Computational Music Analysis
7.0K views | +4 today
Follow
Computational Music Analysis
New technologies and research dedicated to the analysis of music using computers. Towards Music 2.0!
Your new post is loading...
Your new post is loading...
Scooped by Olivier Lartillot
Scoop.it!

'Game-powered machine learning' opens door to Google for music

'Game-powered machine learning' opens door to Google for music | Computational Music Analysis | Scoop.it

Can a computer be taught to automatically label every song on the Internet using sets of examples provided by unpaid music fans? University of California, San Diego engineers have found that the answer is yes, and the results are as accurate as using paid music experts to provide the examples, saving considerable time and money. In results published in the April 24 issue of the Proceedings of the National Academy of Sciences, the researchers report that their solution, called “game-powered machine learning,” would enable music lovers to search every song on the web well beyond popular hits, with a simple text search using key words like “funky” or “spooky electronica.”


Searching for specific multimedia content, including music, is a challenge because of the need to use text to search images, video and audio. The researchers, led by Gert Lanckriet, a professor of electrical engineering at the UC San Diego Jacobs School of Engineering, hope to create a text-based multimedia search engine that will make it far easier to access the explosion of multimedia content online. That’s because humans working round the clock labeling songs with descriptive text could never keep up with the volume of content being uploaded to the Internet.


In Lanckriet’s solution, computers study the examples of music that have been provided by the music fans and labeled in categories such as “romantic,” “jazz,” “saxophone,” or “happy.” The computer then analyzes waveforms of recorded songs in these categories looking for acoustic patterns common to each. It can then automatically label millions of songs by recognizing these patterns. Training computers in this way is referred to as machine learning. “Game-powered” refers to the millions of people who are already online that Lanckriet’s team is enticing to provide the sets of examples by labeling music through a Facebook-based online game called Herd It.

 

“This is a very promising mechanism to address large-scale music search in the future,” said Lanckriet, whose research earned him a spot on MIT Technology Review’s list of the world’s top young innovators in 2011.

 

Another significant finding in the paper is that the machine can use what it has learned to design new games that elicit the most effective training data from the humans in the loop. “The question is if you have only extracted a little bit of knowledge from people and you only have a rudimentary machine learning system, can the computer use that rudimentary version to determine the most effective next questions to ask the people?” said Lanckriet. “It’s like a baby. You teach it a little bit and the baby comes back and asks more questions.” For example, the machine may be great at recognizing the music patterns in rock music but struggle with jazz. In that case, it might ask for more examples of jazz music to study.


It’s the active feedback loop that combines human knowledge about music and the scalability of automated music tagging through machine learning that makes “Google for music” a real possibility. Although human knowledge about music is essential to the process, Lanckriet’s solution requires relatively little human effort to achieve great gains. Through the active feedback loop, the computer automatically creates new Herd It games to collect the specific human input it needs to most effectively improve the auto-tagging algorithms, said Lanckriet. The game goes well beyond the two primary methods of categorizing music used today: paying experts in music theory to analyze songs – the method used by Internet radio sites like Pandora – and collaborative filtering, which online book and music sellers now use to recommend products by comparing a buyer’s past purchases with those of people who made similar choices.


Both methods are effective up to a point. But paid music experts are expensive and can’t possibly keep up with the vast expanse of music available online. Pandora has just 900,000 songs in its catalog after 12 years in operation. Meanwhile, collaborative filtering only really works with books and music that are already popular and selling well.

 

Lanckriet foresees a time when – thanks to this massive database of cataloged music -- cell phone sensors will track the activities and moods of individual cell phone users and use that data to provide a personalized radio service – the kind that matches music to one’s activity and mood, without repeating the same songs over and over again.


“What I would like long-term is just one single radio station that starts in the morning and it adapts to you throughout the day. By that I mean the user doesn’t have to tell the system, “Hey, it’s afternoon now, I prefer to listen to hip hop in the afternoon. The system knows because it has learned the cell phone user’s preferences.”


This kind of personalized cell phone radio can only be made possible if the cell phone has a large database of accurately labeled songs from which to choose. That’s where efforts to develop a music search engine are ultimately heading. The first step is figuring out how to label all the music online well beyond the most popular hits. As Lanckriet’s team demonstrated in PNAS, game-powered machine learning is making that a real possibility.

more...
No comment yet.
Scooped by Olivier Lartillot
Scoop.it!

Talent Shortage Looms Over Big [Music] Data

Talent Shortage Looms Over Big [Music] Data | Computational Music Analysis | Scoop.it

“Big Data is the next big thing in technology, but where are the people who know how to make the most of it?” [And what about music industry?]

