Abstract for "Süsser Die Daten Nie Klingen"

Die Zusammenarbeit zwischen Marken und der Musikbranche verläuft heute zunehmend datenbasierend, eingebettet in eine Story und so, dass beide Seiten profitieren. Wie sich das Verhältnis verändert hat.

Download

Abstract for "Sound mit System"

Wie klingt Birkenstock? Welcher Sound passt zu Vitra – und zu deren Kunden? Robin Hofmann war früher als DJ in der ganzen Welt unterwegs, heute kümmert er sich um das Audiobranding von Marken

Download

Abstract for "Audio Branding mit AI"

Nicht nur das Store-Design, digitale Tools und die Mitarbeiter beeinflussen am Point of Sale das Einkaufserlebnis, sondern auch der Sound. Instore-Music hat die Aufgabe, eine Marke zu repräsentieren. Mithilfe von AI sollen in Zukunft auch äußere Einflussfaktoren wie das Wetter und der Standort den Klang auf der Verkaufsfläche mitbestimmen.

Download

Abstract for "Markenidentität im Zeitalter der Digitalisierung: 10 zukunftsweisende Thesen"

Unternehmen, die ihre Produkte verkaufen wollen, müssen Emotionen wecken. Audio Branding wird für Marken deswegen in Zukunft so unverzichtbar sein wie Markenname, Logo oder Produktdesign. Die Frage ist nicht nur „Wie sehe ich aus?“ und „Wie schreibe ich?“, sondern auch „Wie klinge ich?“

Download

Abstract for "ABC_DJ: l’algoritmo che sceglie la musica giusta"

Fincons Group, società di IT business consulting con 35 anni di esperienza sul mercato internazionale, è parte del consorzio europeo che ha guidato il progetto di ricerca ABC_DJ (Artist to Business to Consumer Audio Branding System), ?nanziato dal programma di ricerca e innovazione di Horizon2020 della Commissione Europea nell’ambito dell’accordo di sovvenzione n. 688122.

Download

Abstract for "¿A qué suena una marca?"

¿A qué suena una marca?
Descubren un algoritmo que elige la canción más adecuada para cada enseña
infoRETAIL.- Canciones para las marcas. El marketing sensorial sigue abriéndose hueco en el sector del retail y gran consumo. El último paso en este sentido lo ha dado el consorcio europeo ABC DJ, que ha desarrollado una herramienta que permite determinar la música más acorde para cada marca, en función de sus valores corporativos.

Este proyecto ha creado un algoritmo para predecir cómo se percibe una canción, basado en la tecnología de machine learning. Esta predicción determina la relación de la canción con 36 términos claves con los que se puede definir una canción en un entorno corporativo: relajada, glamurosa, urbana o feliz son algunos de ellos.

Para el equipo de ABC DJ, los beneficios potenciales de la investigación van en tres líneas: identidades corporativas más fuertes y coherentes, experiencias más significativas para los consumidores y más música para todos.

Download

Abstract for "Así suena tu marca, según el algoritmo de ABC_DJ"

El consorcio europeo ABC_DJ ha desarrollado un algoritmo capaz de identificar los valores que desprende una canción y determinar así el tipo de música que mejor encaje con una marca.

Download

Abstract for "¿A qué suena una marca?"

LA MÚSICA MÁS ACORDE PARA CADA MARCA
¿A qué suena una marca?
Una de las mayores dificultades a las que se enfrentan las marcas hoy día es determinar qué música representa mejor sus valores corporativos. La cuestión no es simple puesto que la percepción musical es subjetiva y las reacciones ante determinados sonidos no tienen porqué ser idénticas. Este es el origen del proyecto ABC_DJ, un consorcio de siete empresas europeas liderado por la Universidad Técnica de Berlín, que ha desarrollado una herramienta que permite definir a través de criterios objetivos la afinidad de una canción con los valores de una determinada marca.

Download

Abstract for "Marken zum Klingen bringen"

Marken zum Klingen bringen
Mittels KI den passenden Soundtrack finden

Haben Sie schon einmal einen Film ohne Musik geschaut? Sie würden sich wundern! Denn auch, wenn Ihnen die Filmmusik während des Zuschauens kaum auffallen mag, sorgt sie am Ende maßgeblich dafür, dass Emotionen geweckt, Spannungen erzeugt und Stimmungen erzielt werden. Dasselbe gilt für alle visualisierten Marketingformate vom Image- und Werbevideo bis hin zum Produkt- und Event-Trailer.

Doch welche musikalische Untermalung passt zur Marke, zum Image, zur Zielgruppe und zum Marketingziel? Eine Frage, die sich kreative Köpfe schon oft gestellt haben mögen – auch und gerade in der MICE- und Eventbranche. Wenn der persönliche Geschmack dabei die Hauptrolle spielt, kann das musikalische Konzept am Ende ganz schnell für Misstöne sorgen. Doch damit soll bereits in naher Zukunft Schluss sein.

Download

Abstract for "Wie muss eine Marke klingen? Auf der Suche nach dem auditiven Code"

Wer seine Marke inszenieren will, braucht mehr als Bilder. Er braucht den perfekten Sound. Den findet man jetzt per Algorithmus. Aktuell gibt es auf dem europäischen Markt jedoch nur zwei marktreife Produkte, die diese automatisierte Auswahl leisten können: den ABC-DJ-Algorithmus und den Sonic-DNA-Algorithmus.

Der Sound einer Marke weckt Emotionen. Doch wer das Klangpotenzial seiner Marke ausschöpfen will, sollte sich besser auf Algorithmen statt auf seinen Geschmack verlassen.
Ein originelles Design, ein ansprechender Name, ein mitreißender Claim - die Seele einer Marke hängt von zahlreichen Komponenten ab. Doch haben Sie sich schon mal gefragt, wie Ihre Marke klingen sollte? Welche musikalischen Eigenschaften Ihre Marke ausmachen, ob sie sich energisch, dramatisch oder ruhig anhört? Sicher: Für Jingles, Audiologos oder andere mit Klang hinterlegte Werbemaßnahmen müssen Marketingverantwortliche sich diesen Fragen gezwungenermaßen widmen. Dann suchen sie auf eigene Faust oder verlassen sich auf die Meinung ihrer Agentur. Allerdings sind selbst die noch so musikalisch versierten Experten nicht in der Lage, den objektiv am besten passenden Sound auszuwählen. Schließlich sind auch sie „nur“ Menschen und hegen – wie jeder andere – subjektive Empfindungen gegenüber Klängen, Künstlern und Musiktiteln.

