CU-SeeMe Microphone Considerations

Bob Dixon (rdixon@stargate.acs.ohio-state.edu)
Thu, 26 Jan 1995 14:32:56 -0500


I have been doing some experimentation with various types of
microphones for use with CU-SeeMe. The goal of this experimentation
is to find a workable system for use in real-world conferences
involving one or a few people at each workstation, none of whom
are particularly familiar with computers or CU-SeeMe.

A typical scenario is 1-4 people sitting at a table, upon which
the computer running CU-SeeMe is placed. This places the people
2-5 feet away from the computer. In order to provide a sufficiently
large image for each to see, the upper right corner of each
relevant window is clicked. But the send and receive audio is more
of a problem.

The internal speaker of the computer is too small and incorrectly
positioned to work well. This can be fixed in a straightforward
way by installing external amplified speakers (several models are
available from Apple, which work fine on both Macs and PCs).

The internal microphone or the one in the video camera is too far
away from the people to work well, and causes feedback problems.
The feedback is manifested as an echo, which in extreme cases grows
continuously in volume until the entire audio system oscillates.

I am using a Mac 7100/66 (Power PC) with a Videospigot card
installed. This information may not apply to other computers
or conferencing systems in detail, but does in general concepts.

A difficulty with using non-Apple microphones with this computer
is the non-standard audio level required. In the audio industry,
audio levels are typically referred to as mike, line and speaker
levels, in order of increasing audio level. I have found that
even line-level devices are not capable of driving the "microphone"
jack sufficiently. Apple mikes which plug in there have built-in
amplifiers, powered thru an extra-long plug pin. The mikes built
into the Videolabs camera I use are also amplified, but even with their
gain controls at maximum the audio level is marginal. So I have used
the speaker output of an audio amplifier, padded down appropriately.
A 10-watt 10 ohm load resistor was placed across the 8 ohm speaker
output to absorb the amplifier output power and provide it an
essentially matched load (to avoid amplifier distortion or damage),
and the output to the computer was taken from the 4 ohm (lowest
voltage) output connection. This provides good quality and controllable
audio, and all other audio sources are connected either to the mike
or line level inputs of the amplifier and work normally.

Four different microphones have been evaluated, with the following
results.

1. A high-quality cardioid hand-held mike. This provides the best
feedback rejection. It has essentially no pickup from the
rear, so that people sitting in front of the speakers and
also talking into the front of the mike do not cause feedback.
The mike has a long cord, and so is easily passed from one
person to another. It can also be mounted on a desk stand
if preferred. It has a switch on it that turns its audio
output on and off. Automatic operation with CU-SeeMe requires
that the push-to-talk icon be deselected, the vox threshold
be set to about 1/5 scale, and the mike switch be left on
all the time.

2. A wireless hand-held mike. This is the kind often used by
stage entertainers. It uses a radio link at 170 mHz to a small
receiver box with whip antenna. This mike is very easily
passed from one person to another, and there is no cord
to get tangled. In other ways it is the same as Mike 1,
but it does not have as much sound rejection off the back
and is hence slightly prone to feedback. I find I use this one
most often for my personal use.

3. A wireless clip-on mike. This is also used by public speakers,
by clipping it onto a necktie or clothing. It is the poorest
of all mikes in terms of feedback, since it is non-directional
and has to have more amplification since it is further from
your mouth. This one uses a 49 mHz radio link. It does have one
advantage over mike 2, in that if someone turns off the mike
switch, the receiver squelches itself silently and nothing
happens. But the mike 2 receiver is poorly designed so that
if it loses the mike signal, a huge roar is sent out. This
just a peculiarity of this specific brand.

4. A hand-held wired true push-to-talk mike. The ultimate way to
positively eliminate feedback is to avoid having more than one
computer able to talk at the same time, and activating the
push-to-talk icon. The only default way to do this is position
the mouse icon on the speaker icon, and then someone has to push on
the mouse button each time someone locally talks. This is
inconvenient and not easily workable with non-familiar people.
Even if they do press the mouse button, they will inadvertantly
move the mouse icon off the speaker icon, rendering everything
non-functional. This problem can be solved by a simple
modification to the mouse. A small hole was drilled in the back,
and a small cable soldered in parallel with the microswitch
inside the mouse. The new cable was taped onto the existing
mouse cable, back to its plug. There a 1/8" female jack was
installed at the end of the new cable. This addition has no
effect on normal useage of the mouse. Using a tape-recorder
or communications-type mike that has a separate push-to-talk
connection brought out to the end of its cable, one can plug it
into the added mouse cable and achieve true push-to-talk
operation with CU-SeeMe. Just position the mouse correctly and
keep it away from everyone. Since the extra cable is fastened to
the mouse cable all the way back to its plug, yanking on the
mike cord cannot move the mouse off-position. The particular mike
I use is an electret condenser mike, which works well since
feedback is not a problem in this case. It is easy for people to
turn the mike on and off. In addition to use of the switch
on the mike, the same added mouse cable can be used to activate
the send function thru use of a separate hand switch or
footswitch. This works well when there is a moderator or
knowledgeable person among the user group.

All the microphone equipment used for this evaluation was obtained from
Radio Shack. Of course alternative equipment could be used.
I would appreciate comments and suggestions. Does anyone know
what the input impedance of the mike jack on a Mac 7100/66 is?
Knowing that, one might be better able to match its impedance
and then require less drive power.

All of this equipment is now operational in my office, and you can see it
or have a demo by connecting to the Ohio State University reflector,
128.146.116.8 .

Bob Dixon
Advanced Technology Group
Academic Technology Services
Ohio State University