjeeling Limited stops at two stations

Wes Anderson is a director who is known for the music in his films. Unfortunately, I'm not sure that his sense of taste transefered to these two promotional stations.

The title of his latest film serves as the name for the first station, The Darjeeling Limited (click to play). It would not surprise me if the four song seeds for this station are featured in the film. Nevertheless, the resulting station is pretty poor since there neither any other seeds nor feedback to flesh out the station. The result here is some blues rock vocals.

The second station, Wes Anderson Films* (click to play), is somewhat more interesting but less coherent. It consists only of eight song seeds, but at least the variety is interesting since it covers tracks by Devo, David Bowie, Cat Stevens and Paul Simon. Again, it's a bit Boomerish, but that's probably appropriate for the topic.

And with that review, I'm officially caught up with the influx of promotional stations after the Internet Radio crisis.

Station Review: SingStar presents two

SingStar seems to be a Karaoke Revolution competitor for the PS2 which is one of the better advertiser fits for Pandora that I've seen.

The first station this promotion offers is SingStar '80s* (click to play). They threw eight hits from the Eighties into a station and walked away without a single look back. Meh. I've heard worse, but, still, they put the 'F' in 'effort' with this one.

They took the exact same approach with SingStar Amped* (click to play). In this case the eight songs seem to be from the Nineties.

Station Review: The Honda Odyssey delivers four

My guess is that it's a car even though the first feature the copy on the Pandora profile page mentions is Bluetooth.

Breaking Free* (click to play) is one of the better designed promotional stations. There are four artist seeds and seven additional song seeds. The idea of connecting the High School Musical soundtrack to REM and They Might Be Giants is an interesting one. The two songs I've heard so far have been strange and good. The station has a little feedback with three thumbs up and three thumbs down.

On the other hand, Truckin'* (click to play) is a much more marginal effort. It's based on four tracks from the Eighties and earlier and minimal feedback (one thumb up and two thumbs down). Yet another boomer nostologia station. Will one of the two remaining stations be Hip Hop? Be still my heart.

Well, the next one, Ticket To Ride (click to play) is a bit moldier (if not Hip Hop), focussing on the Sixties and Seventies. The station is based on six artists and one additional song. You know, The Beatles are surprisingly rare as seeds for promotional station, but they're one of the seeds for this station. The feedback consists of two thumbs up. There are better stations covering this era.

No Hip Hop for the last station, surprisingly. Instead, I've Been Everywhere (click to play) looks at earnest balladiers with Johnny Cash and Dan Zanes and Friends as the artist seeds plus seven additonal song seeds. Like the first station this one is a bit more interesting than most promotional stations. It also has a tiny bit more development than most promotional staions with one thumb up and five thumbs down.

Tuesday, October 23, 2007

Station Review: BET Hip Hop Awards 07

Unlike the vast majority of Hip-Hip stations created for various promotions BET Hip Hip Awards 07 (click to play) is worthy document of the current state of Hip Hop. The station consists of 20 artist seeds and five songs seeds (though only two of the songs are not covered by the artist seeds). There is no development, but if you wanted to evoke that 2007 Hip Hop sound in the future, this would be an excellent set of artists upon which to base a station.

Station Review: The US Air Force tries four recruitment stations

"Patriotism rises in the heart of the American bear," as Fozzie Bear once said. The US Air Force recruiters paid for four stations in their promotion on Pandora.

First up is American Anthems* (click to play). This station is the epitome of bombast and cheese. Based on Lee Greenwood and four patriotic anthems, this station produces a lot of overwrought, whitebread gospel. Heh: "Hey Jude". These is no station development at all as usual.

Next is Air Rock Lounge* (click to play) which has nothing to with Lounge music. Instead, it's a Nickelback station with two addtional song seeds (Dauhgtry and Red Hot Chilli Peppers). Not bad, but there's not much to it since there's no development here either.

Third is a comparatively interesting station, Funk Force 1* (click to play). This station consists entirely of three artist seeds: Mighty Imperials, The Diplomats of Solid Sound, and The Poets of Rhythm. All three groups are comparatively recent funk revival groups and work quite well together. This collection could be the beginning of a good station.

Lastly, we have the obligatory Hip-Hop station, Horizon* (click to play). This one consists of three song seeds, and that's it. A Hip-Hop afficionado would have to tell you whether these tracks go together at all. As it is, it seems unlikely that this station is in any way superior to all the other promotional Hip-Hop stations.

Monday, October 22, 2007

Station Review:: Heavenly Sword carves up three stations

Heavenly Sword is a game for the Playstation 3. The promotion for the game offers up three stations.

First up is Speed Rock* (click to play) which is a generic metal station as you can get being based on Metallica, Motorhead, Slayer and Zeke. As is endemic for these stations, there has been no development whatsoever. If you like metal you cuold create a better station in fifteen minutes.

The second offering from this video game is yet another Indie Rock station, Range Indie* (click to play). The four seeds are Ghostland Observatory, Nada Surf, Okkervil River and The Thermals. In my brief listening to this station I'd say I like the station a bit better that the NikeiD version of the same genre. But again since there is no development, there's nothing here that you couldn't incorporate into your own stations in a minute or two.

Lastly, we have the inevitable urban market Power Hip-Hop (click to play). Excuse me as I roll my eyes, but there are only three artist seeds here (50 Cent, Kanye West and The Notorious B.I.G.) each of which have been featured on most other promotional Hip-Hop stations. Was there any point to creating this station? Of course, as expected, no time was spent developing this poor excuse for a station.

