Slashdot Mirror
The Perception of 'Random' on the iPod
Robaato writes "Stephen Levy writes in the Guardian about the perception of randomness, or the lack thereof, on an iPod set to shuffle." From the article: "My first iPod loved Steely Dan. So do I. But not as much as my iPod did.... I didn't keep track of every song that played every time I shuffled my tunes, but after a while I would keep a sharp ear out for what I came to call the LTBSD (Length of Time Before Steely Dan) Factor. The LTBSD Factor was always perplexingly short." My first iPod shuffle refused to let me delete (sigh) Weird Al's Polkamon off of the flash memory.
Is the iPod Shuffle Playing Favorites?
How about an analysis of the randomness of Zonk dupes. I guess I should be happy it's not a games story.
http://en.wikipedia.org/wiki/Confirmation_bias
You know, instead of wasting your interviewee's time, you could have installed a five song list on your iPod and set it to shuffle. You'd have to carefully mark down the track number being played and listen to it for 100 songs. Do this a few times and make sure you're very methodical about what you do. Wipe the iPod, put five songs on it in order and then listen to a hundred songs "randomly." If you start to see a pattern developing or one song is obviously favored over the other, it will begin to show up.
But on the more technical side, they have to seed the random variable with something. Whether or not it's an internal clock, I'm not sure. Either way, they have to derive a random number and it's possible that their seed isn't good enough or has too few states or is prone to being seeded at the same state, etc. Based on this information, I hate to break it to you but it is very hard to be truly random.
My work here is dung.
My iPod likes Slayer and Marilyn Manson. I guess its posessed.
Well you'll never get true randomness but I used to notice this in mp3 playing software a lot of times particularly winamp seemed to play some tracks a lot more than others when on random.
:)
On Linux I used to use a command line player and a nice structure of directories and symlinks to make the playlists and never used to bother with random.
Now I do most of my work on a Mac, but I also happen to listen to music less now, so random is now random enough for me.
Anyway, slow news day of what, this is the second pointless ipod story I read today on here
I like ipods, I have one but only ever use it on long journeys and no I don't have DRM'd tracks so I didn't care about online music purchases. The ipod just happened to be the one that worked the best (scrollwheel is nice and quick) and having a mac I knew it'd work well.
Probably more random than the percieved randomness of iPod articles that show up on Slashdot
...well, used to. Then I made a different playlist and labeled it as "Upbeat" music. Now It loves Motion City Soundtrack. Now, I like all the music I have on my iPod (duh) but its very noticeable when I hear the same song 3 times in one hour 30 minutes worth of driving. I can tell you that over the last 2 days (4 hours of driving) I have heard Foo Fighters' "Enough Space" 6 times out of 231 songs. Does the iPod sense higher played songs/albums/groups or is its randomness just that awful? 2GB Nano 1g
I recall The Steve personally addressing this issue in one of his keynotes (although not with iPods, but iTunes). People thought iTunes' shuffle feature wasn't random enough. Steve assured everyone that it indeed was completely random, but then announced that iTunes had a new "Smart Shuffle" option. The description in iTunes is "Smart shuffle allows you to control how likely you are to hear multiple songs in a row by the same artist or from the same album." There's a slider with "More Likely" on one end, "Less likely" on the other end, and "Random" right in the middle. Although this feature is in iTunes, it has not yet made it onto iPods.
I personally have had it happen where my iPod is in shuffle mode and I've heard not just two songs in a row by the same artist, but a song plays and then the next song from that album follows it. And that's with a library of over 5,000 songs. Naturally it's more likely to happen on a much smaller Shuffle with a fraction of the songs.
Duct tape is like the Force. It has a light side, a dark side, and it holds the universe together.
I agree that it is extremly difficlut to be truly random. There are some good ways of initializing the seed in such a way that the pseudo random number generator behaves differently. A good way, done in telecom terminals is to measure the noise at the receiver and then use it to seed the random number generator. Since, the noise is truly random, that is a good way to seed the random number generator. If the costs, are not too high, then it might even be a good idea to read noise (or any truly random parameter) whenever required. That would be close to really random, provided, we can map the random parameter into a quantitative parameter without big errors and approximations.
So many iPod owners and none of you have stated the obvious. The iPod will play songs with a higher rating more often.
Nonsense article again, but because it was about iPods it had to be published on the front page.
Thank you Slashdot and all the friends here - my time among you is almost at an end. I will find another news site where I can actually read something that is worth reading. Slashdot used to be a great site, but it's nothing but shit nowadays. Bye bye!
Humans innately seek patterns in things that are random. That's why so many people wear smelly socks because they think they're lucky. Once you identify a supposed "pattern" i.e. non-randomness, you're going to keep noticing instances that fit that pattern, and ignore instances that do not. This is deeply ingrained.
Think about it, if you're at the roulette table and black has come up four times in a row, how likely are you to bet black? Most people would bet red, because, I mean hey, there's got to be a pattern. But (as I'm sure you all can understand) black has the same probability of occurring again as red does.
People have had this complaint about all sorts of playlist randomizers (not just iPod), it's just people seeing what isn't there.
Apple should add another random play mode -- one that acts as it does now, and the other mode that grants every song an equal play count. The only thing that would be random is which order. This way users that have a confirmation bias of their iPod favoring certain songs can no longer be paranoid of Apple conspiracies to promote the songs of {{ artist }} or {{ record_label }}.
...it's good to listen to on the train, you get all sorts of random songs, sort of like your own personal radio station... so long as you like the Beatles, that is. I punished the little unit by removing all the Beatles songs currently on it (there were 5, out of a full gig worth of songs) and instead of every second song being Norwegian Wood (for five hours I got random song, Beatles, random song, Beatles, random song, Beatles...), it started acting more normally.
Recent re-introduction of more Beatles music last week hasn't caused a return to the 60s, either, to my surprise. Yet.
I have noticed a lot more Beethoven than I thought I owned though....
What I want to see is someone who actually scientifically tests the randomness of the suffle. Perhaps you could use Autofill a couple hundred times, and test it to see if all the songs are chosen equally.
Not sure its so much confirmation bias (alone at least), as it is that the odds of NOT playing a song from the same artists over the next X songs shrinks more rapidly than intuition suggests. That is, for example the odds of NOT having a run of X heads or Y tails when flipping Z coins is very, very small.
The article mentions the "how many people does it take to get to a shared birthday thing" - and the point there is that its not that it takes 40 people to get to one with a SPECIFIC birthday but only 40 or so to find two that SHARE a birthday.