Download

Abstract for "Disco in der Damenumkleide"

Der Stuttgarter Audioexperte Robin Hofmann will die Hintergrundmusik in Geschäften per KI so optimieren, dass die Umsätze maximal steigen.
Bei der Einrichtung von Geschäften wird nichts dem Zufall überlassen. Die Anordnung der Gänge, die Einsortierung der Produkte, das Licht und der Geruch, das Gedudel im Hintergrund – alles im Laden ist aus einem ganz bestimmten Grund so, wie es ist: damit die Kunden möglichst viel Geld ausgeben. Wirklich alles? Robin Hofmann hat da seine Zweifel. Er glaubt, dass bei der verkaufsfördernden Wirkung von Musik noch einiges rauszuholen ist.

Download

Abstract for "Algoritmo consegue definir expressão musical de uma marca em 80,1%"

O ABC_DJ é um programa de pesquisa e inovação de grande escala, financiado pela União Europeia, que explora e desenvolve novas formas audio Branding. Uma das novidades desse projeto é a criação de um poderoso algoritmo que escolhe automaticamente a música relevante para uma marca, confiando apenas no conteúdo de áudio de seu banco de dados. Com este software, as marcas e agências de publicidade podem encontrar automaticamente a música certa para qualquer campanha, dando ao planejamento estratégico uma dimensão sonora.

Download

Abstract for "ABC_DJ: The algorithm that turns brand values into music"

Our reactions to music are complex and often personal. So how can a brand accurately choose music that fits its personality and resonates with its audiences? Audio branding agencies typically rely on a medley of creative experience and judgement, scattered research, and consumer feedback to find or create the right music for a brand.  The ABC_DJ research project aims to provide sonic branding agencies with sophisticated tools that support their creative processes.

Download

Abstract for "Algorithmus empfiehlt Marken-Musik"

Wissenschaftlerinnen und Wissenschaftler der TU Berlin entwickeln ein neues Tool für das sogenannte Branding, also die Kreation einer Marke. Diese hängt nicht zuletzt von der perfekten Präsentation der Produkte ab: Dabei beschreibt das Audiobranding die akustische Dimension einer Marke und soll ihre Eigenschaften hörbar machen.

Download

Abstract for "Un algoritmo che sceglie automaticamente la musica"

ABC_DJ ha sviluppato un innovativo software capace di dare consigli musicali. I ricercatori di ABC_DJ hanno creato un potente algoritmo che sceglie automaticamente la musica rilevante basandosi esclusivamente sul contenuto audio delle canzoni stesse, piuttosto che su tag assegnati manualmente. Con questo software, i marchi e le agenzie pubblicitarie possono trovare automaticamente la musica giusta per una campagna, pianificando strategicamente una dimensione sonora.

Download

Abstract for "KI filtert automatisiert passende Musik zu Marken"

FORSCHUNG:  Dem europäischen Forscherteam um das Projekt ABC_DJ ist kürzlich ein Durchbruch in der Musikforschung gelungen. Ab sofort können auf Knopfdruck aus zehntausenden von Musiktiteln die passenden Stücke für Marken gefunden werden.

Download

Abstract for "Innovative music recommendation software to predict brand-fit music"

The ABC_DJ project investigates and develops the future of Audio Branding. Researchers from ABC_DJ have created a powerful algorithm that automatically chooses brand-relevant music relying solely on the audio content of the songs themselves, rather than on manually assigned tags. With this software, brands and advertising agencies can automatically find the right music for any given brand or campaign, giving strategic planning a sonic dimension.

Download

Abstract for "Software Innovador De Recomendación De Música Para Predecir Música Ajustada A La Marca"

El proyecto ABC_DJ investiga y desarrolla el futuro de Audio Branding. Los investigadores de ABC_DJ han creado un potente algoritmo que elige automáticamente música relevante para la marca basándose únicamente en el contenido de audio de las canciones, en lugar de en etiquetas asignadas manualmente. Con este software program, las marcas y agencias de publicidad pueden encontrar automáticamente la música adecuada para cualquier marca o campaña, otorgando a la planificación estratégica una dimensión sónica.

Download

Abstract for "Progressive music recommendation program to predict manufacturer-in shape music"

The ABC_DJ challenge investigates and develops the future of Audio Branding. Scientists from ABC_DJ have established a impressive algorithm that immediately chooses manufacturer-suitable music relying only on the audio content of the songs them selves, relatively than on manually assigned tags. With this program, brand names and promotion organizations can immediately come across the right music for any specified manufacturer or marketing campaign, supplying strategic preparing a sonic dimension.

Download

Abstract for "Muzyczna sygnatura marki"

Muzyka w marketingu
Merchandising wchodzi dzisiaj na wyżyny (choć zdarza mi się wyjść ze sklepu przez muzykę, którą oferują głośniki, albo zbyt intensywny zapach unoszący się już od samego wejścia, ale to moje osobiste zwichrowanie ? ). Adekwatna muzyka, czy zapachy, jak również ułożenie produktów na półkach sklepowych ma wpłynąć na nasze zachowania konsumenckie. Jedne z badań prowadzonych w tym temacie pokazały na przykład, że muzyka może wpłynąć na zainteresowanie i decyzję zakupową wina (francuskie schodziły lepiej gdy w tle grana była muzyka francuska, a niemieckie wtedy gdy niemiecka: North, Hargreaves, McKendrick, 1997).

Download

Abstract for "Audio Branding mit KI optimieren: Marken können den passenden Sound nun generieren"

Download

Abstract for "Wie klingt Leidenschaft?"