Station Review: NIKEiD tromps the SUPER-eGO with two stations

And what could be more id-like than a Hip-Hop station and a Rock station? I have no idea what product is being promoted here, but I assume it's something from Nike.

The Hip-Hop station is NIKEiD Hip-Hop Radio* (click to play). I can't be much of judge of this genre since I never listen to it. I've heard of Kanye West and Rihanna, and I've not heard of the three other seeds Buju Banton, Common and Talib Kweli. Since Pandora is about introducing you to new artists, the fact that I haven't heard of these artists is probably a good thing. In any case, no development time was spent on this station, and so it's still disappointing from that standpoint.

The seeds of NIKEiD Rock Radio* (click to play) are also comparatively obscure (Bloc Party, Interpol, Justice and The Killers plus a redundant song by The Killers), and that's nice as well. It seems to be a refreshing mix of recent Indie Rock. But once again there are no thumbs up or down to guide this station into being a bit more listenable.

Station Review: Four for the Journeyman TV Show

This set is a promotion for a US TV series featuring time traveling, and so the stations are built to reflect different eras.

The late Eighties are the focus of Journeyman Hits 1987-1992* (click to play). The approach was to use five artists of the era and five songs other songs of the era as seeds. It was not a great era to start off with, and the station as usual has very little development: two thumbs up and four thumbs down. (Beyonce gets dissed twice here, hilariously. As far as I can tell, this makes her the only artist to have ever been banned from a promotional station. I'm sure it was inadvertent.)

The next station attempts to encapsulate the next five years in Journeyman Hits 1993-1999* (click to play). Did anyone really need a Spice Girls station? In any case, the Girls join four other artists and two songs as the seeds for this station. The development consists of two thumbs up and two thumbs down. Bland pop at it's most vanilla.

Moving up to the present we get to Journeyman Hits 2000-2007* (click to play). The seeds are six artist and six seeds. Again, it's the tops of the pops like many other promotional stations. Yawn. It was granted a single thumb up for development. You'd be better off building a station off the Billboard lists.

A new low is finally achieved with the last station, Journeyman Hits Time Travel* (click to play). If you really wanted to hear a station based on Tupac, Bon Jovi, Green Day, Madonna, Pink, TLC and U2 you can listen to this station, or, you know, create your own maybe adding some artists you actually like instead. No development time was wasted on this station at all.

Thursday, October 18, 2007

Station Building 101: Moving Ups To Seeds

Nearly a year ago I heard Tim Westergren speak in Berkeley. He mentioned that the list of thumbed-up songs were averaged to form a single seed which could then generate sets. I'm not sure I believed him, and I did not include that fact in my write up of the event. However, a recent update to the Pandora FAQ confirms that this is what the player is doing.

Given that information the question becomes whether we improve the focus of a station by making the best thumbed up songs into song seeds for a station. The idea is that by having many more similar seeds generating song sets, the station should become more consistent.

I examined my list of stations and picked a candidate for improvement. I Sing The Body Electric has always had a bit of an identity crisis. I wanted a station that played world-influenced electronica, but the only two bands I really liked (and of which I was aware) that were doing so were Engma and Afrocelt Soundsystem. And so to have enough bands to get the station going I added Polyphonic Spree and Jon and Vangelis both of which were looking for a home in my stations but neither of which met the ideal I had in mind for the station. It was time for major surgery.

I sorted the list of thumbed-up songs by artist, pressed the "show all" button, and printed out the list. I listened to the sample of each song in the list except for those of the four artist seeds and decided for each song whether to make it a seed, leave it alone or remove it from the list. Starting with the 183 tracks I had thumbed-up to date I made 45 songs into seeds and removed 95 of the tracks since they were more associated with the artists I was eliminating.

As I listen to the resulting station it seems that this a useful procedure; however, the station will need to go through a period of listening in which many new songs being drawn in by the influx of seeds will need to be thumbed down. And, of course, any new songs which are particularly appropriate will need to made into seeds as well as being thumbed up on an on-going basis. It is unclear at this point whether this process of constantly bringing in new seeds will result in a stable, consistent station, but I'm cautiously optimistic at this point.

A final point to consider is whether, in the long run, a station should use any artist seeds at all. My feeling is that as long as it is likely that new tracks by the artist will be made available to play on Pandora, then having artist seeds in a station is good because it will draw those tracks in as they become available. If an artist is no longer producing albums and Pandora has pretty much covered all the artist's albums then you might wish to include only the appropriate tracks by that artist in your stations.

Friday, October 12, 2007

Station Review: Four More For Visa Signature

Another advertiser to re-up for the Pandora promotional station advertising program is Visa Signature. Last December they introduced four of the better promotional stations, and now they're back with four more.

First up is Brit Rock Radio* (click to play). This station is based on Led Zeppelin, The Beatles, The Cure, The Fratellis, The Rolling Stones, The Who and a tune by Oasis. It's sort of a Boomer nostolgia station. Zzzzzzz... At least the station has some development with two thumbs up and five thumbs down. Kudos, at least, to Visa for being unafraid of offending The Ducky Boys.

Next, let's take a look at World Traveler Radio* (click to play). Could it be? A promotional station I actually like? Something must be wrong. In any case, this station is based on five bands, only one of which has any fame at all in the US: Ladysmith Black Mambazo. There are two song seeds as well. Lot's of drums and call-and-response vocals. The station is slightly developed with 2 thumbs up and 2 thumbs down.