 

"Big Data": the ability to acquire, process and sort vast quantities of data in real time. Big Data refers to the idea that an enterprise can mine all the data it collects right across its operations to unlock golden nuggets of business intelligence. And whereas companies in the past have had to rely on sampling, Big Data, or so the promise goes, means you can use your entire corpus of digitized corporate knowledge. It is, by all accounts, the next big thing.

 

"A significant constraint on realizing value from Big Data will be a shortage of talent, particularly of people with deep expertise in statistics and machine learning, and the managers and analysts who know how to operate companies by using insights from Big Data." What the industry needs is a new type of person: the data scientist.

 

"Thirty years ago we didn't have computer-science departments; now every quality school on the planet has a CS department. Now nobody has a data-science department; in 30 years every school on the planet will have one."

 

a data scientist must have three key skills. "They can take a data set and model it mathematically and understand the math required to build those models; they can actually do that, which means they have the engineering skills…and finally they are someone who can find insights and tell stories from their data. That means asking the right questions, and that is usually the hardest piece."

 

It is this ability to turn data into information into action that presents the most challenges. It requires a deep understanding of the business to know the questions to ask. The problem that a lot of companies face is that they don't know what they don't know. The job of the data scientist isn't simply to uncover lost nuggets, but discover new ones and more importantly, turn them into actions. Providing ever-larger screeds of information doesn't help anyone.

 

One of the problems with Big Data is the fact that it has to deal with real data from the real world, which tends to be messy and difficult to represent. Conventional relational databases are excellent at handling stuff that comes in discreet packets, such as your social security number or a stock price. They are less useful when it comes to, say, the content of a phone call, a video, or an email. Out in the real world, most data is unstructured. Handling this sort of real, messy, scrappy data, isn't so simple.

 

"People have been doing data mining for years, but that was on the premise that the data was quite well behaved and lived in big relational databases. How do you deal with data sets that might be very ragged, unreliable, with missing data?"

more...
No comment yet.
Scooped by Olivier Lartillot
Scoop.it!

New Study Shows Which Cities Drive Music Listening Habits [Graphic] - hypebot

New Study Shows Which Cities Drive Music Listening Habits [Graphic] - hypebot | Computational Music Analysis | Scoop.it

“The social media website last.fm provides a detailed snapshot of what its users in hundreds of cities listen to each week. After suitably normalizing this data, we use it to test three hypotheses related to the geographic flow of music. The first is that although many of the most popular artists are listened to around the world, music preferences are closely related to nationality, language, and geographic location. We find support for this hypothesis, with a couple of minor, yet interesting, exceptions. Our second hypothesis is that some cities are consistently early adopters of new music (and early to snub stale music). To test this hypothesis, we adapt a method previously used to detect the leadership networks present in flocks of birds. We find empirical support for the claim that a similar leadership network exists among cities, and this finding is the main contribution of the paper. Finally, we test the hypothesis that large cities tend to be ahead of smaller cities-we find only weak support for this hypothesis.”

more...
No comment yet.
Scooped by Olivier Lartillot
Scoop.it!

Can The Echo Nest Save Music? | Neon Tommy

Can The Echo Nest Save Music? | Neon Tommy | Computational Music Analysis | Scoop.it

The Echo Nest might be one of the few companies to have figured out how to cash in on the music business. Only it has little to do with selling music. Instead, it gives developers the tools to create the next brilliant music discovery app.

 

Jim Lucchese, Echo Nest’s chief executive, believes on-demand apps are the future of the music industry. They let consumers discover their next favorite band with the push of a button. Apps, he says, “are the new Tower Records or the new college radio DJs.” He wants the two-year old company, which was launched by a couple of MIT PhDs, to be the glue that connects app developers with listeners.

 

Echo Nest’s application programming interface, or API, compiles enormous amounts of internet data on how people are talking about music and what songs and artists are popping up on the web. It automatically analyzes everything from the beats-per-minute of 17 million tracks, to which artists are trending across hipster music blogs, to which drummer is described as “funky” the most times. It uses that data to make recommendations.

The platform is the “special sauce” that makes new app ideas possible, said Echo Nest operations director Elissa Barrett.

Echo Nest’s API is like steroids in the hands of developers, allowing them to create apps that they never would have had the power to pull off on their own.

 

Echo Nest creators Brian Whitman and Tristan Jehan met in the PhD program at the MIT Media lab. The Echo Nest has been actively tracking data on the web since 2005 and has amassed acoustic data for over 10 million tracks. Pandora’s Music Genome Project has taken over 10 years to manually curate 1 million tracks.

