DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"> Intelligent Sound Project

Archive

[ Press | Events | News ]


Press archive

2008 June 8 (Sunday)
Interview in Jyllandsposten on privacy issues and intelligent search engines. Link to article (requires login).
 
2008 March 8 (Saturday)
Danish Radio interview with Lars Kai Hansen on brain function and engineering intelligent systems Listen at P1
 
2008 January 15 (Tuesday)
ISOUND's Lasse Mølgaard gave an interview on audio search to the premier Danish radio tech magazine Link to Denmarks Radio.
 
2007 September 12 (Wednesday)
Are the recent Osama statements real? Most likely, however, both video and audio can be faked. Link to Metro Express.
 
2007 April 26 (Thursday)
Børsen TV interview on the MIRocket at the opening of Forskningens Døgn Link for streaming TV.
 
2007 April 12 (Thursday)
Fremtidens hjerner. Interview med Lars Kai Hansen i P1 tema. Link.
 
2007 January 16 (Tuesday)
Mads Christensen laid the foundation of future sound formats. Link
 
2006 December 19 (Tuesday)
"Web 3.0 Internet der læser dine tanker". Kronik i Berlinske Tidende. Link
 
2006 June 25 (Sunday)
I rette rille. Politiken søndag 25. juni 2006 om behovet for at søge i lyd. [1,2]
 
2006 June 11 (Sunday)
Han kortlægger musikkens dna og sælger den på nettet. (Interview i politikken med Tim Westergren, taler ved IMM's lydseminar)
 
2006 June 8 (Thursday)
Computere skal laere at lytte til musik, Jyllands-Posten, link
 
2006 June 2 (Friday)
Danskere vil lave lydens Google, Politiken. link
 
2006 May 4 (Thursday)
Den personlige Juke-BoX, Jyllands Posten
 
2006 May 1 (Monday)
"Verdenspremiere paa DTU's digitale ansigt" Dynamo-DTU Nr. 5.
 
2006 April 3 (Monday)
DTU Avisen: "Digitalt ansigt hjaelper hoerehaemmede"
 
2006 February 28 (Tuesday)
Tue Lehn-Schioeler in TV2 Nyhederne on the project "Talking Faces".
 
2005 November 19 (Saturday)
Stilhed foer fremtiden, Lars Kai Hansen, Kronik in Jyllands Posten Download
 

Event archive

2008 December 8 (Monday)
NIPS 2008. Advances in Neural Information Processing Systems. Link.
 
2008 September 21 (Sunday)
ISCA Tutorial and Research Workshop on Statistical And Perceptual Audition SAPA2008. 21 September 2008, Brisbane, Australia. Link.
 
2008 September 17 (Wednesday)
International and Interdisciplinary Conference on Adaptive knowledge Representation and Reasoning. Link.
 
2008 August 25 (Monday)
The 10th International Conference on Music Perception and Cognition. 25-29 August 2008, Sapporo, Japan. Link.
 
2008 June 24 (Tuesday)
The 14th International Conference on Auditory Display (ICAD), IRCAM, Paris, France.. Link.
 
2008 June 9 (Monday)
The first workshop on Cognitive Information Processing, Santorini, Greece. Link.
 
2008 June 4 (Wednesday)
CANCELLED: NORSIG 2008 Signal Processing Symposium in Copenhagen. Submission deadline Feb 28, 2008. Link.
 
2007 December 7 (Friday)
"Music, Brain & Cognition" - a two-day workshop organized as part of the NIPS Workshops, 7-8 Dec ‘07 Whistler Canada. Link.
 
2007 December 5 (Wednesday)
2nd international conference on Semantics And digital Media Technology (SAMT) - Genova, Italy
 
2007 September 3 (Monday)
EUSIPCO 2007 15'th European signal processing conference Potnan, Poland. Link for more information.
 
2007 August 27 (Monday)
International Computer Music Conference. Venue:Holmen Island in Copenhagen Monday, 27 August to Friday, 31 August. Link
 
2007 July 23 (Monday)
The 30th Annual International ACM SIGIR Conference on Research and Development in Information Retrieval 23-27 July 2007, Amsterdam, The Netherlands link
 
2007 July 11 (Wednesday)
4th Sound and Music Computing Conference 11-13 July 2007,Greece link
 
2007 May 21 (Monday)
Third Intelligent Sound Workshop: May 21-23 2007. Venue: Karlslunde Strand Feriecenter
 
2007 April 26 (Thursday)
Intelligent Sound participates in opening of the national event "Forskningens Døgn". Link for more information.
 
2007 April 11 (Wednesday)
Talk by Stefania Serafin Wednesday April 11, 14:00 "Physical models for sound synthesis"
 
2007 April 1 (Sunday)
The IS Toolbox version 2.0
 
2007 March 15 (Thursday)
AES: Intelligent Audio Environments Conference, Saariselka, Finland. link
 
2007 March 8 (Thursday)
Ingemar Cox seminar: "On ranking the effectiveness of searches" Time and place
 
2006 December 20 (Wednesday)
Digital Music Research Network One-day Workshop Queen Mary University of Londonlink
 
2006 December 15 (Friday)
Deadline for "Special Issue of The IEEE Transactions on Audio, Speech and Language Processing on Music Information Retrieval" link
 
2006 December 9 (Saturday)
Nips Workshop: Advances in Models for Acoustic Processing link
 
2006 December 6 (Wednesday)
1st Workshop on Learning the Semantics of Audio Signals (LSAS) 2006 link
 
2006 December 4 (Monday)
Intelligent Sound MIRocket demo at NIPS 2006 link
 
2006 November 17 (Friday)
Musik og lyd i det offentlige rum link
 
2006 November 15 (Wednesday)
Beyond The Soundbytes Conference, on the furture of the music market. link
 
2006 November 6 (Monday)
International Conference on Computer Music, Tulane University, New Orleans, USA. Submission deadline March 4, 2005.
 
2006 October 27 (Friday)
Audio and Music Computing for Multimedia Workshop in conjunction with ACM Multimedia '06 link
 
2006 October 8 (Sunday)
ISMIR 2006. Conference from Oct 8 to Oct 12 in Victoria, Canada. link
 
2006 September 17 (Sunday)
Music, Interactive Conducting, and Machine Learning, A two-week course link
 
2006 September 6 (Wednesday)
MLSP 2006 conference in Maynooth, Ireland, from Sep 6 to Sep 8. link
 
2006 August 18 (Friday)
PhD course on MIR in Aalborg link
 
2006 August 14 (Monday)
PhD course at DTU: Approximations in machine learning
 
2006 July 24 (Monday)
Summer School in Sound and Music Computing Pompeu Fabra University Barcelona, Spain July 24-28 link
 
2006 July 17 (Monday)
Music and Consciousness, Sheffield link
 
2006 July 1 (Saturday)
Deadline for sound search engine demo
 
2006 June 19 (Monday)
ICAD. The 12th International Conference on Auditory Display, June 19-24, London, UK link
 
2006 June 8 (Thursday)
One day seminar on audio information processing at DTU
 
2006 June 7 (Wednesday)
Peter Ahrendt will defend his phd thesis: "Music Genre Classification Systems – A Computational Approach"
 
2006 June 6 (Tuesday)
Anders Meng will defend his phd thesis: "Temporal feature integration for music organisation"
 
2006 May 22 (Monday)
Intelligent Sound project workshop
 
2006 May 20 (Saturday)
AES 120th Convention, May 20-May 23, in Paris, France, link
 
2006 May 14 (Sunday)
ICASSP 2006, Tolouse in France, May 14-19, 2006. link
 
2006 April 17 (Monday)
Paper submission deadline for ISMIR
 
2006 April 1 (Saturday)
Deadline for first version of MATLAB toolbox
 
2006 March 27 (Monday)
Simon Haykin is giving a talk: "Self-organized adaptive processor for solving the cocktail party problem" at DTU, Building 321, Room 053.
 
