Our personal digital music collections have grown enormously in the last few years. Having thousands of songs on a PC is not at all uncommon now, but actually an everyday experience. But the larger a music collection is, the more difficult it becomes to navigate throughout the collection. Most music players only display the titles saved in lists, and the organisational principles applied are text-based and only allow alphabetical sorting. The music itself or the sound of the individual songs usually isn’t a factor. A large portion of songs contained is therefore never or only seldom listened to.

For this reason, alternative approaches for visualizing music collections and for sorting music is becoming more and more interesting. An exciting example of this is the MusicBox project by Anita Lillie, which developed in connection with her master’s work at MIT and is currently receiving a lot of attention.

The mufin team has also been busy for a long time with the topic of visualizing music collections. In 2007, MAGIX MP3 Maker 14 was one of the first commercial products featuring this technology and had already received plenty of positive feedback.

In November, version 15 of MP3 Maker (German) was published. Besides additional exciting functions like sound-based playlist generation and online music recommendations, a further developed version of visualization (mufin vision) was built in.

mufin vision’s functionality has been clearly expanded. Now you can arrange your music collection in two or three-dimensional space. Individual songs are displayed as points or circles, whereby titles with similar properties lie closer together. As a genuine alternative to the normal list view, mufin vision can apply many functions exactly like in the list: e.g. all titles can be played back by double clicking, right clicking opens the context menu, etc.

The highlight of the new mufin vision is the arrangement of properties. The X, Y, and Z axes, the size of circles for titles, and also the color of the individual songs can be interrelated to arbitrary information (the graphical similarity to Anita’s work really surprised us as a perfect coincidence). Besides sound characteristics like tempo, percussiveness, or audio density, metadata (e.g. publication year) and information about the music collection (e.g. most recently added) can also be assigned. With mufin vision, your music collection can be navigated freely, understood at a glance, and browsed viewed like a map.

MP3 Maker is available now at MAGIX in German, and international versions will be appearing in 2009.

We’ve opened up the mufin Beta website, so now everyone can discover music with their own favorite tunes. Some time has passed since the closed beta was launched, but even if things have been quiet here on our blog we’ve been working hard behind the scenes. We’ve been evaluating user feedback, checking and improving the usability of the page, processing numerous test reports, working on various details and adding many new functions.

Besides many new discovery functions and detailed improvements mufin for Facebook and mufin for iTunes (beta) are now also available. With mufin for Facebook you can add music discovery functions directly to your Facebook profile, just like with mufin for MySpace.

mufin for iTunes shows you how to use the technology of the music discovery engine on your PC to rediscover your music collection. After installation of the program your iTunes music collection is analyzed without requiring an Internet connection or iTunes Shop account. As soon as your music tracks have been analyzed (this may take 1-3 seconds per track) you can create playlists with similar sounding tracks based on one or several songs. Only the sound is relevant for compilation of the playlist. By activating the “Relevance” feature you can also set the threshold of the similarity of the tracks to be included in the various playlists.

Phew, what a start! First of all, thanks to you all of you for taking the time to show your interest in mufin – that’s really awesome. We got some great reviews from all over the Web, like here, here, and here..or here. According to the comments and reviews, most people asked for full track playback, and some responded that the recommendendations didn’t always seem to make sense. We want to offer the full track too, but it simply wasn’t possible to implement this in the time we had available to us, since getting the recommendations right was our first priority. But as you can see, we designed the interface with full track capability in mind, you can create playlists using “More options” and also pop-out the player (that little rectangle in the upper right corner of the player).

You may get some ‘freakomendations‘ like here, but always look out for the musical similarties of the recommended track and not the genre, similar title name, or artist image like ‘nineinchnailsvsbritneyspears’ *cough*. We’re constantly trying to improve the recommendation algorithm and have some great features in the works that will allow you to tailor the recommendations to your individual tastes in music.

On the downside, we had some typical private beta technical problems with album covers and prelistening tracks not loading, but they are fixed now.

Times are changing, because from now on, mufin, your music discovery engine, is live! Tell mufin about your favorite songs and it will respond with songs that sound similar –  mufin only needs to analyze a song’s sound to help you discover new music – popularity, origin, or genre aren’t important. We’ve designed the user interface so that it’s easy to see similarities between the songs, giving you the opportunity to browse our music catalog based on the sound alone. My latest discovery is “Kopfüberkreuz” by Monochrome, a small German band, and I originally started the sound discovery process with Faith No More’s “Star A.D“.

Our new site is still in private beta; we’re still polishing up the site design and usability to make sure that we’ve ironed out any remaining bugs. And of course, we can already incorporate your feedback about what we can improve or what should stay the way it is. Don’t worry if you haven’t gotten an invite code yet – just leave your email here, and we’ll let you know as soon as we have some more seats available onboard. Have fun discovering new music!

Last week researchers in the area of music information retrieval (MIR) from all over the world met for the most important conference in this line of work: the ISMIR. Philadelphia provided a wonderful backdrop for the proceedings - although there was of course not much time to visit the historic sites it has to offer.

The conference was chaired by Dan Ellis from Columbia University and Youngmoo Kim from Drexel University where the event took place. Apart from Sunday which kicked off with some very interesting 3-hour tutorials, the daily routine consisted of an exciting mix of panel discussions, plenary sessions and poster presentations. The relatively small number of attendees made it easy to get in touch with other researchers and get to know each other. I want to thank all volunteers, reviewers, committee members and other people who have helped to make this event so well-organized and substantial.

Over 100 papers had been chosen by the program committee to be presented at the conference. They covered an incredibly wide range of topics; here is an overview:
-    User interfaces, visualization, music organization
-    Music analysis (harmony, melody, timbre, rhythm, etc.) and transcription
-    Programming languages for music analysis and synthesis
-    Music recommendation and recognition
-    Social networks, metadata, tags and blogs

Apart from many interesting poster discussions, the most exciting presentations for me where “Hit Song Science is Not Yet a Science” by Francois Pachet and Pierre Roy and the “Support for MIR Prototyping and Real-Time Applications in the ChucK Programming Language” by Rebecca Fiebrink, Ge Wang and Perry Cook.

In the first mentioned, the researches took an extensive set of musical pieces, each annotated with a set of tags categorized as acoustic, subjective or facts and used their audio classification technology to make the machine learn these features. What was really exciting was that they concluded that semantics and grounding were often not related – many acoustic features proved heard to learn while other, more subjective ones yielded better results. However, their final goal – to check if the characteristics of a popular hit song could be determined by a machine purely based on the acoustic features – proved to be unsuccessful. I think, this is not the worst of news – especially for musicians. Socio-cultural aspects also play a huge part – not to mention the record companies’ ad budget…

The other really exciting presentation was ChucK: This real-time audio programming language starts to get more and more interesting and the two presenters managed to give the audience an introduction to the language, write live code that generated polyphonic sound and demonstrate real-time training and classification of two types of classical music – all of this in less than 20 minutes. I was very impressed! This piece of software manages to make something that is usually very dry, complicated and scientific almost playful. Soon the plugin interface will be out – I can’t wait to look into this further. Kudos to these guys at Princeton University!

So, you see, this has been a very interesting week. We here at mufin are also looking forward to contribute to this scientific community in the coming years. There is a lot of cutting-edge stuff in our lab waiting to be released!