“Since 2005, the Italian Research Conference on Digital Libraries has served as an important national forum focused on digital libraries and associated technical, practical, and social issues. IRCDL encompasses the many meanings of the term “digital libraries”, including new forms of information institutions; operational information systems with all manner of digital content; new means of selecting, collecting, organizing, and distributing digital content…"
The 26th of January Federica presented our joint contribution titled “Applications of Duplicate Detection in Music Archives: from Metadata Comparison to Storage Optimisation”. The work focuses on applications of duplicate detection for managing digital music archives. It aims to make this mature music information retrieval (MIR) technology better known to archivists and provide clear suggestions on how this technology can be used in practice. More specifically applications are discussed to complement meta-data, to link or merge digital music archives, to improve listening experiences and to re-use segmentation data.
This weekend the University Hamburg – Institute for Systematic Musicology and more specifically Christian D. Koehn organized the International Symposium on Computational Ethnomusicological Archiving. The symposium featured a broad selection of research topics (physical modelling of instruments, MIR research, 3D scanning techniques, technology for (re)spacialisation of music, library sciences) which all had a relation with archiving musics of the world:
How could existing digital technologies in the field of music information retrieval, artificial intelligence, and data networking be efficiently implemented with regard to digital music archives? How might current and future developments in these fields benefit researchers in ethnomusicology? How can analytical data about musical sound and descriptive data about musical culture be more comprehensively integrated?
In this presentation we describe our experience of working with computational analysis on digitized wax cylinder recordings. The audio quality of these recordings is limited which poses challenges for standard MIR tools. Unclear recording and playback speeds further hinder some types of audio analysis. Moreover, due to a lack of systematical meta-data notation it is often uncertain where a single recording originates or when exactly it was recorded. However, being the oldest available sound recordings, they are invaluable witnesses of various musical practices and they are opportunities to improve the understanding of these practices. Next to sketching these general concerns, we present results of the analysis of pitch content of 400 wax cylinder recordings from Indiana University (USA) and from the Royal Museum from Central Africa (Belgium). The scales of the 400 recordings are mapped and analyzed as a set. It is found that the fifth is almost always present and that scales with four and five pitch classes are organized similarly and differ from those with six and seven pitch classes, latter center around intervals of 170 cents, and former around 240 cents.
I have contributed to the 4th International Digital Libraries for Musicology workshop (DLfM 2017) which was organized in Shanghai, China. It was a satellite event of the ISMIR 2017 conference. Unfortunately I did not mange to find funding to attend the workshop, I did however contribute as co-author to two proceeding papers. Both were presented by Reinier de Valk (thanks again).
This study is a call for action for the music information retrieval (MIR) community to pay more attention to collaboration with digital music archives. The study, which resulted from an interdisciplinary workshop and subsequent discussion, matches the demand for MIR technologies from various archives with what is already supplied by the MIR community. We conclude that the expressed demands can only be served sustainably through closer collaborations. Whereas MIR systems are described in scientific publications, usable implementations are often absent. If there is a runnable system, user documentation is often sparse—-posing a huge hurdle for archivists to employ it. This study sheds light on the current limitations and opportunities of MIR research in the context of music archives by means of examples, and highlights available tools. As a basic guideline for collaboration, we propose to interpret MIR research as part of a value chain. We identify the following benefits of collaboration between MIR researchers and music archives: new perspectives for content access in archives, more diverse evaluation data and methods, and a more application-oriented MIR research workflow.
This work focuses on applications of duplicate detection for managing digital music archives. It aims to make this mature music information retrieval (MIR) technology better known to archivists and provide clear suggestions on how this technology can be used in practice. More specifically applications are discussed to complement meta-data, to link or merge digital music archives, to improve listening experiences and to re-use segmentation data. The IPEM archive, a digitized music archive containing early electronic music, provides a case study.
The first was a collaboration with Frank Desmet, Micheline Lesaffre, Nathalie Ehrlé and Séverine Samson. The contribution is titled Multimodal Analysis of Synchronization Data from Patients with Dementia. It details a famework to analyze data in an experiment for patients with dementia.
The active listening demo movie above should explain the aim system succinctly. It shows two different ways to provide ‘context’ to audio playing in the room. In the first instance beats information is used to synchronize smartphones and flash the screen, the second demo shows a tactile feedback device responding to beats. The device is a soundbrenner pulse tactile metronome and was kindly sponsored by the company that sells these.