2006 March 5 (Sunday)
ICA 2006 in Charleston, South Carolina, March 5-8, 2006 link
 
2005 November 28 (Monday)
CIMCA 2005, Vienna, Austria link
 
2005 September 11 (Sunday)
ISMIR 2005 in London, 11 - 15 September 2005.
 
2005 August 15 (Monday)
Inaugural Workshop for Intelligent Sound Project in Korsør, August 15 - August 18, 2005.
 

News archive

Sunday, December 28 2008
A digital pledge from John Lennon
The voice and a small video clip of John Lennon as been recreated as a pledge for donations to the one-laptop-per-child project. See the video at Youtube. (lkh)
 
Thursday, November 20 2008
The blogosphere
A technical issue has kept the isound blog down for a while, but it has now been fixed and we are happy to welcome back our readers. (kbp)
 
Thursday, November 20 2008
Muzeeker at WMMP 2008
Intelligent Sound's Muzeeker. Wikipedia powered music search engine is presented at WMMP 2008., The First International Workshop on Mobile Multimedia Processing which is held in conjunction with The 19th International Conference on Pattern Recognition (ICPR 2008). The paper: "MuZeeker a domain specific Wikipedia-based search engine" by S. Halling, M. Sigurdsson, J. Eg Larsen, S. Knudsen and L.K. Hansen, has specific focus on the mobile interface. (lkh)
 
Thursday, July 10 2008
Yahoo's BOSS enables small specialized search engines
With the announcement of BOSS Yahoo provides a search infrastructure (datacenter, crawl, index etc) for folks with great search ideas but limited capital. The new BOSS is a rebranding of the SearchMonkey (Ouch!) "open search" service. The hope is that a new small cap search engine market can develop to become a serious add market. Link to Yahoo Blog. (lkh)
 
Sunday, June 29 2008
ISOUND at Stanford
The 2008 workshop on Modern Massive Data Sets (MMDS 2008) took place at Stanford University June 25-28. The workshop was attended by over 250 researchers from academia and industry. The workshop was sponsored by NSF, Yahoo and Linked-In and more. Jerome Friedman (Stanford) presented his new GPS path algorithm for sparse least squares estimation. Michael Mahoney (Yahoo) presented new community statistics for very large social networks, "whisker" like structures (close nit communities with few links to core) seem to be prevalent. Link to workshop site. Presentations will available as pdf. (lkh)
 
Tuesday, June 17 2008
Science and music
The journal Nature features a series of essays on music. In "The neural roots of music" Laurel Trainor reviews neurological aspects of music: "...Music is built on general, universal features of human sound processing that have deep evolutionary roots. It also incorporates rhythmic, melodic and harmonic structure. Musical structures and styles vary enormously across cultures, and change as continually as languages, yet our biology constrains the possibilities". Link to Nature. (lkh)
 
Sunday, June 8 2008
Wikipedia based Muzeeker in pre-beta
Muzeeker is ISound's music search engine with Wikipedia intelligence. The search engine maps user queries to articles and categories in the English Wikipedia. Once an artist or song is identified a query is issued for Youtube allowing the user to explore new music. The Muzeeker team includes Søren Halling, Magnus Sigurdsson and Lars Kai Hansen. Link to Muzeeker. (lkh)
 
Saturday, May 31 2008
Affective audio-visual computing
PhysOrg reports on a new paper by Yongjin Wang from the University of Toronto and Ling Guan from Ryerson University in Toronto. In contrast to earlier work they fuse audio and visual data for cross-language emotion detection and they "...have developed a system that recognizes six human emotional states: happiness, sadness, anger, fear, surprise, and disgust. Their system can recognize emotions in people from different cultures and who speak different languages with a success rate of 82%". Affective computing will allow search engines to index emotional valence. Link to PhysOrg report. (lkh)
 
Sunday, May 25 2008
More from behind the walls of Gogol
In Anand Rajaraman's blog "Datawocky" he recounts a discussion with Peter Norvig, head of research at Google. It is claimed that Google does'nt use machine learning methods to make decisions about ranking although they work as well as "hand-tuned formula". Apparently the reason is the risk that machine learning could show not robust to unseen outliers. I find this unlikely in lieu of the recent posting by Udi Manber which we referred to yesterday. If it is true that Google has an online quality control system running (at the minute scale...supposedly checking that users make the predicted click-trough) then there is no reason not to use the algorithm that perform better in the present market - with the possibility to switch to a more robust hand-tuned formula if the market swings? A part from this issue we learn other interesting facts, such that Google keeps a 200 parameter characteristic of each web page, one parameter being the PageRank Link to datawocky. (lkh)
 
Saturday, May 24 2008
From inside Google Search Quality
Udi Manber, VP Engineering, Search Quality has posted a interesting comment at the Google blog on the mechanics of their ranking. Days are long gone when PageRank was responsible for Google's high quality ranking. Localization, automated quality control, and web spam filtering are now part of the complex ranking system: "...There are automated evaluations every minute". Manber estimates that more than 1000 engineering man years have gone into developing the ranking system, and promises that he will reveal more details in the future.... Link to blog. (lkh)
 
Sunday, May 4 2008
Can audio CAPTCHAs work?
There is an interesting discussion going on about Google's use of audio CAPTCHAs. CAPTCHA is short for "Completely Automated Public Turing test to tell Computers and Humans Apart". You have most likely met image CAPTCHAs when subscribing to services. CAPTCHAs are used to avoid bots to invade systems like Hotmail, Gmail etc. To help visually impaired Google started using audio CAPTCHAs in 2006. However, in the Wintercore Labs blog ( Link to blog.) Rubén Santamarta provides simple Fourier based tools for hacking Google's audio CAPTCHA. Is the game on? Like in imaging, I expect that audio experts will take the opportunity to develop safer audio CAPTCHA. (lkh)
 
Sunday, April 13 2008
Radiohead takes music interactivity to a new level
By letting users buy individual tracks (bass, drum, vocal) and allowing users to remix and re-distribute Radiohead is breaking new ground in music usage and interactivity. Read more at Wired's Listening Post. You can enjoy user contributed remix, vote for your favourites and get links for mash-up applications at the Radiohead site. (lkh)
 
Sunday, April 13 2008
Cognitive component paper accepted at Cognitive Science Conference
The paper "Is Cognitive Activity of Speech Based On Statistical Independence?" by Ling Feng and Lars Kai Hansen has been accepted for presentation at CogSci 2008 - The 30th Annual Meeting of the Cognitive Science Society. Link to pdf. (lkh)
 
Sunday, April 13 2008
Paper accepted at Cognitive Information Processing Workshop
The paper "On Phonemes as Cognitive Components of Speech" by Ling Feng and Lars Kai Hansen has been accepted for presentation at the 1st IAPR Workshop on Cognitive Information Processing Santorini 2008. Link to pdf. (lkh)
 
Saturday, April 5 2008
Myspace = MiTunes?
In an expected move, three of the big four music companies (Sony BMG Music Entertainment, Warner Music Group Corp and Universal Music) have entered a partnership with MySpace, so that it will be possible to buy DRM free music through MySpace. Link to analysis at eFluxMedia. (lkh)
 