The third station is Blarney Stone Radio (click to play). It's not a bad attempt to create a Celtic station. I would note, however, that my Tir Na Nog station (click to play) covers all the same artists plus many more, and is far more developed. Their station is based on three artists, two songs, one thumb up and one thumb down. Since it has no real development it pulls in a lot of folk and bluegrass which is not remotely Celtic.

Lastly, we take a look at Tokyo Pop Karaoke Radio* (click to play). Again, Visa Signature should be applauded for the mere idea of using a J-Pop station in their campaign. It has five artists, and one thumb down, and so no one stayed up hights perfecting this station. Nevertheless, Puffy is a good place to start any station.

Overall, I'd say that this set of stations is one of the best promotions to date. Much is still lacking in the development of these staions, but at least they are examining genres beyond "what the kids are listening to now".

Station Building 101: Spring Cleaning

In a previous Station Building post I mentioned that I probably should try to insure that no track is a seed or thumbed up on more than one station. I have several stations that are relatively close to one another, and so the idea is reduce the cases of tracks bringing the same songs to multiple stations.

And so I opened each station page and copied the lists of song seeds and thumbed up tracks into one long column of a spreadsheet. I added a column indicating which station the song came from and whether it was an up or a seed. At that point (Aug. 28) I had a list of 3,481 tracks which I then sorted alphabetically.

I then worked my way down the list looking for duplicates. When I found a set, I made decision on which I'd like to keep and copied the other rows in the set to another page. In the end, I found 75 duplicate tracks to eliminate. Sorting this latter list based on the station allowed me to quickly remove the duplicates.

The only downside to doing this task is the fact that afterward you will almost inevitably give a thumb up to tracks that you've cleaned out (particularly if you are listening to mobile Pandora where you cannot see the station). Nevertheless, this technique does help firm up what belongs in each of your stations.

The resulting list of all your seeds and thumbed up tracks is interesting in its own right. Currently, in that strange intersection of what I like and what Pandora has, Loreena McKinnet leads all others at 29 tracks edging out Kraftwerk at 28. Sadly, my two favorite artists, Vangelis and the Swedish/Finnish band Hedningarna (released in the US as The Heathens), are horribly underrepresented at 3 and 0 tracks respectively. (The Jon Anderson and Vangelis collaborations do score another 7 tracks for Vangelis, however.)

Sunday, October 7, 2007

Station Story: Pretty Pop Princesses

This station is probably the closest I'll ever get to having a true, current pop station. I do not listen to currently charting artists. When I was in my teens I definitely turned my nose up at anything popular. I was one of those music snobs who loved a band if I got to know them before they became popular, and scoffed at any band that only became popular because of the marketing machinery of a major label.

Jagged Little Pill is the only album I've ever bought when it was at number 1, as far as I know. I came across Shakira by catching a video of hers on a Spanish language channel years before her first album in English. I mush say that I was turned onto Lavigne by happening on the "Complicated" video on MTV, and, thus, caught her in a rather mundane fashion. I like the fact all three artists participate fully in the writing of their songs (the lyrics, if nothing else).

The additional song seeds for this station come from the period in Pandora history when Quickmix was first implemented. There was a time when you could move tracks while listening to an individual station, but not when you were listening to Quickmix. Thus, I heard these songs on other stations and moved them manually by adding them as seeds to the station.

I try to keep the male lead vocals out of this station. Even by thumbing down every male lead vocal, about ten percent of the tracks do have a male vocalist. Still, I'm pretty satisfied by this station.

Monday, September 24, 2007

Station Review: Comcast Likes Speed

Well, Comcast's promotional station which is shockingly named Comcast Radio* (click to play) has not been developed at all as usual. Nevertheless, I must give it a cautious vote of approval for matching Guns & Roses and AC/DC with Daft Punk and other techno bands. The station is based on four bands and five songs, but it's an interesting mixture, and better than the usual bland and homogenious selection of music.

Station Review: MTV Commemorates the 2007 VMA's with a Station that Contains No Britney Whatsoever

The MTV VMAs Radio* station (click to play) does contain 16 other artists who are not worth listing here. Just imagine a list of every pop artist that's appeared in every other promotional pop station in the last twelve months. No song seeds or thumbs of any kind. A slow jam rap is currently boring me to tears. Lot's of down tempo songs about teh sexy. Bleargh.

Station Review: Southern Comfort Fails To Be Hip By Calling Itself "soco"

But they did pay for two Pandora stations, and they also posted up a cringe-worthy version of every alcohol ad you've ever seen on their profile page. (I swear I thought it was an anti-drinking ad parody until I saw the closing panel and found out what "soco" was.)

The first up is a mostly harmless jam-band station: Jammin' Radio* (click to play). The Dead, Phish, DMB and Widespead Panic form the basis with two additonal song seeds. Ho Hum. Two thumbs up and two thumbs down. There are probably better jam stations out there.

The second station attempts to take you back to the heady days of 2006 with Dance With A Drink In Your Hand Radio* (click to play). Timberlake, Akon, Lily Allen, and Rihanna (and a Pink track) do the obligatory pop thang. One thumb up and two down. This station might be useful for some future ethnomusicologists to write an equally boring paper about early 21st century Pop.