-----
graphically speaking
graphically speaking
The open-source music player I wrote (BSoftPlayer) has a feature called "SmartShuffle". One of the biggest problems with shuffle is that it's difficult to understand when the tracks will change order, and it's difficult to know wheter or not a track is going to be played more than once in a single "cycle". Some shuffle features will play the same track twice before playing through your entire library, and some won't.
With SmartShuffle, the order is randomized, but it remains the same until you "reshuffle".
There must be bias on the internet. I remember reading the same thing not ten minutes ago in TFA... creepy.
I used to carry a bottle of whiskey for snake bite. And two snakes. -Nefarious Wheel
What, we're losing Anonymous Coward? He's been here since the begining, and he wrote half the posts! OH NOES!!!!1!
"Nobody owns the fucking words man." - James Dean
You can get the technical details here: http://en.wikipedia.org/wiki/Birthday_paradox/
The basic gist is that their are far more possible pairs than we'd intuitively imagine. For example, with 20 albums of 20 songs each, the chance of two songs in a row being from the same album is actually:
400/400 * 20/400 = 1/20
Which makes a lot of sense once you sit down and think about it, but is a lot higher than an uneducated guess.
This is the same reason that collision/timing attacks are feasible.
FTFA:Or
Duct tape is like the Force. It has a light side, a dark side, and it holds the universe together.
I think (and have... thunk?) that "randomness" on the ipod is actually a secret R&D weapon in the apple ipod toolkit. From a psychological standpoint alone, what is the value of all other mp3 players being truly (read unadjusted psuedo random) and the ipod being a little less.. that is, what if they, say, mark the number of times you don't let a song play through, but skip it in the first 10 seconds? There are powerful means by which they can onboard build a profile and i have three things to say about that:
1) that is a FUN project for a team of engineers to do and,
2) Why wouldn't they for the HUGE hidden psycological impact it could have in differentiating the player
3) It's closed source so you can't actually tell, so the five songs with-no-user-input model wouldn't work. Another might...
Regardless, i wouldn't expect them to miss the importance such a feature would have. The iPod just keeps the vibe going, while the competition keeps playing country-house-ambient-country-house-ambient
Also, the "sound-check" would be a good place to do some quick BPM detection to have like tempo's play. The new settings for more- or less-random in iTunes almost scream "we are doing something tricky"
Wouldn't you, if you could?
CS majors know the time/space tradeoff, but they never get taught the 3rd, crucial, tradeoff of the set: comprehension!
A long time ago I was dissatisfied by the lack of random in XMMS so I jumped to the source to see what I could do. I think this was my first contribution to a free software project. Anyway, here is what I found: XMMS keeps two copies of the playlist, one that is in the order you set and one that is "shuffled". This has to be clear, all the tracks in the play list are there exactly once in the shuffled playlist.
I can't recall when the shuffled playlist was reshuffled but in was not that often, maybe only when you added or removed tracks. So if you like Smoke on Water but that Ballroom Blitz is just two song after that, too bad, you'll always get Ballroom Blitz soon after you double click on Smoke on Water. Technically speaking, the shuffling was perfect, the random generator was properly seeded and they divided in the right way to prevent loosing entropy. The lack of reshuffling was entirely responsible to the perceived lack of randomness.
So my patch was just that: trigger reshuffling a lot more often. As far as I know this patch was never merged but my copy of XMMS did have the proper random behavior. I don't know if it's the same problem with the iPod. That's something I like with free software: you can fix it!
I wonder if the "random" on an iPod is like the "random" on my Netgear MP101 wireless audio device connected to my stereo system.
/. about not being random. :P
If I select any playlist, and hit "PLAY" when the shuffle option is set to "Random" - then it plays the same order of songs every single time. To get a genuinely "more random" feel to the way it plays songs, I have to select the non-first song at the start of the playlist, then hit PLAY - and then hit "NEXT". After *that* it seems to be relatively random (Except for a pre-disposition to playing Eminem - or so it seems!).
I know it's tough to generate truly random numbers - but lets face it - for a music player - random doesn't need to be "truly random" - merely random enough to avoid getting posted at
How many escape pods are there? "NONE,SIR!" You counted them? "TWICE, SIR!"
(I looked at mine; it was closer to uniform than I'd perceived. There's also a "Skip Count", but it's blank for all my songs.)
Stupid job ads, weird spam, occasional insight at
Because it seems the same songs play at certain points that I drive through in the city.
Or maybe it's not a grand conspiracy at all, but this grand idea called chance.
I got myself a shuffle a couple months ago and put a bunch of songs on it, played it almost every day, at least once a day, sometimes more, for almost 2 months...never noticed a single bias of any kind (other than a small bias towards playing newer songs if you turn it on in shuffle mode). However, I ran my songs through a couple filters to remove all traces of information that I didn't put in (I.E. Ingrained bands, other data apple might put in songs from the ITunes library) so perhaps the shuffle uses a psedu-random generator, weighed towards ITunes songs or something. Or, as someone said earlier, perhaps it's just your perception. If you have 5 songs from a certain band, and 130 songs you'd think it would be 65 songs between them, but that's not even close to true. In a full random generator you'd get bursts of lots of a certain band, followed by breaks between that band's songs. So it would seem like there was very little time between that band's songs, when really you're just forgetting the times that you didn't hear that band for a long time. Or maybe Apple is out to force you to listen to certain bands, 'cause you know they make a penny every time you hear the song...*end sarcasm*...seriously though, what would make Apple program the shuffle to favor certain bands? Maybe some money from those bands, but it would have to be a good amount, and who wants to pay to make sure you listen to them a lot, they don't make money when you hear them either. No one stands to benefit from forcing you to listen to the same band a hundred times so why would they go out of their way to add that to the programming? As a new programmer I'm 99% sure it's harder to program a random generator that's intentionally weighed over one that's not.
There are two kinds of fool One says 'This is old therefore good' Another says 'This is new therefore better'- Dean Ing
My old iPod LOVED Moby.
Karma: Bizzare (mostly affected by varying internal caffeine levels.)
Well, the random is better than Sony's original implementation of 'shuffle' for its CD players. I think it's better now in new players but I used to be able to put in a disc in old players (1993 and earlier) and the track ordering was the same for *every* disc. Eject disc A and put it back in to be played in 'shuffle'? Same order over and over. It was really quite annoying. Disc B? Same 'shuffled' track order. That's quite some effort they put into that algorithm.
I always found that the pattern of "random" on the ipod was pretty easy to pick out. It seems to only like a certain set of music; some stuff gets played all the time while some things never seem to get picked. The interesting thing is that when you put more music on there and seems to bump the whole shuffle order around, so things that were never coming up before all of a sudden start getting picked.
