They formants (=the frequencies where there are amplitude peaks) change over the course of the utterance, as you go from "l" to the "aw" vowel to the "er" vowel to "l".
The dark-shaded regions are the highest amplitude (=loudest, though 'loudness' and 'amplitude' are subtly different for reasons we won't worry about here). Think of the shading on a topographic map: the shading shows where the mountains are. The shading shows which of the frequencies are loudest. In speech, we usually focus on those frequencies between 0 and 10,00 Hz, but since the youngest, healthiest humans can hear up to 20,000 Hz, we sometimes show 0-20,000 Hz, as in the above. There are many different frequencies in speech-many different 'tuning forks' in our analogy-and we need to represent as many of these as we need to describe speech. The x axis is time, so a spectrogram can show things changing over time. The formants are the horizontal stripes in the picture above, called a spectrogram. So, the frequencies (the 'tuning forks') that are loudest are what we call formants.
Hopefully I've been clear so far, because here's where it gets weird. The same pitches are present–the tuning forks are always vibrating–but the loudness of each of the frequency components (each of the tuning forks) changes from vowel to vowel.
The difference between an "ee" and "ah" vowel is that some of the frequencies that are especially loud in "ee" are quiet in "ah" and vice versa. But speech has many more frequency components than just that lowest-frequency component. We hear those changes as changes in the frequency of the voice, like the pitch glide upward when you ask a yes-no question, or the pitch glide downward when you make a statement. We can change the frequency of the so-called 'lowest frequency fork' by changing the tension in our vocal folds (layperson: 'vocal cords'), which causes them to vibrate more slowly or more quickly. If the lowest-frequency tuning fork vibrates at 120 cycles per second, then the tuning forks will be at integer multiples of 120 Hz: 120, 240, 360, up to infinity. If the lowest-frequency tuning fork vibrates at 100 cycles per second, then the tuning forks will be at integer multiples of 100 Hz: 100, 200, 300, up to infinity. Think of it like hundreds of tuning forks playing at once. The production of sounds produced with a relatively open vocal tract (like the vowels in "Laurel" and "Yanny") and some consonants (like the "l", "r", "y", and "n'" sounds in "Laurel" and "Yanny") have infinitely many frequencies in them. On Facebook, the University of Minnesota's Benjamin Munson shared a cogent analysis that he provided to an inquiring reporter, and he has graciously agreed to have an expanded version of his explainer published here as a guest post. Various linguists have chimed in on social media (notably, Suzy J. Laurel perceptual puzzle has been fiercely debated (see coverage in the New York Times, the Atlantic, Vox, and CNET, for starters). What do you hear?! Yanny or Laurel /jvHhCbMc8I This time around, the dividing line is between "Yanny" and "Laurel." A peculiar audio clip has turned into a viral sensation, the acoustic equivalent of " the dress" - which, you'll recall, was either white and gold or blue and black, depending on your point of view.
0 kommentar(er)
