Filtered Speech Experiment The purpose of the filtered speech experiment was to look for three things. The first was to see/hear which frequencies can be eliminated from speech signal and conversely which frequencies can not in order to understand speech. The second purpose was to learn more about sound filters; and the third was to understand what it is like to experience hearing loss.
The instruments we used during this experiment were:
1) Low pass filter of the Coulbourn insrument
2) Tape recorder
3) Tape of two syllable words
4) Speaker (non-human)
The experimenter selected an initial cut-off frequency of 1000 Hz on the low pass filter, and then reduced the frequency by 1000 Hz increments. Below 1000 Hz the frequency was cut in half. (500, 250) We were asked to note the clarity of the signal at each cut off and the changes that occurred. Two words were presented at each level and we were asked to evaluate clarity. The frequency cut off was then added.
10,000 Hz – clear 4000 Hz - clear
9000 Hz – clear 3000 Hz – a little low
8000 Hz- clear 2000 Hz – unclear articulation
7000 Hz – clear 1000 Hz – not clear, sounds mumbled
6000 Hz – clear 500 Hz – sounds muffled
5000 Hz – clear 250 Hz – unintelligible English
The results of the experiment led to the following conclusions:
1) 500,1000 and 2000 Hz are the frequencies needed to really understand speech
2) Consonants were not heard at 1000 Hz because of hearing loss in high frequency
3) Vowels carry the power of the message and syllabic information
4) The difference was heard at the frequency level of 2000 Hz and below.
This experiment really helped me to understand the way a low pass filter works and how difficult
it must be for a person with high frequency hearing loss to understand speech.