Brian B. Monson - firstname.lastname@example.org
National Center for Voice and Speech
University of Utah
136 S. Main Street
Salt Lake City, UT 84101
Brad H. Story - email@example.com
Andrew J. Lotto - firstname.lastname@example.org
Department of Speech, Language, and Hearing Science
University of Arizona
PO Box 210071
Tucson, AZ 85721
Popular Version of Paper 3aMU12
Presented Wednesday morning, May 16, 2012
163rd ASA Meeting, Hong Kong
If you’ve ever played with the "treble" knob on your radio or music player, you know that there is difference in the quality of the sound you hear. When listening to a singing voice, a treble boost makes it “brighter” or even “shrill”. If you turn the treble down it becomes more “dull” or perhaps “muffled”. Sound recording engineers and recording artists know this, too, and they use it to their advantage by changing the amount of treble in a vocal recording to improve the voice aesthetic.
Despite the artistic understanding, treble in singing and speech has been largely ignored by the scientific community -- and for good reason. Treble frequencies (above ~6000 Hz) are beyond the highest notes on a piano. While the human voice creates energy in this range in the form of aspiration noise (from consonants) and very high overtones (from vowels), the energy is much weaker than energy at the low frequencies (Figure 1). High frequencies are also much more difficult to measure and analyze accurately. For these and other reasons, treble frequencies were not included in telephone transmission when telephone technology was developed. For the last several decades, this standard “telephone bandwidth” (300–3400 Hz) has been perpetuated, so your mobile phone still doesn’t transmit treble frequencies.
Figure 1. Spectrogram showing acoustical energy (sound) created by a human voice up to 20,000 Hz. The current cell phone bandwidth (shaded blue) only transmits sounds between about 300 and 3400 Hz. High-frequency energy above 5000 Hz (shaded red) has information potentially useful to the brain when perceiving singing and speech.
What are you missing out on when you talk on your phone? Until recently, treble frequencies were only thought to affect some aspect of voice quality. We now know that it’s more than just quality. In fact, these high frequencies carry a surprising amount of information about a speech sound. Treble frequencies are sufficient for a listener to determine the gender of a talker and at least some of the linguistic message of a phrase. This is likely because of treble differences between different speech sounds. We decided to examine whether treble differences might be able to distinguish different vocal sounds, allowing listeners to extract information from what they hear in the treble. One interesting finding was that normal singing had more treble than normal speech, but loud speech exhibited more treble than loud singing. Why would this be? It suggests that treble may be more important for communication in situations where loud speech is needed, such as in a noisy environment. We also found that treble frequencies travel more directionally out of the mouth towards the front of the talker (whereas low frequencies spread omnidirectionally once they exit the mouth), making treble most advantageous when a talker is directly facing a listener. These are clues that point toward treble frequencies being much more rich in information than scientists have supposed for decades.
What does all of this mean for you? It could explain why you have a hard time understanding someone on your phone, especially when sitting on a noisy train or at a cocktail party. It elucidates why VoIP sounds better than mobile phones (e.g., Skype transmits frequencies up to 7000 Hz, and will hopefully be moving to a higher range soon). If you have high-frequency hearing loss and hearing aids, it could explain why you have a hard time understanding conversation in noisy environments. It may also be why you like certain singing voices better than others. If you are a singer, treble may be a way you could improve your singing voice aesthetic. All in all, treble appears to house a lot of information that we're just now starting to explore. Perhaps scientists could learn a lesson or two from musicians more often.
Monson, B. B., Lotto, A. J., and Ternstrom, S. (2011). "Detection of high-frequency energy changes in sustained vowels produced by singers," Journal of the Acoustical Society of America, 129, 2263-2268.
Monson, B. B. (2011). High-Frequency Energy in Singing and Speech. Doctoral dissertation submitted to University of Arizona.
Monson, B.B. et al. (2011) Lay-language Paper: http://www.acoustics.org/pressroom/httpdocs/162nd/Monson_5aSCb3.html