Station Review: Bud Select Introduces Two New Stations in August

As far as I know, Bud Select is the first company that has paid for a second set of stations. They had an earlier campaign in February of this year which included five stations which, I beleive, I reviewed at This campaign features two stations.

The first is Step Up To Select Radio* (click to play). This is yet another alt-rock station based on the Yeah Yeah Yeah's, four other bands and four other songs. It exibits minimal development with one thumb up and no thumbs down. The couple of songs I've listened to have been synth-heavy whinefests without much interest.

The second is 99 Radio* (click to play), and it is equally insipid. It an techno-y R&B station based on M.I.A., Kanye West, Foxy Brown and six songs. No development whatsoever. I'd recomend Z Huge Rap instead which I found by doing a quick station search on the artists.

September Listening Test Results

It looks like I was a bit hasty in announcing an improvement to the Pandora selection algorithm. My mean satisfaction rating was identical to those of July at 7.41, and, in fact, the median dropped from 7.5 to 7.0. Grrrl Power did do exceptionally well this month, but no other particularly followed suit, and, indeed, Grrl Power topped the charts.

I think advertisers were holding off from investing more in Pandora during the Internet Radio Crisis because I was not seeing many promotional stations this summer. Any reluctance on the advertiser's part seems to be over now because while I was doing this month's listening tests I logged eleven new campaigns, most with multiple stations. Thus, you can anticipate many short Station Reviews over the next week or so as I work through the backlog.

Thursday, September 13, 2007

Station Building 101: Listening Test Mark III

In the previous post, I described the second version of the listening test which I had been doing roughly monthly for the past six months. In the original listening test I only listened to ten songs from each station, and then we learned that Pandora was generating songs in sets of 3 or 4 songs. Thus, I moved to listening to ten sets of songs (30 to 40 sings total).

As I performed the September listening tests, I became slowly convinced that the player was occasionally generating two-song sets. Finally, while testing Pagan Pride (which has an eclectic mix of genres) I came across a case where two folk songs were jammed between a hard rock set and an electronica set.

I e-mailed Tom Conrad, and he confirms that the selection algorithm is now occasionally generating 2-song sets. I've been seeing less value in trying to identify the sets anyway, and so I've decided to switch to listening to 40 songs from each station, and taking ten time the average song score (1 for up, 0 for down, and 0.5 for neither) as the score for the station.

I'm less than halfway through this month's tests, but it appears that the selection algorithm has once again been improved. It would not surprise me if the average station score is up an additional 0.5 since July's increase of a full point. My Grrrlpower station scored an astonishing score of 9.5 with no thumbs-down and four non-thumbed up tracks in 34 songs.

I began to wonder what we might be losing in these changes since my satisfaction has been increasing so consistently. It seems to me that the stations might not be exploring as much new material, and so I've started roughly tracking the number of new tracks that are being played. I can't do it exactly since, clearly, I can not remember every track that has ever played on each station. Nevertheless, a track that already has a thumb-up has clearly been played before, and I'm pretty certainly that if I'm motivated to give a track a thumb up or a thumb down, then it's most likely new since I would have have had the same motivation the first time I heard it. And so I've begun to track the percent of new tracks being played and the percentage of new tracks that get a thumbs-up. So far over five stations the % New has run from 16% to 39% and the % Good|New (% of Good given that a track is New) is all over the map from 23% on Pagan Pride to 83% on O, Wow the Moon.

Friday, September 7, 2007

Station Building 101: Listening Test

The best way to develop a station is to spend time listening to it with intention and focus. I've developed a methodology to do so that allows me to track the development of the station over time and spend some uninterrupted time with each station that I'm actively curating. I came up with this procedure about a year after I started listening, and it's obviously more work than many might be interested in doing. However, I've found it to be quite fun and helpful at understanding what works well in Pandora and defining what I want for particular stations.

The basic idea of my listening test is to listen to ten sets of songs generated by a station and score each song: 1 for thumbs up, 0 for thumbs down and .5 for neither. Please note that it is important when doing a listening test to not click on a thumbs-down until after the songs is over since the player will start an entirely new set at that point and you could miss any remaining songs in the set.

The first time you click on a station after starting up the player seems to start a new set (and so you generally do not have to worry about a partial set at the beginning). To help me identify whether a set is three or four songs, I click on the song page for each song and copy and paste the focus traits into a column of the spreadsheet. I then shade the traits which are common between those songs I believe to be in the same set. On particularly homogeneous stations this step can be hard, and you may need to change you mind and readjust your assessments occasionally. I often will wait on the fourth song of a set to see if it's more similar to the next song.

Once I've identified the sets and scored each song, I then calculate a score for each set as the average score for the songs in the set. For instance, on a three-song set if I had a thumb up, and thumb down and a neither, I'd score that set as a 0.5 ( = (1 + 0 + 0.5)/3). The final score for the station is the sum of the scores across the sets giving a number from 0 to 10. A good score is anything above a 5. An 8 or higher is a great score and was very rare prior to the change in the selection algorithm this summer. Now, you can get at least some of your stations into that range with diligent development.

The scores tend to be pretty volatile, but that is largely because a session of ten sets is not long enough statistically for the average song score to settle down. Nevertheless, a ten-song set is about as long as can be done comfortably at a single setting, and it is long enough to give you some read of the quality of the station.

Monday, September 3, 2007

Stations Story: Loreena and Friends

This station is the oldest station that's still a part of my Quickmix. This station is what Tom Conrad calls "a simple station", that is one that is based on a single artist or song seed.

