Research seminar Sophie Vanvooren - 8 January 2014

Technical Comparison Sennheiser HDA 200/300

In our lab, the Sennheiser HDA 200 headphones are widely used for anything audiology related. They are pleasant to wear for a long time, have a flat frequency response and high passive ambient noise attenuation as they are based on PELTOR hearing protectors.

Sadly, these headphones went out of production. And according to Sennheiser, there is no way that they will resume production. Instead, they recommended their designated successor headphones, the Sennheiser HDA 300.

Official technical information

For both headphones, technical data sheets with information on passive attenuation and sensitivity level (frequency response) are available from the Sennheiser website (HDA 200, HDA 300).

For passive attenuation above 250 Hz, the HDA 300 is worse than the HDA 200, with an average difference of 16 dB for the frequencies range from 500 Hz to 8 kHz.

For the flatness of the sensitivity level, the profile of the HDA 300 has a standard deviation of 9 dB (octave frequencies 125 Hz to 16 kHz), which is nearly twice as much as the 5 dB for the HDA 200.

The HDA 300 headphones weight 490 g, an increase of 50 % over the 330 g of the HDA 200. The transducer impedance of 40 Ohm is reduced to 23 Ohm.

Own measurements

We tried to reproduce the data on passive attenuation and sensitivity level for the two types of headphones. The measurements were done in a sound-proof room with a Brüel & Kjær artificial ear 4153, with 4 speakers creating a diffuse noise sound field of about 70 dB(SPL). For the frequency sensitivity measurements, a continuous sweep was used. One side of the headphones was placed on the artificial ear, while the other side was wrapped around a side of the measurement table.

With this setup, we could observe a smaller difference in passive attenuation of 8 dB (octave frequencies 2 to 8 kHz) for frequencies above 2 kHz instead of 500 Hz. For lower frequencies, no difference was found. We suspect that the difference between our measurements and the official Sennheiser data is caused by our less-than-perfect measurement setup, especially the way we fixed the headphone to the artificial ear.

For the sensitivity level, we measured a standard deviation of 3 dB for the HDA 200 and 9 dB for the HDA 300 (octave frequencies 125 Hz to 16 kHz).

tl;dr (Summary)

Compared to the HDA 200, the HDA 300 headphones have less passive attenuation, a worse frequency sensitivity profile and are much heavier. We are not impressed.

Holyhour Bernhard Laback Tuesday 17 December 2013

Holyhour Christian Lorenzi Friday 13 December 2013

Hearing without Hearing.

By Christian LORENZI, Ecole normale supérieure, Paris, France.

The audiogram reflects the limits of the auditory system's  ability to encode acoustic information  in a soundwave. It is measured typically as the detection threshold for a pure tone (a sinusoid) as a function of sinusoid frequency.  The audiogram provides an indication of how well variations in sound pressure are preserved by the auditory system. However, it does not inform about the fidelity of sound encoding. The question of encoding fidelity is crucial because it indicates “the extent to which variations in the strength of a given supra-threshold auditory feature can convey information about the acoustic signal. A limiting case  of encoding  fidelity is that in which sounds are well detected, but no further  information  about the changes in the sounds parameters is  preserved;  in this case, the encoding fidelity is poor” (Wakefield & Viemeister, 1990). Here, we will review psychophysical evidence that such a “limiting case of encoding fidelity” may be frequently observed for patients with sensorineural hearing loss. In other words, we will show that abnormal encoding fidelity may arise despite normal auditory sensitivity for patients with sensorineural hearing loss. This will be demonstrated for several speech identification tasks performed in the low- or mid-frequency regions where patients with high-frequency sensorineural hearing loss show normal or near-normal auditory sensitivity (Léger et al., 2012a,b,c; Bruce et al., 2013; Stasiak et al., 2013; Goodman et al., 2013). Taken together, these results: (i) confirm that normal auditory sensitivity (as measured clinically by the audiogram) is not sufficient to guarantee robust encoding of certain spectral and temporal suprathreshold auditory features critical in daily listening situations, and (ii) raise the need for using discrimination/identification tasks in auditory screening.

Just published: Efficient hearing screening in noise-exposed listeners using the digit triplet test.

