-
Notifications
You must be signed in to change notification settings - Fork 1
Expand file tree
/
Copy pathindex.html
More file actions
61 lines (60 loc) · 65.5 KB
/
index.html
File metadata and controls
61 lines (60 loc) · 65.5 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
<!DOCTYPE html>
<html lang="en-US">
<head>
<meta charset="utf-8">
<meta name="viewport" content="width=device-width,initial-scale=1">
<title>faroit</title>
<meta name="generator" content="VuePress 1.9.10">
<link rel="stylesheet" href="https://fonts.googleapis.com/css?family=Fira+Sans:300,400,500&display=swap">
<meta name="description" content="">
<link rel="preload" href="/assets/css/0.styles.1138f10f.css" as="style"><link rel="preload" href="/assets/js/app.62bc949a.js" as="script"><link rel="preload" href="/assets/js/2.63207ef2.js" as="script"><link rel="preload" href="/assets/js/1.0551b2b6.js" as="script"><link rel="preload" href="/assets/js/24.6ba64fe2.js" as="script"><link rel="preload" href="/assets/js/5.c1fac5e0.js" as="script"><link rel="prefetch" href="/assets/js/11.4bcf2067.js"><link rel="prefetch" href="/assets/js/12.5463394b.js"><link rel="prefetch" href="/assets/js/13.3cb72fa4.js"><link rel="prefetch" href="/assets/js/14.04835625.js"><link rel="prefetch" href="/assets/js/15.94ee8fd5.js"><link rel="prefetch" href="/assets/js/16.09cfc6e2.js"><link rel="prefetch" href="/assets/js/17.c65e5e35.js"><link rel="prefetch" href="/assets/js/18.b49b2d71.js"><link rel="prefetch" href="/assets/js/19.35b01c9f.js"><link rel="prefetch" href="/assets/js/20.1fbb6ee6.js"><link rel="prefetch" href="/assets/js/21.11b1e553.js"><link rel="prefetch" href="/assets/js/22.a3a80b8d.js"><link rel="prefetch" href="/assets/js/23.cf8a2e28.js"><link rel="prefetch" href="/assets/js/25.f2d993c8.js"><link rel="prefetch" href="/assets/js/3.3d9382a4.js"><link rel="prefetch" href="/assets/js/4.1a61faed.js"><link rel="prefetch" href="/assets/js/6.bb819f12.js"><link rel="prefetch" href="/assets/js/7.f7140dfc.js"><link rel="prefetch" href="/assets/js/8.d818c39a.js"><link rel="prefetch" href="/assets/js/vendors~docsearch.5a59a81b.js">
<link rel="stylesheet" href="/assets/css/0.styles.1138f10f.css">
</head>
<body>
<div id="app" data-server-rendered="true"><div class="theme-container no-sidebar"><header class="navbar"><div class="sidebar-button"><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" role="img" viewBox="0 0 448 512" class="icon"><path fill="currentColor" d="M436 124H12c-6.627 0-12-5.373-12-12V80c0-6.627 5.373-12 12-12h424c6.627 0 12 5.373 12 12v32c0 6.627-5.373 12-12 12zm0 160H12c-6.627 0-12-5.373-12-12v-32c0-6.627 5.373-12 12-12h424c6.627 0 12 5.373 12 12v32c0 6.627-5.373 12-12 12zm0 160H12c-6.627 0-12-5.373-12-12v-32c0-6.627 5.373-12 12-12h424c6.627 0 12 5.373 12 12v32c0 6.627-5.373 12-12 12z"></path></svg></div> <a href="/" aria-current="page" class="home-link router-link-exact-active router-link-active"><!----> <span class="site-name">faroit</span></a> <div class="links"><!----> <nav class="nav-links can-hide"><div class="nav-item"><a href="/#about-me" class="nav-link">
About Me
</a></div><div class="nav-item"><a href="/#scientific-service" class="nav-link">
Scientific Service
</a></div><div class="nav-item"><a href="/#software" class="nav-link">
Software
</a></div><div class="nav-item"><a href="/#datasets" class="nav-link">
Datasets
</a></div><div class="nav-item"><a href="/#publications" class="nav-link">
Publications
</a></div> <!----></nav></div></header> <div class="sidebar-mask"></div> <aside class="sidebar"><nav class="nav-links"><div class="nav-item"><a href="/#about-me" class="nav-link">
About Me
</a></div><div class="nav-item"><a href="/#scientific-service" class="nav-link">
Scientific Service
</a></div><div class="nav-item"><a href="/#software" class="nav-link">
Software
</a></div><div class="nav-item"><a href="/#datasets" class="nav-link">
Datasets
</a></div><div class="nav-item"><a href="/#publications" class="nav-link">
Publications
</a></div> <!----></nav> <!----> </aside> <main aria-labelledby="main-title" class="home"><header class="hero"><img src="./hero2.jpg" alt="hero"> <h1 id="main-title">
Fabian-Robert Stöter
</h1> <p class="description">
Head of Research at Audioshake.ai, Frankfurt, Germany
</p> <p class="social"><a rel="me" href="https://sigmoid.social/@faro" alt="Mastodon"><i class="fab fa-mastodon fa-2x"></i></a> <a href="https://github.com/faroit" alt="Github"><i class="fab fa-github fa-2x"></i></a> <a href="https://scholar.google.com/citations?user=7HsSdqwAAAAJ&hl=en" alt="Google Scholar"><i class="fas fa-graduation-cap fa-2x"></i></a> <a href="mailto:fabian-robert.stoter@inria.fr" alt="Email"><i class="fas fa-envelope fa-2x"></i></a> <a href="https://orcid.org/0000-0002-2534-1165" alt="Orcid"><i class="fab fa-orcid fa-2x"></i></a></p> <!----></header> <div class="features"><div class="feature"><h2><i class="fas fa-music" style="color:#F27405;"></i> Music Processing
