The music business is a cruel and shallow money trench, a long plastic hallway where thieves and pimps run free, and good men die like dogs. There’s also a negative side.
– Hunter S. Thompson
Category Archives: Music
Adding BPM data to mp3s
In my continuing quest to tag my music with metadata to create better playlists, I tried adding beats per minute (BPM) data into my mp3s. Seeing that I have more than 15k songs, I knew I was going to have to find a program that would do it automatically, so I looked around and chose beaTunes. Using the default mix of quality vs. speed, it took at least 36 hours of processing time to do my entire library.
While beaTunes other major feature is the ability to create mixes and song suggestions based on that BPM data, my point was to see if it would be useful in iTunes because that is my preferred application. Through some preliminary testing, I’m not finding that much good out of it. Creating a playlist strictly by BPM, I’m finding that a mix of incorrect BPM analysis (songs tagged at 80 BPM are really probably 160 BPM) and genre hopping (going from rap to rock to soul to metal) is causing the experience to not be as amazing as I had hoped.
There are times when a few songs really do match up well, but you would still really have to either a) know which songs have a wrong BPM and fix them, or b) really know your library enough to spot obvious mismatches and shuffle things accordingly. With over 15,000 songs I don’t think I can do either, so my options are to either find a better BPM tagger, or go back to not caring.
Any suggestions, or recommendations of other BPM taggers?
A fresh start
So, after three years of using Audioscrobbler, I decided (and acted) on a whim to reset my stats on my account at last.fm. It kind of hurts to see all that “work” go away, but it was to the point that it wasn’t moving fast enough for me (I check it a lot), and now that I have my music ratings and taggings in a better place than before, I felt it will be much more accurate than in the past. My biggest pet peeve, which Last.fm can’t really help much, is how artists with more songs recorded get higher rankings than artists who have fewer songs in existance but are more favored. Hopefully with the right ranking system and party shuffle, things should even out.
I was presently surprised to see at least 3 users in my “neighbours” list who used to be there before, when I had 20+ thousand songs in my profile; a fine testament to the last.fm system and algorithms.
Musical baton passing
Dan Hill passed me the
Musical Baton, so I feel obliged, even though I may have just been a sloppy second choice for him. Thanks, Dan.
Total volume of music on my computer
7,360 songs taking up 23.66GB of space. According to iTunes this covers 382 artists, and could last me 16.5 days of continuous listening. Everything is meticulously tagged through Musicbrainz, and rated by hand, so I know these statistics to be terribly accurate.
The last CD I bought
Jonathan Fire*Eater – Tremble Under Boom Lights, purchased used at Amazon. You can always see all my music, because I keep track of such this on the music page of my Web site.
Song playing right now
“You’re Right, I’m Wrong” by Thee Headcoatees, from the album Punk Girls. You can always see every song I’m playing, because I keep track of such things on Audioscrobbler.
Five songs I listen to alot, or that mean a lot to me
I could spend the rest of my life trying to figure out 5 songs that mean a lot to me, so instead I will take the easy, literal route and just look at my iTunes playcount and tell you them in order:
- Converge – Forsaken (album)
- Cursive – When Summer’s Over Will We Dream of Spring (album)
- Descendents – Catalina (album)
- Excuse 17 – Carson (album)
- Born Against – Sendero (album)
Like I said, not necessarily the greatest songs, but it’s all weighted through ratings and party shuffle and iTunes’ (seemingly non-random) randomness… although Catalina has a special place in my heart. (This song made we want to learn how to play guitar. I brought this song on tape with my acoustic guitar to my first day to my teacher (who by coincidence was Bob Mould’s first guitar teacher), and we ended up going through and writing out this punk song in dorky acoustic chord tablature.)
Five people to whom I’m passing the baton
I will (try to) pass it on to George Hotelling, Alf Eaton, Irdial Discs, juniorbonner, and Paul Hammond…all of which are in my Audioscrobbler network.
