Upcoming: Two invited talks on reverse-correlation for high-level auditory cognition

CREAM Lab is hosting a small series of distinguised talks on reverse-correlation this month:

  • Wednesday 22nd March 2017 (11:00) – Prof. Fréderic Gosselin (University of Montreal)
  • Thursday 23rd March 2017 (11:30) – Prof. Peter Neri (Ecole Normale Supérieure, Paris).

These talks are organised in the context of a workshop on reverse-correlation for high-level audio cognition, to be held in IRCAM the same days (on-invitation-only). Both talks are free for all, in IRCAM (1 Place Stravinsky, 75004 Paris). Details (titles, abstract) are below.


Read More

Is musical consonance a signal of social affiliation? Our new study of musical interactions, out today in Cognition

A recently emerging view in music cognition holds that music is not only social and participatory in its production, but also in its perception, i.e. that music is in fact perceived as the sonic trace of social relations between a group of real or virtual agents. To investigate whether it is at all the case, we asked a group of free collective improvisers from Conservatoire Supérieur National de Musique et de Danse de Paris (CNSMDP) to try to communicate a series of social attitudes (being dominant, arrogant, disdainful, conciliatory or caring) to one another, using only their musical interaction. Both musicians and non-musicians were able to recognize these attitudes from the recorded music.

The study, a collaboration between Clément Canonne and JJ Aucouturier, was published today in Cognition. The corpus of 100 improvised duets used in the study is also available online:  http://cream.ircam.fr/?p=575

Read More

[CLOSED] CREAM is looking for a new RA ! (part-time, 5-months)


Assistant de recherche pour passation d’expériences de psychologie cognitive 

Période: Mars à Juillet 2017

Nous cherchons à engager en contrat CDD à temps partiel un ou une assistant/e de recherche pour la passation de plusieurs expériences de psychologie et neurosciences cognitive sur le thème de la voix, de la perception de soi, de la musique, et des émotions, sur la période de Mars à Juillet 2017.

La personne recrutée travaillera en collaboration avec les chercheurs ayant conçu les expériences et sera responsable de la collecte de données, en autonomie. La passation d’expérience se passera au Centre Multidisciplinaire des Sciences Comportementales Sorbonne Universités-INSEAD (6 rue Victor Cousin, 75005 Paris).

Profil souhaité

  • Licence ou master de psychologie expérimentale/neurosciences ou equivalent.
  • OBLIGATOIRE: avoir une expérience de la collecte de données expérimentales en psychologie/neurosciences (avoir fait passer au moins N = 20 participants adultes dans le cadre de ses études, d’un stage ou d’un contrat précédent), incluant la signature de consentement de participation, l’explication du protocole au participant, la veille au bon déroulement technique de l’expérience, la sauvegarde appropriée des différents fichiers de données, et le debriefing avec le participant.
  • FORTEMENT SOUHAITÉ: avoir une expérience en psychoacoustique et/ou en traitement du signal audio.
  • SOUHAITÉ: avoir un intérêt pour la recherche en psychologie/neurosciences cognitive, en particulier dans le domaine des neurosciences musicales, de la conscience de soi et des émotions et/ou avoir un intérêt pour les technologies de la musique et du son.



Contrat: CDD CNRS niveau assistant ingénieur (AI) ou ingénieur d’étude (IE), selon diplôme et experience.

Durée: 5 mois sur la période Mars – Juillet 2017.

Quotité: temps-partiel souhaité (50 ou 60%).

Lieu de travail: Paris (France).

Rémunération: correspondant à 1821.83e mensuel brut (AI) à 2465.67e mensuel brut (IE), selon diplôme et expérience) pour un temps plein.


Envoyer un CV et lettre de motivation détaillant vos expériences précédentes de collecte de données par email à Louise Goupil lougoupil@gmail.com & Jean-Julien Aucouturier aucouturier@gmail.com.