Wem die Musik zur Marke noch fehlt, kann sich ab sofort mittels KI beraten lassen. Der virtuelle DJ bastelt dann den richtigen Sound. Dahinter stecken wissenschaftliche Analysen

Download

Abstract for "Algorithmus empfiehlt Marken-Musik: Wissenschaftlerinnen und Wissenschaftler entwickeln neues Tool für Audio Branding"

Das sogenannte Branding, also die Kreation einer Marke, hängt nicht zuletzt von der perfekten Präsentation der Produkte ab: Die richtige Umgebung, das perfekte Licht, die Einrichtung, all das spielt eine große Rolle und ist umfassend erforscht. Audiobranding beschreibt die akustische Dimension einer Marke und soll ihre Eigenschaften hörbar machen. Die Auswahl der passenden Musik für eine Marke wird bis heute eher „aus dem Bauch heraus“ entschieden. Dabei wird in der Regel auf Musik aus den U.S.-Charts zurückgegriffen. Eine internationale Forschungskooperation, koordiniert vom Fachgebiet Audiokommunikation von Prof. Dr. Weinzierl an der Technischen Universität Berlin, hat jetzt einen leistungsstarken Algorithmus entwickelt, der markenrelevante Musik automatisch auswählt.

Download

Abstract for "Wissenschaftler*innen entwickeln neues Tool für Audio Branding"

Das sogenannte Branding, also die Kreation einer Marke, hängt nicht zuletzt von der perfekten Präsentation der Produkte ab: Die richtige Umgebung, das perfekte Licht, die Einrichtung, all das spielt eine große Rolle und ist umfassend erforscht. Audiobranding beschreibt die akustische Dimension einer Marke und soll ihre Eigenschaften hörbar machen. Die Auswahl der passenden Musik für eine Marke wird bis heute eher „aus dem Bauch heraus“ entschieden. Dabei wird in der Regel auf Musik aus den U.S.-Charts zurückgegriffen. Eine internationale Forschungskooperation, koordiniert vom Fachgebiet Audiokommunikation von Prof. Dr. Weinzierl an der Technischen Universität Berlin, hat jetzt einen leistungsstarken Algorithmus entwickelt, der markenrelevante Musik automatisch auswählt. Dabei setzt der Algorithmus ausschließlich auf den Audioinhalt der Songs selbst und greift nicht auf manuell zugewiesene Tags und Beschreibungen zurück. Mit dieser Software können Marken und Werbeagenturen automatisch und treffsicher die passende Musik für jede Marke oder Kampagne finden.

Download

Abstract for "ABC_DJ - Translating Brand Values into Music"

Felix Haaksman and Andreas Schönrock of HearDis! speak about the ABC_DJ system during Most Wanted: Music, an event organised by the Berlin Music Commission. To listen to the presentation [16:58] please follow this link on SoundCloud.

The Berlin Music Commission (BMC) is the first all-embracing network representing Berlin music businesses. We bring together the potential of Germany’s most important music location. Our members hail from all areas of the music business including live entertainment, music software, club culture, recorded music, and music media.

Download

Abstract for "Marke, wie klingst Du? Wie man Markenbotschaften Gehör verschafft"

Markenerlebnisse mit allen Sinnen umfasst auch das Auditive. Welche Bedeutung Musik in der Erlebniskommunikation spielen kann und wie sie zu einem echten „Asset“ im Kommunikationsmix wird, erklärt Audio Branding-Experte Robin Hofmann in diesem Gastbeitrag. Er ist einer der beiden Geschäftsführer von HearDis!, deren Corporate Sound-Agentur, mit Sitz in Berlin und Stuttgart, bislang unter anderem BMW, Hugo Boss, Porsche Design, Migros und Victorinox akustisch beraten hat.

Download

Abstract for "Instore Music"

[DE] Kunterbunter Musikmix ade! Statt der üblichen Beschallung mit Chartsmusik läuft bei Marken, die mit HearDis! zusammenarbeiten, eine maßgeschneiderte Musikauswahl in den Läden. Die DJs der in Stuttgart und Berlin ansässigen Agentur erarbeiten auf die jeweiligen Markenbotschaften ihrer Kunden abgestimmte Corporate Sounds, wobei die Musik am Point of Sale eine besondere Rolle spielt. Agenturmitbegründer und DJ Robin Hofmann berichtet über seine Arbeit und darüber, was gute Instore Music ausmacht.
[EN] Goodbye motley mix of music! In the case of brands working together with HearDis!, instead of the usual sound a customized selection of music can be heard in the stores. The DJs of the agency based in Stuttgart and Berlin establish corporate sounds tailored to the brand messages of their clients with the music at the point of sale playing a special role. Agency co-founder Robin Hofmann tells us about his work and about what constitutes good instore music.

Download

Abstract for "ABC_DJ – Kompetenz im Bereich Audio Branding"

ABC_DJ steht für "Artist to Business to Business to Consumer Audio Branding System”. Es handelt sich um ein Kooperationsprojekt zwischen sieben auf Musik spezialisierten Unternehmen bzw. Institutionen aus unterschiedlichen europäischen Ländern – von Deutschland über Italien und Frankreich bis Spanien. Die Projektleitung erfolgt durch die Technische Universität Berlin.

Download

Abstract for "HearDis! entwickelt Audio Branding der Zukunft - Das EU-Forschungsprojekt “ABC_DJ“"

Berlin/Stuttgart – HearDis! erforscht und entwickelt seit Januar 2016 die Zukunft der akustischen Markenführung. Im EU-geförderten Forschungsprojekt ABC_DJ arbeiten die Audio-Branding-Experten an Software-Tools, die automatisiert markenrelevante Musikstücke erkennt. Mit dieser Software können Marken und Audio-Branding-Agenturen über einen Algorithmus Playlists passend zur Markenwelt generieren können. Das Projekt wird im Rahmen des “Horizon 2020“ Programms der EU mit einem Gesamtvolumen von 3.5 Millionen Euro gefördert. Insgesamt sind sieben Partner aus fünf europäischen Ländern daran beteiligt.

Download

Abstract for "Akustisches Markenerlebnis am POS"

For the German magazine "Markenartikel", speacialised on brand management, we wrote an article about acoustic branding at a brand's Point Of Sale. Here you can read all about the current situation in Audio Branding and why it is vital to boost European Corporate Sound Agencies.