The company works with 7,000 independent developers, as well as large companies like the BBC, MTV and MOG. So far, the company has produced 160 apps through its platform. The company would not disclose its revenue figures, though it received $7 million in second-round funding orchestrated by Matrix Partners in Oct. 2010.

more...
No comment yet.
Scooped by Olivier Lartillot
Scoop.it!

New Algorithm Captures What Pleases the Human Ear—and May Replace Human Instrument Tuners | 80beats | Discover Magazine

New Algorithm Captures What Pleases the Human Ear—and May Replace Human Instrument Tuners | 80beats | Discover Magazine | Computational Music Analysis | Scoop.it
Imprecision, it turns out, is embedded in our scales, instruments, and tuning system, so pros have to adjust each instrument by ear to make it sound its best. Electronic tuners can’t do this well because there has been no known way to calculate it. Basically, it’s an art, not a science. But now, a new algorithm published in arXiv claims to be just as good as a professional tuner.

 

The new study replaces the human ear’s ability to detect “pleasingness” with an algorithm that minimizes the Shannon entropy of the sound the instrument produces. (Shannon entropy is related to the randomness in a signal, like the waveform of a sound, and is unrelated to the entropy of matter and energy). Entropy is high when notes are out of tune, say the researchers, and it decreases as they get into tune. The algorithm applies small random changes to a note’s frequency until it finds the lowest level of entropy, which is the optimal frequency for it, say the researchers. And setting tuners to follow this algorithm instead of the current, more simple formula, would be a simple fix.
The paper has a graph comparing the results of human (black) and algorithmically tuning (red) as proof of the latter’s effectiveness. Not bad, but entropy-based tuning hasn’t passed the real test yet: a musician’s ear.

more...
No comment yet.
Scooped by Olivier Lartillot
Scoop.it!

Targeted Audio Ads Based On… Human Emotion? - hypebot

Targeted Audio Ads Based On… Human Emotion? - hypebot | Computational Music Analysis | Scoop.it

“Imagine being able to target your advertising messages based on the parameters of a listener’s emotions during a particular song.

 

Moodagent is a service that combines digital signal processing and AI techniques to create music profiles that take into account characteristics such as mood, emotion, genre, style, instrument, vocals, orchestration, production, and beat / tempo. From these characteristics, playlists are created. Moodagent has an enormous database of music in the cloud, in which every track is scored on five attributes: Sensual, Tender, Happy, Angry, and Tempo.

 

Using the advertising capabilities of Mixberry Media’s audio ads technology, coupled with Moodagent’s knowledgebase of the emotional and musical aspects of songs, advertisers can now target their message to distinct emotional profiles.

 

Brands will be able to select a specific song to embody the essence of their message and, as a result, have their ads heard when the listener is enjoying other tracks with the same emotional data and characteristics – allowing advertisers to communicate the core value of their brand as they perceive it and deliver it to users when they’re in a similar mood or state of mind.”

more...
No comment yet.
Scooped by Olivier Lartillot
Scoop.it!

Music Tech Continues Lead Role At Midem 2012

Music Tech Continues Lead Role At Midem 2012 | Computational Music Analysis | Scoop.it

Tech will continue its takeover of the music industry at this year's midem conference taking place January 28 - 31 in Cannes. Music tech startups, sponsors, hackers and bloggers will all be in attendance and capturing as much attention as possible.

 

This year's midem gathering seems designed to emphasize that music tech is not simply a bunch of new communication tools but a transformational force in the industry.

 

Tech-related attendees are expected to include tech companies as diverse as Amazon, Microsoft, Spotify and Facebook as well as The Echo Nest, Webdoc, Soundcloud and ReverbNation.

- Visionary Monday will feature such participants as GigaOm's Om Malik, Angry Birds creator Mikael Hed and Facebook's Dan Rose.

- midem Hack Day will gather 30 hackers for a 48 hour process of creation sponsored by BlueVia. Many will be exploring ideas suggested by midem attendees.

- midemlab will present 30 music tech startups pitching their companies and services to a panel of expert judges.

On the final day, a Bloggers' Wrap featuring Eliot Van Buskirk of Evolver.fm and Sansom Will of Contagious Communications will give the "lowdown on midem’s hottest industry trends."

more...
No comment yet.
Scooped by Olivier Lartillot
Scoop.it!

