Polar.txt (instructions for Tim Sandrik's Polar3)

Instructions for Polar 3.0
8/4/1998
written by: Timothy E. Sandrik

Theory of Operation

Polar 3.0 is written to process ASCII measurement data files from Liberty
Instrument's Liberty AudioSuite. The program will open, read, and process
multiple files and then generate a report in Microsoft Excel that contains the
processed and organized data, and several plots useful in describing radiation
pattern information. The angular resolution handled is 10 degrees, and the
frequency resolution is selectable by the user in points per octave. The
program automatically adjusts for the amount of data the user has taken,
mirroring data where possible, and filling in data where mirroring is not
possible. This means that it is not necessary to make 36 measurements when the
speaker is symmetrical, or when only some portion of the radiation pattern is
being investigated. Users are encouraged to become familiar with the use of
Microsoft Excel to plot data in other ways, and format the plots to best display
the data.

Basic Operation

Open the file Polar 30.xls and click on the button that reads "Click Here to Run
Polar 3.0". This begins the program by presenting file open box for you to use
to browse for a set of polar data files. These files should all be together in
the same directory. Once the files are found click "Open", or press "Enter".
Be sure to make note of the name of the speaker to be processed. The name of
the speaker will be up to the first five characters in the set of files. For
example, the speaker name for the files myspk0.frd, myspk10.frd, myspk20.frd
would be myspk. You will then be prompted for the speaker name. Type this in,
and click "OK" or press "Enter". Next, you will be prompted for the
frequency resolution, in points per octave, the low frequency limit, and the
high frequency limit. The defaults are 3 points per octave(third octave), 20
Hz, and 20,000 Hz, respectively. Type in these, or accept the defaults, and
click "OK" or press "Enter". Now the program will open the files, process and
organize the data, and generate a report.

If errors occur, choose to end the program operation. Check your input files to
be sure they are of the proper format, and are free of errors. The type of
error cited may give you a clue what to check. If the error persists, and seems
to be a problem with the program, then collect what information you can about
the error, and contact the author.

Input file format

The input file should have the extension .FRD, the format generated by LAUD.
The data in the text file should be organized in columns with frequency listed
in the first, corresponding magnitude data in the second, and corresponding
phase data in the third. For example:

Hz     dB         Deg
43.0664     -22.7         16.8
86.1328     -16.4         -42.
129.199     -13.0         -97.
172.265     -12.7         -137
215.332     -16.2         -157
258.398     -17.2         -154

The last 1-3 out of 8 (allowed in DOS) characters of the filename should
correspond to the angle of the measurement. This leaves 5 out of 8 characters
for a "name" of the speaker. The angle should be in increments of 10, be
positive and between 0 and 360. Following is an example of a set of filenames:

myspk0.frd        myspk10.frd         myspk20.frd   
myspk30.frd        myspk330.frd         myspk340.frd
myspk350.frd

Output Report Format and Explanation of Data

The output report contains several forms of raw data. These are:

-On Axis Response
-Average Response (Often called power response)
-On Axis Response/Average Response (Often called DI, or Directivity Index)
--3 dB, -6 dB, -9 dB and -12 dB Coverage Isobars
-Polar Data
-Normalized Polar Data

The data is plotted in 4 plots. Each will be discussed, and some guidance for
interpretation given.

Coverage Plot: This is displays the included coverage angle versus frequency
for the -3 dB, -6 dB, -9 dB, and -12 dB isobars. This plot is intended to
display the width, in degrees, of the main lobe of sound radiated by the
speaker. The -6 dB isobar, for example, would illustrate the angle where the dB
magnitude, at each frequency, first dips below -6 dB. A smooth coverage plot is
usually desired, and would indicate that the frequency response is probably
smooth at many positions around the speaker. Coverage plots fail to describe
off axis lobes, and can be misleading when there is an on axis null. Much like
a smooth on axis response plot, a smooth coverage plot is only part of the
picture.

Responses: This displays three types of data: on axis response, average
response, and on axis response divided by average response. The average
response is often called the axial (horizontal or vertical) power response, but
this is a bit of a misnomer, especially for two dimensional data. It can be
interpreted like power response data, presuming the user understands that it
ignores data outside the plane of measurement. Again a smooth average response
is desired, indicating the frequency response is smooth and consistent at each
point measured.

The on axis response/average response, when expressed in dB is often called
axial directivity index, or axial DI. On axis response/average response is a
measure of the amount of the total energy radiated on axis to the amount
radiated in other directions. This is also a misnomer, technically speaking.
Again, a smooth plot is usually desired, indicating smooth response at each
point measured. The on axis response is self explanatory.

Surface Plot: This is a favorite plot of the author. In one plot it displays
coverage, with detail about nulls and lobes in the off axis response, and is
sometimes easier to read than a multi-band polar plot. Only data in the angles
0-90 and 270-350 is displayed. The default view is from above. This plot is
unusual, and again a series of smooth contours will usually be desired.

Normalized Polar Plot: This is a polar plot displaying polar response
independent of on axis response. It indicates how the response changes relative
to the on axis response at each frequency.

A great amount of data is available in the report, some of which is plotted.
Most of the plots are those typically used in the loudspeaker industry, and are
presented in typically used formats. The above descriptions seem a bit
repetitive, saying that each plot should usually be smooth, but this is true.
As far as the ideal shape of each, that depends on your views on dispersion,
directivity, etc. In the audio industry you'll find many views and opinions.
Polar 3.0 is designed to help you get a look at a few of the other "responses"
that define the character of the speaker.