Download

Abstract for "ABC_DJ Implementation in Territories Outside Europe"

This report lays the foundation for a possible transfer of novel audio branding and in-store music concepts, monetising schemes to Asia and the Americas. A first survey of procedural, legal, technical and cultural requirements for the export of audio branding services is produced.
Being the final analysis of our research, we have dedicated this document to the legal backdrop and regulations of publishing and mechanical copyright for services of background music in countries outside of the European Union. The findings and the success of our algorithmic predictions, especially when considering the conversion of semantic concepts into music, offers a new level of refinement and commercial suitability of playlists with a previously unknown quality and respective client-fit. This justifies focusing the research on this sector.

Download

Abstract for "Agreement on a New Single License Model for In-Store Audio Systems for European Collecting Societies"

The goal of this document is to propose a novel, convenient and adequate in-store license model for both users and collective management organization (CMOs).
To this end we have carried out a case study integrating three different users, from a CMO perspective: A Music Provider, and two end users (a 100 sqm clothing store and bar). The goal is to demonstrate, in a practical way, the procedures these users would need to follow, in order to use music in their businesses in several EU territories. This is accomplished by linking the users’ particular characteristics, to the formalisation of both licenses and payment of the corresponding tariffs, for each competent CMO, in each country.
This case study will shed light on the degree of opacity, difficulty and lack of harmonisation when it comes to music licensing procedures, still displayed by traditional CMOs within the EU.
Henceforward, we will present new licensing and managing models that have been arising in recent years, which advocate for the removal of hurdles and barriers towards one-stop-shop services.
Finally, we will illustrate our proposal for a new licensing model in a didactic and straightforward way. This licensing model is not intended as a finalized formal agreement, since that task must be carried forward and approved by the different European CMOs. Instead, it intends to provide a compact set of principles and guidelines arising out of our three years long research, seeking to provide realistic, practical and viable advice in the short term.
Finally, we will suggest a plausible scaling-up of the repertoire for ABC_DJ’s Music Library Manager, as a result of the conversations held between Lovemonk and independent record labels.

Download

Abstract for "Project Final Evaluation Report"

The ABC_DJ Description of Action identifies the means of final evaluation as “an audio branding demonstrator production,” i.e. an on-site use-test of the tools by a creative agency (HDIS) in the context of creating and implementing an audio brand for a client (PIACENZA) (Part B, pg. 9). Over the lifespan of the project, numerous intermediate tests of particular methods, software and hardware components, and tools were also conducted. This deliverable presents the final on-site test results together with consolidated results of intermediate tests, to the end of providing guidance in evaluating the extent to which the project fulfilled its expected impacts.
• Both expected impacts and comprehensive system performance were evaluated.
• The evaluation methodology comprised a combination of off- and on-site tests.
• Delays in the technical development made some alterations in the planned testing schedule and methodology necessary.
• Usability tests confirmed strong expert interest in the project concept and provided guidance for ongoing improvement of the tools.
• Off-site tests yielded a median overall rating of “good” across all tools, while also providing insight as to remaining weaknesses.
• Comparing usability and off-site test results shows a steady increase in the functionality and usability of the tools.
• On-site baseline and system tests on the provider side indicate moderate overall usability, moderate overall technical stability, and high overall willingness to use the system in the future. Qualitative reviews reveal that:
-- On the one hand, development work must still be done to match the “plug-and-play” functionality of existing basic tools.
-- On the other hand, the system’s advanced capacities make it significantly more inspiring than current tools.
• Final reviews of test results and laboratory performance metrics show that the project fulfilled nearly all its expected impacts
-- The sole exception is the visualisation-oriented Audio Branding Tool, which needs a considerable period of refinement to finalise.
-- In the consortium’s view, this does not detract from the overall impact of the project – particularly as the GMBI-enabled prediction module and the statistics that it generates have proven to be more effective communications tools than visualisations.
The consortium partners have committed to ongoing collaboration in order to realise the project’s promise of next-generation audio branding practices and products.

Download

Abstract for "Final Research Report for Sound Design and Audio Player"

This deliverable describes the work on Task 4.3 Algorithms for sound design and feature developments for audio player. The audio player runs on the in-store player (ISP) and takes care of rendering the music playlists via beat-synchronous automatic DJ mixing, taking advantage of the rich musical content description extracted in T4.2 (beat markers, structural segmentation into intro and outro, musical and sound content classification).
The deliverable covers prototypes and final results on: (1) automatic beat-synchronous mixing by beat alignment and time stretching – we developed an algorithm for beat alignment and scheduling of time-stretched tracks; (2) compensation of play duration changes introduced by time stretching – in order to make the playlist generator independent of beat mixing, we chose to readjust the tempo of played tracks such that their stretched duration is the same as their original duration; (3) prospective research on the extraction of data from DJ mixes – to alleviate the lack of extensive ground truth databases of DJ mixing practices, we propose steps towards extracting this data from existing mixes by alignment and unmixing of the tracks in a mix. We also show how these methods can be evaluated even without labelled test data, and propose an open dataset for further research; (4) a description of the software player module, a GUI-less application to run on the ISP that performs streaming of tracks from disk and beat-synchronous mixing.
The estimation of cue points where tracks should cross-fade is now described in D4.7 Final Research Report on Auto-Tagging of Music

Download

Abstract for "Final Research Report on Auto-Tagging of Music"

Deliverable D4.7 concerns the work achieved by IRCAM until M36 for the “auto-tagging of music”. The deliverable is a research report. The software libraries resulting from the research have been integrated into Fincons/HearDis! Music Library Manager or are used by TU Berlin. The final software libraries are described in D4.5.
The research work on auto-tagging has concentrated on four aspects:
1) Further improving IRCAM’s machine-learning system ircamclass. This has been done by developing the new MASSS audio features, including audio augmentation and audio segmentation into ircamclass. The system has then been applied to train HearDis! “soft” features (Vocals-1, Vocals-2, Pop-Appeal, Intensity, Instrumentation, Timbre, Genre, Style). This is described in Part 3.
2) Developing two sets of “hard” features (i.e. related to musical or musicological concepts) as specified by HearDis! (for integration into Fincons/HearDis! Music Library Manager) and TU Berlin (as input for the prediction model of the GMBI attributes). Such features are either derived from previously estimated higher-level concepts (such as structure, key or succession of chords) or by developing new signal processing algorithm (such as HPSS) or main melody estimation. This is described in Part 4.
3) Developing audio features to characterize the audio quality of a music track. The goal is to describe the quality of the audio independently of its apparent encoding. This is then used to estimate audio degradation or music decade. This is to be used to ensure that playlists contain tracks with similar audio quality. This is described in Part 5.
4) Developing innovative algorithms to extract specific audio features to improve music mixes. So far, innovative techniques (based on various Blind Audio Source Separation algorithms and Convolutional Neural Network) have been developed for singing voice separation, singing voice segmentation, music structure boundaries estimation, and DJ cue-region estimation. This is described in Part 6.

