Polyphonic HMI (Hit song science)

[aquanaut]

Hit Song Science http://www.polyphonichmi.com/index.html

For major and independent music labels
Why do music labels use HSS and what do they use it for?
When music labels are young they don't have much to lose so they take a lot of risks. That's why almost every kind of new and fresh sound and innovative style are first broken by independent labels. As the labels grow or get acquired by the majors they start to have more to worry about on the business side. They have share holders who expect them to make a profit, they have more infrastructure with people working who have families to feed etc. So they don't take nearly as many risks and they tend to stick with what has been working for them. This leads to two major complaints from consumers:
1. All the music on the radio sounds the same.
2. I just bought this CD and I only like two songs.
and a major complaint from unsigned bands:
1. I have a new and fresh style that sounds and feels innovative but I can't get signed.
We help labels address all of these issues by helping them look at their music and their market in new ways. We help them feel more confident in signing and releasing music that may not necessarily sound formulaic because we're able to recognize underlying patterns in the music that can determine if new and fresh sounding music will be enjoyed and purchased by a wide enough audience to make the investment worth while. We help them raise the bar for the quality standard of the music that makes in onto a given CD and ensure that even though not ever song on a CD has to have hit potential it should ideally be enjoyed by those who enjoyed the hit song, that is to say the other songs should have consistent and related mathematical patterns. We help them to see when promoting certain music is not likely to give them a return on their investment and we all know that every label releases more misses than hits. We help the labels avoid wasting money so that their saved money can be invested in developing new, better and more innovative artists and end up on the bottom line.
Is this the end of the line for A&R Executives? Will it now be up to computers to judge music and determine what can be culturally successful?
Absolutely not. This technology is nothing more than a tool to help seasoned and experienced label executives do their jobs better. When we say that our technology is to music what x-rays are to medicine we mean it in more than just to emphasize that the technology is revolutionary for the industry. An x-ray, (or almost any medical lab test) needs a good doctor to interpret it, to know what it is saying and to make a correct diagnosis. It takes a trained ear and someone in touch with the market and society to make educated guesses about what can become a hit. HSS simply adds some scientific and objective data to the mix. The tool will never be used to make decisions on it's own and will never take the place of golden ears and a gut feeling.
Aren't music labels supposed to be looking for new and fresh sounds that would not be anything like past hit music?
Absolutely. Historically, what is pleasing to the human ear has not changed since man began writing music. What has changed are styles, performances, the instruments used and the way music is produced and recorded, but a compelling melody is still compelling and a series of random notes still sounds random to us. That is not to say everything has been invented, however so far, every new style of music that has come into being: country, rock, punk, grunge etc. have all had similar mathematical patterns and the hits in those genres have all come from the same hit clusters that exist today and anything that has fallen outside of those clusters has rarely been successfully on the charts for it's musical qualities. For an explanation of a hit cluster please see our technology explanation.
Isn't music supposed to be art?
Yes, and artistic integrity and creativity are the lifeblood of the music industry and are of paramount importance to our business. Literature writing and movie making are also forms of art, however, the art of story development follows certain rules. Georges Polti, 19th century French author, identified 36 possible plots in literature that would hold the attention of the audience with the right amount of tension and subsequent resolution. Those rules existed before Polti identified them but by identifying them he helped writers become creative within guidelines that would help them connect with their audiences. Every story that is writen, every movie made, from the blockbusters to the intellectual films generally follow these rules of plot. Even the weekly sitcoms follow these rules of plot because they are the only ones that hold our attention as humans.
Polyphonic HMI has identified parameters in music in much the same way that Polti indentified parameters in storylines and since music is more complicated than a storyline it takes pattern recognizing computers to do it. Polyphonic did not invent the patterns, rather they existed anyway. Hopefully by identifying them musicians can become better composers and more insightful and music labels can be left scratching their heads less often when a song doesn't perform in the market the way they expected it to.
Computers have no place in the making of music.
Our computers cannot create music, they can only analyze it. Our computers have not invented anything, rather they've only detected patterns and parameters that already existed. Just as the telescope did not create the universe, it only allowed us to see it in a different way and to make better judgements and insights about it, and just like the telescope has its limitations so does our technology. We've encountered an underlying mathematical "alikeness" between songs. It's real and based in science. There are all sorts of ways that this can be valuable and interesting for the music industry and the listening public. By revealing some before-unseen scientific information about music we can use that to better understand the art and man's desire to be engulphed in compelling sound.
Leave it to the music industry to be so desperate as to use this service.