I wonder what it would take get something like amarok to bump the files around? Maybe a simple function to remove 10 "random" albums and then re-add them would be all that's needed to get a much better shuffle.
Is it possible that human brains just have the gain set a bit higher than simply "random" on connecting temporal causal events? It might be more adventageous to notice connections between events that aren't connected (then dissmiss them) than it is to ignore events that are connected causally.
In other words, it might be better to be a little over paranoid and think that the random shuffle on an Ipod isn't random, that childhood vaccinations cause (insert disease here) than it is to miss the fact that when some people eat fruit A, they die (but not say everyone).
Putting it a bit differently, there's more cost to missing connecting certain dangerous events than there is to miss-identifying harmless events that later turn out to be non-connected.
AccountKiller
What is wrong with an iPod that plays nothing but Steely Dan? I'd buy one if Apple made it.
All Steely Dan, All the Time
There's far worse iTunes/iPod problems to complain about. How about their choices with DRM, difficulties with moving your iPod from one computer (home) to another (work)? Add a third (laptop at home) and then what?
Music shuffling is fine. What needs to be fixed is the randomness of the blocks in the iPod version of Tetris. Without fail, it always starts me out with the red "Z" block, yellow square, red "Z" and the brown "L" block for the first game upon starting Tetris (after this the next game starts with different blocks, unless I quit and reload Tetris).
Seems to me like the programmers used a bad choice of a seed value for the random number generator.
Duct tape is like the Force. It has a light side, a dark side, and it holds the universe together.
I noticed that most people view a uniform distribution as a "random" distribution. As the article states, people impose patterns on things, thus when the average person tries to makes something random they tend to create something with a uniform distribution 'because it spreads the stuff out'. Ah, the fun of counter-intuitive math. :D
"All the darkness in the world can not quench the light of one small candle."
This is just about exactly what "Smart Shuffle" does. It allows you to bias the randomness towards not playing two songs from the same artist or album in a row. Makes it less random to make it feel more random.
Don't thank God, thank a doctor!
I know there will be snarky +5 Funny comments underneath this (as well there should be), but this system to decrease the perceived randomness is actually really similar to the algorithm most radio stations use when programming their music.
There's a simple parameter that's set to control, to within one minute, the amount of temporal separation there must be between playing two songs from the same artist, or the same song twice. The radio algorithm is a little more complicated, since songs aren't in just one big batch like the iTunes library, but in different categories, based generally on the perceived desire of target listeners to hear a given new song, or like and identify with a given older song.
The system is built off the (once literal, now metaphorical) use of index cards: The format clocks say, e.g., at the top of the hour, play a P category song, followed by a B category song, then a G, then an A, etc. You'd have a set of rules, like "don't play the same artist within 45 minutes" or "don't play the same current song within 3 hours", and you'd take the first card in the category that fit all the rules, play it, and move the card to the back of the stack.
Basically, what Apple is doing with that slider is enabling artist separation control, which is completely one of the illusions radio stations (used to) use to convince you they had every song under the sun available to them.
I've noticed it too. There are songs in my MP3 collection that I'm not very fond of that I hear quite a bit of, and many that I like that I never hear on shuffle.
No, it's not! You have "al" four times in that, at least! Do you have something against Al! OMG CONSPIRACY!!!
The point is that true randomness doesn't look random, because randomness necessarily includes the possibility of patterns.
Don't thank God, thank a doctor!
Got on Vans but they look like sneakers!!!
It seemed to get worse as it got older, I even wiped it, took out the hard drive and trashed the partition table on a pc, then rebuilt it and started again, but it didn't change a thing.
Thank god it's now in lots of little bits!
I don't have an iPod but iTunes does tend to group songs by the artist. It's not consistent but easily half the time there will be groupings of songs by a given artist. For some reason the last time it shuffled it grouped all the songs by artist. I'd say it was set to a mode that does that but it's not consistent. It can't be random chance either because it happens too often. If I have two songs by a given artist roughly half the time it'll group those songs together. Odd thing. I'd like it to be much more random. Love the service but it is quirky.
For what it's worth, I've just experimented with XMMS (1.2.10) in random mode, and it seems to be doing this now.
If I double-click a track to play it, then click advance, it was always advancing to a different track, implying that it reshuffled at the point of selecting a track. Simply moving backwards and forwards between tracks left them in a consistent order, however.
Actually it does. There's a counter for the number of times a song has been played through completely. I believe one of the in-built playlists accesses this metadata.
Mind you, as to wether the device uses this information to weight its shuffle function is something I have no idea about.
Maybe iPods are showing us their souls. The inner light is shining through.
How to Download YouTube Videos
another really easy method is to build in a tiny, one-chip radio receiver inside the mp3-player and tune it to a very little-used frequency. The antenna only has to be long enough to get some good white noise. Radio static is more than random enough to seed a mp3 player.
Except for mathematicians and programmers, most think of "random" in a *very* different way from its technical definition. To most humans, saying that a particular sequence is "random" means *guaranteeing* certain things about it. Among them: the same element does not occur back-to-back, EVER, even if there are only a few elements total to choose from. Even more, if there are more than about half a dozen elements, the same element never occurs twice within about five positions. (So if you've got songs 1 through 7 on your iPod, and the first seven played are 5, 3, 7, 2, 4, 1, 6, then the next one has to be 5 or 3, or _maybe_ 7, or it doesn't seem "random" to most people. Yet, the order can't be the same every time through, either.) No element occurs substantially more often than any other element, even over the short term. If the elements have a natural order (e.g., alphabetical), then no three elements that are adjascent in that order can ever occur together in that order, nor should they occur together in the reverse order. (This gets particularly difficult to guarantee when the elements have more than one natural order, e.g., if the elements are people, you can't have three of them in a row by either name or age, or people notice and decide that the order is not random.) Even worse, if the elements can all be categorized into a small number of categories (e.g., by gender), you can't have "too many" from one category in a row. (How many is too many depends on the ratio, but if half of the elements are male and half female, having four of either in a row will make people cry foul, the order is not "random".) If certain elements stand out from the others in some significant way, they can neither occur first nor last. (For instance, if test questions are being drawn from a question bank, neither the easiest nor the hardest question should be first or last; if it is, people will say the order was not random.)
I could go on and on, but what it really amounts to is that when most people say "random" they mean "carefully arranged in a thoroughly mixed-up order". This is almost the *opposite* of what a mathematician or computer programmer thinks the word "random" means.
For this reason, when describing a mathematically-random sequence to an end user, I never EVER use the word "random". I generally call it something like "arbitrary" or "unpredictable". This greatly reduces complaints.