I was first introduced to the music of Loreena McKinnet by a Canadian documentary called The Buring Times which was the second of a three part series called "Woman and Spirituality". An instrumental track called "Tango to Evora" which featured in a dance sequence in the movie. (How many documentaries not about dance actually have dance sequences anyway?) The piece is striking and had me going through the credits in slow motion to discover who had created it. The track was on her recently release album, "The Visit". She became a favorite artist for both my wife and I, and I quickly caught up with her back catalog.

This station introduced me to Luka Bloom and Mary Black, and led me to purchase a CD from each of them. The station, of course, brings in a lot of Enya, but since I used Enya as a seed for my more general New Age station, "O, Wow, the Moon," I do not give thumbs up to her tracks on this station. This station also has a fair amount of overlap with the Celtic station, "Tir Na Nog," and the "Womenfolk" station.

Friday, August 24, 2007

Station Reviews: Two for t-mobile and One for BlackBerry

@Home Radio* (click to play) This station offers recent light rock from four artist seeds and one song seed. Both of these stations created for t-mobile feature a few thumbs-up and, remarkably, a few thumbs-down. It's rare to see any thumbs-down on a promotional station. The few of us who used to talk about such matters over at Pandora Stations conjectured that the station builders do not want to give the impression that their client dislike a particular artist. To me the presence of thumbs-up and down just mean that someone actually spent a little time listening to the station. In this case they probably did not listen long (3 up, 2 down), but, at least they listened which makes the development of this station better than most promotional stations.

As for the results, the station is mostly harmless. I must say, however, I am happy to be listening to a nice cover of Cat Steven's "Trouble" by Elliott Smith while I write this review. I'd say the station is one of the better promotional stations, though it still doesn't come close to being is good as an average station built by a Pandora fan who has learned the value of providing feedback.

On-The-Go Radio (click to play) This station puts out danceable Electronica from three artist seeds and four song seeds. As mentioned above, the station has had at least a minor amount of development featuring one thumb up and three thumbs down. The station is reasonably energetic, but I find this side of the genre repetitive, almost wholly unmelodic, and the rhythm tracks, in particular, predictable and annoying.

BlackBerry presents John Mayer Radio* (click to play) You can create the exact same station by clicking on "Create a New Station" and typing in John Mayer. The people in charge of creating stations for ads do this occasionally, and I completely fail to see the point. (Yes, John, it does piss me off - the station is playing Mayer's "Back to You (Live 2003)") I know that the advertisers want to associate themselves with particular artists, but from the point of view of a Pandora listener, these simple stations with no development what-so-ever add no value at all to the Pandora experience. Why would I want to click through to hear a station I could make for myself in an instant? That being said, I am listening to another nice cover on this station: Jeffery Gaines is doing Peter Gabriel's "In Your Eyes".

Tuesday, August 21, 2007

Station Building 101: Twiddling Your Thumbs

Every time a song comes up on a station, the diligent curator has four choices: thumbs-up, thumbs-down, move it or leave it be. Because I like a wide variety of music, my ideal is to have many tightly focused stations, and use Quickmix to bounce between them. Thus, I have a different goal my mind for each station, and I attempt to use the thumb-decision to improve the focus toward that goal over time.

Frequently, the goal for my station is a genre. The stations that I am currently actively curating cover the following genres: Progressive Rock, New Age, Folk, Bubble Gum, Techno, Ska, Celtic, Early Electronica, World-Music tinged Electronica, Rockabilly and A Cappella. For these stations I use thumbs-down to weed out tracks which I do not consider to be in the genre, thumbs-up to tracks which are exemplars of the genre, and I leave alone those songs that are near a border of the genre, but I would not mind hearing again.

Other stations have a much more diverse range of goals. The goal for The Best Medicine is songs that make me laugh. The goal for Pagan Pride is music that addresses themes of concern to the Neo-Pagan movement or by explicitly Neo-Pagan bands. Pretty Pop Princesses is supposed to generate pop songs by female lead vocalists (I give thumbs down to any male lead vocals). Grrrl Power covers edgier, punkier girl groups, but if a kicking guy group plays, I leave it be. My Ren Faire station focuses on a small sub-genre of folk-rock influenced by early music.

The further these goals are from the information which is captured by genomes, the more frequently the station will stray from the desired goal of the station. For instance, I include as seeds for The Best Medicine all of the available tracks by The Shaggs because their music always makes me laugh. But none of the tracks that the genome identifies as close to those of The Shaggs is remotely bad enough to be funny the way those of The Shaggs are. Thus, I know that including the music of The Shaggs dilutes the focus of the station as a whole; nevertheless, it is important to me that their music crop up occassionally even though I must thumb-down virtual all of the other tracks in sets generated by them.

Frequently, as I explore the music coming from a particular station, secondary goals will arise. For instance, I try to keep Punk out of my Ska station. I don't dislike Punk, but I want my Ska station to play Ska as frequently as possible and there's a lot of Ska-Punk available on Pandora. Some Ska curators want to include that material, but I chose not to. The key to me, therefore, in that case is that I generally will not give thumbs-up to any tracks on that station which do not include

A tertiary and more subtle goal, in general, is to try to keep similar stations distinct from one another. My Loreena McKinnet station, my Celtic station, my Ren Faire station, my New Age station, and my Pagan Pride station are all somewhat close to one another. Thus, I also consider whether a given track would be more appropriate on another station and occasionally will move tracks as necessary.

