Cocktail party effect

The cocktail party effect describes the ability to focus one's listening attention on a single talker among a mixture of conversations and background noises, ignoring other conversations.^[1] The effect enables most people to talk in a noisy place. For example, when conversing in a noisy crowded party, most people can still listen and understand the person they are talking with, and can simultaneously ignore background noise and conversations. Nevertheless, if someone calls out their name from across the room, people will sometimes notice (the "own name effect").^[2]

Another aspect of the cocktail party effect is de-reverberation^{[citation needed]}. The human auditory system seems to ignore most of the reflected sound, because it arrives from other directions than the direct sound.

The auditory system can also switch the direction of attention and turn from one sound source to another^{[citation needed]}.

Binaural processing

The cocktail party effect works best as a binaural effect, which requires hearing with both ears. Persons with only one functional ear are much more disturbed by interfering noise than people with two healthy ears.^{[citation needed]}. However, even without binaural location information, people can, even if with greater difficulty, selectively attend to one particular speaker if the pitch of their voice or the topic of their speech is sufficiently distinctive.

The binaural aspect of the cocktail party effect is related to the localization of sound sources. Experiments have shown^[3] that the auditory system is able to localize at least two sound sources simultaneously and assign the correct sound source characteristics to these sound sources simultaneously too. In other words, as soon as the auditory system has localized a sound source, it can extract the signals of this sound source out of a mixture of interfering sound sources.

It is assumed^[who?] that the auditory system performs a kind of cross-correlation function between both ear signals^{[citation needed]}. A cross correlation function projects signals onto an axis, which corresponds to the time difference between both ear signals. For example, sound with an interaural time difference of 0.3 ms is projected onto the 0.3 ms position of the correlation axis. If multiple sound sources are present, then complex correlation patterns appear. The statistical parameters of these patterns, like mean value and variance, depend on the directions and levels of the sound sources. The auditory system is obviously able to analyze these patterns and determine the signals of a dedicated sound source.

Attempts have been made to simulate the cocktail party effect by technical means^{[citation needed]}. Cocktail party processors have been constructed which can extract the signal of a single sound source out of a mixture of sound sources^{[citation needed]}. There are cocktail party processors, which are based on correlation functions, evaluating interaural time differences, but there are also cocktail party processors for interaural level differences^{[citation needed]}. However, the principles of the human cocktail party effect are not yet fully investigated. Technical cocktail party processors do not yet reach the capabilities of the human auditory system.

Monaural processing

The auditory system does not only use methods for a direction specific signal processing, it also uses monaural effects for noise reduction. If the characteristics of a desired signal are known (like the characteristics of speech) or can be estimated (like expected phonemes at observed mouth movements), then all signal components which do not match the expected characteristics can be suppressed and the disturbing effect of this noise can be reduced.

The human pinna (the external flap of skin and cartilage of the ear) is a directionally-dependent filter that selectively removes particular frequencies, based on the direction from which sound comes. This filter can distinguish sounds from above vs. below, and from front vs. back, even when only a single ear is used.

Control of the direction of attention

In the early 1950s much of the early work in this area can be traced to problems faced by air traffic controllers. At that time, controllers received messages from pilots over loudspeakers in the control tower. Hearing the intermixed voices of many pilots over a single loudspeaker made the controller's task very difficult.^[4] The effect was first defined and named by Colin Cherry in 1953.^[5] Cherry conducted attention experiments in which subjects were asked to listen to two different messages from a single loudspeaker at the same time and try to separate them. His work reveals that our ability to separate sounds from background noise is affected by many variables, such as the gender of the speaker, the direction from which the sound is coming, the pitch, and the rate of speech.^[5]

Also during the 1950s, Broadbent^[6] conducted dichotic listening experiments: subjects were asked to hear and separate different speech signals presented to each ear simultaneously (using headphones). From the results of his experiment, he suggested that "our mind can be conceived as a radio receiving many channels at once": the brain separates incoming sound into channels based on physical characteristics (e.g. perceived location), and submits only certain subsignals for semantic analysis (deciphering meaning). In other words, there exists a type of audio filter in our brain that selects which channel we should pay attention to from the many kinds of sounds perceived. Broadbent proposed that the filter was hypothesized between the sensory buffer and short-term store (what is now called working memory) that prevents overloading memory. This is called Broadbent's filter theory.^[7] There is some empirical evidence to support this theory, though it has been criticized.^[8][9] There is evidence to suggest it is not just dependent on the physical characteristics of the stimulus. There are stimuli, such as someone calling your name,^[10] which will sometimes be processed. Treisman (1960) found evidence to suggest that the filter just increases the threshold of an unattended stimulus, which sometimes can be processed if it is particularly important or has relevant meaning. Deutsch and Deutsch (1963) also offered an alternative to Broadbent's filter theory arguing for late selection depending on whether the stimulus needs to be responded to.

Since the human auditory system can localize and process several sound sources simultaneously, it has to decide, to which of these directions attention shall be paid. The Franssen effect gives some information about how the auditory system works and where the limitations are.