Now, as far as song frequency, I like to rate my tracks on a scale of 1-10, and rig my playlist so that anything under a 6 never plays unless I specifically select it, tracks rated 7 play twice as often as those rated 6, and the frequency keeps going up the higher my rating is. (I only have eight tracks rated as a 10, and they're all things I don't mind hearing back-to-back.) Then if I find a track is playing more often than I like, I figure I rated it too high and cut back its rating.
Cut that out, or I will ship you to Norilsk in a box.
It is analgous to the popularity of Windows - there is no objective reason to like Steely Dan.
But, deep down, you realize, that Steely Dan represents the worst of the music of the 70's and you are having a visceral repulsion to it, like I finally came to realize, every time I hear Stevie Nix...
So, rather than blame your nice, techie, iPod, blame your fallable human self for choosing that music.
If you remove all Steely Dan from your iPod you will no longer have this problem.
This issue is a bit more complicated than you think.
How is this post insightful? It shows clearly that eldavojohn did not even bother to read the article. If he did, he would have noticed the paragraphs about how difficult it is be truly random.
From the article:
Robbin is talking randomness in terms that software can reasonably produce, which is not perfect randomness. True randomness, it turns out, is very difficult to produce. This subject was most famously examined by Claude Shannon, arguably the Father of Randomness.
And from the next paragraph:
And if you're randomising on a computer, you have to introduce a "seed", which is a starting point for the algorithm that mixes up the selections. The seed must draw on some unpredictable input of time that begins outside the computer. Otherwise, the results would be the same over and over again.
Now you can rest easy eldavojohn, you don't have to break anything to Mr. Levy, he already stated the facts you so arrogantly pointed out.
My ipod has this tendency to play tracks from the "Buckaroo banzai /saturn 3" soundtrack. It was a bootleg-only soundtrack that has about 8 BB tracks, and about 20 Saturn 3 tracks. I regret ripping it to my ipod. No matter if I use "shuffle songs" or any custom smart playlists, they ALWAYS show up.
And removing tracks from an ipod was never easy.
Something related to iTunes - a study of the randomness of party shuffle in iTunes. This article does a bit of research and comes up with a function! http://www.omninerd.com/2005/08/25/articles/34
...the OmniNerd article: http://www.omninerd.com/2005/08/25/articles/34
I used to have the same problem on both iTunes and my iPod. However, with some clever playlist structuring, you can get as near to a random experience as possible. Using iTunes SmartPlaylist features is the only way I have been able to stop iTunes (or the iPod) from playing the same handful of songs over and over and over again.
And the question remains, why doesnt the DUPE tag work anymore? I liked that tag. Seeing it meant I could avoid the 500 "OMG! Its a Dupe!11!" comments.
In Soviet Russia the insensitive clod is YOU!
If you read poker forums at all, you'll know that this is a subject that has been beaten to death over and over. So much so in fact that (at least) one poker site's support has a lengthy form letter they send to anyone that complains about "non random" cards:
==================
Firstly, I assure you that our shuffle is entirely random. You can find its description at the first link below, and the results of two independent audits into it at the second:
http://www.pokerstars.com/security.html
http://www.pokerstars.com/rng_audit.html
We will send any real money player's entire hand history to them at any point that they ask for it. We have every real money hand played on our servers and have sent out literally millions of hands. There are many players out there with 100,000 or more hands in their Poker Tracker database. Poker Tracker is a program that allows you to analyse your own play and by extension the randomness of the site in question. If there were anything that deviated from the expected values, these players would have solid evidence to that effect. However, there's nothing on the poker forums. Infact, some players have posted their investigations online, the URLs for which you can find below. I am certain that there are more out there that you can easily find.
http://tinyurl.com/4lfeg
http://tinyurl.com/2nvav
I am sure that many of the players we've sent these hand histories to felt that there were problems with the game. That's the nature of how our brains work. They are very good at picking out patterns, but when faced with randomness, impose patterns of their own on the data. Our brain is also very good at picking out the unusual from the usual. Therefore, you will not take as much notice of the seven times you win your 7/8 shot, but the time you lose it is the time that will stick in your memory.
The nature of poker is that the better a player, the more often he will put his money in when he is ahead. Conversely, the weaker the player, the more often he will put his money in whilst behind. It is therefore natural that the good player will not need to outdrawn his opponent very often, whilst the weaker player has to do it with much greater frequency. Of course, the good player makes more money in the long run, but the weaker player will still win pots from him. It's a fact that in the long run, the good player will outdraw opponents fewer times per hour than he will be outdrawn on himself. This is simply because he is rarely in a situation where he has to outdraw someone to win. However, if you looked at every single time a strong player had a 20% shot to win with all his chips in compared to every single time a weak player had a 20% shot to win, you will find that both players did indeed make their hand 20% of the time.
There are three sets of people who would undoubtably recognise that something was amiss if there were any flaws in the shuffle. The first set is the programmers, the second senior management, and the third the poker specialists (who have access to as many hand histories as they want). All any of these people would have to do is ask for a 100% pay rise with the threat that they will release compelling evidence about the dishonesty. We make our money purely from rake and tournament entry fees. We would immediately lose a lot, if not all of our players as a result of this. The fact that this is not all over the forums, and the fact that I am (sadly) not sitting on a tropical island as I type this mail is a very strong signal that nothing is amiss.
Finally, the offer to send out every real money hand you've played is of course open to you as well. Just let us know and we can send it along. I recommend a tool like Poker Tracker to run your analysis, but there are plenty of other programs out there that do a similiar kind of job.
I hope th
Yes, that's a good point (and true). However, the original article is talking about one particular artist being played more often than expected. You're exactly right in saying that the gap between songs from the same artist in a shuffled list is shorter than most people would expect. The article author, however, is probably wrong in saying that the gap between Steely Dan songs is less than the gap between another artist.
I wrote a CPAN module called Text::Capitalize that includes a function called "scramble_case" that works this way (for when you want capitalization with a "wEiRDly sCRaMbLeD aPpEaREncE").
That's an interesting way of formulating the principle. For my "scramble_case" I weighted the probability against getting a capitalized first letter, because that looks Too Normal. This, by the way, is esentially the conclusion that Stephen Levy comes to... this isn't a bad article at all, though it's a bit verbose, and doesn't get down to the point until two-thirds of the way through. (Which probably makes it Pulitzer Prize material).iPod just plays fine :)
(Ok, sometimes I remember being surprised to see 3 songs from the same artist played consequently. However considering that there are hundreds of songs in total and ~5 from each artist, it would be not-so-random for this to *not* happen in a single run of the playlist. We tend to consider numbers like 444,499,911,101 not random, however it has the same probabilty with all the other 12 digits numbers in a linear distrubition).