TechCrunch | The Echo Nest To Power New Spotify Radio (Which Begins Rolling Out Today)

TechCrunch | The Echo Nest To Power New Spotify Radio (Which Begins Rolling Out Today) | Computational Music Analysis | Scoop.it

Last week, Spotify music service announced that it was redesigning its radio experience from the ground up, offering unlimited stations and unlimited “skips”. And today, Spotify will begin officially rolling out “Radio” to its users on top of its new app platform. But, what Spotify hasn’t been talking about until today is what kind of technology is powering its awesome redesigned Radio functionality.

 

Enter: The Echo Nest, a music intelligence startup whose technology powers many music apps from media companies and independent developers. The Echo Nest is now providing its music intelligence technology to power intelligent radio and radio playlisting within the new Spotify Radio app as it rolls out today across the country.

Given The Echo Nest’s relationship with app developers and record labels (it recently partnered with EMI to open its catalog to app developers), this relationship makes a lot of sense. The Echo Nest will now essentially be powering Spotify Radio, allowing users to create personalized radio stations based around songs or artists in Spotify’s roster of over 15 million tracks.

Partnering with The Echo Nest allows Spotify to enable users to build playlists dynamically around any song or artist for a far deeper radio experience than Spotify has offered previously. As The Echo Nest has one of the more sophisticated playlist engines out there, combining this playlist intelligence with Spotify’s huge catalog and deep social integration should definitely give Pandora pause.

more...
No comment yet.
Scooped by Olivier Lartillot
Scoop.it!

Computer System Recognizes Human Emotion [VIDEO]

Computer System Recognizes Human Emotion [VIDEO] | Computational Music Analysis | Scoop.it
A team of scientists has created a computer system that can recognize human emotion as part of voice recognition.

 

Scientists have created a computer system that attempts to recognize human emotions such as anger and impatience by analyzing the acoustics of one’s voice. Such a system would have obvious implications for perennially frustrating interactive voice response systems, but could be applied to other areas as well.

 

Learn more about the technology in the video.

more...
No comment yet.
Scooped by Olivier Lartillot
Scoop.it!

BBC World discusses MediaMined, sound recognition, and audio search

How do you teach computers to recognise and classify over a million different sounds, often unrecognised and unlabelled before? Click talks to Jay LeBoeuf about sonic search engines. Instead of typing a search term in and seeing a load of returns in text, you could instead play in a sound or tune and it would find you sounds that either match it or resemble it. Jay LeBoeuf discusses how his technology might come to the aid of musicians and filmmakers especially.

 

(starts at 6:45)

more...
No comment yet.
Scooped by Olivier Lartillot
Scoop.it!

Sound (and music), Digested

Sound (and music), Digested | Computational Music Analysis | Scoop.it

Audio engineers have developed a novel artificial intelligence system for understanding and indexing sound, a unique tool for both finding and matching previously un-labeled audio files. Imagine Research of San Francisco, Calif., is now releasing MediaMinedTM for applications ranging from music composition to healthcare.


"MediaMinedTM adds a set of ears to cloud computing," says Imagine Research's founder and CEO Jay LeBoeuf. "It allows computers to index, understand and search sound--as a result, we have made millions of media files searchable."

For recording artists and others in music production, MediaMinedTM enables quick scanning for a large set of tracks and recordings, automatically labeling the inputs.

"It acts as a virtual studio engineer," says LeBoeuf, as it chooses tracks with features that best match qualities the user defines as ideal. "If your software detects male vocals," LeBoeuf adds, "then it would also respond by labeling the tracks and acting as intelligent studio assistant--this allows musicians and audio engineers to concentrate on the creative process rather than the mundane steps of configuring hardware and software."


The technology uses three tiers of analysis to process audio files. First, the software detects the properties of the complex sound wave represented by an audio file's data. The raw data contains a wide range of information, from simple amplitude values to the specific frequencies that form the sound. The data also reveals more musical information, such as the timing, timbre and spatial positioning of sound events.

In the second stage of processing, the software applies statistical techniques to estimate how the characteristics of the sound file might relate to other sound files. For example, the software looks at the patterns represented by the sound wave in relation to data from sound files already in the MediaMinedTM database, the degree to how that sound wave may differ from others, and specific characteristics such as component pitches, peak volume levels, tempo and rhythm.

In the final stage of processing, a number of machine learning processes and other analysis tools assign various labels to the sound wave file and output a user-friendly breakdown. The output delineates the actual contents of the file, such as male speech, applause or rock music. The third stage of processing also highlights which parts of a sound file are representing which components, such as when a snare drum hits or when a vocalist starts singing lyrics.