Hearing screening in occupational medicine is generally based on pure-tone threshold audiometry. However, reliable and valid thresholds can only be obtained in a sound-proof room, using a high-quality, well-calibrated audiometer, and by a well-trained administrator. Thresholds also need to be determined for several audiometric frequencies. This makes the test time-consuming and expensive, which is not ideal for the screening of large populations. A Speech-In-Noise test (SPIN), by contrast, does not have the abovementioned requirements. Because it can be implemented as a quick automated self-test, possibly over the Internet, a SPIN test is highly advantageous for screening purposes. However, its sensitivity for (isolated) high-frequency hearing loss, as typically seen in noise-exposed listeners, was unclear up to present. In this study, the authors investigated the sensitivity and specificity of the Digit Triplet SPIN test for detecting and monitoring (early-stage) high-frequency hearing loss, and its similarity across two different language versions.

Design: One-hundred eighteen noise-exposed workers, representing a wide range from no to severe high-frequency hearing loss, completed the French or Dutch version of the broadband Digit Triplet self-test in an office-like room. Pure-tone thresholds, collected by a professional audiologist in favorable settings, served as the reference. 

Results: The 84 Dutch-speaking participants showed a very strong linear relation between the reference and the Digit Triplet test, with the pure-tone average at 2, 3, 4, and 6 kHz as a strong predictor (R = 0.86) for the speech-reception threshold. The sensitivity and specificity to detect mild high-frequency hearing loss were 92% (61 of 66) and 89% (16 of 18), respectively. The area under the receiver operating characteristic (ROC) curve was very high (≥0.91) for several degrees of high-frequency hearing loss. With a within-subject standard deviation of only 0.8 dB, the Digit Triplet test also had a low measurement error. The results of the 34 French-speaking subjects showed a highly similar trend. 

Conclusions: The Digit Triplet test proves to have a high sensitivity and specificity for detecting different degrees of high-frequency hearing loss. Given its ease of use, this test is very suitable for screening purposes in occupational medicine, and potentially for the screening of adolescents at risk of recreational noise-induced hearing loss.

For more information, please refer to our paper: 

Jansen S., Luts H., Dejonckere P., van Wieringen A. & Wouters J. 2013. Efficient hearing screening in noise-exposed listeners using the digit triplet test. Ear and hearing, 34, 773–8.

Vacancy: Model-based signal processing for bilateral cochlear implants and hearing aids

A 4 year PhD studentship  is available at ExpORL, Dept. Neurosciences, KU Leuven, Belgium on the topic of sound processing for cochlear implants and hearing aids.
A cochlear implant (CI) can restore the hearing of the deaf by stimulating the auditory nerve directly with electrical pulses, using an array of electrodes inserted in the cochlea in a surgical procedure.
Different speech processing strategies have been proposed to optimize binaural hearing with bilateral CIs and bimodal systems (CI and contralateral HA). These strategies aim to improve coding of binaural cues (interaural time and level differences, ITD and ILD) and are generally designed on theoretical bases and psychophysically evaluated. It would be useful to have a functional model to predict to what extent the binaural cues are presented to the auditory system. In this project, we aim to develop a biologically inspired model that yields the available interaural cues based on existing models of the auditory periphery. The firing patterns for the two ears (either electric-electric or electric-acoustic) will be the input of a binaural model.
Candidates need a degree in electrical or biomedical engineering, physics, or similar, with a good background in signal processing. Good English proficiency is required as well as a strong interest in auditory perception and willingness to work in an interdisciplinary team. Knowledge of human neurophysiology and psychology of hearing is a strong plus.
This PhD is part of the ICanHear project (http://www.icanhear.eu/).


More information about the project and our lab can be found at  http://tinyurl.com/modelsigproc
For more information, please contact Prof. Tom Francart (tom.francart@med.kuleuven.be, phone +32 379840). To apply, please follow the procedures outlined on the KU Leuven job site.

Vacancy: sound processing for cochlear implants and hearing aids

Note: this vacancy has been filled. Please refer to the ICanHear website for another similar vacancy in our lab.

A 4 year PhD studentship  is available at ExpORL, Dept. Neurosciences, KU Leuven, Belgium on the topic of sound processing for cochlear implants and hearing aids. A cochlear implant can restore the hearing of the deaf by stimulating the auditory nerve directly with electrical pulses, using an array of electrodes inserted in the cochlea in a surgical procedure. In this project signal-processing models of loudness perception will be used to design sound processing for combined cochlear implant and hearing aid stimulation. This research should lead to an implementation of the algorithm in a commercial device, in close collaboration with the industry.
Candidates need a degree in electrical or biomedical engineering,  physics, or similar, with a good background in signal processing. Good English proficiency is required as well as a strong interest in auditory perception and willingness to work in an interdisciplinary team. Knowledge of human neurophysiology and psychology of hearing is a strong plus.