On Saturday the eight of April I gave a workshop on the ESP 32 micro controller at Newline, the yearly hackerspace conference of Hackerspace Ghent. The aim was to provide a hands-on introduction. The participants had to program to make the ESP execute the following:
Blink
Connecting to wifi
Sending data from the ESP32
Sending data using TCP
Broadcast data over UDP
Broadcast data using OSC over UDP
Broadcasting sensor data over UDP and OSC
Mesh Networking
At the start of the workshop I gave a presention as an introduction.
It was attended by a mix of (Ethno)musicologists, archivists, computer scientists and people identifying themselves as more than one of these categories by varying degrees. This mix ensured a healthy discussion and talks by Frans Wiering, Willard McCarthy, Emilia Gomez, and may more provided ample source material to discuss. These discussions ranged from the abstracts around schemata down to concrete of software tools for archive management.
On a more personal side the workshop did provide useful insights to contextualize my research and help form ideas that can be condensed in my PhD dissertation.
The xOSC board by x-io technologies looks like a very nice solution in many interactive wireless setups. Judging from the specifications and documentation it offers a lot of value. It is basically a small WiFi transmitter with some sensors and a battery attached to it. The board also has some drawbacks. 1) It is expensive at about € 180. This is especially problematic if you need about five or so for your application.2) It seems that it is also hard to add extra sensors via SPI or I²C. 3)The battery needs to be removed to charge, which makes it harder to build into a fixed enclosure. This post describes an alternative based on the ESP32 platform that addresses these shortcomings.
The ESP32 is a micro-controller with a WiFi transmitter which can be programmed using the Arduino environment. Sparkfun has a thing called the ESP32 Thing which contains the ESP32 chip. It can be used to build an xOSC alternative.
It costs about 20$, when you add a battery 5$ and a sensor 20$ (IMU) you end up with a 45$ price tag. The price of course depends on which exact sensor/battery you need for your application. A 500mAh lasts about two hours when sending 66 messages per second over WiFi (using UDP).
The ESP32 Thing supports the Arduino environment which potentially allows you to use all available Arduino libraries and supported sensors. However, some libraries do contain hardware specific instructions which are often not ported yet. Since the hardware is rather new – large scale production started only 3 months ago – not many libraries have been ported. Fortunately a lot of libraries simply work without any changes. At hackaday they have been testing a few: ESP32 and Arduino libraries. I had success with the BNO55 library, it did not need any changes. The OSC library did need some small changes to operate as expected.
The Thing contains a battery charging circuit. Once embedded into an enclosure the battery can stay in place. The software running on the device even keeps running when changing power sources.
Attached to this post you can find modifications to the Andriod OSC library that enable it to run on the ESP32: ESP32-Arduino-OSC-library together with a patch that sends random data over OSC. This should enable you to build an xOSC alternative.
Some drawbacks of the ESP32 is that the supporting software is quite immature. There is a Bluetooth chip on the ESP32 which is currently not supported in the Arduino environment. The setup can be somewhat challenging. The documentation can be improved. Some of the ESP32 Things seem to be unable to connect to old WiFi routers which can be problematic.
Last Saturday, October eight 2016, IPEM was present at the opening event of the digital week. A small video report was made for VRT news, unfortunately our contribution did not make the cut.
Van 8 tot en met 16 oktober 2016 loopt de elfde editie van de De Digitale Week. Plaatselijke organisaties in heel Vlaanderen en Brussel organiseren tijdens deze week diverse laagdrempelige activiteiten waarbij het gebruik van multimedia centraal staat, steeds gratis of zeer goedkoop, en open voor zowel beginners als mensen met wat meer ervaring. Daarnaast loopt er tijdens de Digitale Week een grote publiciteitscampagne die aandacht vraagt voor de thema’s e-inclusie en mediawijsheid.
This weekend IPEM, the research institute in musicology of University Ghent, was present at Parklife 2016. Parklife is a music festival with a special focus on interactive music installations aimed at children. Two of those were provided by IPEM.
The first installation was a trampoline that triggered sounds. Two trampoline were provided with a pressure sensor. An Axoloti provides the sonic feedback. A simple but fun experience especially for younger children.