Thursday, March 27 2008
NYT: Scott de Martinville first recorded sound on April 9, 1860
New York Times reports that an audio recording of "Au Clair de Lune" made in 1860 by Édouard-Léon Scott de Martinville has been re-constructed and that "...David Giovannoni, an American audio historian who led the research effort, will present the findings and play the recording in public on Friday at the annual conference of the Association for Recorded Sound Collections at Stanford University in Palo Alto, Calif." Thomas Edison's first recording of "Mary had a little Lamb" was made seveteen year later in 1877. Link to NYT. (lkh)
 
Friday, March 21 2008
Arthur C. Clarke 1917-2008
Arthur C. Clarke died Wednesday March 19, (aged 90) in Colombo, Sri Lanka. Audio research played a major role in Clarke's most famous novel "2001: A Space Odyssey" and even more so in Stanley Kubrick's movie based on the novel. First of all, in the movie the spaceship computer HAL's self-learned lip-reading competence allows it to "eyes-drop" on the conversation were the crew evaluates the mission objectives. Secondly, HAL's ability to understand and produce speech was based on early Bell Labs research. Link to Bell Labs article. Arthur C. Clarke's work has inspired many other artists including composer Mike Oldfield whose album "The Songs Of a Distant Earth", is based on Clarke's novel of the same title. Link to Wikipedia article. Snippets are available at LastFM. (lkh)
 
Monday, March 17 2008
Free stuff
How can a CDs, DVDs, and AIR travel be free? They are not - one way or the other customers pay. Wired magazine explains the business models that led to the Prince, Radiohead music give aways. For the Daily Mail and Prince the "free" CD was great business according to Wired. Prince earned 18 mill. pounds on expensive concert tickets. Daily Mail lost 700.000 pounds in direct cost, but increased its in brand recognition as "pioneers". Link to Wired article. Although the numbers are quite uncertain it is still an interesting read just how efficient "free stuff" marketing can be and how it is enabled by the low reproduction cost of digital information. (lkh)
 
Tuesday, March 4 2008
Charlatans' album give away
Following the Radiohead and Prince examples, the Charlatans have decided to give away their album "You Cross my Path". The album can be downloaded from Xfm and on the site the Charlatans are quoted "We always knew this album was gonna be given away for free, even before we started writing so we wanted to make it the best album we've ever made. This isn't a case of left over tracks and b-sides, we wanted to give our fans a quality record." Link to Xfm (lkh)
 
Monday, March 3 2008
Torben Bach Petersen new professor in business intelligence
Aalborg University has announced that member of the ISOUND steering group Torben Bach Pedersen has been appointed new professor in Computer Science in the field of business intelligence. Torben Bach Pedersen has already established a strong network in in business intelligence and his research foci are within data warehousing and on-line analytical processing (OLAP). Link to announcement and background (in Danish). Link to Torben's webpage (lkh)
 
Sunday, February 24 2008
Audiosurf wins best audio award at IGF
The simple music controlled game Audiosurf won the audio category at the Independent Games Festial 2008. Link to IGF. The game is a simple driving game where the "shape, speed and mood of each ride" is determined by the music chosen by the user. Link to Audiosurf with Youtube video demos. (lkh)
 
Thursday, February 21 2008
Artist Rihanna tops singles chart based on downloads alone
The Digital Spy reported on Feb 18, that Rihanna's "Don't stop the music" is the first song to top the charts based on downloads alone. The single topped Australian singles chart before it was available in music stores. Link to Digital Spy (lkh)
 
Thursday, February 21 2008
People who listen to oldies are least likely to be Asian
Google's Adsense for Audio allows advertisers to identify useful radio stations for their adds based on an extensive list of demographics features part of which is based on interesting "musical genre to demographics" mapping reavealed in their beginners guide. Link to the guide (lkh)
 
Monday, February 4 2008
ISOUND in San Diego
The Museum of Contemporary Art in La Jolla features a brillant audio centered exhibition "SOUNDWAVES: The Art of Sampling". The web site includes plenty of interesting audio and video materials. Enjoy, for example, the video documentation of French Celeste Boursier-Mougenot's installation consisting of inflatable swimmingpools "... an assortment of porcelain bowls and glass goblets floats in each pool" leaving a weird but highly intriguing audio impression. (lkh)
 
Monday, February 4 2008
ISOUND pilgrimage to Google I
Hosted by Lek-Heng Lim ISOUND paid homage to one of the corner stones of modern search, namely the first Google computer. The machine is on display at Stanford in basement of the William Gates(!) building and features an array of ten harddisks (totalling 40Gb) in a LEGO brick housing.... Link to pictures (lkh)
 
Friday, January 18 2008
Christian S. Jensen new IEEE fellow
ISOUND is proud to congratulate Christian S. Jensen on election to IEEE Fellow. The US engineering organization IEEE has elected Christian S. Jensen for his contributions to database research. The 2008 class counts 295 fellows. Fellow is the highest membership grade of the IEEE. Link to Aalborg Univ announcement. (lkh)
 
Friday, December 28 2007
Warner allows DRM free MP3 at Amazon
After opening DRM free sales in September Amazon now has three of the big record companies on board. The pressure is on Sony. BTW prices are also lower than at iT... Link to story at Gizmodo. Apparently Warners music is sold even without watermarks. (lkh)
 
Saturday, December 15 2007
Google to rival Wikipedia on knowledge sharing?
New York times is reporting on the new "knol" Google service that aims to serve a market similar to Wikipedia's. The "knol" seems to be more formalized than Wikipedia from the sample shown on a Google blog. Link to the offical Google blog "knol" sample. (lkh)
 
Saturday, December 15 2007
Wikipedia to rival Google on web search
New Scientist reports that the distributed search engine earlier announced by Jimmy Wales will air before Christmas in an alpha version based on 500 volunteers running a SETI@home grid spider swarm . Link to New Scientist. (lkh)
 
Wednesday, December 5 2007
DRM problems lurkin for Nokia's Comes with Music?
The excitement is significant over Nokia's new "Comes with Music" scheme that gives the user an all-you-can-eat account with Universal Music. However, Ars Technica now reports that the DRM protection chosen by Nokia will lead to problems related to incompatibility with other main stream services Link to Ars Technica article. (lkh)
 
Saturday, November 24 2007
In the beginning there was a jukebox ....
In this age of playlist liberation it is worth noticing that the first jukebox was installed as early as in 1889 in a bar with the impressive name Palais Royale Saloon in San Francisco. The jukebox is celebrated in Wired magazine. The price of a song was a nickel which interestingly had a value similar to the price of a song in todays web store ($1). Link to the Wired article. (lkh)
 
Friday, November 23 2007
Google still way ahead in search volume
Wednesday the internet information provider "comScore Inc" released the latest figures for US search traffic. Google at 58.5% of the query volume is by a large margin leader of the score board, with Yahoo at 22.9%, and Microsoft at 9.7% in second and third place. The five top search engines had a total volume of 10.5 billion queries in October 2007. Link to the press release. (lkh)
 
Saturday, November 17 2007
RCDR LBL in the air with free(?) music
The RCDR LBL site is now active. The look and feel is that of a music blog, a kind of chaotic MySpace clone. Part of the available audio is for streaming only - with some DRM free downloads. Worst of all, the whole thing is somewhat complex to navigate and at least this user is left confused and needs a few more visits to judge whether this is it. Link to the site. (lkh)
 
