The Cocktail Party Problem

(At)


Imagine you’re at a cocktail party, you’re wearing your best attire for the night, hair fixed, posture straight, ready to mix and mingle. You can hear the music playing softly in the background, mixing with the conversations and the laughter. You decide to stand by the bar, take a sip of your margarita, carefully pushing aside the lemon with your finger (or scotch, whichever you prefer!). You stare into the crowd, and from the distance, someone you find astonishingly attractive starts walking up to you and strikes up a conversation. But for some reason, you can’t understand what that person is saying –you just can’t separate that person’s voice with the rest of the crowd and the music. What do you do next? GO!


a. Panic hysterically and say OMG WTF is WRONG WITH ME!?!
b. Keep your poise and excuse yourself while assuming that you’re the only one who can’t distinguish people’s voices apart from the crowd, and then call a doctor.
c. Complain to the manager that the bartender drugged you.
d. Ask the person “Would you like to go somewhere more quiet?”


“D” would probably be the first thing you would do, and it might get you far (tonight is going to be LEGEN-“D”ahhh you know what I mean). Isn’t it wonderful that we can distinguish one voice from another? Aren’t our auditory systems just awesome?!


Everday , we often use divided attention to determine where the sounds we hear come from (Jones & Yee as cited in Yost, Dye, & Sheft, 1996). We then use selective attention to process a single source of our interest while ignoring the rest of the sounds (Yost, Dye, & Sheft, 1996).


In an experiment simulating a cocktail party (Yost, Dye, & Sheft, 1996)–although it seemed to me it’s not the kind anyone would want to go to –the researchers asked their participants to listen to pre-recorded 46 NU-6 words --NU-6 is the Northwestern University Auditory Test No. 6, and it is used in studies like this that deal with the auditory system, including studies that deal with people who have difficulty hearing (Wilson & Strouse, 2002), the letters of the alphabet, and the numbers 1 to 9 spoken by seven different male talkers.


In one experiment condition, the listener would hear these recordings from speakers situated across the room. In another condition, a microphone was placed where the listener in the previous condition was situated. The microphone was connected to one headphone –only one ear heard the sounds, and the listener was situated in a sound-proof room. In another condition, KEMAR replaced the microphone, and the listener heard the sound through headphones –the two ears could hear the sound. KEMAR stands for Knowles Electronic Manikin for Acoustical Research.


The listeners were asked to enter into a computer all of the words, letters, or numbers that they heard. After all of the utterances were presented, they were presented again in the same order, and this time, the listeners indicated from which loudspeakers the words, letters, and numbers they listed down originated, and they were allowed to listen to it as many times as they wanted. Those that they did not list down were not included in the second task.


The results of the experiment showed that the listeners can identify “more than 90% of the letters and numbers delivered two at a time and over 80% of the letters and numbers delivered three at a time.” It also showed that utterances that were unfamiliar were more difficult to identify, but that the listeners performed well above chance –they didn’t just guess their answers.


The listeners also reported that it was easier to identify where the sound was coming from when they were situated in the room and when KEMAR was used than to identify what was being said on the speakers. They also listened to the sound for a fewer number of times to locate the sound than to identify what was being said.


It was also found that there were no localization cues available for the condition with only one headset (thank your god for your binaural superpowers!). This means that using only one ear to locate where sound comes from is more than difficult! So next time you’re at a party, don’t forget your ears –one is not enough!


Works Cited

Yost, W. A., Dye, J. R., & Sheft, S. (1996). A simulated "cocktail party" with up to three sound sources. Perception & Psychophysics , 58 (7), 1026-1036.

Wilson, R.H., Strouse, A. (2002). Northwestern University Auditory Test No. 6 in multi-talker babble: A preliminary report. Journal of Rehabilitation Research and Development, 39 (1), 105-113.