</h2> <p><span>I have a background in digital signal processing (DSP) and have worked on a wide range of audio and related tasks, including speech- and audio processing, music analysis and music information retrieval.</span></p></div><div class="feature"><h2><i class="fas fa-brain" style="color:#36A8A7;"></i> Audio-AI
</h2> <p><span>I have a profound understanding of deep audio-ml. I am specifically interested in the tasks of <a href="https://github.com/faroit/CountNet"> source count estimation</a> and <a href="https://sigsep.github.io">audio source separation</a>. I am leading the research team at <a href="https://audioshake.ai">Audioshake.ai</a> that created the best performing music separation and lyric transcription models.</span></p></div><div class="feature"><h2><i class="fab fa-pagelines" style="color:#88A61F;"></i> Eco-ML
</h2> <p><span>I was involved in <a href="https://plantnet.org">Pl@ntNet</a> as part of <a href="https://www.cos4cloud-eosc.eu">Cos4Cloud 🇪🇺</a> citizen science project. I was also working on ML for ecoacoustics, analyzing sounds of 🦓 using <a href="https://audiolog.fr">mobile audio loggers</a>.</span></p></div></div> <div class="theme-default-content content__default"><h1 id="about-me"><a href="#about-me" class="header-anchor">#</a> About me</h1> <p><img src="frs.png" width="86" style="float:left;margin-right:1em;"> Since 2021, I'm head of research at <a href="https://www.audioshake.ai" target="_blank" rel="noopener noreferrer">audioshake.ai<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a> working on music-ml research. Before, I was a postdoctoral researcher at the <a href="http://www-sop.inria.fr/teams/zenith/pmwiki/pmwiki.php/Main/HomePage" target="_blank" rel="noopener noreferrer">Inria and University of Montpellier<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a>, France. I did my Ph.D (Dr.-Ing.) at the <a href="https://www.audiolabs-erlangen.de/" target="_blank" rel="noopener noreferrer">International Audio Laboratories Erlangen<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a> (is a joint institution of <a href="https://www.iis.fraunhofer.de" target="_blank" rel="noopener noreferrer">Fraunhofer IIS<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a> and <a href="https://fau.de" target="_blank" rel="noopener noreferrer">FAU Erlangen-Nürnberg<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a>) in Germany supervised by <a href="https://www.audiolabs-erlangen.de/fau/professor/edler" target="_blank" rel="noopener noreferrer">Bernd Edler<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a>. My dissertation titled <strong>«Separation and Count Estimation for Audio Sources Overlapping in Time and Frequency»</strong> can be viewed <a href="https://opus4.kobv.de/opus4-fau/frontdoor/index/index/docId/13114" target="_blank" rel="noopener noreferrer"><i class="fas fa-file-pdf"></i> here<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a>. Before, I graduated in electrical engineering / communication engineering from the <a href="https://www.uni-hannover.de" target="_blank" rel="noopener noreferrer">University of Hannover, Germany<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a>. An extended CV is available <a href="https://github.com/faroit/resume/releases/download/v1.0.2/stoeter_resume.pdf" target="_blank" rel="noopener noreferrer"><i class="fas fa-file-pdf"></i> here<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a>.</p> <h2 id="current-research-interests"><a href="#current-research-interests" class="header-anchor">#</a> Current Research Interests</h2> <ul><li><p><strong>Deep learning on data hubs</strong>: I am interested multi-modal foundation models that can learn the relations between the different modalities to reconstruct or enhance missing or degraded data.</p></li> <li><p><strong>User-centered AI for audio data</strong>: I want to develop new methods and tools for users with domain knowledge to deliver interpretable audio models. Furthermore, <em>evaluation</em> of audio processing tasks is often done in a computational manner, due to the lack of expertise from signal processing researchers in organizing perceptual evaluation campaigns.</p></li> <li><p><strong>Ecological machine-learning</strong>: I want to play a role in reducing the carbon footprint of my work. Reducing the size of datasets speeds up training and therefore saves energy. Reducing the computational complexity of models is an active research topic, with strongly investigated ideas like quantization, pruning or compression. Inspired by current trends in differentiable signal processing, I want to convert deep models so that they can be deployed on edge devices.