One of the key innovations of the new technology is the ability to perform sound-similarity searches. Now, when a musician wants a track with a matching feel to mix into a song, or an audio engineer wants a slightly different sound effect to work into a film, the process can be as simple as uploading an example file and browsing the detected matches.

"There are many tools to analyze and index sound, but the novel, machine-learning approach of MediaMinedTM was one reason we felt the technology could prove important," says Errol Arkilic, the NSF program director who helped oversee the Imagine Research grants. "The software enables users to go beyond finding unique objects, allowing similarity searches--free of the burden of keywords--that generate previously hidden connections and potentially present entirely new applications."


While new applications continue to emerge, the developers believe MediaMinedTM may aid not only with new audio creation in the music and film industries, but also help with other, more complex tasks. For example, the technology could be used to enable mobile devices to detect their acoustic surrounding and enable new means of interaction. Or, physicians could use the system to collect data on such sounds as coughing, sneezing or snoring and not only characterize the qualities of such sounds, but also measure duration, frequency and intensity. Such information could potentially aid disease diagnosis and guide treatment.

"Teaching computers how to listen is an incredibly complex problem, and we've only scratched the surface," says LeBoeuf. "We will be working with our launch partners to enable intelligent audio-aware software, apps and searchable media collections."

more...
No comment yet.
Scooped by Olivier Lartillot
Scoop.it!

EMI Music opens its catalog up to developers to create apps for its artists

The initiative provides access to music from the famous Blue Note Records jazz label, and a catalog of thousands of songs from acts such as Culture Club, Shirley Bassey and The Verve. Developers will be able to make use of The Echo Nest’s vast database of information about songs, from simple things like tempo and genre, to complex data about the ‘mood’ of songs. There’s also access to dynamic playlist APIs, open source audio fingerprinting, audio analysis, and remix software.

more...
No comment yet.
Scooped by Olivier Lartillot
Scoop.it!

Instant Mix for Music Beta by Google

Instant Mix uses machine hearing to characterize music attributes such as its timbre, mood and tempo.

 

Music Beta by Google allows users to stream their music collections from the cloud to any supported device, including a web browser. It’s a first step in creating a platform that gives users a range of compelling music experiences. One key component of the product, Instant Mix, is a playlist generator developed by Google Research. Instant Mix uses machine hearing to extract attributes from audio which can be used to answer questions such as “Is there a Hammond B-3 organ?” (instrumentation / timbre), “Is it angry?” (mood), “Can I jog to it?” (tempo / meter) and so on. Machine learning algorithms relate these audio features to what we know about music on the web, such as the fact that Jimmy Smith is a jazz organist or that Arcade Fire and Wolf Parade are similar artists. From this we can predict similar tracks for a seed track and, with some additional sequencing logic, generate Instant Mix playlists from songs in a user’s locker.

Because we combine audio analysis with information about which artists and albums go well together, we can use both dimensions of similarity to compare songs. If you pick a mellow track from an album, we will make a mellower playlist than if you pick a high energy track from the same album.

more...
No comment yet.
Scooped by Olivier Lartillot
Scoop.it!

Million Song Dataset Challenge

Million Song Dataset Challenge | Computational Music Analysis | Scoop.it

"You have the full listening history for 1M users, and half history for 100K users, and your goal is to predict the missing half.

 

The Million Song Dataset Challenge aims at being the best possible offline evaluation of a music recommendation system. Any type of algorithm can be used: collaborative filtering, content-based methods, web crawling, even human oracles! By relying on the Million Song Dataset, the data for the competition is completely open: almost everything is known and possibly available.


What is the task in a few words? You have: 1) the full listening history for 1M users, 2) half of the listening history for 110K users (10K validation set, 100K test set), and you must predict the missing half. How much easier can it get?

 

The most straightforward approach to this task is pure collaborative filtering, but remember that there is a wealth of information available to you through the Million Song Dataset. Go ahead, explore! If you have questions, we recommend that you consult the MSD Mailing List."

more...
No comment yet.
Scooped by Olivier Lartillot
Scoop.it!

‘Cocktail party effect' identified in the brain - health - 18 April 2012 - New Scientist

‘Cocktail party effect' identified in the brain - health - 18 April 2012 - New Scientist | Computational Music Analysis | Scoop.it

“Now we can visualise the workings of the brain as it picks out one voice from many others...

 

The researchers collected enough information to develop an algorithm that turned this brain activity into a spectrogram - a read-out of all the properties of sound, and the time frame in which they occurred.

This enabled the pair to tell when a person hears a certain word. When the team watched the brain activity of the volunteers as they listened to the word, they could see the neurons alter their activity as they tuned into each frequency.