Download

Abstract for "Report on Collecting Schemes Europe (v2.1)"

This document is the second of two dealing with the legal regulation and the management of the Intellectual Property (IP) rights involved in audio branding processes and is complemented by document D7.2 - (Analysis of European IP Laws for Audio Branding Context).
There are no reports or compendia dealing with the management of IP rights from the point of view of audio branding. Thus, we have conducted surveys among the Collective Management Organizations (CMOs) with competences in in-store music, with the objective to clarify the rights management process.
The current document compiles details on how CMOs collect tariffs, identify owners and distribute the revenues originated by the usage of musical works in audio branding processes.
We carried out a survey among 91 CMOs in the Europe; however, in spite of great and long-lasting (16 months) efforts by our side, in the end just 34 were willing to answer our questionnaire. The information we gathered points to a lack of communication and coordination between CMOs both at European and local levels. Answers to very simple questions that should have borne a resemblance to each other, have in fact turned out to be diverse and contradictory. We find that this places users and music providers at a disadvantage and it does not help to improve the in-store music field. Monitoring of usage is deficient and the remuneration to right holders through COMs is not accurate.
However, there are some good news too, as some CMOs are moving forward in several countries in order to develop innovative solutions. For instance, some are sharing collecting bodies in order to simplify payments; this is known as a “one-stop shop”. In the case of mechanicals rights, at least one CMO has signed a deal with a background music provider that covers all European countries, which could turn out be a very useful model for other providers in the future.

Download

Abstract for "Common Definitions in Audio Branding Processes"

Technical, business and artistic terminologies and user languages in the highly interdisciplinary domain of audio branding vary greatly depending on the groups involved. To minimise internal and external misconceptions on any level, the ABC_DJ project partners developed an internal terminology standard concerning music, business and advertising semantics is created, built upon professional usage of a range of specific terms. This document is the public version (v2.0) of the original D2.1 deliverable from March 2016.

Download

Abstract for "Survey Report: Audio Branding Support Systems"

Existing tools for use in audio branding are surveyed and typical core work steps are defined. Particular attention is paid to professional metaphors in use and intuitive usability which support audio branding communication, workflows, automation, monitoring and maintenance. Furthermore, design of UIs which give representation support is examined in detail. Results are arranged into concrete requirements and recommendations for the project's tool developments.

Download

Abstract for "Analysis of European IP Laws for Audio Branding Context (v1.3)"

In depth analysis of legal regulations and the management of the Intellectual Property (IP) rights involved during audio branding processes. This document exclusively deals with legislation and case law, it furthermore provides a general overview of collective management. Details on how Collective Management Organisations collect tariffs, identify owners and distribute the revenues originated by the usage of works in audio branding processes will be part of an upcoming public document (June 2017).

Download

Abstract for "Music Providers Landscape Europe"

The document provides for a comprehensive overview of the landscape of European music creators and their respective multipliers in an audio branding and in-store music context. This includes an investigation of music management procedures as well as audio file formats and qualities in use.

Download

Abstract for "Methods and Datasets for DJ-Mix Reverse Engineering"

DJ techniques are an important part of popular music culture. However, they are also not sufficiently investigated by researchers due to the lack of annotated datasets of DJ mixes. Thus, this paper aims at filling this gap by introducing novel methods to automatically deconstruct and annotate recorded mixes for which the constituent tracks are known. A rough alignment first estimates where in the mix each track starts, and which time-stretching factor was applied. Second, a sample-precise alignment is applied to determine the exact offset of each track in the mix. Third, we propose a new method to estimate the cue points and the fade curves which operates in the time-frequency domain to increase its robustness to interference with other tracks. The proposed methods are finally evaluated on our new publicly available DJ-mix dataset UnmixDB. This dataset contains automatically generated beat-synchronous mixes based on freely available music tracks, and the ground truth about the placement, transformations and effects of tracks in a mix.

Abstract for "Popular Music as Entertainment Communication: How Perceived Semantic Expression Explains Liking of Previously Unknown Music"

Our contribution addresses popular music as essential part of media entertainment offerings. Prior works explained liking for specific music titles in ‘push scenarios’ (radio programs, music recommendation, curated playlists) by either drawing on personal genre preferences, or on findings about ‘cognitive side effects’ leading to a preference drift towards familiar and society-wide popular tracks. However, both approaches do not satisfactorily explain why previously unknown music is liked. To address this, we hypothesise that unknown music is liked the more it is perceived as emotionally and semantically expressive, a notion based on concepts from media entertainment research and popular music studies. By a secondary analysis of existing data from an EU-funded R&D project, we demonstrate that this approach is more successful in predicting 10000 listeners’ liking ratings regarding 549 tracks from different genres than all hitherto theories combined. We further show that major expression dimensions are perceived relatively homogeneous across different sociodemographic groups and countries. Finally, we exhibit that music is such a stable, non-verbal sign-carrier that a machine learning model drawing on automatic audio signal analysis is successfully able to predict significant proportions of variance in musical meaning decoding.

Download

Abstract for "A Computational Model for Predicting Perceived Musical Expression in Branding Scenarios"

We describe the development of a computational model predicting listener-perceived  expressions of music in branding contexts. Representative ground truth from multi-national online listening experiments was combined with machine learning of music branding expert knowledge, and audio signal analysis toolbox outputs. A mixture of random forest and traditional regression models is able to predict average ratings of perceived brand image on four dimensions. Resulting cross-validated prediction accuracy (R²) was Arousal: 61%, Valence: 44%, Authenticity: 55%, and Timeliness: 74%. Audio descriptors for rhythm, instrumentation, and musical style contributed most. Adaptive sub-models for different marketing target groups further increase prediction accuracy.

