January 31, 2014 at 3:59 pm

SoundClouders Love Commenting on ‘The Drop’ in Dubstep Songs

For those not yet familiar with the somewhat maligned dubstep genre, A) It consists of more than Skrillex (pictured above), B) Some of it’s actually pretty good, and C) The “drop” is the part of the song when everybody goes wild because the song just turned inside out and started making crazy bass noises.

SoundCloud, which aspires to be “the YouTube of audio,” includes a nice time-based commenting feature that lets people not only say what they think about a song, but when they thought it. When it comes to dubstep, one part of the song elicits more comments than any other: “the drop.” When that happens, it makes people comment. How do we know this?

It’s all due to the research of one Nicola Montecchio, who found a correlation between the drop and SoundCloud comments at a joint meeting of NEMISIG (North East Music Information Special Interest Group) and HAMR (Hacking Audio and Music Research) at Columbia University last weekend. (Montecchio is music retrieval scientist at The Echo Nest, publisher of Evolver.fm.)

Here’s a graph Montecchio made for Subkill’s “BIG Project VIP” – one of the tracks chosen because it has lots of comments. The blue line represents the concentration of comments on SoundCloud. The green line represents the songs’ loudness.

The key to reading these: Any time you see a deep “V” in the green line, that’s a “drop,” because dubstep gets quieter before trying to overwhelm the listener’s nervous system. As you can see, most of the comments happen right after the first drop:

dubstep the drop comments

Interestingly, the second drop elicits almost no bump in comments. If the drop is going to make someone leave a comment, it’s going to happen on the first one.

Here’s another dubstep song on SoundCloud with enough comments to analyze — Astronaut’s “Rain (DJ Swoon Remix).” Once again, the most comments happen right after the drop:

The same holds true for Creation & Subceptron’s “Pop The Trunk [clip]“:

“I did the same on approximately 100 dubstep tracks, and it turns out that comments are clearly clustered around drops,” explained Montecchio.

(Top image courtesy of Flickr/NRK P3; graphs courtesy of Nicola Montecchio)