The pair used the algorithm to investigate the cocktail party effect in the volunteers. The pair were able to follow which speaker each volunteer was listening to, just by monitoring their brain activity - the first time this has been done. After the key word was spoken, the spectrogram showed that the volunteer's auditory cortex was responding only to a single voice rather than a combination of the two. The algorithm also enabled the team to tell when listeners mistakenly focused on the wrong speaker, as the translated brain activity in the spectrogram represented a sentence spoken by the other voice.

"I've never seen anything like this before," Because Mesgarani and Chang looked at brain areas involved in assessing the vocal characteristics of a speaker, rather than just those involved in processing sound, they were able to show that the brain can rapidly enhance a voice with certain characteristics to single it out from others.

The researchers hope the algorithm could help to replicate the cocktail party effect in voice recognition systems, which struggle to decipher speech in a noisy room.”

more...
No comment yet.
Scooped by Olivier Lartillot
Scoop.it!

The Netflix Tech Blog: Netflix Recommendations: Beyond the 5 stars (Part 1)

The Netflix Tech Blog: Netflix Recommendations: Beyond the 5 stars (Part 1) | Computational Music Analysis | Scoop.it

This official post from Netflix was written by Xavier Amatriain, who used to work on Music Information Retrieval, before working for Netflix.

 

Here is a presentation of that blog post from theverge.com:

“Netflix has taken to its blog to post part one of an exploration into its recommendation engine. The company offers little in terms of hard data and algorithms (those are promised for part two), but the writeup does highlight just how many different things are "recommendations" on the website. It's not just the "Top 10" list that the algorithms need to make up, it's all of those different genres on the front page, which order to put the films in on each row, and what movies are similar to other titles. It's not just some static list, either; Netflix says that it focuses on keeping results fresh and diversifying where and what order it recommends movies. If your entire household watches on the same account, movies tailored for different family members will pop up, providing options for everyone. And those quirky recommended genres like "Imaginative Time Travel Movies from the 1980s" look like they're here to stay — the company's data shows that there's a direct correlation between how likely a member is to stick with the service and how high up the page it places those specific genres.

The company says that its customers are so confident in the system at this point that 75 percent of all movies watched by members come from recommendations. If all of this talk of recommendation engines reminds you of the Netflix Prize competition that started in 2006, you'll be interested to hear that the million dollar-winning entry hasn't been implemented in any way. It turns out that it would've been too time consuming and costly to use the system on Netflix's massive scale, and by the time the prize was awarded the company was already shifting its focus away from DVDs and towards streaming, something that requires a different sort of recommendation tool altogether.”

http://www.theverge.com/2012/4/8/2934375/netflix-recommendation-system-explained

more...
No comment yet.
Scooped by Olivier Lartillot
Scoop.it!

Music's Data Problem May Be Depriving Artists Of Significant Revenue - hypebot

For nearly 100 years, performing rights organizations have tracked the music played on the radio, then the television, and now the internet. Their goal: to figure out who should get paid.

These organizations - ASCAP and BMI are the big ones - have traditionally relied on the radio, television, and internet music companies they monitor to report what they played, and to how many people, then they cross-reference that with random sampling.

 

In 2012, there is no longer a need for either of those ancient approaches. Back when I did college radio, we used to write down each song we played to submit them to these PROs, and to a great extent, that is still how they work. To borrow a phrase from the old Six Million Dollar Man television show, “we have the technology” to fix this: audio fingerprinting, which can identify every song and snippet of a song that plays on every radio station, television channel, and streaming radio company. Why guess when you can know?

 

This is why we’ve been intrigued by TuneSat, which actually sets up televisions and computers, and feeds them into other computers. The computers actually identify what is being played, rather than counting on broadcasters and webcasters to report things accurately.

I saw TuneSat’s Chris Woods explain what his company does at a MusicTech Meetup in Brooklyn last month, after which I posed a question: “Why don’t ASCAP and BMI use this technology, or simply buy TuneSat outright?” My question was met with knowing guffaws. Someone else in the audience piped up, “Where do we start?”

Woods went on to explain that those organizations are too slow, too mired in the past, and “not nimble enough.”

 

Identifying music on broadcasts would seem to be a perfect application of “big data” — analyzing all media to find the songs and pay the pipers. But to Woods, it clearly wasn’t being used properly.

“I can tell you for a fact that they have never used technology to report the use of my music on any of the broadcasts,” he said. “They have had the technology to do so since 2005, and it’s now 2012, so something’s not right here. It doesn’t take a rocket scientist to figure that out. I don’t know what the real issue is — maybe they’re too big, or slow to adapt to new technology, or maybe it represents exposing their formulas or how they collect and distribute royalties…

more...
No comment yet.
Scooped by Olivier Lartillot
Scoop.it!

