Ultrasound frames and videos

The black-and-white, blurry images and videos are recordings from an ultrasound device. To make these recordings, an ultrasound probe is held in place on the underside of the speaker’s jaw, between their chin and their neck. The probe produces an image of a vertical plane directly above it. In these images, the front of the mouth is on the right and the root of the tongue is toward the left. Brighter areas indicate a rapid change in density; the surface of the tongue appears as a white line.

These images are the basis for the SSANOVA diagrams, each of which represents the average of 5-7 repetitions of the same sound by one speaker.

SSANOVA diagrams

Each SSANOVA diagram (such as the one below) compares the position of the tongue when the velarized, or non-palatalized, version of the consonant is pronounced, to the position of the tongue when the palatalized version of the same consonant is pronounced. In these figures, the front of the mouth is to the right. The green lines represent the tongue’s average position when pronouncing the palatalized consonants, and the black lines represent the tongue’s position when pronouncing the velarized consonants.

The green line and the black line represent averages, derived from processing several raw ultrasound frames. Because ultrasound images are “noisy” and processing is involved, it’s possible that the “real” average position of the tongue is not exactly where the green and black lines have been drawn. Statistics tells us, though, that we can be confident that the actual tongue position falls within the range marked by the dotted lines on either side of each solid line (the “confidence interval”). You’ll notice that the confidence interval is larger at the end of each line than at its middle. This is because ultrasound imaging is more reliable for the tongue body (the middle of the image) than for the tongue tip or root (the edges of the image), as well as the nature of regression analyses.