Pondering The Future Of Automated Rock With The Men Who Make It Happen

Photo by tantek courtesy of flickr Creative Commons.

Photo by tantek courtesy of flickr Creative Commons.

I really enjoyed my trip to C2SV in San Jose a couple weeks ago. I got to watch amazing sets by the Lemonheads, Thee Oh Sees and the Stooges, check out the Dirty Ghosts, Talky Tinas and Curious Quail who were all new to me, play a set there myself, and listen to Jello Biafra give his views on the future of music distribution, at a panel with the guys involved in some of the more recognizable online music services. It was while waiting for the latter event to get started that I had one of the weirder conversations I’ve ever had about what rock and roll is all about.

Because C2SV is part music festival and part technology convention, there’s a different crowd in addition to the one you usually at rock festivals, one that gets very buzzy about emerging capabilities that can be capitalized. Everybody with a big idea is there trying to sell it, hoping the guy checking out their booth is a venture capitalist with coins to spare.

There were a handful of booths in the convention center where the panels were being held, and the only one that looked remotely interesting had a banner over it reading Beyond The Realms Of Metal, with a pretty metal-looking dude hanging around underneath it. I needed to kill a few minutes so I wandered over. He was busy with his laptop, not seeming too interested whether anybody was checking out his booth or not. At this point I noticed a sign on his table for something called HITLAB and a stack of flyers advertising a contest for new bands in which the winner would to “Go To The Grammys.”

My mind struggled to make a connection between these various pieces of information. Why would any self respecting metal band want to pledge allegiance to the Grammys, an organization whose idea of a good metal band is Jethro Tull?

But I asked the guy, so, are you HITLAB or Beyond The Realms Of Metal or are you gonna help me Go The Grammys or what. He perked immediately, and answered “Yes!’ with a big smile. “Yes to all of the above!”

Now, at this point I should pause to say, I wish I had this conversation on tape, but I don’t, and none of the following “quotes” should be taken as literal. I’m attempting to paraphrase all of this from my best recollection of how it went down. There was more than one point where I was having trouble following the guy, because it was a Silicon Valley tech convention and at times you could hear them speak a different language.

First we established that Beyond The Realms Of Metal was the name of a radio show he had on a nearby college station, so, OK. That explains that.

Next, the Grammy thing is a battle of the bands contest where emerging bands can play at events during Grammy week and showcase their talents for industry reps in LA, so, fine, I understand that too. We do that sort of thing all the time around here.

But then came the interesting part.

HITLAB, he answered, is a community of artists and musicians, working with UNPRECEDENTED TECHNOLOGY to open the horizons for all kinds of music to flourish and find new audiences, who could then pay to download the band’s music on the spot. It sounded like a combination music discovery service, marketplace and PR agent.

“Like a CDbaby or Bandcamp kind of thing? What’s the unprecedented technology?” I wanted to know.

And at this point he told me about DHS – Dynamic Hit Scoring.

I can’t begin to paraphrase the entirety of what I heard next, so I’ll defer to the description given on the company’s own FAQ page:

How does it work?
The program takes six years’ worth of Billboard Hot 100 singles and compares the uploaded song to those hits based on many factors. As a result, it can predict the probability of it becoming a hit and what previous hits it closely resembles.

Basically, the guy said, you can visually compare your sing to Lady Gaga’s current top ten hit and see how it measures up. If you get a bad score, go re-write the chorus and see if the score improves.

I then asked him, if I’m an extreme metal band, not that interested in the Top 40, why am I concerned about this? Why would I want to?

Everybody’s trying to find a way to get noticed, he said, but can’t be sure if what they’re doing is really up to snuff. You need to have some way to figure this out objectively. This is the UNPRECEDENTED part of it. It’s NEVER BEEN DONE BEFORE. The people who are in it now are on the CUTTING EDGE. This is the STATE OF THE ART.

I finally cut the guy off and said, “I hear what you’re saying… and I’m starting to throw up in my mouth a little. I don’t want to think that the future of music is going to be dictated by robots that identify patterns matching up to last year’s hit records.”

