Franssen Effect Research Articles

Auditory Spatial Illusion – A Psychoacoustic Study in Young Adults

The Franssen illusion, or Franssen effect (FE), is one of the auditory spatial illusions. Few studies have explored the FE, and the mechanisms underlying it remain unknown. The present study was conducted to clarify the FE occurrence with different tasks and presentation modes in young adults. It also sought to investigate possible neurophysiological similarities between interaural time difference (ITD) cue processing and FE perception. FE perception was evaluated using two different tasks and two presentation modes (i.e., insert phones and loudspeakers). Sound reflections (reverberation) were presented in the diffuse field (loudspeaker mode). ITD performance was investigated using different stimuli delivered via insert phones. No significant difference between the two FE perception tasks was found ( F 1,25 = 0.138, p = 0.713). However, the FE perception showed a significant difference between the two presentation modes (F 1,25 = 434.03, p < 0.001). Spearman’s correlation did not reveal a significant relationship between FE perception and ITD scores ( p > 0.05). The current findings show the importance of reverberation in the FE occurrence. Also, the non-significant correlation between the results of the behavioral binaural temporal resolution test and FE perception in young people with normal temporal resolution may indicate that room reflections (reverberation) complicate the ability to process ITDs (rather than poor ITD processing for the “steady state” portion of signal).

Sensitivity to an Illusion of Sound Location in Human Auditory Cortex.

Human listeners place greater weight on the beginning of a sound compared to the middle or end when determining sound location, creating an auditory illusion known as the Franssen effect. Here, we exploited that effect to test whether human auditory cortex (AC) represents the physical vs. perceived spatial features of a sound. We used functional magnetic resonance imaging (fMRI) to measure AC responses to sounds that varied in perceived location due to interaural level differences (ILD) applied to sound onsets or to the full sound duration. Analysis of hemodynamic responses in AC revealed sensitivity to ILD in both full-cue (veridical) and onset-only (illusory) lateralized stimuli. Classification analysis revealed regional differences in the sensitivity to onset-only ILDs, where better classification was observed in posterior compared to primary AC. That is, restricting the ILD to sound onset—which alters the physical but not the perceptual nature of the spatial cue—did not eliminate cortical sensitivity to that cue. These results suggest that perceptual representations of auditory space emerge or are refined in higher-order AC regions, supporting the stable perception of auditory space in noisy or reverberant environments and forming the basis of illusions such as the Franssen effect.

Reverberation enhances the relative potency of onset cues for sound localization

Onset dominance in spatial hearing has been repeatedly demonstrated through studies of the precedence effect, Franssen effect, and temporal weighting in sound localization. Such effects are often cited as evidence that onset dominance enhances spatial perception, particularly in reverberant environments where post-onset cues are degraded by echoes and reverberation. Most studies, however, have presented sounds in anechoic space or via headphones. Thus, the actual effects of reverberation on onset dominance are not well understood. This study adapted the approach of Stecker & Hafter (2002; JASA 112:1046-57) to measure temporal weighting functions (TWF) for freefield localization of Gabor click trains (4 kHz; 200 Hz click rate) in the presence or absence of simulated reverberation. Reverberant conditions simulated a 10m x 10m room with modestly absorptive walls (α = 0.5). For each potential source location, direct sound and all reflections up to 13th order were synthesized at correct azimuths, intensities, ...

Threshold of the precedence effect in the presence of interfering noise.

Recent work suggests that the precedence effect involves three distinct phenomena: (1) a transient onset that enhances early spatial cues relative to those arriving shortly after; (2) the Franssen effect, in which localization/lateralization estimates established at onset are maintained throughout a long‐duration stimulus; and (3) an ongoing precedence effect, in which leading cues throughout a long stimulus determine localization and lateralization. The goal of the present study was to determine whether the three phenomena are influenced to differing degrees by interference from other sources. Brief and longer trains of noise bursts that alternated between 0‐ and 500‐μs interaural time delays were created to elicit the three types of phenomena. Masked thresholds for the stimuli beginning with 0‐μs interaural delays were nearly identical to those beginning with 500 μs across the different types, indicating no precedence effect for detection threshold. Listeners’ ability to discriminate the 0‐μs‐leading an...

The Franssen effect illusion in budgerigars (Melopsittacus undulatus) and zebra finches (Taeniopygia guttata)

The properties of the Franssen effect (FE) were measured in budgerigars and zebra finches. To elicit the FE, listeners are presented with a signal which has been split into a transient component, carrying an abrupt onset and ramped offset and separated in space from the sustained component which has a slowly rising onset and longer duration. When these two signals are played under certain conditions, the perception is that of a long-duration steady state tone being played at the location of the transient. The birds were trained using operant conditioning methods on a categorization task to peck a left key when presented with a stimulus from a left speaker and to peck a right key when presented with a stimulus from a right speaker. Once training was completed, FE stimuli were presented during a small proportion of trials. The FE was measured at speaker separations of 60 degrees and 180 degrees in both echoic and echoic-reduced conditions. Both species of birds exhibited the FE, although to varying degrees, across conditions. These results show that nonmammals also experience the FE illusion in confusing listening situations in a manner similar to mammals, suggestive of similar auditory processing mechanisms.