The second installation was more involved. It consisted of a bike – controlled by a first participant – that provided the speed of falling blocks that a second participant had to step on. When the second participant triggered the blocks on time a melody appeared. The video above makes it more clear.
The aim of the thesis was to design and develop a system to automatically synchronize streams of incoming sensor data in real-time. Ward followed up on an idea that was described in an article called Synchronizing Multimodal Recordings Using Audio-To-Audio Alignment. The extended abstract can be consulted. The remainder of the thesis is in Dutch.
For the thesis Ward developed a Max/MSP object to read data from sensors together with audio. Also provided by Ward is an object to synchronize audio and data in real-time. The objects are depicted above.
I have given a presentation at the the Newline conference, a yearly event organized by the Hackerspace Ghent. It was about:
“In this talk I will give a practical overview on how to connect hard- and software components for musical applications. Next to an overview there will be demos! Do you want to make a musical instrument using a light sensor? Use your smartphone as an input device for a synth? Or are you simply interested in simple low-latency communication between devices? Come to this talk! More concretely the talk will feature the Axoloti audio board, Teensy micro-controller with audio board, MIDI and OSC protocols, Android MIDI features and some sensors.”
During the presentation the hard and software components were demonstrated. More concretely an introduction was given to the following:
This morning, the 30th of October 2015, I gave a lecture on Music Information Retrieval in general and two MIR-tasks in particular. The two more detailed tasks were tone scale analysis and acoustic fingerprinting.
During the lecture some live demonstrations were done with Panako and Tarsos. Also some examples from TarsosDSP were used. Excerpts of the music used is available here, this is especially interesting if you want to repeat the demos. Sonic visualizer, Music21 and MuseScore were also mentioned during the lecture.
Kelsec Systems developed a nice sensor for measuring running impact, the TgForce Running Impact Sensor. The sensor comes with an IOS application but has no available counterpart on Android. To interface with the sensor on Android I needed to create some glue code. The people of Kelsec Systems were kind enough to mail some documentation about the protocol and with that information I got to work.
The TgForce Sensor Android code is available on GitHub, together with some documentation which is available below as well:
TgForce Impact Running Sensor Andoid API
The TgForceSensor repository contains Android code to interface with the TgForce Impact Running Sensor. The TgForce sensor is a Bluetooth LE device that measures tibial shock. It follows the
Bluetooth LE standards and is relatively easy to interface with.
This repository contains Android code to interface with the device. The protocol is encoded in the source code and is documented in the readme.
For the opening of the Sport Science Laboratory – Jacques Rogge of University Ghent I have created a demo of a system to visualize running impact. The demo can be seen starting at 45s in the video below.
Aangezien heel wat joggers met muziek trainen, wilden onderzoekers van het IPEM (het onderzoekscentrum van de afdeling Musicologie, Vakgroep Kunst-, Muziek-, en Theaterwetenschappen aan de UGent) nagaan of het tempo van muziek de pasfrequentie tijdens het lopen kan beïnvloeden. Eerdere studies hadden al aangetoond dat muziek een motiverend effect kan hebben op sportprestaties en dat een hogere pasfrequentie blessurepreventief kan werken.
Een neerslag van het onderzoek is te lezen in het artikel Spontaneous Entrainment of Running Cadence to Music Tempo. Het persbericht werd goed opgepikt door de media en ook de lokale televisiezender AVS vertoonde interesse. Een cameraploeg kwam langs en dit resulteerde in volgend verslag. In het verslag spelen mijn vriendin en ikzelf een figurantenrol. De hoofdrol is weggelegd voor Dieter.
The Mi Band is a bracelet with some sensors, three RGB leds and a vibration motor. It is marketed as an activity tracker and notifier. It is a neat little device that communicates via Bluetooth LE and has a battery life of around 30 days. It would be nice if it could be used for whatever purpose you want but alas, its API is not very open. This blog post gives pointers to useful resources and tips to make it work with your own code.
I would advise against installing the official Mi Band app, if you want to use it with custom code. The app upgrades the firmware to the latest version and it seems that Xiaomi is obfuscating the protocol more and more with each version. I was able to send vibrate and led commands to a Mi Band with firmware version 10.0.9.3. With the previously mentioned sources and the flow described to the right the device reacts to commands. I used an Android device. The flow:
Pair with the Mi Band in the Android Bluetooth setting.
