Tuesday, June 22, 2010

SnoutCast #21: "The Trouble with Music Clues"

Because we can't stop trying new things: this week's episode was recorded more or less live, with minimal editing, and we limited our topic discussion to twenty-three minutes.


[ Download mp3 - 26 MB ]

00:00 - promo teaser: GC Summit 2010 videos now online!
01:05 - "audible"
02:49 - live(ish) from Portland
04:15 - the Shinteki Decathlon 6 music clue, generally
06:47 - burying the lede, QED
08:36 - the garden-variety music clue: identify song, use meta-data
10:14 - a clever variation from Wonka(?) which required close listening
12:46 - "the internal clue"
13:55 - the limitations of musical meta-data
14:28 - we'll just call this "the pants premise"
17:42 - the two-part music CD clue from Paparazzi(?)
20:10 - the DASH 1 music puzzle
21:31 - DeeAnn's European Television Adventure
22:15 - #w00tstock
23:32 - it's all about the music, man
26:00 - upcoming events: Shinteki SF Scramble, Puzzle Pursuit San Diego, Can't Stop The Serenity (nationwide)
27:27 - The End

Music: instrumentals from "Code Monkey" and "A Talk with George" by Jonathan Coulton

[ Subscribe to SnoutCast / iTunes link ]

CKL DeeAnn

3 comments:

Skott said...

I think the trouble with music clues is that there's no way to make the data extraction fun or interesting. Say, for example, you're trying to get the song title. There are four main ways of doing it:

- Know the song title already
- Use some sort of song ID software
- Google the lyrics
- Phone a friend

If you can do the first, what's the point? You may as well just have been given a list of songs. If you need to do the remaining three, then it's just busy work.

Shinteki tried to eliminate the song ID software possibility by using short clips, but all that did was make IDing songs even more difficult for non-Beatles fans/aficionados. This is pretty frustrating; I know Wei-Hwa specifically designed a clue as revenge for pop song ID puzzles (see http://www.gooooogol.com/puzzles/player-solution.html) .


And although there's limited data in a song, there's enough to cause confusion. For the first hunt Jonathan and I ever hosted, we had clips of earworm-style songs with the title cut out, the theory being that your mind would kind of automatically fill it in. Only one out of five teams solved it. I later showed it to Ian, who said he was going to have to make a lot of notes... song title, band, clip length, song length, beats per minute, words in the clip, etc. Kind of realized at that point how much of a pain a song clue could be.

The DASH 1 music puzzle was probably the best possible way to do it. Even if you didn't know the name of a song, it was usually in the lyrics and could identified using the word-length indicator. Song title identification wasn't the focus of the clue, thus the challenge of doing it was minimized and the focus was clearly shown to be something else, which was a pretty fun aha.

lahosken said...

"And although there's limited data in a song, there's enough to cause confusion."

Yeah. When I heard "too little metadata", I did a double-take.

Greg said...

What about puzzles that rely on spotting a commonality in the lyrics themselves? The first (and only so far) Game I played in had one where all of the clips contained a NATO word. It meant that actually listening to the music was an effective path to solving it, and they could give us the song titles to show us that we didn't need to identify. (Granted, we used this knowledge to search the lyrics and extract NATO letters from the full song lyrics. :)