Those of you who have read my earlier blogs, How Effective Is the iPad as a Therapy Tool, Minimal Pairs App, and Is the iPad Worth the Investment, have read that voice recognition technology is highly inaccurate. The technology works best when it recognizes the user and when it analyzes large communication segments, such as words in phrases. The finer the analysis needed the less accurate the technology becomes. This means that voice recognition technology is highly unreliable when it comes to analyzing sounds in words, syllables and in isolation. I have examined two apps that target sound production. One relies on the adult to determine whether or not the phoneme is correctly produced and rewarded. For the other, voice recognition is an integral part. This difference is critical.
Speech Stickers: This
app was developed by speech pathologist Carol Fast. This app targets the
following sounds: /m/, /p/, /b/, /t/, /d/, /n/, /k/, /g/, /s/, /ʃ/. These
phonemes are practiced as either CV (consonant-vowel) or VC (vowel consonant).
The vowels used in the app are: /ʌ/, /ɒ/ (shown as /a/ in the app), /u/,
/i/ and /o/. The workings of this app are straightforward. The app opens to a screen on which one can tap on play, options or instructions. Tap on play and a screen, with the first phoneme /m/, is displayed at the top of the screen accompanied by the narrator saying /m/. At the bottom of the screen are five faces: a girl, boy, alien, cat and monster. The child selects one of the faces. When the face is tapped, it appears on the middle of the screen and /m/ is said again. On the right side of the screen is a vertical bar, or meter, with numbers 0-5 in ascending order. The goal is for the child to produce the sound five times. Each time the child produces a sound, the adult either taps on the green check mark, for a correct repetition or red X if incorrect. Either the check mark or the X must be tapped for the face to show up on the meter and for the sound to be produced again. This level of control allows the adult to carefully monitor the child’s production and give feedback before the game can progress. Once the five repetitions are noted as either correct or incorrect, a new screen, with 10 objects (rocket, hiking boot, bomb, shark head, bus, train, tractor, wand, balloon and a dragon), appears at the bottom of the screen. The child taps on the picture of choice and watches as the object performs a short action.
If one wishes to switch sounds or work on the sound in syllables, one taps twice on the phoneme at the top of the screen. (I like the idea of tapping twice to keep the quick tapping fingers of children from making the changes.) This opens a screen showing all the targeted apps as well as the sound in CV and VC. One can switch phonemes or select to work at the syllable level by tapping on a specific syllable. There is a scoring feature for each screen. As noted above, the child’s production is marked as correct or incorrect. The number of correct is tallied and shown as a percentage for only that screen. There is no database to refer to at another time. Also, the app can be used only in vertical position.
I like that care is given to the production of the target sounds; their clarity in isolation and syllables is good overall. The graphics and animations are basic.This app can be used with children who are at the beginning levels of learning a new sound. I can see the app being used as a fun diversion from basic sound and syllable repetition that is typical at the early stages of phoneme learning.
Ages: 3-5 years
Ratings: +++
Developer website: Serioustree.com
Cost: $14.99
Tiga
Talk: This app offers 23 phonemes designed to elicit the sounds in isolation.
One taps on the target sound one wants the child to say. A screen appears with
an animated animal producing a sound, a picture of an item beginning with the
targeted sound, the letter for the sound and a sound meter whose lights move with the speaker's volume. On the left side of the meter is a
mouth icon. Tap on the mouth icon and the child is able to view a mouth moving
as it produces the targeted sound. The animal instructs the child to
say the sound. Once the child has said a sound at least two times, he is
praised and stars explode on the screen. The child is then told, “You’re doing
very well. Let’s play a game to celebrate.” There are four game choices:
Driving, Balloon Pop, Smash and All, which offers a selection of the three
games. The child is asked to say the
target sound. Choose Drive and a vehicle moves closer to its destination each
time a sound is said. Balloons cover a vehicle in Balloon Pop; balloons pop, to
reveal a vehicle, each time the child says a sound. Each time a child says a
sound in Smash, tires, hay bales, logs, boxes, rocks or tires are thrown aside
to reveal a vehicle. I used this app
with one of my clients who does not have an articulation problem. He enjoyed
repeating the sound after the animated character told him it was his turn. He
also enjoyed playing the game.
The website for Tiga Talk states,
“Game play rewards are based entirely on participation, not
accuracy, so the child is constantly getting positive feedback
as long as they are trying to make sounds!” This is important to keep in
mind before purchasing the app. A child who cannot yet produce a particular
sound in isolation, consistently, should be
supervised when he uses this app. Here is why. Let us say that the target
sound for the game is /d/. Whether I said /d/, /g/, /v/, or “muh,” the app
recognized the sound as the correct production. This leads me to disagree with website's statement
that Tiga Talk “can improve speech clarity and articulation.” Unsupervised, the child will
get rewarded whether he says the target sound or another sound. This happened
for every sound the app offered.
Another word of warning: a fast talking child, eager to move
the boat across the lake, let us say, may be able to say the incorrect sound a
few times (thereby self-reinforcing an incorrect sound production) and will receive positive feedback from
the game before the adult might have a chance to intervene. I am concerned that
this has the potential to make the adult’s task more difficult. The child might
wonder why the app said, “Great job,” but the adult indicated that the sound
was not correctly produced. I would have preferred that this app not rely on
voice recognition, but rather on adult feedback as Speech Stickers does. Also,
visuals for mouth movements, is, like voice recognition, a tricky feature to
put into apps. I found that the mouth movements, except for bilabials,
difficult to see and easy to misinterpret.
Tiga Talk has serious limitations as a speech app. For children who can easily say the sounds without error, it can be a fun game that children will enjoy.
Ages: 3-6 years
Ratings: +1/2
Developer website: tigatalk.com
Cost: $4.99
No comments:
Post a Comment