Saturday, November 10 2007
DRM free music sales higher
Ars Technica reports today that DRM free music now outperforms protected music sales by four to one! The sales figures come from 7 Digital, and relates to sales in the UK. Link to article. The increased focus on digital music, higher sales for both single tracks and albums, and less trouble with DRM thus easier legal sharing, will once more increase the focus on music navigation and search. (lkh)
 
Saturday, November 3 2007
ISOUND paper in NIPS 07 workshop
The NIPS workshop "Modeling the Structure of Music and its Effect on the Brain" will feature a paper "Discovering Music Structure via Similarity Fusion" by Jeronimo Arenas-Garcýa, Emilio Parrado-Hernandez, Anders Meng, Lars Kai Hansen and Jan Larsen. Link workshop. (lkh)
 
Saturday, November 3 2007
P2P users buy more music
Industry Canada has performed a detailed analysis of the impact of P2P downloads on the music market. From the abstract: "..The report, prepared by University of London researchers, Birgitte Andersen and Marion Frenz, found that music downloads have a positive effect on music purchases among Canadian downloaders but that there is no effect taken over the entire population aged 15 and over." Link to report. (lkh)
 
Saturday, October 27 2007
ISOUND paper and software for bi-clique community detection
Sune Lehmann, Martin Schwartz, and Lars Kai Hansen have submitted a paper "Bi-clique Communities" on community detection in bi-partite networks. The paper generalizes the notion of k-clique communities to bi-networks. As in k-clique community detection the new scheme finds overlapping communities, i.e., nodes can be part of more than one community. A new efficient software tool has been developed for finding bi-clique communities. Link to paper in arxiv. (lkh)
 
Thursday, October 25 2007
ISOUND paper on community detection in social networks
A paper on "Deterministic modularity optimization" by Sune Lehmann and Lars Kai Hansen has been accepted for European Physical Journal B. Mean Field Annealing is suggested as a tool for finding social network communities according to Newman's modularity measure"Q": An interesting aspect of the modularity measure is that it automatically determines the number of communities. Link to paper. (lkh)
 
Thursday, October 25 2007
ISOUND paper at WASPAA 2007
The paper "Linear Regression on Sparse Features for Single-Channel Speech Separation" by Mikkel Schmidt and Rasmus Kongsgaard Olsson is presented this week at the IEEE Workshop on Applications of Signal Processing to Audio and Acoustics (WASPAA) in New Paltz, NY. Link to paper. From the abstract: "We formulate a linear regression model for estimating each speaker based on features derived from the mixture. The employed feature representation is a sparse, non-negative encoding of the speech mixture in terms of pre-learned speaker-dependent dictionaries." (lkh)
 
Wednesday, October 17 2007
MySpace and Sony in add-supported music deal
A posting in Yahoo! News announced Tuesday that MySpace and Sony have struck a deal that will allow MySpace users to enjoy Sony's catalogue while Sony and MySpace will share the revenue from the adds shown by MySpace in relation to the music. Yahoo! News noted "...Sharing advertising revenue with Web sites that air its videos for free to consumers has been one of the new business models that music companies are exploring" Link. (lkh)
 
Wednesday, October 17 2007
MySpace with Skype enabled conversation
Reuters report today that MySpace and Skype have entered a partnership that will give automatic Skype access to Myspace accounts. This will enable a real-time speech based community beyond slow text chat. BBC says "With the service, which starts from the end of November, the two firms say they are creating the world's largest online, voice-connected community." Is it time for speech summaries to keep friends updated with other friends-of-a-friend conversations? (lkh)
 
Monday, October 15 2007
Hard hats on
There is a new large scale music experiment going on at the site Deadairspace. Radiohead has decided to offer free download of their new album In Rainbows. It is yet unclear if this initiative is really contageous... Prince started,...who will be the next big name to give away music, is this a viable new business model? Is the music business entering the post-record-company era? Read WIRED's Eliot Van Buskirk on the Radiohead action: The Listening Post. (lkh)
 
Tuesday, October 9 2007
Google and IBM announce new cloud computing center
Google and IBM have announced that they together will fund a new research center focusing on cloud computing. Cloud computing refers to services in which the computing takes place in the network and typically in the form of grids involving distributed large scale computer facilities, ie., like the search service provided by Google. The center will contain a small scale cloud initially equipped with 1600 processors. Link to a New York Times article. Link to a Wiki article on cloud computing. (lkh)
 
Sunday, September 30 2007
First evaluation of Amazon's DRM free MP3
The TidBits blog provides a first evaluation of the Amazon MP3 service. Adam Engst's conclusion goes "...I don't think Amazon MP3 will be putting the iTunes Store out of business by any stretch of the imagination. It's competitive, thanks to the lack of DRM, low prices, and ease of shopping, but it's clumsier than using iTunes... The good news is that by releasing an online music store that doesn't suck, Amazon has given Apple some real competition, and where there's competition, there's innovation." Link to the TidBits article. (lkh)
 
Tuesday, September 25 2007
Amazon DRM free MP3 in beta
Amazon today announced their DRM free MP3 service: "To put it plainly, we're music nerds who love bargains and want to be able to transfer our digital music between our computers, portable players and CD-Rs with no strings attached, so we worked to build a place where we would want to shop. We hope you're as excited about this as we are...." Link to announcement at the Amazon Earworm's Blog. (lkh)
 
Friday, September 14 2007
Large vocabulary parrot's last words
The famous talking bird, parrot Alex, 31, whose vocabulary exceeded 100 words, died in Boston on Sep 6. The bird's last words were: “You be good, see you tomorrow. I love you.” Link to article in New York Times. (lkh)
 
Tuesday, September 4 2007
Two ISOUND papers at EUSIPCO 2007 in Poznan
The session Signal Processing for Music at the 15'th European Signal Processing Conference (EUSIPCO 2007) in Poznan, Poland, features two ISOUND papers: "A Framework for Analysis of Music Similarity Measures" by Jesper Hojvang Jensen, Mads Grasboll Christensen, and Soren Holdt Jensen and "Learning and Clean-Up in a Large Scale Music Database" by Lars Kai Hansen, Tue Lehn-Schiøler, Kaare Brandt Petersen, Jeronimo Arenas-Garcia, Jan Larsen and Søren Holdt Jensen. Link to EUSIPCO 2007. (lkh)
 
Thursday, August 30 2007
Speak softly to your plants (or they will modify genetically)
Mi-Jeong Jeong of the National Institute of Agricultural Biotechnology, Korea, and coworkers, claim in a new paper in the Springer journal Molecular Breeding, that they have found "sound-responsive genes". The article says "..Under both light and dark conditions, sound up-regulated expression of rbcS and ald. These are also light-responsive genes and these results suggest that sound could represent an alternative to light as a gene regulator." Link to abtract. (lkh)
 
Monday, August 27 2007
First Monday paper on a new music industry
In a paper in the August issue of First Monday Robert L. (Bob) Frost proposes "... a systematic model of disintermediation in the recorded–music business. Were such a model successful and tied to digital distribution, prices to consumers would fall considerably, artist compensation would rise, online piracy would drop, and the information–feedback loops necessary for signaling consumer tastes back to artists would become far more efficient as information asymmetries were mitigated." Link to First Monday paper. (lkh)
 
Thursday, August 23 2007
Speaking to an intelligent phone
A new text processing tool in combination with speech to text transcription, from US company Vlingo promises an easier access to mobile services. Link to article in TechnologyReview. A Vlingo whitepaper "Revolutionizing Voice UI for mobile" gives more detail. According to the whitepaper the software features "...Hierarchical Language Model Based Speech Recognition: We have replaced the constrained grammars with very large vocabulary (millions of words) Hierarchical Language Models (HLMs)..." (lkh)
 