More information about the project and our lab can be found at  http://tinyurl.com/bimodal.
For more information, please contact Prof. Tom Francart (tom.francart@med.kuleuven.be, phone +32 379840). To apply, please
include your CV and motivation letter.

Prevalentie en functionele gebruik van gebaren als ondersteunde communicatie bij volwassenen met een verstandelijke beperking

Het SMOG-gebaar "kers"

Op ExpORL wordt onderzoek gedaan naar het gebruik van gebaren als ondersteunde communicatie (OC) bij volwassenen met een verstandelijke beperking (VB). Hieruit blijkt enerzijds dat ongeveer 25% van de volwassenen met een VB in 50% van de Vlaamse voorzieningen, gebaren gebruiken. Anderzijds vonden we dat hun communicatieve mogelijkheden serieus onderschat worden als we hen testen met klassieke taal- en communicatietests.

Wie soms in contact komt met personen met een VB, heeft vast wel al van SMOG gehoord. SMOG staat in deze context niet voor de welbekende luchtvervuiling, maar voor Spreken Met Ondersteuning van Gebaren. SMOG is een gebarensysteem dat in de jaren ’80 specifiek ontwikkeld werd voor personen met een VB die moeite hebben met verbale communicatie. Hieronder verstaan we zowel moeite met het begrijpen van gesproken taal, als moeite met het zelf spreken. Bij SMOG wordt de gesproken taal gecombineerd met gebaren, waarbij we enkel de kernwoorden in de gesproken zin gaan ondersteunen met een gebaar. Het is dan ook een key word signing (KWS) systeem, waarvan er in elk land wel een lokale variant bestaat. SMOG wordt in heel wat scholen en voorzieningen in Vlaanderen gebruikt, maar sinds de jaren ’80 is er zelden onderzoek naar gedaan.

Het onderzoeksproject op ExpORL is, gezien de weinige informatie die voorhanden is over KWS in het algemeen en SMOG in het bijzonder, exploratief van aard. We richten ons in eerste instantie enkel op volwassenen met een VB. In een eerste fase hebben we bekeken welke volwassenen met een VB nu precies SMOG gebruiken in Vlaanderen. Door vragenlijstonderzoek zijn we te weten gekomen dat SMOG gebruikt wordt in ongeveer 50% van de Vlaamse voorzieningen voor volwassenen met een VB. In deze voorzieningen gebruikt ongeveer 25% van de cliënten zelf actief SMOG.

Een deel van deze groep SMOG-gebruikers zijn we nader gaan bekijken. We vonden 48 volwassen SMOG-gebruikers bereid om deel te nemen aan een observatiestudie. Hierin probeerden we het dagdagelijkse, functionele SMOG-gebruik te objectiveren, en te bekijken welke factoren dit SMOG-gebruik beïnvloeden. We namen een testbatterij af bij onze deelnemers om hun communicatieve en motorische vaardigheden in kaart te brengen. Vervolgens filmden we hen in een spontaan gesprek. De eerste resultaten van dit onderzoek tonen aan dat de klassieke taal- en communicatietests onvoldoende de communicatieve mogelijkheden van onze SMOG-gebruikers in een spontaan gesprek weergeven. De verbale mogelijkheden (= gesproken taal) relateren zeer sterk met de testscores, maar de nonverbale mogelijkheden (= gebarengebruik) helemaal niet. Het gevaar bestaat dus om de functionele, dagdagelijkse communicatieve mogelijkheden van een SMOG-gebruiker, op basis van zijn/haar testscores op klassieke taal- en communicatietests, enorm te onderschatten.

In een volgende fase van het onderzoek worden ook de motorische vaardigheden gerelateerd aan de functionele communicatie, en wordt die functionele communicatie verder kwalitatief geanalyseerd. We willen immers ook te weten komen welke gebaren precies gebruikt worden, en voor welke communicatieve functies dit gebeurt. Gebruiken SMOG-gebruikers vooral gebaren om commentaar te geven op gebeurtenissen, of om vragen te stellen, of voor sociale interactie? Op deze en nog veel meer vragen hopen we binnenkort een antwoord te kunnen formuleren!

Meer informatie? stien.meuris <at> med.kuleuven.be

Referentie: 

Meuris, K., Maes, B., and Zink, I. (2013). The use of key word signing in residential and day care programs for adults with an intellectual disability. Manuscript submitted for publication.