But you see, said Mr. Metalhead, this is going to INSPIRE creativity, because it’s never been done before, and it’s UNPRECEDENTED. By now his boss has come over, apparently having sniffed out a turd in the punchbowl. Yes, they are telling me, this is going to REPLACE the music industry as we know it, by giving people a chance at success. And the really great thing about it is that it’s NEVER BEEN DONE BEFORE.

“You’re removing all the magic from the most magical of the arts!”, I cry. Both are almost giddy to be faced with this level of rejection of their product, because it allows them to go into their long-form pitch, which rephrases everything we’ve just heard about unprecedented inspirational capability, with additional messages about how “the old system is broken and it’s never coming back”.

Now from my own experience, this is somewhat true at the core, but the only purpose of stating this truth is to give credence to the idea that we the people need bots to tell us what to listen to, and that artists need bots to tell them what to write.

Bots are already choosing some of our music. This topic came up in the panel discussion, in which one of the group noted that algorithms driving Last.fm and Pandora that predict and recommend other artists based on the ones you’ve liked in the past, had once suggested he listen to Air Supply based on the fact he liked Crowded House. Evidently, the two share a certain number of tempos and chord progressions that the machine picked up on. If that’s as precise as the machine can get, is it even worth turning on?

At present there seems to be no good bot predictor of artistic sensibility, outside of those companies that pay people – supposed experts in their field, like the very people on that panel – to come up with playlists and spreadsheets that say, if you like Pigmeat Markham, you may also enjoy Wynonie Harris. That human element, which Emusic tried to inject in its early days with contributed playlists and album recommendations from critics and artists, is still imperfect and tuned to the biases of whatever expert is doing the picking, but at least it feels authentic. Not surprisingly, the panel members agreed when I brought up this algorithm-based music selection during the Q&A and asked if it’s just old-fashioned of me to find the idea kind of sinister. What you need, they all agreed, is a Cool Music Buddy who does the research and knows enough stuff about music to point things out that you would never have found on your own. (I mean, of course! They’re certainly not about to disagree and say a robot could do their job.)

But if you want to find the good tunes in today’s ocean of online content, it’s going to be expensive to pay a human being to troll through it all and make recommendations. This is where the bots come in. They can analyze cross-buying patterns (people who bought this also bought this) and make suggestions based on that, but that won’t help an emerging artist who hasn’t had a lot of sales yet. And so, isn’t there some potential for a system that could analyze the waveforms of my favorite songs and pick out somebody I’ve never heard before that also makes songs with shifting tempos in minor keys with screaming vocals and badass drumming?

I presume this is the magical and inspiring part of the equation that the sellers were trying to communicate. I could reach users of this service, if my music matches the measurable sonic traits of what’s already popular. Or should I choose, I could alter my music to be more like a popular act like Mastodon, and make sure their pattern recognition software produced a match, so my tune would float to the top of the list of recommended songs every time someone listened to Mastodon. It’s kind of how Ebay sellers use combinations of words to maximize their ability to be found in searches, except applied to the content of the auction itself.

I can make music that sounds like the kind of music people are already buying… and this might lead to SUCCESS. What an UNPRECEDENTED idea!

I haven’t decided yet if I should pay $30 to have one of my songs analyzed, or instead sign up for the $19.99 monthly fee that, their FAQ page suggests, a lot of people sign up for and never use. One thing they didn’t tell me at the convention center is that this is something to do with Senegalese rapper Akon and that part of the benefit resulting from submitting your songs is the opportunity to have a record deal with Akon and sing with him. Now there’s something for an extreme metal band to aspire to.

You can read more about it in the ad copy that makes up their Wikipedia page:

HITLAB is a place for people who love music to discover new talents, share their treasured finds with their friends and decide who the next big thing will be! HITLAB is a music community. You can find new artists, listen to their songs, buy their music, tell your friends about your new favorite bands and help musicians possibly live off of their art?

I suspect that question mark at the end of “art” in unintentional, and yet, it says so much. 

This entry was posted in Music, Technology and tagged , . Bookmark the permalink.

Leave a Reply