Thursday, August 23 2007
Four papers at MLSP 2007
The Machine Learning for Signal Processing conference, August 27-29, 2007, in Thessaloniki, Greece, features four ISOUND related papers: 1)Multiplicative Updates For The LASSO by Morten Mørup, Line H. Clemmensen; 2)Shifted Non-Negative Matrix Factorization by Morten Mørup, Kristoffer H. Madsen, Lars K. Hansen; 3)Unveiling Music Structure Via PLSA Similarity Fusion by Jerónimo Arenas-García, Anders Meng, Kaare Brandt Petersen, Tue Lehn-Schiøler, Lars Kai Hansen, Jan Larsen; 4) Wind Noise Reduction Using Non-Negative Sparse Coding by Mikkel N. Schmidt, Jan Larsen, Fu-Tien Hsiao. Link to the conference programme. (lkh)
 
Friday, August 17 2007
Universal's DRM free music with Google music search?
Forbes.com reports that Universal's DRM free music sales will happen through conventional outlets, such as Amazon, and also through a new Google backed music search facility in the context of the company gBox Inc. Link to Forbes.com. (lkh)
 
Sunday, August 12 2007
ISOUND Paper opens the ISMIR 2007 conference
The annual conference of the International Society for Music Information Retrieval opens in Vienna on Sep 23, with a presentation by ISOUND's François Deliège. The paper is titled Fuzzy Song Sets for Music Warehouses, and is co-authored by Torben Bach Pedersen. (lkh)
 
Friday, August 10 2007
Universal to sell DRM free music
The New York Times reports that another major label now will start DRM free digital music sales. This time it is Universal Music, it seems like we wont see an increased price tag in contrast to the EMI iTunes deal? Link to NYT. (lkh)
 
Tuesday, August 7 2007
ITMAN-ISOUND PhD course in Machine Learning
Monday Aug 13 to Friday Aug 17, ISOUND will host a PhD course at DTU. Teachers include Jeronimo Arenas-Garcia (Univ Carlos III, Madrid, picture) and Joaquin Candela Quinonero (Microsoft, Cambridge). Link for more information. (lkh)
 
Thursday, July 26 2007
ISOUND featured in the Informer
The summer 2007 issue of the Informer - the quarterly newsletter of the British Computer Society's Information Retrieval Specialist Group (IRSG) - features an article by ISOUND's Maria M. Ruxanda on the status of the project. Link to the Informer. Maria participated in the IRSG's Summer School on Multimedia Semantics in Glasgow, July 15-21, 2007. Link to the summer school web site. (lkh)
 
Sunday, July 1 2007
The artist that may again be known as Prince...
British Mail on Sunday will give away the new Prince album "Panet Earth" with the newspaper. Link for more information at The Daily Mail. The Music industry has expressed concern over the move as it will erode the basic business model. The Guardian writes: "The singer had signed a global deal for the promotion and distribution of Planet Earth in partnership with Columbia Records, a division of music company Sony BMG. A spokesman for the group said last night that the UK arm of Sony BMG had withdrawn from Prince's global deal and would not distribute the album to UK stores." Link to Guardian article. (lkh)
 
Saturday, June 30 2007
iPhone: what about music?
The much hyped iPhone has hit the US market and is expected to arrive in EU later this year. From a sound point of view we are interested in iPhone's music related features! BBC reports on the first reactions and give some numbers. Link to BBC's report. First thing we note is that the iPhone allows six hours of browsing and seven hours of video watching on its battery, hence Pandora, MySpace, or YouTube listening pleasure. However, it is available in two versions with four or eight gigabytes of memory, only, thus no iPod replacement. BBC quotes initial web speed frustrations. On the positive side: Among the announced accessories is a bluetooth headset Link to engadget.com. Then another downer: BBC says that the device "costs either $499 or $599 and buyers must also commit to a two-year contract with AT&T that will cost them a minimum of $59.99 per month". Finally, it appears that the iPhone will not support iTunes Sic! Link to Yahoo News' report. (lkh)
 
Monday, June 25 2007
Voice mail as text message
CallWave of Santa Barbara, CA, has launched as service "Vtxt" based on speech recognition and message summarization that allows spoken messages to be transmitted under the 140 character sms constraint. In an interview with Wired. Callwave scientist Anthony Bladon is qouted "We provide only the information that answers the questions, 'Do I need to act on this now? Do I need to call this person back?'..." (lkh)
 
Saturday, June 23 2007
Economist on pluggd and search in podcasts
Economist's Technology Quarterly magazine features an article on speech recognition and related business opportunities. The article introduces Alex Castro the entrepreneur behind "pluggd" who deliver search in podcasts. In the pluggd blog Castro points to their chunking technology, that seem to share features with ISOUND's "speaker turn" concept implemented in "Castsearch". Link to Technology Quarterly. Link to pluggd. Link to Castsearch. (lkh)
 
Thursday, June 21 2007
ezmo- New music sharing opportunity
Norwegian owned ezmo has developed a new business model for music sharing that seems to comply with copyright rules. The site allows upload and streaming to a limited number of friends. In a way you create your own radio station. Link to ezmo. (lkh)
 
Wednesday, June 20 2007
The Sound of Hillary is Celine Dion
Hillary Clinton let her presidential campain signature song decide by popular vote. She had suggested U2's "Beautiful Day" among others, but a voter suggested the Celine Dion song "You and I" and it won the race getting more than 200.000 votes. Listen to the song. (lkh)
 
Tuesday, June 19 2007
Blind boy navigates by echo location
Ben Underwood -the Sonar boy- is a blind kid who masters navigation by sonar, rather like bats albeit using sound in the audible spectrum. This article in Slate.com gives a bit of perspective. Nothing superhuman about echo location. Ben is just a little better at something, that we can all learn. (lkh)
 
Saturday, June 16 2007
ISOUND researcher speaks at IKT Rådet
The prime ICT advisory body for the Danish science ministry "IKT Rådet" met Thursday June 14. ISOUND's Mads Græsbøll Christensen presented his view on the key challenges for Danish IT. Link to the agenda (in Danish). (lkh)
 
Monday, June 11 2007
RCDR LBL to give away music for free
Yet another attack on the current music business model is brewing in New York. According to the New York Post: "Internet entrepreneur Peter Rojas plan to launch an online-only record label that will offer its music for free and generate revenue only through advertising and sponsorships...". Link to NYP article.. Peter Rojas will create RCDR LBL in collaboration with the existing label Downtown Records who have "Cold War Kids" among their artists. (lkh)
 
Friday, June 8 2007
Lasse Mølgaard and Kasper Winther wins computer science prize
The Danish Society for Computer Science's prize for the best Master Thesis 2006 was awarded to ISOUND's Lasse Lohilati Mølgaard and Kasper Winther Jørgensen for their thesis "Tools for Automatic Audio Indexing" on the spoken document search engine castsearch. Link to Castsearch. The prize was installed in 1991 and is co-sponsored by the industry board Dansk IT. Link to "Dansk Selskab for Datalogi". (lkh)
 