TechCrunch | The Echo Nest CEO On What Big Data Means To The Music Industry

TechCrunch | The Echo Nest CEO On What Big Data Means To The Music Industry | Computational Music Analysis | Scoop.it

(featuring a 10-min video interview)
“The Echo Nest is possibly the hottest music data company around right now. They've signed deals with Nokia, EMI, Clear Channel, Spotify, and most recently, Vevo. 

So chances are if you enjoy music, The Echo Nest has something to do with what songs you’re recommended.

Knowing this, Techcrunch Jordan Crook couldn’t resist sitting down with CEO Jim Lucchese to chat out what the music industry will look like in the next couple years, and how The Echo Nest may shape it.

Lucchese believes that the songs you listen to say something about your identity, and that music services have a huge problem ahead of them in the form of millions of listeners and millions of digital music titles. Being the middle man between such huge pools of information is nearly impossible without a deep understanding of the music itself.

But Lucchese believes that the real shift will come by way of understanding the listener, too. We’re getting to a point now where music can be analyzed and categorized in a number of different ways, but little is known about why someone would enjoy Nicki Minaj and Florence + The Machine at the same time. That’s what The Echo Nest is trying to figure out, and it would seem that the company is doing so ahead of the rest of the industry.”

more...
No comment yet.
Scooped by Olivier Lartillot
Scoop.it!

Midemlab 2012 Finalists: Music Discovery, Recommendation, and Creation (Part 1) | Evolver.fm

Among other finalists:

 

“You know how the likes of Shazam only recognises recorded music? WatZatSong claims to recognise songs you sing or hum into your computer, notably by asking its online community.”

 

“WhoSampled.com “allows music fans to explore the DNA of their favourite music”, by tracking songs over the past thousand years, no less! Direct comparisons of, say, how Kanye West sampled Daft Punk are just a click away.”

 

“All of the finalists will present their services to an expert panel featuring SoundCloud, MOG, Sony, Music Ally and more — at midem Saturday January 28. The winners will be revealed at midem’s Visionary Monday, January 30.”

more...
No comment yet.
Scooped by Olivier Lartillot
Scoop.it!

Pop Hit Prediction Algorithm Mines 50 Years of Chart-Toppers for Data

Machine-learning engineers from the University of Bristol think they might have the master equation to predicting the popularity of a song.

 

They use Echo-Nest musical features (tempo, time signature, song duration, loudness, how energetic it is., etc.). By using a machine-learning algorithm, the team could mine official U.K. top-40 singles charts over the past 50 years to see how important these 23 features are to producing a hit song.

Musical style doesn’t stand still, and the weights have to be tweaked to match the era. In the ’80s, for example, low-tempo, ballad-esque musical styles were more likely to become a hit. Plus, before the ’80s, the “danceability” of a song was not particularly relevant to its hit potential.

Once the algorithm has churned out these weights it’s simply a case of mining your proposed song for these exact same features and working out whether they correspond to the trends of the time. This gives you a hit-prediction score.

The team at Bristol found they could determine whether a song would be a hit and, with an accuracy rate of 60 percent, predict whether a song will make it to top five or if it will never reach above position 30 in the chart.

 

Predicting pop songs through science and algorithms has certainly been done before, with varying levels of success.

Researchers at Tel Aviv University’s School of Electrical Engineering mined popular peer-to-peer file sharing site Gnutella for trends, and have a success rate of about 30 percent to 50 percent in predicting the next music superstar. The secret? Geography.

Meanwhile, Emory University neuroscientists went straight to the source and looked at how teenage brains reacted to new music tracks.

 

And then there’s Hit Song Science. It uses an idea similar to Bristol University’s equation, by using algorithms to analyze the world of popular music to look for trends, styles and sounds that are a popular amongst listeners. At the website Uplaya, wannabe hit-makers can upload a track and get a score. The higher the score, the better your song is.

Well, the more catchy it is, at least. The algorithm gives “I Gotta Feeling” by The Black Eyed Peas a hit score of 8.9 out 10, for example.

Bristol’s study differs from previous research because of its high accuracy rate and the time-shifting perception to account for evolving musical taste. Tijl De Bie, senior lecturer in Artificial Intelligence, said, “musical tastes evolve, which means our ‘hit potential equation’ needs to evolve as well.”