</p></li></ul> <h2 id="press-media-interviews"><a href="#press-media-interviews" class="header-anchor">#</a> Press/Media Interviews</h2> <ul><li><i class="fas fa-podcast"></i> <strong>02/2023</strong> <a href="https://ici.radio-canada.ca/ohdio/premiere/emissions/jusquau-bout/episodes/683554/rattrapage-du-vendredi-27-janvier-2023" target="_blank" rel="noopener noreferrer">"l'intelligence artificielle et du droit d'auteur", Radio-Canada (French)<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a></li> <li><i class="fas fa-podcast"></i> <strong>12/2022</strong> <a href="https://www.deutschlandfunkkultur.de/jahresrueckblick-und-vorausschau-ki-musik-und-metaverse-dlf-kultur-e70bf447-100.html" target="_blank" rel="noopener noreferrer">"Jahresrückblick und Vorausschau: KI Musik und Metaverse", Deutschlandfunk Kultur (German)<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a></li> <li><i class="fas fa-podcast"></i> <strong>02/2022</strong> <a href="https://ici.radio-canada.ca/ohdio/premiere/emissions/jusquau-bout/episodes/605268/rattrapage-du-vendredi-11-fevrier-2022/1" target="_blank" rel="noopener noreferrer">"L'intelligence artificielle au profit des stems musicaux", Radio-Canada (French)<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a></li> <li><i class="fas fa-podcast"></i> <strong>12/2021</strong> <a href="https://www.deutschlandfunkkultur.de/recycling-von-songs-wie-ki-neue-musik-generiert-dlf-kultur-90e01124-100.html" target="_blank" rel="noopener noreferrer">"Recycling von Songs: Wie KI neue Musik generiert", Deutschlandfunk Kultur (German)<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a></li></ul> <h1 id="scientific-service"><a href="#scientific-service" class="header-anchor">#</a> Scientific Service</h1> <h2 id="editing"><a href="#editing" class="header-anchor">#</a> Editing</h2> <ul><li><strong>Journals</strong>: Topic Editor for ML-Audio for the <a href="https://joss.theoj.org/papers/edited_by/@faroit" target="_blank" rel="noopener noreferrer">Journal of Open Source Software<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a>.</li></ul> <h2 id="reviewing"><a href="#reviewing" class="header-anchor">#</a> Reviewing</h2> <ul><li><strong>Journals:</strong> <a href="https://joss.theoj.org/papers/reviewed_by/@faroit" target="_blank" rel="noopener noreferrer">Journal of Open Source Software<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a>, <s>IEEE <a href="https://ieeexplore.ieee.org/xpl/RecentIssue.jsp?punumber=6570655" target="_blank" rel="noopener noreferrer">Transaction in Audio, Speech and Language Processing<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a>,</s> <s><a href="https://ieeexplore.ieee.org/xpl/RecentIssue.jsp?punumber=97" target="_blank" rel="noopener noreferrer">Signal Processing Letters<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a></s>, <a href="https://www.eurasip.org/" target="_blank" rel="noopener noreferrer">EURASIP<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a>,</li> <li><strong>Conferences:</strong> <a href="https://www.ismir.net" target="_blank" rel="noopener noreferrer">ISMIR<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a>, <s><a href="http://ieeeicassp.org/" target="_blank" rel="noopener noreferrer">ICASSP<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a></s>, <a href="https://eusipco2020.org/" target="_blank" rel="noopener noreferrer">EUSIPCO<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a>, <a href="https://www.dafx.de/" target="_blank" rel="noopener noreferrer">DAFx<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a></li></ul> <h2 id="student-supervision"><a href="#student-supervision" class="header-anchor">#</a> Student Supervision</h2> <ul><li><a href="https://www.linkedin.com/in/laura-ibnz" target="_blank" rel="noopener noreferrer">Laura Ibáñez Martínez<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a>, Master Student, Co-Supervision: Master thesis: "MIDI-AudioLDM: MIDI-Conditional Text-to-Audio Synthesis Using ControlNet on AudioLDM" <a href="https://lauraibnz.github.io/docs/TFM.pdf" target="_blank" rel="noopener noreferrer"><i class="fas fa-file-pdf"></i> Thesis<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a>, Summer 2023</li> <li><a href="https://twitter.com/joimort" target="_blank" rel="noopener noreferrer">Johannes Imort<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a>, Master student, RWTH Aachen (Germany), Internship <em>"Sound Activity Detection"</em>. Winter 2022</li> <li><a href="https://onedas.github.io/" target="_blank" rel="noopener noreferrer">Jinsung Kim<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a> and Yeong-Seok Jeong, Master students, Korea University, (Winter 2022/2023) Internship on <em>"Unsupervised Music Separation"</em> (Summer 2022).