An important part of keeping similar stations distinct is to follow the following simple rule: do not give thumbs-up to a track which is the seed of another station. In fact, if I create a new station, I will go to the station-page of adjacent stations, expand the thumbs-up list, and remove all the thumbs-up for tracks which are covered by the seeds of the new station. If I like a track, I generally do not mind hearing it coming from several stations, but I want that track drawing in new music on only one station.

Following this idea to its logical conclusion, I probably should attempt to insure that particular tracks are thumbed-up on at most one station. I have not done so, but it's the kind of project that I could see myself doing periodically.

Clearly, I do not listen to Pandora while actively attending to these standards all the time. But when I do click on the player page, these are the considerations which cross my mind as I consider the decision to thumb or not to thumb.

Sunday, August 19, 2007

Station Story: Dance Dammit

Over three months after creating my first station, I finally felt the need to create a second. The intent was to create a station that would play up-tempo dance numbers across a variety of genres. I put in four song seeds of some favorite songs with no real commonality other than the tempo.

This station taught me the lesson that Pandora is not good at isolating common traits between the suggested tracks. It did introduce me to Brave Combo, and Batmobile, but, overall, the station was a disappointment since not all the tracks were up-tempo. I stopped developing the station this past January after giving it an initial listening test.

Saturday, August 18, 2007

Is that Pandora in your pocket, or are you just happy to see me?

I thought that I had another year to go on my previous cell contract, but once Pandora had survived the deadline for royalty crisis, I checked to be sure. To my surprise, I was free to switch, and so I started getting Pandora on the Go a few weeks ago.

The first decision was which of the dozen available phones to choose. I started thread over at the Head-Fi forums eliciting any opinions about the audio quality of the possible phones. No one replied. And so I researched the phones over at Phone Scoop. The only significant difference in the specifications that I could see that might be relevant was that the Sanyo M1 had about 20 times the internal memory at one gigabyte than the other phones. Sprint had discontinued the phone a couple months ago (three or four of the phones listed by Pandora are discontinued), but several on-line merchants were still selling it.

All the phones required a two year contract, and you must have a Power Vision plan to have the unlimited internet time that Pandora requires. I went with the cheapest talk-time plan at $30/mo., and the cheapest Power Vision plan at $15/mo. A subscription to Pandora is also required at $3/mo. My previous plan was $40/mo., and so I'm spending about $8/mo. more to get Pandora everywhere Sprint can reach.

Unfortunately, Sprint's reach does not include my house. The nearest cell apparently ends about 100ft away from my driveway. Even with good reception, the PCS call quality is not as good as I was used to with T-Mobile. On the other hand, the customer service efficiently cleared up a set-up problem, and I like the operating system better than that which was on my T-Mobile RAZR. After the initial setting up, Pandora is just the push of two buttons away.

The M1 does connect and download swiftly. It comes with a 2.5mm (male) to 3.5mm (female) headphone adapter/microphone. However, the wire on the adapter is thin and a bit too long. I wanted a male-to-male adapter and ordered one from Crutchfield. Unfortunately, there is more than one kind of 2.5mm jack. Most cellphones have four contacts (left out, right out, microphone in, and ground) and most 2.5 mm adapters only have three. Using the wrong adapter means you'll only get sound out of one channel. I then purchased the one Pandora suggests, and it is shorter and sturdier. The remaining downside is that there are four connections between the phone and my ears: the 2.5 mm jack to the phone, the adapter to a lead, a lead to my head-amp, and the head-amp to my phones.

The sound quality is not as good as that from a computer. The M1, at least, has a noticeable amount of distortion, and particularly sibilant highs. The stereo definition and width of the soundstage are actually pretty good, though, even on the Sennheiser MX500 buds I use on the go. These audio problems are clearly the fault of the M1 since I use the same components through my computer at work, and the sound is definitely better on the computer.

A different audio problem is a fault of the Pandora player, however. Momentary interruptions of receptions can be heard as gaps in the music resulting in occasional blips and blurts. If you've ever had internet problems on your computer, you know that the Flash version of the player will almost always continue to the end of a song before it stops. That is, the Flash version uses the computer's memory to buffer the song in the background while the song is playing. The cellphone player does not buffer at all: when the reception stops, the music stops. I wish that the cellphone version of the player would avail itself of the available memory, but I must admit that the other phones have less than 50 megabytes to work with. (To compare, the Flash version requires 250MB to work at all.)

My only other complaint with the cell player is that when listening to your Quickmix, there is no display of which station a particular songs is coming from. I prefer to give my feedback in context of the station which is playing the song. That is, even if I like a particular, say, punk-pop song, I would like to give that song a thumb-down if it's coming from my Ska's The Limit station because I'm trying to weed out the ska-punk from that station.

In all other respects, the cell player is phenomenal. It was incredibly easy to set it up to access my Pandora account and stations. The interface is easy and intuitive, and it allows you to do everything the flash version does except edit your station pages (which I would not want to do without a thumbpad). One button is clearly marked thumbs up, another thumbs down. Skipping is the right direction button, reviewing the previously played songs is the left direction button, and changing stations is the up direction button. Extraordinarily clean and well designed. Great job, Tom.

