August 13, 2003
Speakers, Musicians Adjust Tempos To Mimic What They Just Heard
Melissa Jungers, Caroline Palmer, and Shari Speer, of Ohio State University have discovered that musicians and speakers adjust the speed and, in the case of the speakers, even the pattern of phrase breaks to match what they just heard.
In one study, the researchers used 16 experienced adult pianists. The pianists sight-read two melodies to establish their preferred performing rate. They then alternated listening to computer-generated melodies and playing different melodies on the piano. They heard 10 melodies and played 10 melodies. The participants were led to believe they were participating in a memory test – they were given no direction about how slow or fast to play.
The melodies they heard were performed at relatively slow or relatively fast rates. When they played, the pianists were provided notations that did not include bar lines or time signatures so there would be no indication of meter or any indirect indication of a rate at which the tune should be played.
After listening to a slow melody, the pianists played their melody slower (an average of 6.8 seconds) when compared to when they had first heard the faster melody (an average of 5.3 seconds). The pianists preferred rate – the rate they played before they heard other melodies – was in between those two times, at 6.1 seconds.
The second study was designed to be as close as possible to the first study – except that the researchers examined speech rather than music. In this study, 64 native English speakers spoke 10 short sentences (6 to 7 words each). First, speakers read two sentences aloud from a computer screen to measure their preferred speaking rate. Next, speakers alternated listening to and reading sentences. In some cases, they listened to sentences spoken at a fast rate, in other cases at a slow rate. The participants were not told how fast or slow to speak – they were simply told to try to remember the sentences for a later memory test.
For 43 of the 64 speakers, they spoke faster after they heard a sentence spoken at a fast tempo and spoke slower when they heard a slower speaking delivery.
When they heard the slow speakers, their sentences averaged 1.81 seconds long, while when they heard the fast speakers their sentences averaged 1.72 seconds long. Their preferred rate of speech averaged 1.80 seconds per sentence.
In addition, this study showed that speakers were also influenced by the pattern of phrase breaks in the recordings they heard. If the recordings had the largest pause in the sentence after the verb, for example, that is where the speakers tended to put their biggest pause.
It would be interesting to play music at different rates and then see if people will speak at different speeds as a result.
Unpublished measurements by myself (on unsuspecting subjects) certainly showed that driving speed is affected by the music the driver is listening to. In one case, changing from a slow ballad to "We will rock you" by Queen resulted in a speed increase of 20 km/h. This was too much really and the chortles from the recording experimenters in the back seat notified the subject as to what was going on.
Pat, brilliant experiment. This idea obviously calls out for being done more systematically with a lot more people. Researchers have to use the same section of road with a lot of people who didn't know the real reason for the trip.
It seems to me that this could be done very easily in a systematic way with a bunch of people who not only don't know the real reason for the trip, but also don't know there's a study of anything being done by anyone, although this might be a shade intrusive:
From http://abcnews.go.com/sections/scitech/DailyNews/cybershake030131.html :
"The billboards contain antennas that are sensitive enough to determine what radio stations are tuned in on the cars that pass by.
"Our technology can capture up to 80 percent of all the traffic passing this unit, whether that's 1,000 cars or 10,000 cars," says Phyllis Neal, the chief executive officer of Mobiltrak, which created the system.
Once the system determines what station — and therefore the type of music — is being tuned in by the majority of the car traffic below, ads can be targeted to best suit those marketing and demographic research data of those station's listeners."
If a more-precise system could be devised to detect the station from individual motorists' antennas, a team could check the station of those motorists along a given stretch were listening to, and periodically measure their speed every x seconds. This could later be compared to time-indexed recordings of the area's radio stations.
For that matter, the system as it exists now--coupled with a LOT of radar guns--could probably extract the same results in aggregate. "80% of the passing cars were listening to Sweet 98.5, and when Beddingfields' 'Get Through This' came on following a Celine Dion ballad, 72% of the cars sped up an average of 6 mph..."
Disclaimer: I'm not a scientist or professional researcher of any sort. Just thinking out loud, err, on computer.
Max, That's wild. Yes, if you can watch the same car at different points on its trip and know its speed and the song playing on the radio you could discover the relationship with car speed.
You could also discover indications of what kinds of music increase the probability of accidents. Picture the notice that would come in the mail from a car insurance company:
Dear Customer, due to your observed recent shift in listening style from Chopin and Bach to Guns 'N Roses and Led Zeppelin we find it necessary to increase your car insurance premium by 30%. If you wish to avoid further premium increases please avoid listening to rap music. Also, we have observed that you have an XM Radio subscription. We can promise you a 40% fare reduction if you opt to restrict your listening habits to their 14 certified low accident rate channels. If you desire to benefit from an immediate premium reduction reply to this message with the title "Low Risk Radio Premiums" to authorize us to remotely update the firmware of your car radio to restrict your listening choices to only certified low accident rate music.
D'oh! Here I am, minding my own business, just having fun devising a hypothetical study set-up, and suddenly I turned over control of my radio to the auto insurance industry.