</li> <li>Michael Tänzer, PhD student, Fraunhofer IDMT (Germany), (Summer 2021), Internship on audio tagging.</li> <li><a href="https://synergy.st-andrews.ac.uk/cbd/person/lm354/" target="_blank" rel="noopener noreferrer">Lucas Mathieu<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a>, Master student, AgroParistech (France), Master thesis <em>"Listening to the Wild"</em> (03/2020). Theoretical research on self-supervised learning using data from animal-born loggers (<a href="https://muse.edu.umontpellier.fr/" target="_blank" rel="noopener noreferrer">MUSE project<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a>). Lucas was accepted as a PhD student after master thesis.</li> <li><a href="https://www.linkedin.com/in/clarajacintho/" target="_blank" rel="noopener noreferrer">Clara Jacintho<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a> and Delton Vaz, Bachelor Thesis, PolyTech Montpellier (France), <em>"Machine Learning for Audio on the Web"</em> (12/2019). Research on web based separation architectures. Resulted in a paper submitted to the <a href="https://webaudioconf2021.com" target="_blank" rel="noopener noreferrer">Web Audio Conference 2021<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a>.</li> <li><a href="https://www.audiolabs-erlangen.de/fau/assistant/mack" target="_blank" rel="noopener noreferrer">Wolfgang Mack<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a>, Master Thesis (FAU Erlangen-Nürnberg, Germany), <em>"Investigations on Speaker Separation using Embeddings obtained by Deep Learning"</em>, (05/2017), Wolfgang was accepted as PhD student after master thesis.</li> <li><a href="https://ca.linkedin.com/in/ecmjohnson" target="_blank" rel="noopener noreferrer">Erik Johnson<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a>, <a href="https://www.daad.de/rise/en/" target="_blank" rel="noopener noreferrer">DAAD<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a> Research internship, (Carleton University, Canada), <a href="https://github.com/craffel/mir_eval/pull/199" target="_blank" rel="noopener noreferrer"><em>"Open-Source Implementation of Multichannel BSSEval in Python"</em><span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a> (03/2014).</li> <li><a href="https://nils-werner.github.io/" target="_blank" rel="noopener noreferrer">Nils Werner<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a>, Master Thesis, (FAU Erlangen-Nürnberg, Germany), <em>"Parameter Estimation for Time-Varying Harmonic Audio Signals"</em>, (02/2014), Nils was accepted as PhD student after master thesis.</li> <li><a href="https://www.linkedin.com/in/jeremy-hunt-94204b85/" target="_blank" rel="noopener noreferrer">Jeremy Hunt<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a>, <a href="https://www.daad.de/rise/en/" target="_blank" rel="noopener noreferrer">DAAD<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a> research internship, (Rice University, USA)</li> <li>Bufei Liu, Master, Research Internship (Shanghai University, China), 2014.</li> <li><a href="https://www.idc.tf.fau.de/person/aravindh-krishnamoorthy" target="_blank" rel="noopener noreferrer">Aravindh Krishnamoorty<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a>, Master, Internship, 2014</li> <li>Ercan Berkan, Master Thesis, (Bilkent University, Turkey), <em>"Music Instrument Source Separation"</em>, 3/2013</li> <li><a href="https://www.linkedin.com/in/shujieguo" target="_blank" rel="noopener noreferrer">Shujie Guo<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a>, Master, Research Internship, (FAU Erlangen-Nürnberg, Germany)</li></ul> <h2 id="teaching"><a href="#teaching" class="header-anchor">#</a> Teaching</h2> <h3 id="graduate-programs"><a href="#graduate-programs" class="header-anchor">#</a> Graduate Programs</h3> <ul><li><strong>2024</strong>: Guest-Lecture: "Is music separation interesting in the age of generative AI?", Music Information Retrieval (MIR) program, Master-2, Telecom-ParisTech</li> <li><strong>2021</strong>: <a href="https://audiolabs-erlangen.de/fau/professor/mueller/teaching/2021s_dla" target="_blank" rel="noopener noreferrer">Guest-Lecture: Selected Topics in Deep Learning for Audio, Speech, and Music Processing<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a>, Music Source Separation, University of Erlangen (Germany).