Therefore the auditory system can only localize a sound source if the directional information is reliable (like at the onset of a specific sound). If no reliable directional information is available, the auditory system seems to hold the last information as long as no better information is available:

The cocktail party effect can occur both when we are paying attention to one of the sounds around us and when it is invoked by a stimulus which grabs our attention suddenly.^[10]

This phenomenon is still very much a subject of research, in humans as well as in computer implementations (where it is typically referred to as source separation or blind source separation). The neural mechanism in human brains is not yet fully clear.

See also

• Auditory processing disorder
• Auditory scene analysis
• Cognitive science
• Echoic memory
• Language processing
• King–Kopetzky syndrome
• Cocktail party
• Dream speech
• Spatial hearing loss

References

1. ^ Bronkhorst, Adelbert W. (2000). "The Cocktail Party Phenomenon: A Review on Speech Intelligibility in Multiple-Talker Conditions" (pdf). Acta Acustica united with Acustica 86: 117–128. http://eaa-fenestra.org/products/acta-acustica/most-cited/acta_86_2000_Bronkhorst.pdf. Retrieved 2010-04-18. 
2. ^ http://www.csun.edu/~vcpsy00h/students/arousal.htm
3. ^ Slatky, Harald (1992): Algorithms for direction specific Processing of Sound Signals - the Realization of a binaural Cocktail-Party-Processor-System, Dissertation, Ruhr-University Bochum, Germany
4. ^ Sorkin, Robert D.; Kantowitz, Barry H. (1983). Human factors: understanding people-system relationships. New York: Wiley. ISBN 0-471-09594-X. http://www.amazon.com/Human-Factors-Understanding-People-System-Relationships/dp/047109594X. 
5. ^ ^{a b} Cherry, E. Colin (1953-09). "Some Experiments on the Recognition of Speech, with One and with Two Ears". Journal of Acoustic Society of America 25 (5): 975–979. doi:10.1121/1.1907229. http://scitation.aip.org/getabs/servlet/GetabsServlet?prog=normal&id=JASMAN000025000005000975000001&idtype=cvips&gifs=yes&ref=no. Retrieved 2010-07-10. 
6. ^ Broadbent, D.E. (1954-03). "The role of auditory localization in attention and memory span". Journal of Experimental Psychology 47 (3): 191–196. doi:10.1037/h0054182. PMID 13152294. http://www.sciencedirect.com/science?_ob=ArticleURL&_udi=B8JB9-4NRM2HM-C&_user=10&_coverDate=03%2F31%2F1954&_rdoc=1&_fmt=high&_orig=search&_sort=d&_docanchor=&view=c&_searchStrId=1396365743&_rerunOrigin=google&_acct=C000050221&_version=1&_urlVersion=0&_userid=10&md5=0c0effd3d5eef4e3e8a32957c2bf6c2f. Retrieved 2010-07-10. 
7. ^ Broadbent, D.E. (1958). Perception and communication. New York:Pergamon.
8. ^ Norman, Donald A. (1969-02-01). "Memory While Shadowing". Quarterly Journal of Experimental Psychology 21 (1): 85–93. doi:10.1080/14640746908400200. PMID 5777987. http://www.informaworld.com/smpp/content~db=all~content=a776269794. Retrieved 2010-07-10. 
9. ^ Lachter J, Forster KI, Ruthruff E (October 2004). "Forty-five years after Broadbent (1958): still no identification without attention". Psychol Rev 111 (4): 880–913. doi:10.1037/0033-295X.111.4.880. PMID 15482066. http://www.sciencedirect.com/science?_ob=ArticleURL&_udi=B6X04-4DKK3N7-3&_user=10&_coverDate=10%2F31%2F2004&_rdoc=1&_fmt=high&_orig=search&_sort=d&_docanchor=&view=c&_searchStrId=1396377237&_rerunOrigin=google&_acct=C000050221&_version=1&_urlVersion=0&_userid=10&md5=a621c3972159cc39266da3716c8811ea. 
10. ^ ^{a b} Moray, Neville (1959-02-01). "Attention in dichotic listening: Affective cues and the influence of instructions". The Quarterly Journal of Experimental Psychology 11 (1): 56–60. doi:10.1080/17470215908416289. http://www.informaworld.com/smpp/content~db=all~content=a791213717. Retrieved 2010-07-10. 

Further reading

• Potts GF, Wood SM, Kothmann D, Martin LE (October 2008). "Parallel perceptual enhancement and hierarchic relevance evaluation in an audiovisual conjunction task". Brain Res. 1236: 126–39. doi:10.1016/j.brainres.2008.07.104. PMC 2652515. PMID 18723003. http://www.pubmedcentral.nih.gov/articlerender.fcgi?tool=pmcentrez&artid=2652515. 
• McLachlan N, Wilson S (January 2010). "The central role of recognition in auditory perception: a neurobiological model". Psychol Rev 117 (1): 175–96. doi:10.1037/a0018063. PMID 20063967. http://psycnet.apa.org/index.cfm?fa=search.displayRecord&uid=2009-25263-003. 

External links

• Audio examples of a cocktail party processor
• Handbook for acoustic ecology
• Dichotic Listening Experiment
• Early and Late Selection Theories of Attention by Nick Milton (May 1994)
• Neurofunctional studies in birds and bats
• A Primer on 'The cocktail party problem' by Prof. Josh McDermott in Current Biology
• [1]