A computer tells you just what you want him to say. Nature and physic still tell us how a song's spectrum has to be to result pleasant, music studies tell you what's the right chord sequence to generate a certain mood. What Hit Song Science do is to synthetize choices that smart producers and audio engineers used in past. But it isn't creative, it can be useful for the instant present, not for the future. Imagine a similar software done for paints in 1800. And imagine a young Picasso using it. Sorry for my bad english :]

I do not like...having anything Artistic evaluated in a synthetic or scientific way...whether it is Music- Art-
Literature-Films-Business Ideas-etc.
Too many songs do not make it through "Taxi" & its
evaluations by music industry experts...Too many good performers do not make it through American Idol...Too
many Songs are thrown away in Music contests.
I like the "Dark Horse" that no one can predict!!!

I can understand people's reservations and confusion.

NeZoomie -- the software DOES NOT create music, it only evaluates existing mp3s based on those algorithms.

"UFO JIM" -- this is about business, and business will gravitate towards anything that provides reliable metrics and valuable predictions. I think Bruce is definitely on to something, and Hit Song Science will continue to become an integral part of the changing face of the music biz.

I feel like I trust my own taste (and ears) more than any amount of hype or digital readouts, but I'm also used to the fact that I hear more and hear differently than a lot of people. HSS is a useful service and good equalizer, but I'm interested to see if anyone else is offering something similar. The only thing they need to fear is competition doing it better for less -- and it looks like a lot of the research behind these MIR engines is publicly available.

Bruce, this was a very succinct summation of a concept you've referenced before -- definitely useful and appreciated.

I never said that hms does music, but it judges it. Well a software CAN'T predict what we'll like in the future. Quite simple.

My advice to artists (and another point I strongly disagree with Bruce about): don't - whatever you do - concentrate on single-song sales!

The reason is simple: singles sell songs. What you want to do is sell the artist.

Compare: the number of successfuly charting singles with the number of long-life artists.

Does the phrase one-hit-wonder ring a bell?

The only way to build a performing career is to make potential fans aware of a large number of songs. That's how Led Zeppelin or Pink Floyd were able to build their careers without selling ANY singles.

Thanks for the compliment, Bruce.

First off: I think HSS is a useful tool. I certainly plan to drop them some cash. I plan to drop some cash by We Are Listening as well. However, I think you have to be skeptical about it. Computers don't listen to music any more than they buy it. What I fear is - in the current trend of technocratic voodoo-thinking - that labels are going to start churning out material based purely on HSS ratings. All I'm saying is: don't read too much into it.

Second, I really don't think that either Zeppelin or Floyd got into the studio based on the quality of the songs they were to record. Pink Floyd certainly made most of their music IN the studio. The point I was endeavouring to make is that an artist becomes big as much by virtue of the quantity as the quality of his or her output. You said it yourself - most of the people buying or sharing music do it for single songs. There are two ways of approaching the issue: either you cater to this trend and concentrate on selling one song (which works great for a certain type of artist and always has in the singles market) or you attempt to build an identity as an artist - this means quantity. In other words: if you have a choice of spending the same money on great production of one song or good production of several - I say go with the latter.

(And never underestimate the power of the underground aesthetic... :) )

The problem may be that I base my comments on the general impression I get from reading a large number of your posts. They have given me much food for thought and I hope to profit by reading what you have to say. Kudos to you there. However, the general strategic advice you give often grates against my good business sense. It may be that I am simply a fossil... ;)

Discussing HSS with some friends last night, I started wondering: are the algorithms of HSS a self-fulfilling prophecy? Given how much radio formats have homogenized in the past decade, has that simplified the question of "What Sounds Like a Hit" to the point it's essentially an automated process?

I was talking with an intern at Sony who told me he was surprised to discover there's only 3 genres of music -- Heavy, Urban, and Adult. Bruce Springsteen, the Dixie Chicks and Radiohead all fit into the same overall genre, and it's all mixed and mastered the same. From dance-rock to death-metal, everything is mixed and mastered the same.

Wether a hit song is from Nashville or the Hit Factory in NYC, it's usually major-key, upbeat, drum-heavy and shares the same overall dynamics, breaks and structure.

I really wonder if HSS represents a death spiral of engineered Catchy. As the formats grow more and more constricted, as the genres become more and more alike, this could become a cycle that feeds itself into redundancy and oblivion.

I think what will keep HSS "honest" is the unpredictable singles -- the absurdly lush soundscapes of Gnarls Barkley's "Crazy" or the downright atonal drone of Mims and 50 Cent beats -- which get fed into the HSS database.

This is all based on speculation, though...I think I need to sit down with the data and the charts to test my intuition that the top 40 is a much less diverse body of work these days than it used to be.

I greatly dislike this whole idea - a music judging system that self-proclaims its the future...it reminds me of a completely unconstructive Garageband.com contest (Garageband I should state - which in comparisson is fantastic, as it is a contest whereby artists can gain feedback and positions in the site's charts). But to try and claim this is where the music industry is going is awful. What about artists who do not instantly come across as great to listeners (or curators) - but then would after a few more listens. This is the case with so many FM chart hits. This system doesn't take circumstances like this into account. Besides that, I think the advice you give to artists is narrow-minded, the greatest songs are built on sincerity - not gathering as much tear-it-apart information as possible to restructure a better formula for the song. I think the process your referring to creates songs that may have a catchy melody, but will be unlikely to make classics.

You seem to be under the impression that you are in a position to offer this kind of advice. But anyone offering advice to artists who are really knowledgeable enough to do so, would know that great music comes from a different source in each different individual. Sometimes outside advice is a part of that, sometimes not - it's not for you to so sternly dictate it has to be, here on your site.

Some of the best artists I have ever heard have extremely DIY sounding recordings, but that is intended as part of their image and musical persona. It's impossible to do justice to music as an art form by applying your particular system of analysis to it - great artists and songs are simply too diverse for this kind of system to really allow for.

And how about completely new experimental genres that fall outside of the conventional framework that this system was built around? This would clearly not be a site for truly new, inspiring music, doing very different things to what's gone before, rhythmically, melodically, tonally. If The Celestial Funnel were built, relied upon and used centuries ago, I doubt we would have half of the intriguing new genres of music we now have.