</li> <li><strong>2020</strong>: <a href="https://www.polytech.umontpellier.fr/partenariats/stages-et-projets" target="_blank" rel="noopener noreferrer">Research Internship<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a> (Master, Stage 5), PolyTech Montpellier</li> <li><strong>2018, 2019</strong>: Introduction to Deep Learning, Master 2, PolyTech Montpellier</li> <li><strong>2016</strong>: <a href="https://github.com/audiolabs/APSRR-2016" target="_blank" rel="noopener noreferrer">Reproducible Audio Research Seminar<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a>, University of Erlangen (Germany)</li> <li><strong>2014-2016</strong>: Multimedia Programming , Highschool Students, University of Erlangen (Germany)</li> <li><strong>2013-2016</strong>: Lab Course, <em>Statistical Methods for Audio Experiments</em>, Master Students, University of Erlangen (Germany) <a href="https://www.audiolabs-erlangen.de/content/05-fau/professor/00-mueller/02-teaching/2016s_apl/LabCourse_StatsMethods.pdf" target="_blank" rel="noopener noreferrer"><i class="fas fa-file-pdf"></i> Course Material<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a>.</li></ul> <h3 id="talks"><a href="#talks" class="header-anchor">#</a> Talks</h3> <ul><li><strong>2026</strong>: Invited industry talk, Music Information Retrieval (MIR) program, Master-2, Telecom-ParisTech, Paris (France)</li> <li><strong>2025</strong>: Invited research talk and Q&A hosted by <a href="http://sap.ist.i.kyoto-u.ac.jp/members/yoshii/" target="_blank" rel="noopener noreferrer">Kazuyoshi Yoshii</a>, Kyoto University, Kyoto (Japan)</li> <li><strong>2023</strong>: Invited talk: "Music Source Separation: Is it solved yet?", ParisTech, Paris (France) <a href="https://adasp.telecom-paris.fr/resources/2023-02-09-talk-stoter/" target="_blank" rel="noopener noreferrer">Event page</a> <a href="https://faroit.com/slides-paristech23" target="_blank" rel="noopener noreferrer">Slides<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a></li> <li><strong>2020</strong>: Invited talk at AES Symposium <a href="https://www.aes.org/events/2020/learning/" target="_blank" rel="noopener noreferrer">"AES Virtual Symposium: Applications of Machine Learning in Audio"<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a> titled "Current Trends in Audio Source Separation". <a href="https://sigsep.github.io/AES2020_CurrentTrendsInSourceSeparation.pdf" target="_blank" rel="noopener noreferrer">Slides (PDF)<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a> <a href="https://www.youtube.com/watch?v=AB-F2JmI9U4" target="_blank" rel="noopener noreferrer">Video<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a></li> <li><strong>2019</strong>: Invited talk at a conference <a href="https://www.lebesgue.fr/content/sem2018-deeplearning" target="_blank" rel="noopener noreferrer">“Deep learning: From theory to applications”<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a> titled “Deep learning for music unmixing”. <a href="https://www.lebesgue.fr/video/2879" target="_blank" rel="noopener noreferrer">Video<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a> <a href="">Slides</a></li> <li><strong>2019</strong>: Tutorial at <a href="http://eusipco2019.org/" target="_blank" rel="noopener noreferrer">EUSIPCO 2019<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a>: <em>"Deep learning for music separation"</em>. <a href="https://sigsep.github.io/tutorials/"><i class="fad fa-presentation"></i> Slides</a> <a href="https://sigsep.github.io/tutorials/"><i class="fas fa-link"></i> Website</a></li> <li><strong>2018</strong>: Tutorial at <a href="http://ismir2018.ircam.fr/pages/events-tutorial-01.html" target="_blank" rel="noopener noreferrer">ISMIR 2019<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a>: <em>"Music Separation with DNNs: Making It Work"</em>. <a href="https://sigsep.github.io/tutorials/"><i class="fad fa-presentation"></i> Slides</a> <a href="https://sigsep.github.io/tutorials/"><i class="fas fa-link"></i> Website</a></li></ul> <h3 id="other-ressources"><a href="#other-ressources" class="header-anchor">#</a> Other Ressources</h3> <ul><li><a href="https://sigsep.github.io" target="_blank" rel="noopener noreferrer">sigsep.io<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a> - Open ressources for music separation.</li> <li><a href="https://github.com/faroit/awesome-python-scientific-audio" target="_blank" rel="noopener noreferrer">awesome-scientific-python-audio<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a> - Curated list of python packages for scientific research in audio.</li></ul> <h1 id="software"><a href="#software" class="header-anchor">#</a> Software</h1> <h2 id="open-unmix"><a href="#open-unmix" class="header-anchor">#</a> <img src="/assets/img/pytorchlogo.d98658a0.svg" width="20rem"> open-unmix <a href="https://devpost.com/software/open-unmix"><span class="badge success" style="vertical-align:top;" data-v-15b7b770>Winner: Pytorch Global Hackathon 2019</span></a></h2> <iframe width="100%" height="300rem" src="https://www.youtube-nocookie.com/embed/IxLnoy-GzqI" frameborder="0" allow="accelerometer; autoplay; encrypted-media; gyroscope; picture-in-picture" allowfullscreen="allowfullscreen"></iframe> <p>Open-Unmix, a deep neural network reference implementation (<a href="https://github.com/sigsep/open-unmix-pytorch" target="_blank" rel="noopener noreferrer">PyTorch<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a> and <a href="https://github.com/sigsep/open-unmix-nnabla" target="_blank" rel="noopener noreferrer">NNabla<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a>) for music source separation, applicable for researchers, audio engineers and artists. Open-Unmix provides ready-to-use models that allow users to separate pop music into four stems: vocals, drums, bass and the remaining other instruments.