Doesn't the shuffle have the best sound quality of the set? (look it up through google)
I do the same thing with winamp.
One of the playlist window options is "randomize list". I just hit that a few times, and then listen through the playlist "straight". Even better is that I can go backward through the same list, and can see what "random" song is coming up next.
(Unknown Chinese)\(Unknown Chinese)\(Unknown Chinese 1).01.mp3
The rest of the shuffle seems to be random enough, I haven't noticed any bias.
I've always wondered about that.
T.
This is a something I've noticed with my iRiver H10; when it chooses a next track "randomly," it seems to be chosen as an offset from the song that just played. For example, I'll often simply select the very first artist (beginning with A) and the very first song. From then on, I'll hear almost entirely artists that begin with the letter A. sometimes it makes it to an artist that begins with B, or back to Z, but it hardly ever hits any other letter.
Pretty annoying.
Beatles pay 5 to 1
Moby pays 2 to 1
Slayer and Marilyn Manson pay 10 to 1.
The doors pay nothing
Based on this thread I would be lining up the suckers and making some real money.
With a non-iPod mp3 player. It's played the same song 3-4 times in a row on occasion. (Combined with the unfortunate fact that the "shuffle" only works when actually waiting for the song to end, not when switching through tracks, this can be frustrating.)
But I'm not convinced it wasn't a fluke. Unlikely things do happen, and the song is no more or less likely to come up after the last nine repetitions.
If so, then this subjective problem could still be solved simply by slightly disadvantaging the songs or artists that have played recently, so they're less likely to repeat within a short time. iTunes has a similar option, but that does the opposite (making artists more likely to be played in groups).
I have an iRiver H320. The stock firmware for that player doesn't do shuffling properly at all. Every time you play something shuffled, it's in the same order.
The Rockbox firmware solves that, of course.
too important to be left to chance."
Sony Ericsson MP3 phones (at least the D750i/K750i/W800i) have a bizarre feature.
If you set a playlist to shuffle, it fixes the shuffle order, and if you turn off the the phone, after you turn it on again it finishes playing the current song (good behaviour) then restarts the playlist from the first song (bad, bad behaviour).
Took a while to figure this out, but I always seemed to hear Iron Maiden first thing in the morning. In my case I put it down to confirmation bias for a while, as what I heard when I started the phone was always different. But I finally realised that the second (and subsequent) song I heard in the morning was the one that was the same.
New practice - do not turn the phone off at night, or randomly skip a large number of tracks at the first sign of "Different World"
Does the play count for each song in iTunes also reflect the number of times the song has been played on the iPod?
:)
In that case, can't someone just replace the mp3's on his/hers iPod with 1-second versions and then have the iPod play away for a while. I own an iPod and the thought of it liking some songs/artists more than others has struck me as well, but I've just assumed it's because I react to some songs more than others.
If the playcount shows up as even this discussion could be ended once and for all.
www.freshpilot.com
Someone please take away Zonk's privileges! Everything he posts is just complete trash.
Well. My iPod has a song on it, called "Herfst en Nieuwegein". Nieuwegein is a small city near Urecht. It already happened to me twice that that particular song played exactly during my passing of the small city in question.
;-)
First you think "somebody's playing with me, that's sooo unlikely to happen!".
But when you do the calculations it's not so unlikely to happen as you think. 4GB iPod, aout 500 songs (with my encoding of choice). Say 5 minutes per song. Passing Nieuwegein takes me about 7 to 10 minutes, and when the song will play during that time it will be OK. So on average that's about 3 "pickings" (is that the right word) out of 500 songs.
A chance on average of about 3 in 500, or a chance of 0,6%.
As we all know chances like that pop up 9 times of of 10.
The earlier Wiki articles on Confirmation Bias are right on the money. I will further emphasize that the user is complaining because of the "excess Steely Dan", which indicates it is a *weaker selection* for his playlist.
If he had liked Slayer, he should have trumped how pleased "his iPod liked him, and gave him his daily dose of Slayer".
What Steve Jobs picked up on, is that true randomness includes clumps. Seeds aside, there simply occur odd mini-patterns in a true random event.
What the user *wants* is "Something Different*. Therefore the best algorithms are ones that *exclude* that same album, and if desired, that same artist, and then shuffle the *remaining* choices.
I have never been a fan of shuffle. I prefer to take some care and design a playlist properly so the order it comes out in is solid, and if I made a mistake and included a song that wears out too easily... just nuke it. It only takes a few button clicks to start in the *second half* of the playlist too.
My first Journal Entry ever, in 8 years! http://slashdot.org/journal/365947/aphelion-scifi-fantasy-horror-poetry-webzine
It's a psychological thing. Like walking by a streetlight and having it go on or off. If it happens again within an hour or so, you're SURE it had something to do with you. Even if it's the same streetlight. Even if you know when the bulbs get old, they overheat and cycle on and off. Every week or so I hear two Bob Dylan tunes back to back on my iPod. Not too surprising, I only have about 60 tunes on there and Senor Zimmerman "sings" four of them.
The problem is, the iPod is TOO random. True randomness guarentees the possibility that a single song can be played 10 times back to back to back to...to back. And nobody wants that when listening to music! The solution? Random for music should mean something different than random. The next track should be different than the one currently playing. Different genre, different artest, different tempo, different key, different . And which different xxxx should not be random either. After a death metal song, I don't want sludge metal (well, I do, but not on random mode). The iPod should pick a song that is truely different than the one I am currently listening to. The point it, in user applications, nobody every wants real randomness. So why do we keep getting it?
Everybody I've recommended a shuffle to says the same thing - "but I don't want to play my music randomly." There's a perception from the name of the player that all it does is play music on "shuffle"! So when people perceive it as being not random enough.... a little ironic, no? Maybe Apple should have just called it the "nugget" or "nibble."
Excellent. Have you submitted a feature request to Apple?
/. :-)
http://www.apple.com/feedback/ipod.html
Believe it or not, I don't think that the Apple product managers read
For this reason, when describing a mathematically-random sequence to an end user, I never EVER use the word "random". I generally call it something like "arbitrary" or "unpredictable". This greatly reduces complaints.
Computers are eminently predictable. Give them the same inputs and they'll always return the same outputs. You can try to fake an entropy pool, but if you replay the inputs you will get the same outputs. There is no randomness in computers. Unless the iPod is keeping some history of user inputs or time, or some outside-the-iPod change, then the mathematical formula that decides the order of the songs will return the same sequence again and again.
Recently while mowing my grass I noticed that my iPod seemed to have a certain affinity for Sting. However, since I have just about every Sting album ever released I figure it's bound to show up a bit more often than some other artists. What I thought was odd was that it played like 4 Sting songs right in a row. The possibility of that happening seems remote considering I have around 3,000 songs on my iPod right now.