Friday, June 8 2007
ISOUND at Sound Days
The Copenhagen area sound and music enthusiasts were invited for a two day arrangement by ‘Sound Forum Øresund’ June 7-8. Link to the Sound Days Blog. The program featured presentations by the radio montage pioneer Stephen Schwartz, Karsten Kjems of Sonic Branding, and Richard Widerberg & Zeenath Hasan, who gave an IMPROVe your mobile phone concert. ISOUND presented and demo'ed the MIRocket/WinAmp real-time music recommendation system and the Castsearch spoken document search engine. (lkh)
 
Tuesday, June 5 2007
Swedish speaking paper
Research teams at the Mid Sweden University forestry industry research program Fiber Science and Communication Network (FSCN) are working on the fourth generation of paper products that can communicate with computers. This integrates paper with the digital world. The new paper is engineered to support higher densities of print and has new structures supporting computational features, such as build-in speech and other media capabilities. Link to the press release. Will your mailbox be full of commercials with speech and music? (lkh)
 
Sunday, June 3 2007
Two ISOUND papers at EUSIPCO 2007
The two papers "A Framework for Analysis of Music Similarity Measures" by J. Højvang Jensen et al., and "Learning and Clean-Up in a Large Scale Music Database" by L.K. Hansen et al., will be presented at EUSIPCO on Wednesday Sep 05, 2007, in the lecture session "Signal Processing for Music Analysis". (lkh)
 
Wednesday, May 30 2007
CBS buys Last.fm
The music recommender system Last.fm with more than 15 mill users world wide has been sold to the US media company CBS for $280 mill. Link to the CBS press release: "Founded in 2002, Last.fm creates communities around music by bringing together listeners, artists and music. Through its unique "scrobbling" technology and social recommendation engine, Last.fm builds a comprehensive profile of each user’s musical taste and leverages each user’s song list to make highly personalized recommendations, connect users who share similar tastes, provide custom radio streams and other music related community features." (lkh)
 
Monday, May 28 2007
Powered by sound — stove could help reduce poverty
The SCORE project aims for a wood-powered generator capable of both cooking and cooling food and thus address the energy needs of rural communities in Africa and Asia, where access to power is limited. The SCORE press release goes on: "Thermoacoustic principles involve the use of sound waves. The wood is burned to produce heat. This then goes into a specially shaped pipe which produces areas of high and low gas pressure in such a way as to generate sound...". Link to the SCORE project with audio samples from the Sterling engine. Initial development of the technique was sponsored by Ben & Jerry's Ice Cream. (lkh)
 
Saturday, May 26 2007
Lifelog: Why don't you just record everything?
The New Yorker brings an in depth portrait of Gordon Bell, the "total recall" researcher behind DARPA's Lifelog and Microsoft's MyLifeBits projects to store everything you do, hear, or see. Link to the New Yorker article. Audio issues have been investigated by Dan Ellis as part of the related Microsoft "Digital Memories" project. Link to Ellis' Personal Audio page. (lkh)
 
Friday, May 25 2007
Babies can detect language from visual cues
Whitney Weikum, Univ. British Columbia, Canada, has tested 4-6 month old infants and found that they were able to discriminate between english and french from visual cues alone. Link to the Science paper. (lkh)
 
Thursday, May 24 2007
Three days of ISOUND workshop
The 2007 workshop took place in Karlslunde and was blessed by sun, fine food, and extra-ordinarily intelligent sounds, delivered by invited speakers including Dan Ellis, Stefania Serafin, Esben Skovenborg, and Deliang Wang, and researchers from the ISOUND groups. The group picture was shot in Roskilde at the Viking ship museum's wood shop. Link to Ellis' invited talk. (lkh)
 
Thursday, May 17 2007
Amazon to sell DRM free MP3 in 2007
In a press release Amazon.com has announced that it will "..launch a digital music store later this year offering millions of songs in the DRM-free MP3 format from more than 12,000 record labels...". Will pricing be like for DRM protected music? Link to Amazon's press release. Thanks to Kaare Brandt Petersen for passing this to ISOUND. (lkh)
 
Monday, May 14 2007
Visualizer for social software
Social software or "web 2.0" evolves rapidly and researchers aim to characterize sites according to basic dimensions such as "Identity" - a way of uniquely identifying people in the system; "Presence" - a way of knowing who is online, available or otherwise nearby; "Groups" - a way of forming communities of interest, etc. Gene Smith has developed a honeycomb "radar" to visualize a given service's focus. Follow this link for more information. (lkh)
 
Monday, May 14 2007
Program for the 2007 ISOUND workshop
The workshop takes place in Karlslunde Strand Feriecenter, May 21 to May 23, 2007. Link to the program. Speakers include Dan Ellis and Deliang Wang. (lkh)
 
Thursday, May 10 2007
Audio for Google Earth
New Scientist reports that Google Earth is soon to be augmented by local audio clips from the Wild Sanctury collection: "..sound recordings from all over the world, including bird and whale song and the crackle of melting glaciers". Link to NewScientistTech. (lkh)
 
Saturday, May 5 2007
Two Neural Computation theory papers from ISOUND
Two theory papers from ISOUND appeared in the April issue of Neural Computation: Mads Dyrholm, Scott Makeig, and Lars Kai Hansen "Model Selection for Convolutive ICA with an Application to Spatiotemporal Analysis of EEG". Neural Comp. 2007 19: 934-955. Link to pdf. Rasmus Kongsgaard Olsson, Kaare Brandt Petersen, and Tue Lehn-Schiøler "State-Space Models: From the EM Algorithm to a Gradient Approach". Neural Comp. 2007 19: 1097-1111. (lkh)
 
Friday, May 4 2007
I just listened to horse hemoglobin
Researchers at UCLA have found a new way to audify genes that includes rhythmic features. "..A rhythm has been encoded into the musical sequence according to the organism’s codon distribution used in the genome-encoded protein sequence." Link to the gene2music web site with audio examples. (lkh)
 
Thursday, May 3 2007
Pandora shuts down service for non-US accounts
TechCrunch reports that Pandora has decided to shut down its services to non-US accounts based on IP filtering . Link to TechCrunch. Thanks to Harish Saini for passing this link to ISOUND. (lkh)
 
Sunday, April 29 2007
Will Arctic Monkeys downloads set new record and show decline of the album playlist?
The midweek charts indicate that ALL songs on the new Arctic Monkeys album will end up in the top-100 list based on singles download. This is another indication that consumers want songs - not push playlists. Link to scenta article. (lkh)
 
Saturday, April 28 2007
Westergren interview on Pandora's current and future business model
AlterWeb brings an interview with Pandora founder Tim Westergren, who iterates the strong critique of the new internet radio license fees in the US. He also reveals aspects of the current business model: ".. Until this ruling, the rate that we paid was 1.17 cents per listener-hour, which equates to about 0.076 cents per song streamed to each listener". Westergren also notes that Pandora is not yet profitable, with a projected break even in two years. Link to interview. (lkh)
 
Saturday, April 28 2007
Internet Radio Equality Act to rescue web radio?
Ars Technica reports that Rep. Jay Inslee (D-WA) "will introduce a bill that can overturn the recent ruling requiring webcasters to pay a flat rate per song streamed, rather than the traditional percentage of their profits". Link to Ars Technica. This issue is discussed further by The SaveNetRadio coalition made up of artists, labels, listeners, and webcasters. Link to SaveNetRadio. (lkh)
 
Sunday, April 22 2007
Intelligent Sound participates in the opening of Forskningens Døgn
This year the opening of the national science event "Forskningens Døgn" takes place at Pakhus 11, Dampfærgevej 2, 2100 København Ø. Intelligent Sound will present automatic music recommendation and search in spoken documents. Link for more information. (lkh)
 