KEYWORDS: Music information retrieval, listener modelling, recommendation systems, audio branding, machine learning

Download

Abstract for "Towards a Common Terminology for Music Branding Campaigns"

Download

Abstract for "Methods and Datasets for DJ-Mix Reverse Engineering"

DJ techniques are an important part of popular music culture. However, they are also not sufficiently investigated by researchers due to the lack of annotated datasets of DJ mixes. Thus, this paper aims at filling this gap by introducing novel methods to automatically deconstruct and annotate recorded mixes for which the constituent tracks are known. A rough alignment first estimates where in the mix each track starts, and which time-stretching factor was applied. Second, a sample-precise alignment is applied to determine the exact offset of each track in the mix. Third, we propose a new method to estimate the cue points and the fade curves which operates in the time-frequency domain to increase its robustness to interference with other tracks. The proposed methods are finally evaluated on our new publicly available DJ-mix dataset. This dataset contains automatically generated beat-synchronous mixes based on freely available music tracks, and the ground truth about the placement of tracks in a mix.

Download

Abstract for "New Music Recommendation Algorithm Facilitates Audio Branding"

While marketing continues to work primarily in a visual domain, holistic brand experiences and multisensory marketing techniques are becoming ever more important. Audio branding is one such technique. The European Commission-funded research and innovation project ABC_DJ has developed an algorithmic music recommender system specifically for use in audio branding. The system’s core function is to assist marketers in determining musical brand fit, i.e. whether listeners will evaluate a given piece of music as fitting the perceived attributes of a given brand, and target group preference, i.e. whether listeners belonging to a particular target group will like a given piece of music. This article describes the ABC_DJ project’s aims, outputs, commercial applications, and status as an asset for European SMEs facing the ongoing evolution of the point of sale into a point of experience

Download

Abstract for "Local AM/FM Parameters Estimation: Application to Sinusoidal Modeling and Blind Audio Source Separation"

Abstract: This letter extends our recently introduced method [D. Fourer, F. Auger, K. Czarnecki, S. Meignen, and P. Flandrin, “Chirp rate and instantaneous frequency estimation: Application to recursive vertical synchrosqueezing,” IEEE Signal Process. Lett., vol. 24, no. 11, pp. 1724–1728, Nov. 2017] which was designed to estimate instantaneous frequency and chirp rate of linearly modulated signals. Now, we propose to derive several new estimators related to our previous ones which provide in the time-frequency plane all the signal parameters of the investigated model: amplitude, frequency and their local modulations (AM/FM). Our estimators are first introduced and compared in terms of statistical efficiency with theoretical bounds and with other state-of-the-art estimators. Then, they are used to improve spectral analysis applied to audio sinusoidal modeling. Finally, they lead to a new source separation technique based on coherent amplitude and frequency modulation that is evaluated on real-world music signals.
Index Terms: time-frequency analysis, sinusoidal modeling, source separation, audio processing.

Download

Abstract for "UnmixDB: A Dataset for DJ-Mix Information Retrieval"

DJ techniques are an important part of popular music culture but are also not sufficiently investigated by researchers due to the lack of annotated datasets of DJ mixes. This paper aims at filling this gap by introducing a publicly available DJ-mix dataset. It contains automatically generated beat-synchronous mixes based on freely available music tracks, and the ground truth about the placement of tracks in a mix. Each mix is generated in several variants
with different effects and time-stretching methods applied. Possible applications are to test novel methods for track identification in mixes, or for automatic annotation and deconstruction of recorded mixes for which the constituent tracks are known.

Download

Abstract for "A Heuristic Algorithm for DJ Cue Point Estimation"

This article treats an aspect in a larger research agenda to understand DJ practices, which are an important part of popular music culture: We present a heuristic algorithm that estimates cue points where tracks should cross-fade in a DJ mix. We deduced statistics and heuristics from a list of rules provided by human experts, and from a database of example tracks with given cue regions. We then created an algorithm for cue-point estimation based on rich automatic annotations by state of the art MIR methods, such as music structure segmentation and beat tracking. The results were evaluated quantitatively on the example database and qualitatively by experts.

Download

Abstract for "Single-Channel Blind Source Separation for Singing Voice Detection: A Comparative Study"

We propose a novel unsupervised singing voice detection method which use single-channel Blind Audio Source Separation (BASS) algorithm as a preliminary step. To reach this goal, we investigate three promising BASS approaches which operate through a morphological filtering of the analyzed mixture spectrogram. The contributions of this paper are manyfold. First, the investigated BASS methods are reworded with the same formalism and we investigate their respective hyperparameters by numerical simulations. Second, we propose an extension of the KAM method for which we propose a novel training algorithm used to compute a source-specific kernel from a given isolated source signal. Second, the BASS methods are compared together in terms of source separation accuracy and in terms of singing voice detection accuracy when they are used in our new singing voice detection framework. Finally, we do an exhaustive singing voice detection evaluation for which we compare both supervised and unsupervised singing voice detection methods. Our comparison explores different combination of the proposed BASS methods with new features such as the new proposed KAM features and the scattering transform through a machine learning framework and also considers convolutional neural networks methods.

Download

Abstract for "Fast and adaptive blind audio source separation using recursive Levenberg-Marquardt synchrosqueezing"

This paper revisits the Degenerate Unmixing Estimation Technique (DUET) for blind audio separation of an arbitrary number of sources given two mixtures through a recursively computed and adaptive time-frequency representation. Recently, synchrosqueezing was introduced as a promising signal disentangling method which allows to compute reversible and sharpen time-frequency representations. Thus, it can be used to reduce overlaps between the sources in the time-frequency plane and to improve the sources’ sparsity which is often exploited by source separation techniques. Furthermore, synchrosqueezing can also be extended using the Levenberg-Marquardt algorithm to allow a user to adjust the energy oncentration of a time-frequency representation which can be efficiently implemented without the FFT algorithm. Hence, we show that our approach can improve the quality of the source separation process while remaining suitable for real-time applications.