The guitars sound good, now give me about 10db more on the cow bell.
The other day my iPod played Rachid Taha's cover version of "Rock The Casbah" followed immediately by The Clash's original version.
I've definitely seen this seemingly non-random randomness in my own iPod, but not quite to the degree that the author has. I have definitely noticed that although the "shuffle songs" does seem to randomize everything, it will DEFINITELY favor 2 or 3 particular artists after each shuffle. If I happen to start a particular mix, after about the 7th or 8th song or so, I will definitely hear higher-than-"normal" repetition of 2 or 3 of the artists that I've heard in the first opening segments.
As to whether or not this is a good thing - depends on who it decides will be its favorites for the day (or for that particular "shuffle session".
Mine LOVES Jimmie Buffet!
I don't particularily care for his stuff, but it can'tgo 5 minutes without playng a Jimmie Buffet Song!
My ipod likes Greenday, but my itunes liked coldplay... I think they don't like each other.
When I sync up my ipod using the autofill it almost always loads up pretty much every coldplay song I have, but when I'm playiung the songs on my ipod it sticks to wathever greenday songs I have on there. Who knew that two pieces of technology could have such different tastes?
Look at a scatterplot of 2D (pseudo)random numbers. It naturally has overdense regions and underdense regions. (By comparison, a scatterplot of "quasirandom" numbers appears more uniform; see more here.) This has nothing to do with poor pseudorandom number generation, psychological perception of nonexistent patterns, etc. — random numbers really do cluster at times. This led British citizens to believe that V-bomb strikes in South London had a particular pattern to them when they were really dispersed randomly.
Wouldn't this solve the problem completely? Every song that gets selected for play is removed from a stored list. Once the last song is played off the stored list, start again. Man, I'd...I'd...I'd even consider giving up my 78's for this feature.
I come here for the love
...I think we're all dumber for having read that.
How about a shuffle that actually does what I want: play all the songs in the library (or playlist) just one time each in a random order? Yah, sure, I can manually sort my playlist... but that's only good for one pass.
BTW, I think the confirmation bias being linked to over and over and over again is confirmation that those chuckleheads are biased towards being theoretically correct about random-approximating procedures in automata... instead of focusing on what the users actually want.
--
Break the rules. Keep the faith. Fight for love.
Acording to Apples web site, the ipod shuffle holds 240 songs. That means there are 240! orders they could be played in. That's about 4x10^468 acording to my calculator. If we start with a 64 bit random seed, that's only about 2x10^19. Since the random seed will determine the playing order, there is no way the shuffle can be random with out some form of entropy generator which would cost too much to be worth while.
That said, most of what people claim as evidence that it's not random are really just coincidences.
Since you can hear the same song again, it's actually 1* 20/400.
In Soviet Russia... iPod shuffles YOU!
(Sorry.)
-David
The iPod is programmed with software that prevents the user from making irrational decisions, for example the iPod knows a decisions like "Should I delete Weird Al's Polkamon?" is an irrational decision. This helps the user focus on more useful activities like practicing dance moves for Apple commercials and music video auditions. Makers of The Zune plan on taking this technology to another level with blue-screen-saver's. When attempting to delete tracks like Weird Al's Polkamon from the Zune, the user will be greeted with a vibrant blue-screen followed by error codes and a dialog for updates. These are not true error codes, this is the Zune prompting the user to rethink her decision to erase Weird Al's Polkamon, blue-screening will be an inherent function of the zune that will change the way we think about mp3 players.
"Say, Bill. Would you rub some of this powder on my lips?"
> There is no randomness in computers.
I don't know what's available on the iPod hardware, but in a desktop scenario, I generally multiply the pseudorandom number from whatever standard rand function is available by the hundredth-of-a-second value off the system clock and then use modular arithmetic to get the result in the range I need. This is almost certainly not cryptographically sound, but it is good enough to seem unpredictable to most humans, so it's a good start for things like shuffling a playlist or pulling questions from a test bank.
You generally still want to fudge the results in various ways to make them appear more "random" to the user, though. It's easy to seem unpredictable, but you have to work a bit more to achieve "random". (Note here that I am using "random" here in the sense end users generally understand it, *not* in the math/programmer jargon sense. Even when I do various fudgings to make the results seem more "random", I still avoid using the _word_ "random", because that word sets the users' expectations pretty high, and I'm usually not that confident in my fudging algorithms.
Cut that out, or I will ship you to Norilsk in a box.
It may be superstitious, but sure is a freaky thing to experience. When I worked late nights at the office, I would walk a secluded path to the parking lot that was a bit out of the way (but the end of the path was closer to my car). There was a particular light that often turned off when I walked by it. I always just chocked it up to the fact that I walk out of the office usually at the same time, but even a sensible person like me gets surprised a few times by something like that. :)
I can't help but wonder, though, if my presence didn't have some sort of effect (for example, my footsteps causing a harmonic vibration to reach the light mechanism, which caused it to reach a state of change sooner than it would have all by itself). The reason why I think that is because I can't recall ever seeing the light go off before I arrived at it. Maybe I did and my brain rejected the memory of it. o_O
Are you sure you don't have that same toon in your iPod more than once? That alone would explain more often replays.
"It's the height of ridiculousness to say for those 9 lines you get hundreds of millions."
True randomness is the stuff of magic to humanity. Randomness has inspired just about everything "mystical" or magic in our understanding of the universe. Look at the discussions raging about creationism vs. evolution. People believe God created the universe becasue humans can't wrap their minds around the idea that in a random univers, all things have a chance of happening. What we want from our randomness isn't "all things", its one-each of everything. The one-each randomness is what human s truly believe, deep, deep down inside, to be random.
Steve Levy should know better, but I can't think of a better way to illustrate the differences between real randomness and the human perception of randomness.
Steve, its random already.
Here will be an old abusing of God's patience and the king's English.
Apple is peopled by Human Interface idiots. It would be easy to add a "Shuffle" as well as a "Random" (especially to something NAMED SHUFFLE).
Shuffle means "shuffle the entire list of songs and play them in that order".
Random means "pick a random song, from the entire pool, and play that, everytime a song is played".
You can get an annoying Madonna song 5 times in a row with random, but once and only once per list with shuffle.
Shuffle is almost always better with any list shorter than [some amount to be determined].
Fucking iTunes will shuffle, but I don't believe it does random (well, party-fucking-"shuffle" randomizes), the fucking iPod will randmonize but it won't fucking shuffle.