Our customers are the music labels but the beneficiaries of our work include the people who create, perform and consume music. The more healthy the industry is the more opportnities there are for everyone who hopes to make a career in music. No matter what music artists make they'd like to at least be able to support themselves and have their music enjoyed by a certain audience, and we all know how many are out there wanting to be tomorrow's superstars. Our technology does not take into high consideration the actual "acustic sound" of music. That is up to the A&R people, but it does take into consideration the underlying mathematical patterns that humans have found enjoyable throughout time. There's not just one or two hit song formulea. There are many and no matter what a song sounds like it may have many attributes that will make it enjoyable to a large enough audience to merit investment and promotion.
The music industry has taken a beating in the past few years. We are looking for solutions to address the concerns and criticisms we've been subject to. Many music label executives have lost their jobs and will never get them back in an industry that is shrinking. Today's executive knows we cannot continue with business-as-usual and still keep our jobs. We are demonstrating that Hit Song Science is a very good tool. Take some time to understand it. Experiment with it and see the results for yourself. We think our technology could be a wonderful and real revolution for everyone from the labels to the garage band.
Doesn't this technology just ensure that all music will end up sounding the same? How can you help labels feel confident in promoting new-sounding music?
Absolutely not. We are acutely aware that due to the fact that we compare new music to past hits it seems like we are looking for the same sorts of sounds. While we CAN do that it is not an inherent trait of our technology. We predicted the success of Norah Jones at a time when nothing in our past (five year) hit database sounded anything like her.
Patterns in music that are pleasing to the human ear have not changed much, if at all, since the times of the classical composers. For example, the dictionary describes melody as a series of notes strung together in a meaningful sequence. Why are some sequences meaningful to us and make up a beautiful song and other sequences just sound like random notes? While the number of possible melody patterns combined with all of the other variables in recorded music allow for a seemingly infinite number of combinations. The patterns that we find pleasing haven't changed a lot. What have changed are the instruments used, the styles, the way the music is performed etc.
Our technology can actually help music labels gain confidence in promoting music that does not sound formulaic because we can show you that the which of the music in question will be successful and which of that music will show you a return on your investment. Keep in mind that most labels liked Norah Jones' music but did not feel they would make any money promoting it. Our technology can help during those kinds of evaluations.
Will your computers be spitting out digi super-songs that mesmerize children and steal their souls?
Our technology cannot create music. It is only capable of looking at music that historically we (the collective masses) have said is good and compare the patterns in it to those in new music. For this application we are using a database of past chart hits from either the US or the UK. We also use a database of what we call classic hit from 1956 to the present time. The technology is capable of recognizing the commercial value in catchy pop songs as well as in never-heard-before new sounds, since what we tend to find enjoyable in them are timeless patterns that may be genetically or culturally engrained.
What if, as a label, we're not looking for Billboard Hot 100 hits but rather have a more urban or other genre focus?
We can and do create custom databases for music labels. For example we can meet with an A&R team or a single A&R professional and create a database filled with past and recent hits that are the kinds of song with the kind of success your label is looking for. By doing this we would be able to detect which new music is most similar and therefore most likely to emulate the same performance.
How does the technology work?
We use artificial intelligence applications as well as other methods to analyze the underlying mathematical patterns in music. Our technology does something called spectral deconvolution which is a fancy way of saying that we can isolate and separate many musical events that occur in a song. Some of these event are patterns in melody, harmony, chord progression, brilliance, fullness of sound, beat, tempo, rhythm, octave, and pitch. We then compare the patterns in new music to patterns in recent chart hits and to patterns in classic hit songs going back to 1956. By doing this combined with other mathematical calculations we've been able to develop a highly accurate and scientific tool. While this may sound like science fiction and that the results can't possibly be meaningful please keep in mind that we are already working for major labels on both side of the Atlantic who trust and use our results to make decisions and who understand many of the points made in this document. Please see our explanation of the technology for more information on how it all works.
Will do your reports tell the labels?
Music labels get an indication of the hit potential of a song in the current market as well as a lot of in-depth information and analyses on the music. In our reports you can see a lot of nuances and gain insight into some of the music that will helps you make business decisions and evaluations that can then help you determine release order, promotion investment, market timing and where the breaking line is for you to see a return on your investment.
What is considered a good score?
We use a logarithmic scale where the distance between 4 and 5 is half the distance as that between 5 and 6 and that distance in turn is half that between 6 and 7 etc. (illustration below)