Download

Abstract for "Towards Extraction of Ground Truth Data from DJ Mixes"

DJ techniques are an important part of popular music culture but that are so far not very well researched because of the lack of annotated databases of DJ mixes. We first offer an overview of the necessary components to annotate recorded mixes automatically, which are: fingerprinting to obtain the tracklist, alignment to determine where in the mix each track starts and what tempo changes are applied to achieve beat-synchronicity, unmixing to estimate the fade curves for volume, bass and treble, and content and metadata analysis to derive the genre and social tags attached to the music to inform about the choices a DJ makes when creating a mix. Most of these components have been addressed by recent MIR research, except the alignment part for which we will give a first attempt using multi-scale correlation and dynamic time warping.

Download

Abstract for "High-Level Chord Features Extracted from Audio can Predict Perceived Musical Expression"

We investigated the relationship between high-level chord features and the perceived semantic and emotional expression of musical pieces in the context of music branding. Therefore, we first developed high-level chord features based on musicological considerations and novel MIR technologies. Inter alia, these features represent the number of chords, the proportion of major/minor chords, and the frequency of certain cadences and turnarounds. The validity of these features for predicting listeners’ perceived musical expression beyond genre information was subsequently tested by means of data from two online listening experiments, where musical expression of 549 music titles had been rated on four factors, Easy-going, Joyful, Authentic, and Progressive. Results show that in all four models chord features significantly improved prediction results. Most important features turned out to be those representing the number of (unique) chords and the proportion of minor chords. Implications of results are discussed, and future work is outlined.

Download

Abstract for "Estimation locale des modulations am/fm: applications à la modélisation sinusoidale audio et à la séparation de sources aveugle"

English translation: Abstract – In this paper, several new amplitude (AM) and frequency modulation (FM) estimators designed for linear modulated signals (linear chirps) are investigated. First they are compared together in terms of statistical efficiency with other state-of-the-art methods. Second, they are used to improve sinusoidal modeling of audio signals. Finally, a new source separation technique based on coherent amplitude and frequency modulation is proposed and evaluated on musical signals.

Résumé – Dans cet article, plusieurs nouveaux estimateurs de modulation d’amplitude (AM) et de fréquence (FM), dédiés aux signaux linéairement modulés sont étudiés. Premièrement, ces estimateurs sont comparés entre eux, puis avec d’autres méthodes issues de l’état de l’art en termes d’efficacité statistique. Deuxièmement, nous proposons d’appliquer ces nouveaux estimateurs à la modélisation sinusoïdale des signaux audio. Enfin, une nouvelle technique de séparation de sources s’appuyant sur les modulations cohérentes d’amplitude et de fréquence est proposée et évaluée sur des signaux musicaux.

Download

Abstract for "Automatische Vorhersage musik-induzierter Attributassoziationen im Kontext von Audio-Branding"

Musik wird häufig zur strategisch geplanten Kommunikation einer Markenidentität an die anvisierten Zielgruppen verwendet, beispielsweise am „Point-of-Sale“ oder in audio-visuellen Werbemitteln. Nachdem bislang zur Musikselektion in diesen Szenarien vor allem auf die praktische Expertise von Audio-Branding-Experten zurückgegriffen wurde, wird in dem vorgestellten europäischen Forschungsprojekt ein Softwaretool entwickelt, welches automatisch zur Marke passende Titel aus vorhandenen digitalen Musikarchiven auswählt.

Download

Abstract for "Names and Titles Matter: The Impact of Linguistic Fluency and the Affect Heuristic on Aesthetic and Value Judgements of Music"

It has been shown that titles influence peoples’ evaluation of visual art. However, the question of whether song titles affect listeners when evaluating music has not yet been studied. By using two well-known cognitive heuristics, two experiments were conducted to investigate title effects on aesthetic and value judgements of popular music. Experiment 1 (N=48) focused on linguistic fluency. The same song excepts were presented with easy-to-pronounce (fluent) and difficult-to-pronounce (disfluent) names. Experiment 2 (N=100) studied the affect heuristic. The same music excerpts were presented with positive (e.g., Kiss), negative (e.g., Suicide), and neutral (e.g., Window) titles. In both studies, aesthetic and value judgements of popular music were significantly influenced by titles. Participants in Experiment 1 evaluated the same music more positively when presented with fluent names compared to disfluent names. In Experiment 2, presenting the music with negative titles resulted in the lowest aesthetic and value judegments. Moreover, song excerpts presented with neutral and negative titles were remembered significantly more often than positive titles. Finally, a comparison of the music presented with and without titles indicated that music excerpts were more liked in the presence of titles than in their absence, regardless of the emotional content of the titles. The present research shows different ways in which aesthetic and value judgements of music can be influenced by song titles. Results suggest that like any other human judgement, evaluations of music also rely on heuristic principles that do not necessarily depend on the aesthetic stimuli themselves.
Keywords: aesthetics, musical judgements and preferences, song title, fluency, affect heuristic

Download

Abstract for "The diffusion of music streaming services in Germany between 2012-2015 and its impact on habitual audio media repertoires of the normal population"

During the past years, European countries have experienced the diffusion of various new digital music streaming services (Spotify, Deezer, etc.). At the same time, YouTube has developed into a global jukebox for discovering new music (Lepa, Hoklas & Weinzierl, 2014). Since these services afford (mobile) trans-media access to a huge library of music with differing gadgets at differing places in different social contexts (Lepa & Hoklas, 2015), this may radically change everyday music listening practices. But do these potentials really affect the listening habits of the ‘normal population’? Or are only members of the ‘digital generation’ truly adopting the various affordances of mobile streaming technologies into their everyday life, while mainstream users simply ‘stick to their habits’ and ignore the new offerings or treat e.g. Spotify as if it were just a virtual harddrive or just another radio channel? The presentation aims to shed light on these question by presenting results from a population-representative panel analysis testing for the amount of stability and change in German audio repertoires during the introduction of new music streaming services between 2012-2015. Furthermore, it will report findings of regression analyses explaining observed changes by milieux of origin, gender, age, birth cohort and changing life-circumstances.