Really, Steve "I invented everything at Apple" Jobs doesn't care about this, nor about you, nor about anything other than his ego. So I offer a reward to anyone who can tie this to his ego so something gets done.
And if someone could explain this to Steven "I'm myoptically focusing on one part of the picture to exclusion of all else" Levy I'll be grateful. No reward, but I'll be grateful.
(without freedom)
my password really is 'stinkypants'
I'd much rather have an iPod that supported smart playlists fully than one that played videos.
...
:(
I have a set of playlists based on ratings and a few arbitrary decisions about how often I wantto hear certain genres.
best: rating is 5 stars or rating is 4 stars and genre is not like "classical".
good: rating is 4 stars or playlist is "weird but good".
Then I have a second set of playlists like this:
day: Playlist is "best" and last play time is not within the past 24 hours.
twoday: Playlist is "good" and last play time is not within the past 48 hours.
week: Playlist is "weekly" and last play time is not within the last 7 days.
fortnight: Playlist is "fortnightly" and last play time is not within the last 14 days.
month: Playlist is "monthly" and last play time is not within the last 30 days.
Then a meta-list:
Playlist is "day" or playlist is "twoday" or playlist is "week" or playlist is "fortnight" or playlist is "month".
THAT list goes into party shuffle.
That produces a mix that's rarely predictable enough to be boring but still predictable enough to be comfortable. But the iPod can't handle it.
I'm guessing the LTBSD was hey... nineteen?
THANK YOU! I know a lot of geeks who understand that text book random could have to 2 values that are ordered next to one another, can also be next to another in a "random" sampling. What they can not fathom in their wildest dreams is that, that kind of random is not what people are looking for. They'll say, "But if you don't let the values be next to each other then it's not really random." And I have to keep reiterating that it doesn't matter. I don't want to hear 3 Crystal Method songs back to back(even when they are in different albums) when I have 500 hundred songs from other artists that I want to hear.
(Which is almost on topic)
Back before I was using iTunes, I had a combination of a somewhat odd directory structure and a simple Perl script controlling mpg123 to handle my "random" shuffle.
The way it worked was this:
A command line option specified the depth of the repeat stack. The repeat stack was where the file names of the last x played songs were stored (FIFO). The repeat stack was used to prevent the same file name from being played more than once every x songs.
Another command line option specified the starting place in the directory structure.
When run, the script looked in the the starting folder and built a list of any '.mp(2|3)' files *and* any subdirectories then choose randomly from the list. If it selected a directory it would move to that directory and choose again, if it selected a song file it would pass it to mpg123, put the file name in the repeat stack and then start the main loop again (going back to the starting directory). If it selected a song file whose file name was on the repeat stack, it skipped the song and went back to the main loop. If a directory had no song files it would build a list of subdirectories and then choose one of them -- if it ran itself into a dead end, it would go back to the top of the loop and the main directory.
So, yes, if the repeat stack was larger than the number of available song files, it would eventually just sit there and do nothing.
To 'organize' my song file collection what I did was have my mp3s mostly in directories and subdirectories by genre, group, album but with some directories arranged more haphazardly and containing mixes of song files and sub directories.
Obviously, the more buried the song file was in the directory structure, the less likely it would play. So some of the stuff I had collected, but didn't always want to hear, so I would bury those files. If I was in the mood for a specific genre I'd just start at that genre's place in the directory structure. If I was going to let it run for a long time and didn't want to hear the same song twice, I'd just make the repeat stack really long.
It worked well, and I was experimenting with the option of a true random walk (i.e. don't return to the top of the directory structure each time and allow it to choose to go "up" in the directory structure as well as down) and the announcing of the song by taking the text of the tags and creating a mp3 file and playing it (or at random intervals saying "This W C R A Z Y, the time is X XX").
But then they released iTunes, and I got fat and lazy.
-- I browse at +5 with stripped sigs
I have no clue as to why,but for whatever reason all the mp3 players I've seen tend to prefer certain songs. Maybe it's the randomising algorithm, Maybe it's just a quirk.As long as my Sandisk keeps holding up as good as it has I'll be happy to let it have it's likes and dislikes.
ACs don't waste your time replying, your posts are never seen by me.
Let's not forget that microprocessors aren't capable of generating true randomness. An external source of entropy is required, such as a Lava Lamp, decaying Cesium-137 atoms, or something else..
Now, I gotta say it would be pretty cool if these iPods use a vibration-sensing mechanism to gather enough entropy to seed the PRNG!
--Weasel
[BEGIN PGP PUBLIC KEY]: X5O!P%@AP[4\PZX54(P^)7CC)7}$EICAR-STANDARD-ANTIVI
Except for mathematicians and programmers, most think of "random" in a *very* different way from its technical definition. To most humans, saying that a particular sequence is "random" means *guaranteeing* certain things about it. Among them: the same element does not occur back-to-back, EVER, even if there are only a few elements total to choose from.
With this sort of definition, you will probably find that "random" means something different to just about everybody. It sounds like what you want in music sequencing is something that is a little more equally distributed-- that is, all the artists are equally distributed over the total playing time of the entire shuffle sequence. Or maybe what you really want is to have all the "moods" equally distributed. Or tempo, or genre, or... To achieve that sort of thing you probably want some kind of permutation and not a complete dice-roll, with just a dice-roll to get you to the first choice in the permutation so sequential runs won't get you the exact same sequence... Still, people will then complain because every time they hear Dylan song A it's always followed by Britney Spears' song B, as that's something that they'll notice before long. That suggests a more complex form of permutation combined with a dice-roll is in order.
Randomness or apparent-randomness is not as simple as many people think. I've been programming computers for over 20 years now, and I've seen a lot of code written by programmers who clearly didn't understand randomness very well-- in the "early days" you'd be amazed at the really bad random number generators that were built into many languages. It wasn't uncommon to just walk a pointer through all of memory and pick up whatever bytes were found as a random number source, which produces really abysmal results. Also, it can make a surprising difference in the quality of the results of programs dependent on them. There's a lot of non-intuitive things that come into play almost immediately when you delve into the subject. (see This Book for some juicy ones). And while I don't consider myself an expert on the subject I have gained a respect for the complexity of seemingly simple problems in randomness-- and avoid inventing ad-hoc randomizing solutions.
One example of subtlety is the easy-to-implement "swap" method of mentioned in related threads above doesn't really give you very good shuffling if you only pass over the deck once, and gets worse the shorter the card deck is. I've always preferred sorting an array of random data along with the indices and utilizing the indicies as the shuffled deck, though that too could have some non-obvious quirks-- certainly it's very dependent on the quality of the random number generator being used.