Tuesday, April 17 2007
Review: Six Internet Radio Sites Help You Discover New Music
informationweek.com reviews six Internet radio sites. How well do they deliver the music you want to hear, and how well do they help you find new music Pandora, Last.fm, Slacker, TagWorld, Live365.com, and LAUNCHcast are tested: link . (tls)
 
Sunday, April 8 2007
David Byrne on the license issue for Web radio
Talking Heads singer David Byrne is worried that costs may explode for web radio with the new US license rules. Currently his station has about 40.000 listeners a month and costs him about $2000/month. Link to Byrne's blog. (lkh)
 
Friday, April 6 2007
Denmark tops Global Information Technology Report rankings
The World economic forum's report measuring 67 ICT features in more than hundred countries puts Denmark at the top with Sweden and Singapore in second and third place. Link for more information. (lkh)
 
Tuesday, April 3 2007
EMI and iTunes promote DRM free distribution
It was announced Monday that iTunes and EMI will sell DRM free music. Initially at a higher price and in better quality. Link to press release. (lkh)
 
Monday, April 2 2007
Aloha ICASSP 2007
Hawaii will see an ISOUND invasion April 15-20, 2007. Enjoy the following papers: "Analysis of music similarity measures" Jesper Højvang Jensen, Mads G. Christensen, Søren Holdt Jensen. "Castsearch - Context Based Spoken Document Retrieval" Lasse Lohilati Mølgaard, Kasper Winther Jørgensen, Lars Kai Hansen."The Multi-Pitch Estimation Problem: Some New Solutions" Mads G. Christensen, Petre Stoica, Andreas Jakobsson, Søren Holdt Jensen. "On the relevance of spectral features for instrument classification" Andreas Brinch Nielsen, Sigurdur Sigurdsson, Lars Kai Hansen, Jeronimo Arenas-García. "On affine non-negative matrix factorization" Hans Laurberg and Lars Kai Hansen. (lkh)
 
Sunday, April 1 2007
Cognitive component paper accepted at Cognitive Science Conference
The paper "Cognitive components at different time-scales" by Ling Feng and Lars Kai Hansen has been accepted for the conference CogSci2007 taking place in Nashville, Tennessee at the Gaylord Opryland on August 1-4. (lkh)
 
Tuesday, March 27 2007
Are CD/albums about to go extint?
This article in the New York Times discuss the future of the push album, i.e., the producer defined playlist. It is speculated that albums will survive in certain genres, but most likely not in pop music. (lkh)
 
Monday, March 5 2007
Web radio stations worry over new music royalty fees
According to RAIN (Radio and Internet News Letter) the US Copyright Royalty Board (CRB) has announced its decision on Internet radio royalty rates, rejecting all of the arguments made by Webcasters and instead adopting the "per play" rate proposal put forth by SoundExchange. Will Pandora survive? Link for more information. (lkh)
 
Tuesday, February 27 2007
Jobs' thoughts on DRM, iTunes and the music business
iTunes' Steve Jobs shares his thoughts on Digital Rights Management (DRM) in relation to the new music sales channels. (lkh)
 
Thursday, February 8 2007
No players in NY traffic
New York state Senator Carl Kruger proposes a ban on Ipods in the street. Three people have been killed by "Ipod oblivion". Kruger says:"If you want to listen to your iPod, sit down and listen to it...". More at Arstechnica. (lkh)
 
Tuesday, January 30 2007
The worst sound is .... vomiting
A study, set up by Trevor Cox, Salford University, sought opinions on 34 sounds at the website www.sound101.org to learn what makes certain noises so objectionable. (lkh)
 
Friday, January 26 2007
Mads Græsbøll Christensen gets elite award
The Danish science ministry has awarded Mads Græsbøll the prestigeous young elite researcher award 200.000 DKK. (lkh)
 
Friday, January 26 2007
Morten Mørup gets elite travel award
The Danish Science Ministry has presented Morten Mørup with the prestigeous 250.000 DKK elite researcher travel award. (lkh)
 
Thursday, January 18 2007
Doubling of digital music sales in 2006
IFPI report: "Record labels have become digitally literate companies, selling an estimated US$2 billion worth of music online or through mobile phones in 2006 (trade revenues), almost doubling the market in the last year. Digital sales now account for around 10% of the music market as record companies experiment and innovate with an array of business models and digital music products, involving hundreds of licensing partners." (lkh)
 
Wednesday, January 17 2007
Rasmus Olsson's blind signal separation paper appears in JMLR
The cocktail party problem is solved by convolutive ICA using a Kalman filter representation with harmonic excitation. The paper "Linear State-Space Models for Blind Source Separation" by Rasmus Kongsgaard Olsson and Lars Kai Hansen has appeared in Journal of Machine Learning Research (pdf) (lkh)
 
Tuesday, January 16 2007
Sune Lehmann cover story in Nature
Sune’s research concerns mathematical modeling of social networks within, e.g., scientific citations and music. As an example he has analyzed scientific citations patterns in collaboration with two researchers from the University of Copenhagen. The team has discovered a new way of evaluating research performance measures. The results are published in Nature and were announced as a cover story of the Dec 21, 2006 issue. (lkh)
 
Thursday, January 11 2007
End of the music industry?
BBC reports that the new charts will feature a song by Koopa, an unsigned band only distributed via download. Has the powerlaw finally been broken? (lkh)
 
Monday, January 1 2007
Happy New Year 2007
Happy New Year to all Isounders and friends! We are glad that the Isound server is up and running again after a fatal crash on Christmas Eve. The past year 2006 was fantastic for the project with many new results, papers, prices and demos at the ISMIR and NIPS conferences. At the end of 2006 we have said goodbye and good luck to Kaare, Tue, and Kasper. Kaare and Tue have started Epital and Kasper is off for Yale. Visiting researcher Jeronimo Garcia-Arenas is back in Madrid after a fabulous visit. Jeronimo has left a lasting fingerprint on the project and we look forward to the continued collaboration. The Isound postdoc positions will be filled by a number of medium time visitors in 2007 to further boost the basic research programme. Our programmers Bart and Elvis will continue to support and develop the demos. (lkh)
 
Tuesday, December 5 2006
Intelligent Sound at NIPS 2006
Monday evening the WinAmp plugin MIRocket was demoed at the 20th NIPS conference in Vancouver. Tuesday, Dec 5, Jeronimo Arenas-Garcia will present the KOPLS musical genre estimation approach. (lkh)
 
Thursday, November 30 2006
New professor
We are happy to congratualte Søren Holdt Jensen with his promotion as professor. Søren will head the new section for Multimedia Information and Signal procesing at Aalborg University (jl)
 
Tuesday, November 21 2006
Dont raise your voice in Groningen
Multimedia surveillance has been installed in Groningen. Dutch police has already arrested three suspects based on detected aggressive voice. Microphone surveillance systems to be installed in subways etc. (lkh)
 
Thursday, November 16 2006
Beyond the soundbytes
Peter Jenner's report for MusicTank on the music market predicts new business models: “In the future it is likely that the consumer will want access to a huge amount of music and, to have a sense of ownership, purchase a much smaller quantity of music. Perhaps pricing in the future needs to be based on this distinction, which is in the mind of the consumer, rather than in its technological distinctiveness.” (lkh)
 