He added: “Indeed, we have found the hit potential of a song depends on the era. This may be due to the varying dominant music style, culture and environment.”

more...
No comment yet.
Scooped by Olivier Lartillot
Scoop.it!

Software that spots patterns of emotional speech

Software that spots patterns of emotional speech | Computational Music Analysis | Scoop.it

Researchers are teaching computers how to spot deception in people’s speech. Cues include loudness, changes in pitch, pauses between words, ums and ahs, nervous laughs and dozens of other tiny signs that can suggest a lie.

 

A small band of linguists, engineers and computer scientists, among others, are busy training computers to recognize hallmarks of what they call emotional speech — talk that reflects deception, anger, friendliness and even flirtation.

 

The technology is becoming more accurate as labs share new building blocks, said Dan Jurafsky, a professor at Stanford whose research focuses on the understanding of language by both machines and humans. “The scientific goal is to understand how our emotions are reflected in our speech,” Dr. Jurafsky said. “The engineering goal is to build better systems that understand these emotions.”

 

But homing in on the finer signs of emotions is tougher. “We are constantly trying to calculate pitch very accurately” to capture minute variations, he said. His mathematical techniques use hundreds of cues from pitch, timing and intensity to distinguish between patterns of angry and non-angry speech.

His lab has also found ways to use vocal cues to spot inebriation, though it hasn’t yet had luck in making its computers detect humor — a hard task for the machines, he said.

more...
No comment yet.
Scooped by Olivier Lartillot
Scoop.it!

System that recognizes emotions in people's voices could lead to less phone rage

System that recognizes emotions in people's voices could lead to less phone rage | Computational Music Analysis | Scoop.it
An emotion-recognizing computer system has been designed to make the use of automated telephone services less stressful.

 

A team of scientists have created a computer system that is able to recognize the emotional state of a person speaking to it, so that it can alter its behavior to make things less stressful.

The system analyzes a total of 60 acoustic parameters of users' voices, including tone, speed of speech, duration of pauses and energy of voice signal. The scientists designed the system to look for negative emotions in particular, that would indicate anger, boredom, or doubt.

 

Down the road, perhaps it might someday be combined with a system being developed at Binghamton University, that identifies computer users' emotional states by looking at their faces.

more...
No comment yet.
Scooped by Olivier Lartillot
Scoop.it!

Intelligent Audio Technology | Imagine Research

Intelligent Audio Technology | Imagine Research | Computational Music Analysis | Scoop.it

Imagine Research adds a set of ears to cloud computing and mobile devices. They create software that hears, understands, and labels sounds, making media files searchable and enabling innovative workflows.

more...
No comment yet.
Scooped by Olivier Lartillot
Scoop.it!

Echofi helps you find artists on Spotify that fit your tastes

Echofi helps you find artists on Spotify that fit your tastes | Computational Music Analysis | Scoop.it

Echofi uses the Echo Nest API, as does the Twitter Music Trends app. 

The Echonest Platform is a music intelligence platform that currently has 220 apps built on top of its API, and has which aggregates data about popular music. The company says that is has collected five million data points on thirty million songs and 1.5 million artists.

Music has gone social, and companies like Spotify are just getting started.

more...
No comment yet.
Scooped by Olivier Lartillot
Scoop.it!

Discovering Talented Musicians on YouTube with Acoustic Analysis

We wondered if we could use acoustic analysis and machine learning to pore over videos and automatically identify talented musicians.

First we analyzed audio and visual features of videos being uploaded on YouTube. We wanted to find “singing at home” videos -- often correlated with features such as ambient indoor lighting, head-and-shoulders view of a person singing in front of a fixed camera, few instruments and often a single dominant voice. Here’s a sample set of videos we found.

Then we estimated the quality of singing in each video. Our approach is based on acoustic analysis similar to that used by Instant Mix, coupled with a small set of singing quality annotations from human raters. Given these data we used machine learning to build a ranker that predicts if an average listener would like a performance.

While machines are useful for weeding through thousands of not-so-great videos to find potential stars, we know they alone can't pick the next great star. So we turn to YouTube users to help us identify the real hidden gems by playing a voting game called YouTube Slam. We're putting an equal amount of effort into the game itself -- how do people vote? What makes it fun? How do we know when we have a true hit? We're looking forward to your feedback to help us refine this process: give it a try*. You can also check out singer and voter leaderboards. Toggle “All time” to “Last week” to find emerging talent in fresh videos or all-time favorites.

Our “Music Slam” has only been running for a few weeks and we have already found some very talented musicians. Many of the videos have less than 100 views when we find them.

more...
No comment yet.