Done and Liked, Featured, Journal Paper, Publication »

[19 Mai 2014 | Comments Off | 302 views] The preprint of our singing intonation paper is now available! Enjoy! Abstract:
This paper presents a study on intonation and intonation drift in unaccompanied singing and proposes a simple model of reference pitch memory that accounts for many of the effects observed. Singing experiments were conducted with 24 singers of varying ability under 3 conditions (Normal, Masked, Imagined). Over the duration of a recording, approximately 50 seconds, a median absolute intonation drift of 11 cents was observed. While smaller than the median note error (19 cents), drift was significant in 22% of recordings. Drift magnitude did not correlate with other measures of singing accuracy, singing experience or with the presence of conditio…

Done and Liked »

[29 Apr 2014 | Comments Off | 1,361 views]
Well, it’s not really news, but I thought I might say it again, since I hadn’t really written a dedicated blog post on Segmentino: Segmentino segments songs into segments. Also I finally got round to making a lil Segmentino page on Isophonics: The repository, software builds and all can still be found on the Segmentino page on SoundSoftware. That’s it. Check it out!

Done and Liked »

[11 Apr 2014 | Comments Off | 468 views]
I had this code lying around for a long time, and nothing happened to it, so I thought I might just as well give it away so that people can try it out. It is based on my thesis work and employs the Dynamic Bayesian Network (DBN) described in this paper and my PhD thesis). It is not a masterpiece of software engineering and it uses unspeakable amounts of memory (songs of about 6 minutes can use >10GB!, but your standard 4 minute pop song should be doable on today’s laptops). But it still works. You need Matlab to run it, and the matlab binary should be in your path. To test, run ./ testFileList.txt testout/ on OS/X or ./ testFileList.txt testout/. So for people who missed the first link, the code can be downloade…

Done and Liked, Featured »

[3 Apr 2014 | Comments Off | 463 views] Today I presented Tony, the melody annotation software developed by Chris Cannam, George Fazekas and myself, at this year’s Conference of the Society for Education, Music and Psychology Research. Excitingly, this is now the much improved version 0.5, and Chris managed to put new builds of Tony up on the project’s Download page. I had a relatively small audience, maybe 25 people, but reception was generally good, with some people being very excited. Not sure what this means. I hope some people will start using Tony soon. I uploaded my slides here. The conference itself is different from what I’m used to usually. Obviously, it’s focused on education, and there’s actually quite a lot of technology going on. So it’s quite funny to be at a music technology conference where there is essentially no-one I know! Thanks to Chris and George, but also, for their invaluable input: Rachel Bittner, Justin Salamon and Juan Pablo Bello at NYU.…

Conference Paper, Publication »

[28 Jan 2014 | Comments Off | 374 views]
Abstract: We introduce the Audio Degradation Toolbox (ADT) for the controlled degradation of audio signals, and propose its usage as a means of evaluating and comparing the robustness of audio processing algorithms. Music recordings encountered in practical applications are subject to varied, sometimes unpredictable degradation. For example, audio is degraded by low-quality microphones, noisy recording environments, MP3 compression, dynamic compression in broadcasting or vinyl decay. In spite of this, no standard software for the degradation of audio exists, and music processing methods are usually evaluated against clean data. The ADT fills this gap by providing Matlab scripts that emulate a wide range of degrada…

Conference Paper, Publication »

[28 Jan 2014 | Comments Off | 347 views]
Abstract: In the past decade, non-negative matrix factorisation (NMF) and probabilistic latent component analysis (PLCA) have been used widely in automatic music transcription. Despite their successes, these methods only guarantee that the decomposition converges to a local minimum in the cost function. In order to find better local minima, we propose to extend an existing PLCA-based transcription method with the deterministic annealing EM (DAEM) algorithm. The PLCA update rules are modified by introducing a “temperature” parameter. At higher temperatures, general areas of the search space containing good solutions are found. As the temperature is gradually decreased, distinctions in the data are sharpened, resu…

from me to you »

[7 Jan 2014 | Comments Off | 252 views]
Just a quick reminder that you can still apply for a funded PhD position in Signal Processing and Data Mining Tools for the Analysis of Musical Evolution, as previously posted. Job description is here (at

Done and Liked »

[6 Jan 2014 | Comments Off | 473 views]
I’d already sent it out to the world on Twitter and the music-ir list, but here’s just a quick note to say that there’s quite a lot of new software out from me and the C4DM at large. Four different things! Here’s approximately what I wrote on the music-ir list: Segmentino — this is my 2009 segmentation method, which we have recently implemented as a Vamp plugin. This version is what ran at MIREX this year. BeatRoot Vamp Plugin — this is a C++ implementation of Simon Dixon’s trusty beat tracker (originally implemented in Java). This is the second serious beat tracker QM offers (the other one comes with the qm-vamp-plugins). pYIN is “probabilistic YIN” — my generalisation of the time-tested YIN algorithm for monophonic pitch tracking. Also implements YIN, but pYIN is more robust than YIN alone. Also implements a note tracker. Tony is a tool for melody annotation. A little brother of Sonic Visualiser, which is simpler but specialised on note editing. A collaboration between Chris Cannam, George Fazekas and me — WARNING: still a prototype (don’t expect too much). All these are also linked from my Software page.…

Done and Liked »

[25 Nov 2013 | Comments Off | 526 views] I’m very lucky to be able to offer a funded PhD position for next year, in combined MIR and evolutionary music analysis. So if you know someone who’d be a good fit, let them know! They will be based here at the Centre for Digital Music, I will supervise, and Prof. Armand Leroi at Imperial College will advise on evolutionary matters. This should be one of the most fun PhDs in the universe (if that’s your idea of fun—it is mine!). You can find the serious description on the website. Go for it!

Featured, from me to you »

[31 Okt 2013 | Comments Off | 599 views] I’m about to embark on a tour promoting our new album, well: paper, called “The Audio Degradation Toolbox and its Application to Robustness Evaluation” (pdf here). You can read all about the toolbox, listen to examples and get the source code on My little tour first leads me to New York, where I’ll give a little presentation at LabROSA, Dan Ellis’s lab at Columbia University. Only two nights though, then I’ll be off to Curitiba in Brazil, for the ISMIR conference. There I’ll be presenting the ADT “officially” with a poster. Tian Cheng, my first PhD student, has some interesting stuff to tell abo…