Tuesday, November 14 2006
Supercomputers Top 500
There is a new list out with the most potent supercomputers world wide. Number one in the list Lawrence Livermore Natl. Lab.'s Blue Gene eserver is reported to deliver a blazing 280 Linpack teraflops! Pride of Denmark: The DCSC cluster at Univ. Copenhagen has dropped to the grim 481'th place after having peaked at the 203'th place in 2005. (lkh)
 
Wednesday, November 8 2006
Tuomas Virtanen PhD Thesis
A doctoral thesis "Sound Source Separation in Monaural Music Signals" by Tuomas Virtanen's can be found here. Demonstrations related to the thesis are also available. The thesis covers work on signal processing and unsupervised learning methods for source separation in one-channel music signals. (tls)
 
Tuesday, November 7 2006
New challenges ahead
After a little more than a year as a postdoc in the Intelligent Sound project, the time has come for for me to seek new challenges. Together with a colleague, I have startet a machine learning company called Epital and we are aiming to apply if not all then a greatest hits of algorithms on commercial problems. First of all I would like to thank everybody in the project for a time of personal development with good discussions, interesting collaborative publications and intriguing demos. It has most certainly been a pleasure and I am looking forward to continuing the good relations in the future. (Kaare Brandt Petersen)
 
Thursday, November 2 2006
Maarten Grachten PhD Thesis
A phd thesis entitled "Expressivity-aware Tempo Transformations of Music Performances using Case Based Reasoning" By Maarten Grachten Pompeu Fabra University is published. The thesis (pdf) can be found here. (tls)
 
Wednesday, November 1 2006
Hit Prediction with machine learning last week in the New Yorker
An article in the New Yorker last week describes how the company platinumblueinc discovers the hits of the future. (tls)
 
Thursday, October 26 2006
Court ruling against Allofmp3
A Danish court has today ordered the internet service provider Tele2, to block access to the russian internet website Allofmp3. Allofmp3 are selling music without the accept of the artists and has lately been subject to a number of accusations. According to russian law, however, is it not illegal to sell music without the permission of the artists and the record companies are now seeking alternative ways to get at Allofmp3, such as the ISPs. Read more at The Register (kbp)
 
Wednesday, October 25 2006
Spar Nord Fonden Price to Mads Græsbøll Christensen
Mads G. Christensen, postdoc in the Intelligent Sound project, receives the prestigeous Spar Nord price for his PhD thesis ""Estimation and Modeling Problems in Parametric Audio Coding" Friday Oct. 27. The thesis is based on ten published papers, four of which are published in IEEE journals. For more information see Mads' homepage at Aalborg University. (lkh)
 
Tuesday, October 24 2006
Pictures from ISMIR availible
A Flikr archive with images from the conference is availible here (tls)
 
Tuesday, October 24 2006
Music Retrieval: A Tutorial and Review
A review of music information retrieval methods by Nicola Orio (University of Padova) is availble here . The review contains chapters about music language, the role of the user, music processing (features etc), systems for music retrieval and evaluation. (tls)
 
Monday, October 16 2006
CD cover war dadaism
It is potato holiday in Denmark. Time for some good clean family entertainment (lkh)
 
Tuesday, October 10 2006
MIRocket presented at ISMIR
The MIRocket plugin was presented at the ISMIR conference today. The purpose of the plugin is to gather data for MIR research, if you want to try the plugin check out the box to the left. If you want access to the data send an email to mirocket@intelligentsound.org . (tls)
 
Wednesday, October 4 2006
ExtremeTech on music recommendation
Who provide the best music recommendations? ExtremeTech has run an informal test of eight services: MusicStrands(8), Liveplasma(7), UpTo11.net(7), Audiri(8), Pandora(9), Mercora(7), Yahoo LAUNCHcast Radio(5), Last.fm(9). As the "marks" (10-scale) indicate Pandora and Last.fm come out on the top (lkh)
 
Sunday, September 17 2006
Show North your music and he will tell you who you are
Andrew North conducts a web based investigation of relations between musical taste and personality. He has found numerous associations between participants’ musical preferences and various aspects of their interpersonal relationships, living arrangements, beliefs, and criminal behaviour. It is concluded that participants’ musical preferences provide a meaningful way of distinguishing different lifestyle choices. Go here for more information and the test (lkh)
 
Friday, September 15 2006
Launch of pre-beta-test
Today we launched the internal pre-testing of the Winamp plugin. It is not a beta yet, and is only for people within the project and close allies. But today the plugin was launched with a set of new features and a corresponding website to track the status of the system, the users, and so on. The thing has the working title MIRocket (The Music Information Rocket), which is very good name for jokes on sci-fi and ... other things. The pre-beta test will helps us identify bugs, load issues and user needs and hopefully we are able to make a more stable and possibly also more simple release sometime in the near future. For now, we are gaining experience about the performance and what problems one should expect. (kbp)
 
Thursday, September 14 2006
Zuner or later
The looks of Microsofts Zune music player aimed at the iPod digital music market has been released. According to BBC News the music player should be able to play video and include a radio. "The digital music entertainment revolution is just beginning," said J Allard, a Microsoft vice president of design and development, who is in charge of the Zune line of products. (lkh)
 
Monday, September 4 2006
Paper for the NIPS conference
The paper "Sparse Kernel Orthonormalized PLS for feature extraction in large data sets", by J. Arenas-Garcia, K. B. Petersen, L. K. Hansen has been accepted for NIPS 2006 (Neural Information Processing Systems). NIPS is a selective conference (acceptance rate 24%) in the areas of machine learning, computational neuroscience and signal processing. The new method presented in the paper is applied to musical genre estimation. (kbp)
 
Friday, September 1 2006
A Friday afternoon with playlist research
Friday afternoon were some people from the project and related friends gathered for the ultimate informal test on playlists: Does the music make sense while having some snacks and a quiet Friday-beer? Tue Lehn-Schioler was setting up a small test environment and played seed songs and the closest ten neighbours to the audience. Notes were carefully written down and will be taken into account in the next few weeks before the pre-launch of the alpha-version due some time here in September. We are all looking forward to that, of course, and enjoying the life as guinea-pigs until then.

(kbp)
 
Wednesday, August 30 2006
Commercial catalog goes online
Vivendi Universal, the world's biggest music group, has signed a deal to make its music catalogue available on a free legal downloads service in US and Canada. They plan to make money by carrying adverts on the site. Link (Anders Meng)
 
Sunday, August 20 2006
ISMIR technical program available
The program for ISMIR 2006 The 7th International Conference on Music Information Retrieval has been posted. The conference takes place Oct 8-12. The ISOUND demo is presented Monday Oct 9, and two ISOUND presentations are scheduled for Thursday morning. Link (lkh)
 
Saturday, August 19 2006
Google gets into music statistics
Thursday August 17, CNET News reported that Google is gettign into the market of music tracking and recommendattion. A small application on your computer keeps track of the music you listen to and sums it all up in an anonymized report on Googles website. (kbp)
 
Tuesday, August 15 2006
The Matrix Cookbook
A short notification to technical researchers in the field of signal processing and machine learning on sound: The Matrix Cookbook now has got its own domain: matrixcookbook.com. The Matrix Cookbook is a free desktop reference on matrices and various standard results regarding differentiation etc. (kbp)
 
Friday, August 4 2006
Musical Chairs
A new massage chair from Singapore based OSIM synchronizes the massage program with features of the user's MP3 music. "Its in-built microchip detects the music's frequency and amplitude and translates them into synchronized massage for your pleasure". A mock-up demo is shown at the retailers website Musical genre seems to be an important feature here. We send a kind warning to all the Slayer fans out there: Don't try this at home! (Lars Kai Hansen)