The ‘Franssen’ illusion for short duration tones is preattentive: A study using mismatch negativity

When a tone burst is divided into two parts, an onset transient and a sustained tone smoothly fading on, and these parts are delivered to two stereophonically located loudspeakers in a room, a listener gains the impression that the whole sound is coming from the loudspeaker that actually emits merely the transient. Due to this auditory illusion known as the ‘Franssen effect’ (FE), the physical and the perceived lateralizations of the sustained sound become different. A two-block mismatch negativity (MMN) paradigm was used to investigate the stage of auditory processing at which this segregation would take place. In one block, standard stimuli were 100 ms, 1 kHz tone bursts emitted by one of the loudspeakers, and deviant stimuli were their split version, with the sustained part switched to the other loudspeaker. In the other block, the roles of the two stimuli were swapped. A room acoustics software was used for generating the signals to a headphone. The responses recorded from 10 subjects displayed no MMN, although the same stimuli but without the transients evoked prominent MMNs. This indicated that the mechanism underlying this illusion modifies the neural representation of the stimulus with FE in such a way that it becomes similar to that of the stimulus without FE before reaching the input of the preattentive mechanism indexed by the MMN. Considering the possible relationship of this illusion to the precedence effect and also the relevant electrophysiological findings in the literature, we conclude that the primary auditory cortex is the most plausible site of the mechanism leading to the FE.

Cats exhibit the Franssen Effect illusion

The Franssen Effect (FE) is a striking auditory illusion previously demonstrated only in humans. To elicit the FE, subjects are presented with two spatially-separated sounds; one a transient tone with an abrupt onset and immediate ramped offset and the other a sustained tone of the same frequency with a ramped onset which remains on for several hundred ms. The FE illusion occurs when listeners localize the tones at the location of the transient signal, even though that sound has ended and the sustained one is still present. The FE illusion occurs most readily in reverberant environments and with pure tones of approximately 1-2.5 kHz in humans, conditions where sound localization is difficult in humans. Here, we demonstrate this illusion in domestic cats using, for the first time, localization procedures. Previous studies in humans employed discrimination procedures, making it difficult to link the FE to sound localization mechanisms. The frequencies for eliciting the FE in cats were higher than in humans, corresponding to frequencies where cats have difficulty localizing pure tones. These findings strengthen the hypothesis that difficulty in accurately localizing sounds is the basis for the FE.

Cocktails for Franssen: Asynchronous transient bias and other attentional factors in auditory localization

The cocktail-party paradigm was applied to nonspeech signals. Pure-tone stimuli were used in an extension of the Franssen effect, an auditory illusion wherein the location of a slow-onset signal is perceived to be the same as a simultaneous, contralateral sudden-onset signal. Listeners heard simultaneous sudden-onset (transient) and contralateral slow-onset (steady-state) tones in a reverberant environment with a second delayed transient from a third azimuthal location. Results showed that the Franssen effect was either maintained or ‘‘reset,’’ but not reduced. The ongoing steady-state tone was perceived either at the initial-transient and then delayed-transient location, or at the initial transient location only. None of the listeners showed a location bias to the delayed transient tone when its frequency differed from the initial signal frequency or was replaced with noise. In additional conditions based on the ‘‘false Haas effect,’’ consonant–vowel pairs representing transient and steady-state signals were segregated contralaterally in a reverberant space. Results showed no resemblance to the Franssen effect. In general, results indicated that the role of attention is fundamental to the localization of an ongoing stimulus. [Work supported by NIH.]

Reverberation and the Franssen effect

The Franssen effect, in which the location of a sudden-onset (transient) tone can occlude the location of a contralateral slow-onset (steady-state) tone, has been previously shown to occur only in a reverberant space. The nature of the reverberation required—how much and what kind—for the effect to occur has yet to be determined. To explore the role of reverberation in the Franssen effect, listeners were asked to identify the location(s) of transient/steady-state pure tones with frequencies ranging from 250–4000 Hz in a variety of real and virtual contexts: (a) limited (single reflective panel) reverberant enclosure; (b) discrete reflections in an anechoic room; (c) mannequin recordings of a reverberant enclosure presented over headphones; (d) impulse-response filtered signals presented over headphones; and (e) simulated 3-D reverberation via consumer-grade software. Results indicate that the Franssen effect can be a simple, effective means to judge the verisimilitude of a virtual environment. [Work supported by NIDCD.]

The role of masking in the Franssen effect

The sudden onset of a sound which slowly decays at one loudspeaker can occlude the presence and location of the same sound presented simultaneously with a slow rise time at another loudspeaker. This phenomenon, known as the Franssen effect, has been shown in previous research to be specific to low-frequency pure tones in reverberant rooms. To examine the possible mechanisms involved, listeners heard pure-tone transient/steady-state signal pairs at frequencies ranging from 250–4000 Hz from either one loudspeaker or two contralateral loudspeakers in an eight-speaker array. Signals were masked with Gaussian noise. Using a two-interval, forced-choice tracking procedure, thresholds for detecting the steady-state signals were measured. The transient tone was present in both intervals. The results showed increases across frequencies in threshold of approximately 6 dB for two-source (Franssen) conditions compared to single-source conditions. In an auxiliary experiment, signals were masked with tones of differing frequencies. For Franssen conditions, masking was dependent on the interaction of signal and masker frequencies. Both results are discussed in terms of onset dominance, nonecho suppression and free-field masking. [Work supported by NIDCD.]