With regards to music shuffling-- I don't think it's a very good idea in the first place, as you'll always be likely to set tunes with completely different moods against each other which will be likely to grate even if you like both tunes. It's better to take the time and manually make some good playlists, IMHO-- they don't pay radio station program directors the big bucks for nothing...
> I think the lesson here is that you should never use a mathematically
> random algorithm for [a]esthetic purposes.
You can use a standard rand() function as a starting point, but you have to fudge the results considerably. Also, using rand() is not the only way to achieve random. I have a program that generates twenty-question quizzes from a large bank of questions. For some things (e.g., selecting which questions to use) it uses rand() as a starting point and then fudges in various ways (e.g., tossing out questions whose answer references the same place in the source text as a question that has already been selected), but once it has selected which twenty questions it's going to use, it goes into decidedly-not-mathematically random mode to determine how to order them. Basically, it extracts the four special questions (two finish questions and two reference questions) and places one of them in each of four five-question quarter-quizzes, then it *sorts* the normal questions according to the order of their references and puts the first one in the first quarter, the second in the second quarter, and so on (wrapping around to the first quarter after the fourth). It does then randomorder each quarter and subsequently does a little more fudging (e.g., to make sure that the first and last questions of the quiz are not special questions)...
> For my "scramble_case" I weighted the probability against getting a
> capitalized first letter, because that looks Too Normal.
Yes, that's exactly the sort of thing I'm talking about. If half or more of the words started with a capital letter, the scrambling wouldn't look "random" enough. It's a real paradigm shift from the meaning of random in computer science 101. It's the same reason my quiz generator must guarantee that the first question and the last question of the quiz are neither one ever a finish nor a reference question, even though with a mathematically random arrangement they (each) would be about a fifth of the time, over the long term. But if it happens even once, people complain that the quizzes aren't random and we should get a computer to do them. Most of the quizzers assume they're arranged by hand, but that would be a lot of work. However, even with all of the fudging I have built into the program, I still have to look over the quizzes before each rally and check that nothing went wrong and the quizzes are "random" enough. It just wouldn't do to have e.g. five "Why" questions in a row or some such oddity.
Cut that out, or I will ship you to Norilsk in a box.
I think the birthday paradox has something to do with it, but there still seems to be something wrong with the "random play" mode of iTunes. And random mode on my Squeezebox is even worse. With iTunes, I was able to help things by pushing the "smart shuffle" preference toward the "less likely" end. I have the smart shuffle preference one notch away from the very end of "less likely", and something still seems to be wrong. Supposedly, this setting would make iTunes play almost everything in my song library before repeating a tune. Doesn't seem to be working that way. I still hear a lot of repeats. And I have over a half-dozen songs (out of 176) that seem to never get played by the random shuffle selector, e.g. "Strip My Mind" by the Red Hot Chili Peppers. I've been trying to discern a pattern for what is happening. I don't have any compelling statistical proof, but the following seems to be the case: * When a particular song is playing, iTunes seems to be biased toward songs that are nearby on the play list (in alphabetical order). I have all 176 songs in a single large "album", and I get the impression that when it shuffles to the next song it seems to take small jumps around the alphabet based on the first character of the song name. * When you play the shuffle with a particular song, it seems as if the identity of that song sets a randseed. There seems to be not only repeating songs, but also to some extend repeating song patterns. None of the above is scientific, and I could be totally wrong. As some other posts have pointed out, it seems like an algorithm that shuffles the songs like a deck or cards and then deals the entire deck works better. Using quasi-random algorithms rather than pseudo-random algorithms might also be worth trying. I would also be interested in an algorithm that could look at the audio properties of two MP3 files and attempt to determine if playing the two songs in sequence would sound good (i.e. something that could avoid playing punk rock right after Yanni). Maybe something with a tuning knob what I could use to indicate my "mood".
I have about 50% CDs, 20% out of the iTunes store, and the other 30% from various free sources -- very, very few off of Napsterish sources, honest, but the rest from artist's web sites, free samples, other cuts from Salon or other places that give away a free song or two. And when I play music from my iPod, I never recognize a thing? It seems to me that most of the time, I'm surprised, even when I'm hearing something familiar. (Oh, listen to that guitar lick. That lyric says THAT?) That kind of thing.
No matter how random something is, the human brain will always try to find patterns in it, and if there isn't one, we make something up. Constellations, cloud formations, potato chips and grilled cheese sandwiches...
...but with that said, I have an iPod that loves 80's hair band metal.
There is a very big difference between the mathematical definition of randomness and what people mean when they say they want randomness. When you design a product and say it is random, then you're bbetter off trying to fit in with what people expect when they say random, rather than a strict mathematical definition.
Engineering is the art of compromise.
> Yes, but if it repeats that sequence every 6000 years, it really doesn't matter.
True, but...
> Modern pseudorandom alorithms are very good.
Not in this context. Pseudorandom algorithms approximate *mathematical* randomness.
> This completely ignores the issue at hand, which is that randomness is not
> what people want anyway.
Exactly. People think they want "random", but they don't want mathematical random. They want things carefully and thoroughly mixed up in an approximately _even_ distribution with no proximate repeats and no naturally-ordered subsequences of any significant length, among other things.
Cut that out, or I will ship you to Norilsk in a box.
I guess my Shuffle must be broke, then.
Because if all it does is shuffle the entire playlist why do I often get the same song seperated by just one or two other songs?
I'm not daft enough (yet) to put dupes on the shuffle, and it happens too regularly to be explained purely as the join between one instance of playing the playlist and another.
Good point, but if we're talking about the same problem (how many people do you have to get before it's likely that two share a birthday), then the number's only 23. Once you reach 40, then the probability of a birthday match is about 90%. More here.
Ceterum censeo subscriptionem esse delendam.
For several years, I have been using random/shuffle modes as my primary listening method. Things are probably getting better now, but for a long time random just didn't work well on most consumer devices. This was always something that I wished was covered better in reviews, but it's very time consuming to test correctly, and since it was rarely reviewed, I think most companies felt that as long as they had some sort of random mode, it didn't matter how good it was.
In one particularly memorable instance, I was listening to a CD on random in my car. This CD had about 30 tracks on it (a TMBG album), and when I played it on random, I consistently got two consecutive tracks in a row. e.g. it would play 4,5,18,19,10,11,7,8,12,13,etc. This only happened with this CD, but I think it showed an interesting failure mode of their random algorithm.
Computers don't make mistakes. What they do, they do on purpose.
The link a year ago was to my column in Newsweek. This current one is to a book excerpt (which references that column and the reaction to it, and the ensuing events) from my just-released tome, "The Perfect Thing: How the iPod Shuffles Commerce, Culture, and Coolness." Hope that clears it up, folks!