Offre – Assistant de Recherche CREAM – PDF

Read More

[CLOSED] Two new research internship for 2017: psychoacoustics of singing voice, and datamining of baby cries

CREAM is looking for talented master students for two research internship positions, for a duration of 5-­6months first half of 2017 (e.g. Feb­‐June ’17). UPDATE (Jan. 2017): The positions have now been filled. 

The first position mainly involves experimental, psychoacoustic research: it is examining the link between between increased/decreased pitch in speech and singing voice and the listener’s emotional response. It will be supervised by Louise Goupil & JJ Aucouturier, and is suitable for a student with a strong interest in experimental psychology and psychoacoustics, and good acoustic/audio signal processing/music production skills. See the complete announcement here: [pdf]

The second position is a lot more computational: it involves building audio pattern recognition tools in order to datamine a large corpus of audio recordings of human infant cries for acoustical patterns informative of the babies’ development of linguistic/communicative abilities. It will be supervised by JJ Aucouturier, in collaboration with Kazuo Okanoya and Yulri Nonaka from the University of Tokyo in Japan. It is suitable for a student with strong audio machine learning/music information retrieval skills and programming experience in Matlab or (preferably) Python. See the complete announcement here: [pdf]

Applications: send a CV and cover letter by email (see announcement). Interviews for selected applicants will be held in December’16-January’17.

Read More

CREAM Lab in Japan 1~8/11

Exciting news: Laura, Pablo, Emmanuel and JJ from the lab will be “touring” (the academic version thereof, at least) Japan this coming week, with two events planned in Tokyo:

If you’re around, and want to chat, please drop us a line. 「日本にきてとてもうれしい!!」

Read More

[CLOSED] Research internship (EEG/Speech/Emotion)



CREAM is looking for a talented master student interested in EEG and speech for a spiffy research internship combining Mismatch Negativity and some of our fancy new voice transformations technologies (here and there). The intern will work under the supervision of Laura Rachman and Jean-Julien Aucouturier (CNRS/IRCAM) & Stéphanie Dubal (Brain & Spine Institute, Hopital La Salpétrière, Paris).



See the complete announcement here: internship-annonce

Duration: 5-­6months first half of 2017 (e.g. Feb­‐June ’17).

Applications: send a CV and cover letter by email to Laura Rachman & JJ Aucouturier (see announcement). Interviews for selected applicants will be held in November‐December 2016.











Read More

The overdub corpus

A corpus of free pop music recordings, mixed by a professional sound engineer in several variants used to experiment with a listener’s feelings of social cohesion. Each track is available in X variants

  1. Single lead voice + orchestra (PBO) – drums
  2. Single lead voice + vocal overdub (double-tracking) + orchestra (PBO)
  3. Single lead voice – overdub + orchestra + drums
  4. Single lead voice + randomly detuned PBO – drums
  5. Single lead voice + randomly desynchronized PBO

The corpus is made available under a Creative Commons licence, from archive.org.

Mixed by Sarah Hermann (CNSMDP, Paris) at IRCAM, July 2016.



Read More

Bullying the doc: stronger-sounding patients get more 911 attention

Our team published a new paper this week, in which we test the influence of patients’ tone of voice on medical decisions.

In the line of our recent real-time emotional voice transformations, we manipulated the voice of (fake) patients calling a 911 phone simulator used for training (real) medical doctors, to make them sound more, or less, physically dominant (with deeper, more masculine voices corresponding to lower pitch and greater formant dispersion, see e.g. Sell, A. et al. Adaptations in humans for assessing physical strength from the voice. Proceedings of the Royal Society of London B: Biological Sciences, 277(1699), 3509–3518 (2010) – link).

We found that patients whose voice signalled physical strength obtained a higher grade of response, a higher evaluation of medical emergency and longer attention from medical doctors than callers with strictly identical medical needs whose voice signaled lower physical dominance.

The paper, a collaboration with Laurent Boidron M.D. and his colleagues at the Department of Emergency Medicine of the Dijon CHU Hospital/Université de Bourgogne, was published last tuesday in Scientific Reports (link, pdf).

Read More