[AAVSO-DIS] RE: Photometry Software

[Richard Berry]

Hi Michael--

> I can certainly share your concern over using 1 comp star to get an accurate
> measurement. When I first started in photometry about 3 years ago, I
> remember going almost stark raving mad over my "erroneous" measurements of
> SS Del. I was consistently getting the same amount of errors from the comp
> star with a 14.5 chart mag. When this chart was revised, it was determined
> that the 14.5 comp was actually 13.64....

Keith touches on an important point, and perhaps *the* important point: if the
magnitudes of the comparison stars on the chart are not reliable, then accurate
photometry is not possible. However, you can do very precise photometry. This
seeming paradox occurs because two key terms in measurement are easily confused.
These are accuracy and precision.

Accuracy is the departure of a measurement (or measurments) from the actual
value; precision is the internal consistency of a set of measurements.

A measurement can be precise without being accurate. Measurements that are
inaccurate but precise are good for finding the time of minimum of an eclipsing
binary. You don't care whether you know the actual magnitude, but you do want
precise measurments so you can get a good time of minimum. Thus you can measure
EBs and cataclysmic variables with a single comp star and do a good job.

Measurements can also be accurate without being precise. If you measure a very
faint star, the statistical variation in the number of photons can lead to a
large scatter in the magnitude measured on an individual image. An individual
measurement is therefore not precise. However, if you average a large number of
measurements together, the measurement may prove to be accurate. Although
measurements varied, the values clustered about the actual value. This is the
rationale behind the AAVSO's visual observing programs.

In the case where the magnitudes given for the comp stars are accurate but
imprecise (i.e., the chart magnitudes have random errors about the true
magnitudes), then using multiple comp stars may tend to average out the errors
in the individual comp stars. If they are precise but inaccurate, these same
thing occurs: multiple comp stars may tend to average the errors, unless, of
course, all the errors are in the same directions.

It is also important to remember that color errors that did not matter in the
days of visual observing are significant with CCD photometry. If passbands for
either the chart stars or for your CCD stars are not standard, your CCD results
won't agree with the chart stars even if the chart stars are both accurate and
precise.

--Richard