I did not see an HVAC anemometer that was suitable for such low flow
as in a tracker bar tube; if there is one, it is probably expensive.
There are a few relatively inexpensive ways to measure relative flow,
which can be calibrated to absolute flow if really necessary, but
in player work all you really care about is comparisons between
alternatives and/or balancing the flow between similar components.

I have a large collection of Rotameter flow meters that I bought
surplus.  I have at least two of them set up for player work but rarely
use them today.  With patience it is probably possible to find similar
deals on eBay.  I am not at home and will not be for a while or I could
check to see the flow ratings of the tubes that I found useful for
player work.  These are the type I have but I paid more like $5 each
for them, not $500:  http://www.mcmaster.com/#rotameters/=locs1q 

You can determine the flow rate that you need pretty easily by the
flowing method.  Determine the volume of a large turkey roasting bag.
These are both strong and flexible and make a good volume reference.
You can use any other bag but they are usually sticky or too thick.

There are many ways to determine the volume of the roasting bag but the
easiest is to fill it with water.  Put two tubes into the bag, one that
will be just like the one you use for the flow rate testing and another
as a temporary exhaust.  _Over the sink_ fill the bag to capacity and
then measure the volume or the weight on a scale and convert to cubic
feet or inches as you prefer.

Water weighs 62.428 pounds per cubic feet which is why you want a
sturdy turkey bag, not a grocery store produce bag.  Get a clean dry
bag and install a single tube in it and make sure it seals tightly.
It is probably best to use a large diameter rigid tube for the
connection to the bag and to spread the pleats around neatly, then
clamp with O-rings or tightly wrapped and tied cord if you are good at
knots.  You can also use a hose clamp but put something under it to
protect the bag.

Using a series of reducers, you can connect the turkey bag to any
player piano component and measure the flow by timing how long it takes
to completely exhaust the bag after filling it to the level at which
you measured the volume.  Be sure to test for leaks first.  Just
inflate the bag fully and block the tube.  Press on it and listen for
leaks.  You can submerge it in a tub of water to test for small leaks
or use diluted dish soap for bubble testing.

If you only need it occasionally you can use just the bag and a
stop watch for the cheapest possible flow meter, or you can use the
information obtained by connecting it to something typical that you
would want to measure, to find a suitable meter on eBay.

Rotameters are rated for specific gasses so try to find something with
a gas density similar to air.  The best would be a meter for dry air or
nitrogen.  Other gases such as argon and carbon dioxide are close
enough for non-critical absolute measurements.  Don't get anything
really strange unless you check to see that the density is close to
that of nitrogen.

Another way to measure flow is to measure the pressure drop across
an orifice.  The only problem with this is that you will probably need
a very sensitive pressure meter for low flow measurements.  Magnehelic
gauges can be found on eBay in the low pressure range of under 1 inch
water column.  Get the most sensitive you can find because you can
always increase the size of the orifice used for testing to reduce the
sensitivity of the flow meter but you can't necessarily go the other
way without affecting the flow itself if required pressure drop is too
high.  You can calibrate this meter with the turkey bag.  Just put the
two in series with a constant flow source exhausting the bag through
the meter.

You can also use an electronic pressure meter, which is actually my
favorite method.  Strain gauge pressure sensors are cheap now and can
be found all over the place.  They come in different qualities, with
different sensitivities, ranges, accuracy and temperature compensation.
The most expensive (which you do not need) are individually calibrated
for linearity.

I have at least a few digital flow meters that I have made using gauge
amplifiers and computer data acquisition interfaces.  The first used
the computer serial port but the newer ones use USB.  This company used
to have introductory models for less than $50, and the web site still
says something about "as low as $49", but I can't find the $49 model:
http://www.mccdaq.com/solutions/USB-Multifunction-DAQ.aspx  The one
I have came with free basic charting software.  I imagine this is still
the case.

If I did not have so many other options for measuring flow and so little
need to do it, I would make an Arduino interface for a flow meter.  If
anyone wants to do this, contact me because I have all the needed code
from other applications to stick something together pretty easily.  All
that is needed is a pressure sensor in the 5 psi or so range for about
$5.  (I have about 50 of these if anyone is having difficulty finding
one) and a "Teensy 2" microcontroller development board for $14.95.

The Teensy can act as a virtual joystick.  Connect it to the computer
and install a simple program on it and it will send joystick axis
levels to the PC.  You can connect as many as four sensors since
there are four joystick axes.  I just wrote a little PC program that
visualizes four joystick levels with bar graphs.  This would be fine
for relative measurement.  The easiest way to scale the system would
be by adjusting the measuring orifice, best done with a needle valve.

I also have a PC application that converts the joystick levels into
MIDI events and passes them to any MIDI device.  This is not really
the best way to record data but it is already written so I would tend
to use it myself instead of writing something else.  If there is enough
demand for this, I would be willing to add features to my simple
visualization program to allow scaling and or logging to a file in
tabbed text.  This would allow charting in Excel.

Best regards, Spencer Chase
Garberville, Calif.
http://www.spencerserolls.com/