And so each morning, I put on my buds, turn on the head-amp, and start the player. It loads as I drive down the hill, generally kicking in before I reach the first stoplight. The reception ends on BART under Oakland. I click the enter key to launch the player again when we emerge from the tunnel. I listen to another tune through West Oakland, and then lose the reception a second time in the Transbay Tube. As we pull into the Embaracdero I press the enter key again for the short walk to my building. I switch over to the computer when I'm at my desk. I switch back to the phone over lunch, and again when I head home. From 7:00 to 5:00 the buds are in my ears except for meetings and bathroom breaks. An unending, inter-venous connection to beautiful music. I love my mobile Pandora.

Thursday, August 16, 2007

Station Review: Malibu Barbardos Beach Radio*

One of the things that Pandora does is build stations for advertisers which are then incorporated into the ads that frame the player. I like to review these stations as they crop up. Because I now listen to Pandora on my mobile phone, I am now a subscriber. Subscribers do not normally see the ads, and so I asked Tom Conrad for the ability to do so, and he helpfully sent me this URL.

On deck today is Malibu Barbados Beach Radio (click to play). This station was built to sell a brand of coconut flavored rum. The station definition consists solely of 5 artists with no thumbs whatsoever. The bands work fairly well together with an up-to-date reggae, blues, indy feel. Nevertheless, as usual the lack of development is disappointing, and the station is below average even for the typical mediocrity of the advertising stations.

Wednesday, August 15, 2007

Station Building 101: Artists vs. Songs

I frequently use both artists and songs as seeds. I will use a set of consistently similar artists to form the base of station, and then add any individual songs by other artists of which I'm aware that might also fit the desired theme of the station.

However, one must be careful when adding songs as seeds to any station. The fact is that there are songs which have a Backstage page, and will come up as a positive result of a search, but which have not been analyzed and are not available for play on Pandora. When you add such a song, the Player behaves exactly as if you had added the corresponding artist to the station. And so the problem is that you may think you adding a single song which fits the theme by an artist who does not usually do that type of song while, in fact, you could be adding the entire otherwise irrelevant corpus of that artist.

In order, to verify that that a song is available for play, check its Backstage page and verify that it has a list of "Similar Songs". Of course, if you've already heard the tune on Pandra, then it's available for play and (modulo any rare cases where, say, a studio version has been analyzed but a live version has not) perfectly safe to add to your station.

Another advantage of artist seeds is that Pandora is constantly adding new music, and by using artist seeds you are much more likely to hear new music by that artist as it becomes available.

Sunday, August 12, 2007

Station Story: Ecclectica

On January 12, 2006 I read a story about Pandora in the East Bay Express. When I went home that night I created my first station. As I look at the dates for all the thumbs up I gave to the station during its ten-month life, I am reminded that when I first encountered Pandora I had no good way to listen to it. My house is a loft, and my computer shares a space with the living where my seven-year old generally watches TV in the evening. Thus, my first experience of Pandora was on the tiny, poor speakers of my home computer.

I had to have my Pandora after that first night, and so I quickly researched the world of Head-Fi, the audiophile approach to headphones. I decided to get Beyerdynamic DT880's for headphones, and a great, inexpensive head-amp, the Go-Vibe v3. (I got the 2nd-to-last v3 produced, and later I got a v4 for work and, now, the commute. Looks like the v7 will be the last that James Delgarno will produce. If you have any desire for a head-amp, buy this amp when it becomes available.). And so it wasn't until Feb. 11 when all the equipment was gathered that I was able to start listening to Pandora on a daily basis.

I started Ecclectica with songs by Cake, Madness, Puffy AmiYumi, and Dar Williams. The next month I expanded the range of the station even further with tracks by Orchestral Manoeuvres In The Dark, and Loreena McKinnet. That first day introduced me to The Sounds which means they were the first band I discovered with Pandora and later bought a CD. As you can see I like a wide range of music, and Ecclectica was the beginning of a beautiful relationship to Pandora.

The station eventually strayed further into reggae than I wished, but the eclectic range of the songs kept me quite happy until the advent of the Quickmix feature in October. I then disassembled it into several other stations, and kept it around only for the sake of nostalgia.

Saturday, August 11, 2007

Station Building 101: Getting Started

A good Pandora station is one which frequently plays music that you like. That goal is subjective, and, furthermore, the tools that Pandora provides to achieve that goal are indirect at best. Thus, particularly discriminating or compulsive listeners can find building a station in Pandora to be frustrating. With a little bit of understanding and patience it needn't be.

The first thing that you need to understand is that Pandora considers ALL aspects of the genome when assessing the similarity of tracks. That is, the player tends to weight fairly equally qualities like tempo, instrumentation, voice registry, lyrical content, and song structure. If you try to create a station which only focuses on any one of those qualities, you are pretty much doomed to failure. For instance, you could construct at station with every song you know that has a 5/4 time signature, and Pandora will almost completely fail to find other songs which have a 5/4 time signature.

On the other hand, because Pandora weights all these musical qualities fairly equally, it does a fantastic job at finding songs within fairly narrow genres. The reason Pandora works well for genres is that tunes within a genre tend to employ similar tempos, instrumentation, lyrical content and so on. Thus, a good place to start is to pick some genre you like and try to create a station for that genre.

You might think that using the smallest number of tracks as seeds for a station would result in the most focused station, but, unfortunately, if you look at the Backstage page for each track you add and do not agree that the six listed similar songs are what you are looking for, then you will not be satisfied with your station. If your station has few seeds, then you will hear the listed similar songs a lot.