</p> <p>Demo Separations on <a href="https://sigsep.github.io/musdb" target="_blank" rel="noopener noreferrer">MUSDB18<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a> Dataset:</p> <iframe width="100%" height="490rem" src="https://d2cowzs755i94n.cloudfront.net" frameborder="0" allow="accelerometer; autoplay; encrypted-media; gyroscope; picture-in-picture" allowfullscreen="allowfullscreen"></iframe> <p><a href="https://sigsep.github.io/open-unmix/" class="button"><i class="fas fa-link"></i> Website/Demo</a> <a href="https://github.com/sigsep/open-unmix-pytorch" class="button"><i class="fab fa-github"></i> Code</a> <a href="https://joss.theoj.org/papers/10.21105/joss.01667" class="button"><i class="fa fa-file-pdf"></i> Paper</a> <a href="https://anr.fr/fr/actualites-de-lanr/details/news/open-unmix-un-logiciel-open-source-issu-du-projet-anr-kamoulox-pour-demixer-la-musique/" class="button"><i class="fas fa-link"></i> ANR Blog (french)</a> <a href="https://devpost.com/software/open-unmix" class="button"><i class="fad fa-rocket-launch"></i> Pytorch Hackathon</a></p> <h2 id="countnet"><a href="#countnet" class="header-anchor">#</a> CountNet</h2> <video width="100%" controls="controls"><source src="https://www.audiolabs-erlangen.de/content/resources/00-2017-CountNet/rnn_demo.mp4" type="video/mp4">
Your browser does not support the video tag.
</video> <p>CountNet is a deep learning model that estimates the number of concurrent speakers from single channel speech mixtures. This task is a mandatory first step to address any realistic “cocktail-party” scenario. It has various audio-based applications such as blind source separation, speaker diarisation, and audio surveillance.</p> <p><a href="https://github.com/faroit/countnet" class="button"><i class="fab fa-github"></i> code</a></p> <h2 id="musdb-museval"><a href="#musdb-museval" class="header-anchor">#</a> <i class="fab fa-python"></i> musdb + museval</h2> <p>A python package to parse and process the <a href="https://sigsep.github.io/musdb" target="_blank" rel="noopener noreferrer">MUSDB18 dataset<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a>, the largest open access dataset for music source separation. The tool was originally developed for the <a href="sisec18.unmix.app">Music Separation task</a> as part of the <a href="https://sisec.inria.fr/" target="_blank" rel="noopener noreferrer">Signal Separation Evaluation Campaign (SISEC)<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a>.</p> <p>Using <code>musdb</code> users can quickly iterate over multi-track music datasets. In just three lines of code a subset of the MUSDB18 is automatically downloaded and can be parsed:</p> <div class="language-python extra-class"><pre class="language-python"><code><span class="token keyword">import</span> musdb
mus <span class="token operator">=</span> musdb<span class="token punctuation">.</span>DB<span class="token punctuation">(</span>download<span class="token operator">=</span><span class="token boolean">True</span><span class="token punctuation">)</span>
<span class="token keyword">for</span> track <span class="token keyword">in</span> mus<span class="token punctuation">:</span>
train<span class="token punctuation">(</span>track<span class="token punctuation">.</span>audio<span class="token punctuation">,</span> track<span class="token punctuation">.</span>targets<span class="token punctuation">[</span><span class="token string">'vocals'</span><span class="token punctuation">]</span><span class="token punctuation">.</span>audio<span class="token punctuation">)</span>
</code></pre></div><p>Now, given a trained model, evaluation can simply be performed using <strong>museval</strong></p> <div class="language-python extra-class"><pre class="language-python"><code><span class="token keyword">import</span> museval
<span class="token keyword">for</span> track <span class="token keyword">in</span> mus<span class="token punctuation">:</span>
estimates <span class="token operator">=</span> predict<span class="token punctuation">(</span>track<span class="token punctuation">)</span> <span class="token comment"># model outputs dict</span>