In your code, connect to the paired device. Save the device address, you will need it later.
Send a pair command to the device. This is part of the Mi Band protocol and has nothing to do with the previous Bluetooth pairing. If all goes well it reacts with a 2. See here
Send user info. This step is crucial and not trivial. The user info needs to be encoded in a certain way and is CRC’d with the device address. The following is an example implementation of the Mi Band user info encoding
Now you can send vibrate or other commands.
Some notes: the self-test command works without the set user step. For Android the Mi Band protocol implementation in Java works well. To check the firmware version of the device, call the get device info characteristic. The last bytes, interpreted as an integer, define the version info. For my device it is 10.9.3.2:
Another note: the set user info needs to be called with a 1 as type the first time the band is used. This is done with new UserInfo(20111111, 1, 32, 180, 55, "NM", 1) with the Android sdk by GitHub user pangliang. This sets and overwrites the user info. The next times you do not want to overwrite the info and the type needs to be zero.
The article titled “Synchronizing Multimodal Recordings Using Audio-To-Audio Alignment” by Joren Six and Marc Leman has been accepted for publication in the Journal on Multimodal User Interfaces. The article will be published later this year. It describes and tests a method to synchronize data-streams. Below you can find the abstract, pointers to the software under discussion and an author version of the article itself.
Synchronizing Multimodal Recordings Using Audio-To-Audio Alignment An Application of Acoustic Fingerprinting to Facilitate Music Interaction Research
Abstract:Research on the interaction between movement and music often involves analysis of multi-track audio, video streams and sensor data. To facilitate such research a framework is presented here that allows synchronization of multimodal data. A low cost approach is proposed to synchronize streams by embedding ambient audio into each data-stream. This effectively reduces the synchronization problem to audio-to-audio alignment. As a part of the framework a robust, computationally efficient audio-to-audio alignment algorithm is presented for reliable synchronization of embedded audio streams of varying quality. The algorithm uses audio fingerprinting techniques to measure offsets. It also identifies drift and dropped samples, which makes it possible to find a synchronization solution under such circumstances as well. The framework is evaluated with synthetic signals and a case study, showing millisecond accurate synchronization.
The algorithm under discussion is included in Panako an audio fingerprinting system but is also available for download here. The SyncSink application has been packaged separately for ease of use.
To use the application start it with double click the downloaded SyncSink JAR-file. Subsequently add various audio or video files using drag and drop. If the same audio is found in the various media files a time-box plot appears, as in the screenshot below. To add corresponding data-files click one of the boxes on the timeline and choose a data file that is synchronized with the audio. The data-file should be a CSV-file. The separator should be ‘,’ and the first column should contain a time-stamp in fractional seconds. After pressing Sync a new CSV-file is created with the first column containing correctly shifted time stamps. If this is done for multiple files, a synchronized sensor-stream is created. Also, ffmpeg commands to synchronize the media files themselves are printed to the command line.
This work was supported by funding by a Methusalem grant from the Flemish Government, Belgium. Special thanks goes to Ivan Schepers for building the balance boards used in the case study. If you want to cite the article, use the following BiBTeX:
@article{six2015multimodal,
author = {Joren Six and Marc Leman},
title = {{Synchronizing Multimodal Recordings Using Audio-To-Audio Alignment}},
issn = {1783-7677},
volume = {9},
number = {3},
pages = {223-229},
doi = {10.1007/s12193-015-0196-1},
journal = {{Journal of Multimodal User Interfaces}},
publisher = {Springer Berlin Heidelberg},
year = 2015
}
A microcontroller fitted with an electret microphone and a microSD card slot. It can record audio in real-time together with sensor data.
Synchronized streams in Sonic Visualizer. Here you can see two channel audio synchronized with accelerometer data (top, green) and balanceboard data (bottom, purple).
The synchronized data from the two webcams, accelerometer and balanceboard in ELAN. From top to bottom the synchronized streams are two video-streams, balance-board data (red), accelerometer-data (green) and audio (black).
Conceptual drawing used as a basis for the SyncSync application. A reference stream (blue) can be synchronized with streams one and two. It allows a workflow where streams are started and stopped (red) or start before the reference stream (green).