BPI Radar launched
I would like to interrupt your normal non-reading of this blog with a post, stating that the BPI Radar has been launched. Yes, the UK has it’s own version of the RIAA, and thus gets its own tool. It searches the UK version of Amazon and everything, so you get to see what those wacky Brits are listening to, as well as learn about musical genres that other places don’t have (or care about so much.)
An indie label milestone
There are now over 20 albums in the top 100 hot sellers on Amazon.com (that’s one-fifth!) that are not released by RIAA members. This is the highest point since I have been tracking it, which started in October 2003. (Here is my report page, if you want to follow along.)
Dear Lazyweb, please make my old iPod as cool as an iPod photo
Dear Lazyweb, please make an iPod hack that will display the album art on my older iPod, using ASCII characters. Thanks in advance.
Thoughts on playlists and iPod usage
Random personal notes on this “article” in the New York Times on how people use playlists and randomization:
- This entry is an excellent instant reaction to the piece. Why are such simple functions (shuffle, smart playlists) so unknown and mysterious?
- Mr. Angus (although somewhat fruity) is creating playlists for specific purposes. Good!
- Smart playlists will not solve the randomization or conspiracy theory problem. Smart playlists are based on completely objective data. The only way you could never have an “incorrect” song play is by providing/accessing metadata about uses and meanings of particular songs.
- Saving the uses and meanings of songs is, oddly enough, exactly the kind of stuff that I am doing with Mixmatcher. As people add songs to playlists, it is relating that song with some sort of subjective metadata. Over time, you will be able to get an understanding of what a song means, along with its possible uses based on the playlists it has been added to, without having to know the song beforehand.
- I really like the idea of TuneTags, which is basically creating adhoc metadata (ala del.icio.us) for songs. Let people tag them however they want to describe them, and let the aggregation sort out what is the most popular way of tagging. (“Kill them all and let a Norse God sort ’em out!”) I think the missing link with this idea is how to use this newfound data: What is the best way to make a mix for Mr. Angus bike workout? Create a smart playlists based on songs tagged with “upbeat” and “biking”? Are people really tagging things that way?
- I really like the possibilities of Mixmatcher or TuneTags for disovering new music. There are literally millions of songs that are published every week. There is no way to even begin picking the relevant (although not necessarily always “good”) songs out of that haystack. Even if only a small percentage of people tagging or sorting songs, I think it would still be providing enough data to make the songs useful to rest of the population. It would at least level the playing field, instead of 99% of those songs being obscure in a month.
- I think tagging ala del.icio.us might end up being easier and thus more popular, but that alone wouldn’t be giving people the final end product they desire. Even if you do give people another way to access the music, you still don’t have the things that mixtapes do well: structure and specific song selection, often done by a trusted individual.
- I don’t want to say it, but Audioscrobbler is another app that I don’t think fulfills its potential and/or does a very good job of its mission priorities (if I am to go by what their site tells me) of a) building my musical profile, b) matching me to people with similar tastes, and c) personalized music recommendations. Part A is being handled very well, but Part B and Part C are basically one section or page of each user profile. Does anyone know of any apps built using their data?
(This reminds me of one of my favorite jokes: Two women are eating at a restaurant. One says, “This food is terrible,” and the other woman says, “And such small portions!”)
This making sense to anybody?
Mixmatcher is faster now
I finally got a break from mudding and taping drywall, and I made Mixmatcher about a billion times faster. Turns out I only had it on 10, thus promptly turned it up to 11.
Netflix, Open up or die…
This feature from Engadget is nice, but I just wanted to nitpick, because the particular line I have an issue with is now being quoted elsewhere. The quote says:
Open up. Google did it, Amazon did, Apple did it, Netflix— expose your API so people out there can use www.netflix.com the way they want to, in new ways, in ways you haven’t imagined.
I don’t mean to break the euphoria, but although they may have developed a seemingly neat thing or two, they do not have an open API for the iPod or iTMS. (See also: iTunes Music Store wins a Webby when it’s not even a Web site)