scores <span class="token operator">=</span> museval<span class="token punctuation">.</span>eval_mus_track<span class="token punctuation">(</span>track<span class="token punctuation">,</span> estimates<span class="token punctuation">)</span>
<span class="token keyword">print</span><span class="token punctuation">(</span>scores<span class="token punctuation">)</span>
</code></pre></div><p><a href="https://github.com/sigsep/sigsep-mus-db" class="button"><i class="fab fa-github"></i> musdb</a> <a href="https://github.com/sigsep/sigsep-mus-eval" class="button"><i class="fab fa-github"></i> museval</a></p> <h2 id="hackathon-projects"><a href="#hackathon-projects" class="header-anchor">#</a> Hackathon Projects</h2> <h3 id="demask"><a href="#demask" class="header-anchor">#</a> DeMask <span class="badge success" style="vertical-align:top;" data-v-15b7b770>1st Place</span></h3> <p><strong>Event</strong>: 2020 PyTorch Summer Hackathon –
<strong>Collaborators</strong>: Manuel Pariente, Samuele Cornell, Michel Olvera, Jonas Haag</p> <iframe width="100%" height="300rem" src="https://www.youtube-nocookie.com/embed/QLf10Uqu8Yk" frameborder="0" allow="accelerometer; autoplay; encrypted-media; gyroscope; picture-in-picture" allowfullscreen="allowfullscreen"></iframe> <p>DeMask is an end-to-end model for enhancing speech while wearing face masks — offering a clear benefit during times when face masks are mandatory in many spaces and for workers who wear face masks on the job. Built with Asteroid, a PyTorch-based audio source separation toolkit, DeMask is trained to recognize distortions in speech created by the muffling from face masks and to adjust the speech to make it sound clearer.</p> <p><a href="https://devpost.com/software/asteroid-the-pytorch-based-source-separation-toolkit" class="button"><i class="fas fa-link"></i> DevPost Website</a></p> <h3 id="git-wig"><a href="#git-wig" class="header-anchor">#</a> <code>git wig</code> <span class="badge success" style="vertical-align:top;" data-v-15b7b770>Winner</span></h3> <p><strong>Event</strong>: 2015 Midi-Hackday Berlin,
<strong>Collaborators</strong>: <a href="https://nils-werner.github.io/" target="_blank" rel="noopener noreferrer">Nils Werner<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a>, <a href="https://www.audiolabs-erlangen.de/fau/assistant/lopez" target="_blank" rel="noopener noreferrer">Patricio-Lopez Serrano<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a></p> <img src="https://camo.githubusercontent.com/c4eeb051705f27dd1d531ace4c540e8fc3954d36/687474703a2f2f692e696d6775722e636f6d2f446453686271322e6a7067" height="238rem" align="left"> <img src="https://user-images.githubusercontent.com/72940/28498777-0454202e-6fa6-11e7-8a3c-9c85506013fa.gif" height="238rem"> <p>Why can't we have version on control for making music? In this hack, we merged <code>git</code> with a terminal based music sequencer, calling it <code>git wig</code>. We also created a suitable, diffable sequencer format to compose music. Finally, we realized <code>git push</code> by bringing this feature into a hardware controller.</p> <p><a href="https://github.com/RocketScienceAbteilung/git-grid" class="button"><i class="fab fa-github"></i> git grid</a> <a href="https://github.com/RocketScienceAbteilung/git-wig" class="button"><i class="fab fa-github"></i> git wig</a></p> <h3 id="deepfandom"><a href="#deepfandom" class="header-anchor">#</a> DeepFandom <span class="badge success" style="vertical-align:top;" data-v-15b7b770>1st Place</span></h3> <p><strong>Event:</strong> 2016 Music Hackday Berlin.
<strong>Collaborators</strong>: <a href="https://www.audiolabs-erlangen.de/fau/assistant/lopez" target="_blank" rel="noopener noreferrer">Patricio-Lopez Serrano<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a></p> <iframe width="100%" height="300rem" src="https://www.youtube-nocookie.com/embed/uPb56-UfFRI" frameborder="0" allow="accelerometer; autoplay; encrypted-media; gyroscope; picture-in-picture" allowfullscreen="allowfullscreen"></iframe> <p>DeepFandom is a deep learning model that learns the <a href="soundcloud.com">Soundcloud</a> comments and predicts what <em>YOUR</em> track could get as comments and where they are positioned on the waveform.</p> <p><a href="https://devpost.com/software/deep-fandom" class="button"><i class="fas fa-link"></i> Website</a></p> <h3 id="magiclock"><a href="#magiclock" class="header-anchor">#</a> Magiclock</h3> <img src="https://cloud.githubusercontent.com/assets/72940/16177370/e492626a-362a-11e6-9f66-2291040f98c1.gif" width="100%"> <p>Magiclock is an macOS application that uses haptic feedback (also called Taptic Engine™) to let you <strong>feel</strong> the MIDI clock beat from your Magic Trackpad.