Furthermore, pruning by thumbing down from a small set of seeds only results in a different small play-list from which the player will still deviate. One station builder added the individual track of Abbey Road and gave over three thousand thumbs down with no thumbs up, and was still giving thumbs down at the same rate as at the beginning.

For a station to work well there seems to be needed a minimum number of tracks as seeds and thumbs-up. I find that having at least four artist seeds is a good place to start, and if a genre is fairly homogeneous (like, say, Ska), then the more, the merrier.

If you do not know a lot of bands within the genre you're creating, you can generally find more by using Pandora's station search function to find what other listeners have grouped together. You will be looking for stations clearly focused the same genre. You will find some stations have every band and the kitchen sink (particularly the ones started prior to the release of Pandora's Quickmix feature), and you will need to ignore those stations. If you haven't heard of a band, it's generally a good idea to look at the Backstage artist page for the band to confirm that it is indeed within the genre.

Once you've got a good base of artists for your station, you will naturally discover other artists within the genre as you play the station. Be aware of scope creep, however. A single thumbed-up reggae track can infest your station with reggae (which is only a bad thing if you were aiming for some other genre). Do not be afraid to give a track a thumbs down even if you like it, if it does not fit the kind of music that you want the station to play as a whole.

And there you go. Pick four similar artists, add them to a station, and enjoy.

Friday, August 10, 2007

What we know about Pandora

How exactly Pandora works is, of course, proprietary information that helps keep Pandora in business by selling ads and subscriptions. Nevertheless, Tim Westergren (CEO) and Tom Conrad (CTO) have revealed bits and pieces of how the player works some of which is not widely known.

The Genomes The simple way to understand the Music Genome Project is that there is a set of traits which are assessed for each song in the database. More accurately, however, there are several such sets. The fact is that the initial genome did not make useful distinctions for certain genres like Electronica and Rap, and so they built more detailed genomes tailored to these genres as they entered into these genres. There is no genome for Classical music, and it remains unclear how high on the list of company's priorities is the creation of a Classical genome. Or, perhaps, it's proven particularly hard to create an effective genome for Classical.

The Focus Traits The traits listed on most song pages are NOT the genome traits. The listed traits are called "focus traits" and are calculated for each song after the genome traits are assessed for a track. The evidence is fairly clear that there is a different set of focus traits for each genome. For instance, Electronica tracks tend to list many more focus traits for each song than, say, a rock track. It also should be noted that around 5% of the songs which play on Pandora list no focus traits at all.

Song Assessment Pandora buys CD's to acquire the license to play the tracks of that CD on the player. However, when one of the trained Pandora musicologists tackles a CD they do not, in general, assess the genome for every song on the CD. Thus, even though a track may have a page in the database and even a sample, the track may not be available to play on Pandora because it's genome has not been assessed. If and only if a song lists "Similar Songs" on its page will that track be available for play on Pandora. The company will go back and add a song if enough people search for it; however, if one of your favorite tracks was passed over when they first covered a CD and that group does not have a following on Pandora, then it's unlikely that the track will ever appear on Pandora.

Song Selection The non-Quickmix player generates songs in sets of 3 or 4. The player appears to select a "keystone" song for a set from the list of the seeds (the Artists or Songs you selected for the station), the list of thumbed-up songs, or (rarely) one of the songs which are considered "similar" to one of the songs in those lists. The keystone can play in any position in the set (1st, 2nd, 3rd or 4th), and the other songs in the set will be particularly similar to that song (often these songs will appear in the list of "Similar Songs" for that song on the keystone's song page). Since June 20, 2007, the song selection algorithm has increased the chance that songs from the lists of seeds and thumbs-up will be among the additional songs in a set, and so it is more difficult to identify the keystones when listening. It appears that when a Quickmix plays several songs in a row from the same station, the songs are from an identically generated set of 3 or 4 songs.

Thursday, August 9, 2007

Links To Prior Posts

I've written several previous posts about Pandora on my main blog.


My first post about Pandora was an enthusiastic description of the service written a couple of weeks after I came across it. It's interesting that many of my technical wishes have been met, but the seemingly easier wish to build a community has not really gotten there. They have recently expanded onto facebook, but I'm sceptical.

Dr. Pandora Love

Pandora can make you confront your musical prejudices. I wrote this post in reaction to hearing a Cyndi Lauper tune on Pandora.

Shaggs on Pandora

I wrote a brief post on my delight that The Shaggs were available for play on Pandora. They have become a delightfully intractable part of my novelty station, The Best Medicine.

Pandora Shame, Pt. 2

Upon giving a thumb up to a Diana DeGarmo song, I begged forgiveness from the Lord. The track became one of the seeds of my Grrrl Power station.

Pandora Town Hall Meeting San Francisco, July 19, 2006

I attended a meeting with the Pandora folks, and wrote a thorough recap of the evening.

Pandora Town Hall, UC Berkeley, Dec. 4, 2006

I attended another town hall six months later, and provided another recap of the event.

July Listening Tests

I posted up a graph of my listening test results for the last six months. This graph shows the evidence that they made an important improvement to the selection algorithm in June. Indeed, I received an e-mail from Tom Conrad that they had made the change on June 20.

And that's all I've written so far in my blog on the topic.


Pandora is a free internet radio service that plays commercial-free radio which is tailored to your taste. Now that the fan community site has died, I've decided to start a blog to continue to talk about Pandora, and preserve some of what we have learned about building stations on Pandora over the last year and a half.