We tend to say that a score of 7.00 and higher has solid hit potential and scores below 4.00 do not. There are some instances in which a low score (below 4.0) can actually mean there is solid hit potential but to determine that there are other variables in the report such as recentness and market tendencies which we clearly show. While given the right kind of promotion and enough public repetition is can be argued that any song can become a hit of some sort, we help labels identify those songs that will rise in the charts and sell in the stores with less resistance than other songs. So generally speaking you want to shoot for a score over 7.00.
BUT there is a major caveat: There are three factors to making a hit song:
1. The song must be good from an A&R perspective. That is it must sound like a hit song to human ears.
2. It must have optimal mathematical patterns. (that's where this service comes in).
3. It must be promoted well and with an appropriate artist.
So even if a song gets an optimal score from us it does NOT mean that it is a great song. But with a high score you have one of the three requirements for a hit song. The rest is up to you.
What do big hits typically score?
As stated above we tend to use 7.00 and higher as a score for a hit song because that's where they tend to score. There have been hits that score a little lower but the promotion has tended to be more aggressive. Some big hits score very low on the HSS scale but more than make up for that low score in other aspects of analyses that a label can do on your music with us if you happen to be negotiating a deal.
Does a high score mean that a song is better and a low score mean that a song is worse?
No. A high score means that a song is mathematically similar to recent hit songs and a low score means it is dissimilar. These scores have meaning when it comes to success potential in today's market but is not meant to mean a song is good or bad. For example, when tested for today's market some really great classic hits from the 60's 70's and 80's score very low and would most likely not become hits today with their original production or chord progression. That does not mean that they are not good songs and it is quite possible that if produced more in line with today's sounds they could score much higher.
Who is Polyphonic HMI?
Founded in 2002, Polyphonic HMI is a wholly owned subsidiary of Grupo AIA, a leading artificial intelligence developer for business since 1988. HSS is a proprietary, artificial intelligence driven, music analysis system that can be used to predict hit probability of new music. Meet the management team.
Everyone right down to our tech leaders are music people. The technology was developed over a period of years by our parent company, Grupo AIA (www.aia.es) by scientists who have PhDs in sciences like mathematics, physics, acoustical engineering and other mind-boggling subjects. Polyphonic HMI has its own team of scientists but who's first love is music. We are the music industry taking control of the technology and putting it to use in increasing the quality of the music being released. That is something everyone can agree upon as being a good thing.
Can't any song become a hit with enough marketing?
We predict the hit potential of the song based on its own music attributes and merits. In theory, songs with a high hit potential are pieces that have a greater likelihood of "making it" with fewer promotional efforts and resources.
What about songs that are hits based on their lyrical content such as hip hop or The Angry American by Toby Keith?
Our analysis program does not take into account the lyrical content of songs. So yes, sometimes the lyrics can be the most powerful thing in a song and propel it up the charts. However good lyrics coupled with music that has optimal mathematical properties for today's market will have it much easier.
Are songs with a very low HSS Scale rating bad songs? Should they be taken off the CD?
No. Just because they do not rate highly with the HSS and the hit database does not mean they should be taken off the CD. Nor does this mean that they will not be enjoyed by the people who purchase the CD. That's why it is so important for this information to be in the hands of a seasoned A&R professional. The 'human' criteria and feel that is needed when creating a CD cannot be substituted by artificial intelligence however, it can be enhanced and that is what we do.
What is an affinity value?
The affinity value is a rating that shows how closely related the mathematical patterns in one song are to another. The lower the affinity value between two songs the more closely related they are. When the same recording of a song is compared against itself the affinity value is zero. The algorithm takes into account all aspects of a song including beat, melody, harmonic variation, tempo, pitch, octave, fullness of sound and beat. This means that two different recordings of the same song (a song from a studio and a live version) will not have an affinity value of zero but normally have close affinity values.
You can see how closely related your song is to past hits by looking at the affinity value. The lower the affinity value the more closely related the song is. An affinity value under 1.00 denotes close similarity while an affinity value of over 1.00 denotes similarity but less obvious and more distant similarity.
How do you come up with the HSS ratings?
This is done through a series of mathematical calculations based on the number of past hit songs related to the song being tested. The greater the number of past hits with lower affinity values, the higher the hit potential of the new song. A rating of 7.00 or greater should be considered to be a very strong track and very closely related to what is currently popular in the market.
What else does this technology do?
Polyphonic HMI will also be using the technology at the production level of music creation. By teaming up with veteran artists and producers the application adds to the creative process by isolating the important attributes of the music which will allow new sounds and styles to flourish. We will be working with many veteran and Grammy winning producers. If you are a producer and would like more information about this please email us.

Copyright © 2002, 2003, 2004 Polyphonic HMI S.L.

Heard also that Norah Jones was signed after being analysed by that software

http://www.polyphonichmi.com/index.html

[mosc]

Hey, I have a program that will tell you which stocks to buy and sell. Artifical intellegence, of course.