Franssen à la phones

According to current explanation, the Franssen effect reflects a competition between localization cues from the onset of a tone and cues from the steady-state portion. For a sine tone in a room, the steady-state interaural time and level cues (ITD and ILD) are implausible because of room reflections. Therefore, onset cues dominate, and a listener localizes the tone at its onset position regardless of the true origin of the steady state. To test this explanation we used headphones to simulate the simplest kind of Franssen stimulus: The ITD was fixed throughout the tone. At onset, the ILD was zero. After onset, the ILD grew so as to contradict the ITD in the steady state. Equal numbers of trials were done with abrupt and slow onsets. The listener’s task was to localize the steady-state part of the tone. A Franssen analog was identified by comparing responses to abrupt-onset and slow-onset tones. Results were compared with the standard Franssen effect in a large room. Clear evidence of a Franssen effect, with its dependence on plausibility, appeared in the headphones experiments, but the data were less dramatic than data obtained in the room. [Work supported by the NIDCD.]

Franssen effect as a function of carrier waveform, carrier level, and room acoustics.

The Franssen effect was studied as a function of the frequency of the source and echo sounds, the difference in frequency and level between the source and echo sounds, and as a function of the amount of room reverberation. A modification of the single‐interval procedure used by Hartmann and Rakerd [J. Acoust. Soc. Am. 86, 1366–1373 (1989)] was used to measure the degree to which the source and echo sounds were fused in terms of determining the location of the perceived sound. The amount of fusion was a nonmonotonic function of frequency, decreased as the frequency and level difference between the echo and source increased, and is greater for a room with more reverberation. The results will be discussed in terms of the precedence effect, listener’s prior experience, and the plausibility hypothesis. [Work supported by a Program Project Grant from NIDCD and a grant from the AFOSR.]

Lateralization mechanisms and Deutsch's octave illusion

The octave illusion described by D. Deutsch [Hearing Research and Theory, edited by J. V. Tobias and E. D. Schubert. Vol. 1. pp. 100–142 (1981)] is a phenomenon observed when two continuous sinusoidal tones bearing an octave relation (e.g., 400 and 800 Hz) are presented dichotically such that the ear/frequency relationship is switched at regular intervals (every 250 ms). The illusion consists of the impression, reported by a majority of listeners, that the tone with the higher frequency in one ear (usually the right) alternates with the low frequency in the other ear. The present paper attempts to account for the mislocalization component of the illusion, based on two recognized properties of the binaural system: (1) In order for a stimulus to be lateralized, the dichotic information must be in the same band(s) [e.g., L. A. Jeffress, Foundations of Modern Auditory Theory, edited by J. V. Tobias (Academic, New York, 1972), Vol. 2, pp. 349–368]; and continuous pure tones may be mislocalized based on information not deriving from their onset [i.e., the Franssen effect, W. M. Hartmann and B. Rakerd, J. Acoust. Soc. Am. 86, 1366–1373 (1989)]. It will be shown that these two properties, together with some basic peripheral processing characteristics, can lead to lateral shifts in the dichotic image, consistent with the illusion. [Work supported by the Department of Veterans Affairs.]

Localization of sound in rooms. IV: The Franssen effect.

The Franssen effect is an illusion that causes human listeners to make large errors in localizing a sound source. This paper describes steps taken to convert the illusion into an experiment in order to study the localization precedence effect as it operates in rooms. The results of the experiment suggest that there are two components to the illusion: The first is the inability of listeners to localize a sine tone in a room in the absence of an onset; the second is the obscuring of modulation cues by the irregular transient response of a room. Experiments show that the Franssen effect fails completely in an anechoic environment, as expected if the effect depends upon the implausibility of steady-state cues in a room. The Franssen effect also fails when the spectrum of the sound is dense.

The Franssen effect and the localization plausibility hypothesis

The Franssen effect is obtained with two loudspeakers in a room. If a sine tone is abruptly turned on at the left loudspeaker, then slowly faded off while the right loudspeaker is slowly faded on, a listener will judge that the tone continues to come from the left loudspeaker, even though the left loudspeaker is not sounding at all. Our own studies of localization of sound in rooms have led to a principle of localization called the “plausibility hypothesis.” One of the predictions of this hypothesis is that in an anechoic room the Franssen effect should fail [B. Rakerd and W. M. Hartmann, J. Acoust. Soc. Am. 78, 524–533 (1985)]. Experimental studies are reported using the Franssen stimulus both in an ordinary room and in an anechoic room. The results of the experiment support the prediction. [Work supported by the National Institutes of Health.]