NTID HyperSonic Speakers Project
From ICELabWiki
Contents |
[edit] Preliminary Notes
See Hypersonic Speaker Report.
[edit] Outcome
"Outcome is a statement of the desired outcome - how we’ll know when the project is complete."
[edit] Plans
Ed & Andrew are working on the scenarios and questionnaire, due Thursday. We will correspond with Larry Scott about the questions.
Bill & Larry will get in touch with Al Smith to find out when the smart classroms are available (because of construction through 18 August). They will also get in touch with George to ask how we can get an audio signal corresponding to the video wall.
Ed & Andrew will now...
- Put 2 scenarios into a Word document
- Our working title is "The Hypersonic Sound Project"
- Make Questionnaires look like something people would fill out
- Have nice-looking header, "collaboration with..."
- Have a liquid-scale for [strongly] agree/disagree/etc.
- Age, hearing loss, scenario experienced
- Email the document to the group, solicit questions.
[edit] Scenario 1: Classroom & Entertainment
In this room, we'll have two local sound reinforcement systems:
- A normal speaker system (In the NTID Smart Classrooms, they cover the entire room at equal sound pressure level)
- The Hypersonic speaker (which provides additional reinforcement to a small area -- in this case, a few seats)
We want to see if the Hypersonic speakers can help Deaf/HH students understand speech without adversely affecting anyone nearby. Our questionnaire will include subjective and objective questions.
- Subjective questions:
- Deaf/HH subjects: "How well could you understand? Was it better before or after we changed the sound levels?"
- Hearing subjects: "Did you notice anything unusual during the test? Was it too loud or difficult to understand?"
- Objective questions: "What did they just say?" (after a short sound clip)
Our subjects may include 10 people with average hearing, 10 people with hearing aids, and 10 people with cochlear implants.
We will set up the room for 10-20 students. The hypersonic speaker will point at one cluster of chairs, where hearing-impaired subjects will sit.
In each test, we will either play a recorded clip or deliver a speech in person[1]. We will tell subjects that at some point during the speech/clip, we're going to change something about the sound (but they won't be told what). At some pre-arranged point in the speech, we will turn the hypersonic speaker on or off.
[1]: Should this clip be video? Audio-only? Do we want to use a real speaking person, and how much should lip-reading matter? Whatever it is should be a lecture or conversation.
- Perhaps some clips should start with the speaker on, some leave it off, and some remain consistent throughout the test.
- Perhaps the first clips should be al-on and all-off, then switch part-way through a later clip so people can experience the change from regular to hypersonic reinforcement.
- For midway-through changes, ask if it was easier to hear before or after the change.
- We should be able to turn the speakers on and off without causing any "pops". Perhaps this means we need to change its volume without actually turning it on and off during the test. In any case, the "active" volume level should be consistent; we want as little room as possible for people to "fudge" the sound level with an accidental twitch.
Ed suggests we conduct the test in several 10-subject passes: half-HH and half-hearing.
[edit] Home-entertainment
Same as above, but with short films instead of lectures/conversations.
[edit] Scenario 2: Video Wall
The goal of this is to determine the performance of the speaker in an always-on informal setting. We will use the video wall in the Student Development Center (near The College Grind, downstairs from the dining commons)
Decide where to put the cone of sound (probably pointed down at one table), then ask people to self-describe hearing ability. Leave it there for 2 weeks, and collect responses?
There isn't any sound there presently, so it would be difficult to compare this sound to other methods... but it would be good for testing where the sound bounces in such a large area, how easy it is to hear and how much it bothers passers-by.
We will leave the speaker focused on one area (table?) for a few weeks, and collect survey responses with a "suggestion-style drop-box".
Can we reasonably put it there and hope students will reply? Probably not.
[edit] Questionnaire
- How well could you hear [without the speakers]?
- How well could you hear [with the speakers]?
- (these questions would really be "How well could you hear the first half? the second half?")
Supply recordings of various phrases, with varying amounts of background noise. Instead of "how well could you understand", ask "what did the person say?"
I was thinking to make the questionnaire as simple as possible. Questions more along the lines of:
- Could you hear the person talking?
- Did you understand what the talker was saying?
- I'd like to ask at least some "What did the person say?" questions, so we can compare how well they say they understood to where we measure their understanding. EdBrannin 17:18, 20 June 2007 (EDT)
- Was the speaker loud enough? too loud?
others
- What signal to noise ratios are we talking about for the background noise?
- I hadn't really thought about that -- just ideas like "people talking in silence", "people talking with background music", etc. EdBrannin 17:18, 20 June 2007 (EDT)
We should also include some demographic and hearing-ability questions:
- Age/Sex?
- Are you:
- Deaf?
- Hard of hearing?
- Hearing-impaired?
- Hearing (not impaired)?
- Some sort of ID number/code or made-up name so we can match different papers to the same person (if we ask more than one batch of questions).
- I recommend we assign colors to groups (like red and blue for deaf and hearing), and number within those groups. "Blue-5" is easier to remember than "15". EdBrannin 16:30, 21 June 2007 (EDT)
- Yes getting some demographic information would be good. We could ask them if they self-identify themselves as deaf, hard of hearing, hearing impaired. Many individuals with a hearing loss don't know their specific dB hearing loss, or they "translate" it to a %. So the self-identification is probably the best way to go. The other suggestions look good. --Larry
[edit] History
"History is for actions that have been completed, or notes of things that happened that were connected with this project. Time-stamping with ~~~~ may be a good idea."
- Copied today's meeting notes onto this wiki page. EdBrannin 14:45, 19 June 2007 (EDT)
[edit] Notes
"Notes is for any other information. In the case of things like these blog entries, the notes will contain the actual article as I’m working on it."
We have up to three speakers to work with. One is in CASCI, and the others are with Chas and (probably) George Slack.
It may be difficult to access the smart classrooms with summer construction in LBJ. We may wind up dong it in CASCI instead.
(The classroom & entertainment settings seem to be more a matter of content than room arrangement; we'll probably share a venue for both modes)
[edit] Update from Bill re getting students for the study.
I spoke with Rhonda S and can obtain a list of all students who are registered at NTID and RIT this summer, along with Coop students. She suggested I contact Sara S of VPs office to let them know I will be contacting students. Once we have the IRB OK, I will let Sara know and we are good to email students re the projects. Good suggestion to work with Rhonda!