SyncSink Synchronize media files. A user-friendly interface to synchronize media and data files. First a reference media-file is added using drag-and-drop. The audio steam of the reference is extracted and plotted on a timeline as the topmost box. Subsequently other media-files are added. The offsets with respect to the reference are calculated and plotted. CSV-files with timestamps and data recorded in sync with a stream can be attached to a respective audio stream. Finally, after pressing Sync!, the data and media files are modified to be exactly in sync with the reference.
Multimodal recording system diagram. Each webcam has a microphone and is connected to the pc via USB. The dashed arrows represent analog signals. The balance board has four analog sensors but these are simplified to one connection in the schematic. The analog output of the microphones is also recorded through the DAQ. An analog accelerometer is connected with a microcontroller which also records audio.
Two streams of audio with fingerprints marked. Some fingerprints are present in both streams (green, O) while others are not (red, x). Matching fingerprints have the same offset, indicated by the dotted lines.
This post explains how to do real-time pitch-shifting and audio time-stretching in Java. It uses two components. The first component is a high quality software C++ library for audio time-stretching and pitch-shifting C++ called Rubber Band. The second component is a Java audio library called TarsosDSP. To bridge the gap between the two JNI is used. Rubber Band provides a JNI interface and starting from the currently unreleased version 1.8.2, makefiles are provided that make compiling and subsequently using the JNI version of Rubber Band relatively straightforward.
However, it still requires some effort to control real-time pitch-shifting and audio time-stretching from java. To make it more easy some example code and documentation is available in a GitHub repository called RubberBandJNI. It documents some of the configuration steps needed to get things working. It also offers precompiled libraries and documents how to compile those for the following systems:
This post describes how to decode MP3’s using an already compiled ffmpeg binary on android. Using ffmpeg to decode audio on Android has advantages:
It supports about every audio format known to man. Three channel flac, vorbis with 32 bit samples, … do not pose a problem.
Extracting audio from video container formats is supported. Accessing the first audio stream from mkv, avi, mov,… just works.
Decoding audio frames is more efficient using native code than often buggy Java decoders.
Resampling and downmixing is supported. If you want to resample incoming audio to e.g. 44.1kHz and only want single channel audio this is easily achievable.
The main disadvantage is that you need an ffmpeg build for your Android device. Luckily some poor soul already managed to compile ffmeg for Android for several architectures. The precompiled ffmpeg binaries for Android are available for download and are mirrored here as well.
To bridge the ffmpeg binary and the java world TarsosDSP contains some glue code. The AndroidFFMPEGLocator is responsible to find and extract the correct binary for your Android device. It expects these ffmpeg binaries in the assets folder of your Android application. When the correct ffmpeg binary has been extracted and made executable the PipeDecoder is able to call it. The PipeDecoder calls ffmpeg so that decoded, downmixed and resampled PCM samples are streamed into the Java application via a pipe, which explains its name.
With the TarsosDSP Android library the following code plays an MP3 from external storage:
new AndroidFFMPEGLocator(this);
new Thread(new Runnable() {
@Override
public void run() {
File externalStorage = Environment.getExternalStorageDirectory();
File mp3 = new File(externalStorage.getAbsolutePath() , "/audio.mp3");
AudioDispatcher adp;
adp = AudioDispatcherFactory.fromPipe(mp3.getAbsolutePath(),44100,5000,2500);
adp.addAudioProcessor(new AndroidAudioPlayer(adp.getFormat(),5000, AudioManager.STREAM_MUSIC));
adp.run();
}
}).start();
This code just works if the application has the READ_EXTERNAL_STORAGE permission, includes a recent TarsosDSP-Android.jar, is ran on one of the supported ffmpeg architectures and has these binaries available in the assets folder.
UGent
2018.IRCDL-duplicates-presentation.pdf and 2018.IRCDL-duplicates-article-author-version.pdf
I have contributed to the 4th International Digital Libraries for Musicology workshop (DLfM 2017) which was organized in Shanghai, China. It was a satellite event of the 
The 25th anniversary edition of the 
On Saturday the eight of April I gave a workshop on the 
From 27 to 31 March 2017 I have attended a workshop on 
The xOSC board by 
I have given a presentation at the the 
The 