Abstract for "Predicting Musical Meaning in Audio Branding Scenarios"

Abstract: The aim of the presented study was to test the feasibility of predicting the music-induced activation of branding relevant semantic associations. In order to achieve this, we aim to predict brand attributes (such as ‘rebellious’, ‘reliable’, ‘sentimental’ or ‘confident’) based on a variety of lower and higher order acoustic and musical features and based on the moderating influence of the individual differences of music listeners. Furthermore, we aim on developing a prototype system for automatic music recommendation within the audio branding domain.

Download

Abstract for "Development and Evaluation of a General Attribute Inventory for Music in Branding [Poster Presentation]"

Abstract: The aim of the presented work was to collect and evaluate the newly developed General Music Branding Inventory (GMBI). It consists of a list of branding-relevant attributes that can be reliably communicated with music. We therefore conducted a mixed-methods project that consists of three interrelated studies: a focus group consisting of music consultants constructed and defined an initial list of relevant branding attributes that in their view can be communicated through music (Study 1). This list was then evaluated and reduced in size in a subsequent online survey of branding experts (Study 2). Finally, several thousand participants rated the fit of the remaining items to a broad selection of music excerpts (Study 3).

Download

Abstract for "Music Structure Boundaries Estimation Using Multiple Self-Similarity Matrices as Input Depth of Convolutional Neural Networks"

In this paper, we propose a new representation as input of a Convolutional Neural Network with the goal of estimating music structure boundaries. For this task, previous works used a network performing the late-fusion of a Mel-scaled log-magnitude spectrogram and a self-similarity-lag-matrix. We propose here to use the squaresub-matrices centered on the main diagonals of several self-similarity-matrices, each one representing a different audio descriptors. We propose to combine them using the depth of the input layer. We show that this representation improves the results over the use of the self-similarity-lag-matrix. We also show that using the depth of the input layer provide a convenient way for early fusion of audio representations.

Download

Abstract for "Entwicklung eines Systems zur automatischen Musikempfehlung im Kontext des Music Brandings"

Abstract: In den vergangenen Jahren haben Audio Branding und Music Branding, d.h. Maßnahmen akustischer Markenführung im Bereich des Marketings, zunehmend an Bedeutung gewonnen. Das EU-geförderte Forschungsprojekt ABC_DJ zielt darauf ab, Audio-Branding-Agenturen mit Softwarelösungen auszustatten, die eine Identifikation markenrelevanter Musikstücke aus großen Archiven, etwa zur automatischen Generierung von Playlists, ermöglicht. Zu diesem Zweck soll ein statistisches Modell entwickelt werden, das den semantischen Ausdruck von Musiktiteln basierend auf deren akustischen Charakteristika sowie dem sozialen und kulturellen Kontext der Hörer vorhersagt. Im ersten Schritt wurde dafür zunächst das sogenannte General Music Branding Inventory (GMBI) entwickelt, welche ein umfangreiches, psychometrisch optimiertes Fachvokabular von Markenattributen dargestellt, wie es im Kontext des Music Branding verwendet wird. Daraufhin wurde eine Online-Studie mit 3485 Personen aus drei europäischen Ländern durchgeführt. Im Rahmen dieser Studie wurden die Versuchsteilnehmer gebeten, Auszüge aus verschiedenen Musikstücken mit Hilfe der GMBI-Attribute zu bewerten. Insgesamt wurden für den Versuch 183 Musikausschnitte aus 61 verschiedenen Stilistiken mit einer Länge von jeweils 30 Sekunden verwendet. Diese Ausschnitte wurden zudem in Hinblick auf 50 verschiedene Audio Features analysiert. Im Rahmen des Beitrages werden erste Ergebnisse der Studie, insbesondere zum Zusammenhang zwischen musikalisch-akustischen Eigenschaften und dem wahrgenommenen semantisch‐emotionalen Musikausdruck, dargestellt. Darauf basierend werden ferner erste Entwürfe des Vorhersagemodells skizziert.

Download

Abstract for "Objective characterization of audio signal quality: Applications to music collection description"

In this paper,  we propose a set of audio features to describe the quality of an audio signal.  Audio quality is here considered as being modified by the chain of processes/effects applied to the individual instrument tracks to obtain the final mix of a musical piece.   Thus,  the quality also depends on the mastering processes applied to the final mix or the signal degradation caused by MP3 compression. To evaluate our proposal,  we created a large set of artificial mixes and also used  real-world  studio  mixes.   Using  unsupervised  and  supervised classification methods, we show that our proposed audio features can detect the processing chain. Since this processing chain applied in professional studio has evolved over the  years,  we  use  our  audio  features  to  directly  predict  the decade during which a music track was recorded.

Download

Abstract for "Scale and Shift Invariant Time/Frequency Representation using Auditory Statistics: Application to Rhythm Description"

Abstract: In this paper we propose two novel scale and shift-invariant time-frequency representations of the audio content. Scale-invariance is a desired property to describe the rhythm of an audio signal as it will allow to obtain the same representations for same rhythms played at different tempi. This property can be achieved by expressing the time-axis in log-scale, for example using the Scale Transform (ST). Since the frequency locations of the audio content are also important, we previously extended the ST to the Modulation Scale Spectrum (MSS). However, this MSS does not allow to represent the inter-relationship between the audio content existing in various frequency bands. To solve this issue, we propose here two novel representations. The first one is based on the 2D Scale Transform, the second on statistics (inspired by the auditory experiments of McDermott) that represent the inter-relationship between the various frequency bands. We apply both representations to a task of rhythm class recognition and demonstrates their benefits. We show that the introduction of auditory statistics allows a large increase of the recognition results.

Download

Abstract for "Towards Automatic Music Recommendation For Audio Branding Scenarios"

Abstract: Within the MIR community, most prediction models of musical impact on listeners focus on mood or emotional effects (perceived or induced). The ABC_DJ project investigates the associative impact of music on listeners from the specific perspective of music branding that surrounds us in our everyday lives. We present a general concept for applying automatic music recommendation within this domain. Creating a scientifically validated basic terminology for communicating brand attributes and human emotions in this field is the key challenge. As a first result, we introduce the Music Branding Expert Terminology (MBET), a comprehensive terminology of verbal attributes used in music branding, upon which a prediction model will be developed to facilitate automatic music recommendation in the context of music branding.

Download