</p> <p><a href="https://github.com/faroit/magiclock" class="button"><i class="fab fa-github"></i> Code</a></p> <h2 id="other-software-contributions"><a href="#other-software-contributions" class="header-anchor">#</a> Other Software Contributions</h2> <ul><li><i class="fab fa-python"></i> <a href="https://github.com/faroit/stempeg" target="_blank" rel="noopener noreferrer">stempeg<span><svg xmlns="http://www.w3.org/2000/svg" aria-hidden="true" focusable="false" x="0px" y="0px" viewBox="0 0 100 100" width="15" height="15" class="icon outbound"><path fill="currentColor" d="M18.8,85.1h56l0,0c2.2,0,4-1.8,4-4v-32h-8v28h-48v-48h28v-8h-32l0,0c-2.2,0-4,1.8-4,4v56C14.8,83.3,16.6,85.1,18.8,85.1z"></path> <polygon fill="currentColor" points="45.7,48.7 51.3,54.3 77.2,28.5 77.2,37.2 85.2,37.2 85.2,14.9 62.8,14.9 62.8,22.9 71.5,22.9"></polygon></svg> <span class="sr-only">(opens new window)</span></span></a> - read/write of STEMS multistream audio.</li> <li><i class="fab fa-js"></i> <a href="https://github.com/audiolabs/trackswitch.js/">trackswitch.js</a> - A Versatile Web-Based Audio Player for Presenting Scientifc Results.</li> <li><i class="fab fa-js"></i> <a href="https://github.com/audiolabs/webMUSHRA">webMUSHRA</a> - MUSHRA compliant web audio API based experiment software.</li> <li><i class="fab fa-python"></i> <a href="https://github.com/sigsep/norbert">norbert</a> - Painless Wiener filters for audio separation.</li></ul> <h1 id="datasets"><a href="#datasets" class="header-anchor">#</a> Datasets</h1> <h2 id="musdb18"><a href="#musdb18" class="header-anchor">#</a> <img src="/assets/img/sigsep.9a00297a.png" width="20rem"> MUSDB18</h2> <p><img src="https://sigsep.github.io/assets/img/musheader.41c6bf29.png" alt=""></p> <p>The <em>musdb18</em> is a dataset of 150 full lengths music tracks (~10h duration) of different genres along with their isolated drums, bass, vocals and others stems. It is currently the largest, publicly available dataset used for music separation. <em>MUSDB18</em> serves as a benchmark for music separation tasks.</p> <p><a href="https://sigsep.github.io/datasets/musdb.html" class="button"><i class="fas fa-link"></i> Website</a> <a href="<https://paperswithcode.com/dataset/musdb18>
" class="button"><i class="fas fa-link"></i> Paperswithcode</a></p> <h2 id="libricount"><a href="#libricount" class="header-anchor">#</a> LibriCount</h2> <p><img src="https://www.audiolabs-erlangen.de/content/resources/00-2017-CountNet/teaser.svg" alt=""></p> <p>The dataset contains a simulated cocktail party environment of [0..10] speakers, mixed with 0dB SNR from random utterances of different speakers from the LibriSpeech <code>CleanTest</code> dataset.
All recordings are of 5s durations, and all speakers are active for the most part of the recording. For each unique recording, we provide the audio wave file (16bits, 16kHz, mono) and an annotation <code>json</code> file with the same name as the recording.</p> <p><a href="https://denumerate.app" class="button"><i class="fas fa-link"></i> Listening Experiment</a> <a href="https://zenodo.org/record/1216072" class="button"><i class="fas fa-link"></i> Download</a></p> <h2 id="muserc"><a href="#muserc" class="header-anchor">#</a> Muserc</h2> <iframe width="100%" height="300rem" src="https://www.youtube-nocookie.com/embed/yOKvqz2jZgM" frameborder="0" allow="accelerometer; autoplay; encrypted-media; gyroscope; picture-in-picture" allowfullscreen="allowfullscreen"></iframe> <p>A novel dataset for musical instruments where we recorded a <strong>violin cello</strong> that includes sensor recordings capturing the finger position on the fingerboard which is converted into an instantaneous frequency estimate. We also included professional high-speed video camera data to capture excitations from the string at 2000 fps. All of the data is sample synchronized</p> <p><a href="https://www.audiolabs-erlangen.com/resources/muserc" class="button"><i class="fas fa-link"></i> Website</a> <a href="https://zenodo.org/record/1560651" class="button"><i class="fas fa-link"></i> Download</a></p> <h1 id="publications"><a href="#publications" class="header-anchor">#</a> Publications</h1> <p><a href="https://scholar.google.com/citations?user=7HsSdqwAAAAJ&hl=en" class="button"><i class="fas fa-graduation-cap"></i> Google Scholar</a> <a href="https://www.zotero.org/faroit" class="button"><i class="fas fa-graduation-cap"></i> Zotero</a></p></div> <div class="footer"><a href="https://github.com/faroit/website"><i class="fab fa-github"></i> Code for this Website</a></div></main></div><div class="global-ui"></div></div>
<script src="/assets/js/app.62bc949a.js" defer></script><script src="/assets/js/2.63207ef2.js" defer></script><script src="/assets/js/1.0551b2b6.js" defer></script><script src="/assets/js/24.6ba64fe2.js" defer></script><script src="/assets/js/5.c1fac5e0.js" defer></script>
</body>
</html>