Skip to main content

Visual observing of Eclipsing Binaries - Still useful?

9 posts / 0 new
Last post
kwierzchos's picture
Visual observing of Eclipsing Binaries - Still useful?

Dear all,

I noticed that most of the legacy eclipsing variables have very few CCD recent estimates in the AID. Some of them appear to have been observed once in the past three or four years.

In the eclipsing binary section page it is written: There have been some recent papers that dismissed all ToMs derived from visual observations, even from dates before any photoelectric data was available. This unfortunately means that new eclipse times deduced from visual observations are of much less interest to researchers than they used to be.

It is obvious that CCD observations can give times of minima that are orders of magnitude more precise than thoose derived from visual observations. But I'd like to know if visual estimates of eclipsing binaries that will give us ToM are still useful. My guess is that if a legacy star did not had any estimates in the past years, even a visual timing is better than nothing right?

Thank you


SET's picture
Visual observing of Eclipsing Binaries - Still useful?

Hi Kacper,


I still observe eclipsing binary stars visually. There seems to be a difference of opinion on this issue. For the past 5 years I have been sending my observations to the Japanese group, and they publish them in their yearly publication.


Chris Stephan   SET

No, visual timings are not

No, visual timings are not really useful, except in a historical sense. It's also useful to observe a single star, obtain the full visual light curve, and use it to determine the type of variability, the period, and so forth.

But scattered visual timings are no longer valued by researchers. See my article on "The Visual Era of the Eclipsing Binary Program in JAAVSO vol. 40, 2012, page 180, where I summarized the visual program's history and the reasons for its extinction.

nmi's picture
Great question!

In my opinion, any and all observations are needed as we never know who can observe and in what way. However, this question causes me to ask others: which stars are best observed in Visual, V, Multi-color, transformed? Can the staff of the AAVSO offer advice without insinuating that one type is better than another? I've often heard that it depends on the purpose of the study, OK, what's being studied? I've often asked myself why a professional astronomer used my data when I see it listed in Sara's e-mail report? Aside from formal campaigns, the AAVSO data archive is part of a feed forward process. Is it possible to receive feedback from the community on the types of data needed by the professional? This gets to the latest professional collaboration efforts by the staff. What's happening? And one last question, when will we see the staff astronomer hired?

It's easy to ask questions, much harder sometimes to answer them. I want to thank the staff for all the work they do.
Clear skies,

kwierzchos's picture
Dear all, 

Dear all, 

Thank you for your answers. After researching the topic online I found that indeed the  opinions are varied. David, I read your article and found it very interesting. You mention that "But scattered visual timings are no longer valued by researchers".

A quick search in the literature returns several recent papers where visual ToM are used, for example:

All of them from 2016, and there are more. 

Best wishes




Utility of visual observations

The answer has to be "Yes".  Although I am not an EB observer, it is clear to me that more observations lead to better estimates of the periods of EBs.  One can see that by looking at the expression for the Carmér-Rao bound for the estimation of frequency, which is inversely related to period.  The CR bound for frequency estimation based on the Fourier transform depends inversely upon the SNR and the number of observations.  Now, in the context of variable star observing, visual observations can be modeled as having lower SNR than, for example, CCD observations, but what the expression for CR bound says is that one can trade SNR against number of observations to achieve the same effect.




lmk's picture
Need many more

"visual observations can be modeled as having lower SNR than, for example, CCD observations, but what the expression for CR bound says is that one can trade SNR against number of observations to achieve the same effect."

This is a very good point! AAVSO studies have shown that long term visual LPV observations, when combined from many observers and observations, results in very good accuracy and precision, due to the "law of large numbers". In fact, there is really no theoretical limit to the accuracy obtainable, only a practical one of getting enough observers.

Unfortunately, the numbers of visual observations are not increasing in step with the total and CCD observations, maybe even declining? So, the most essential feature needed for visual success - more participation - is taking the hardest hit.

I get the impression that most new observers are going directly to imaging, skipping the visual phase completely, due to this preconceived notion that visual is "inaccurate" and/or "useless" in today's technology. While CCD is more sensitive and precise, their accuracy vs. visual is quite debatable, especially given the rather superficial expertise with CCD (and getting caught by all its subtle "gotcha's") that the newer observers possess.


HRHA's picture
slightly different take on visual observing

From going through recorded data of solar observations back to 1612 it is useful to re-sample the visual observations for different time periods. This is because of the sparse number of observers in the older data.  

As an example here are visual observations for CE Tau that come from the AID. The raw observations from the new light curve generator show some sparse data during the last 10 - 15 years.  And then weekly re-sampled observations using some R code looks somewhat different.  

There seems to be differences for the most recent ~ 10 years or so of visual observations looking at the LCG vs the weekly re-sampled data. Why this would be I don't know. Perhaps there is some value to re-sampling the visual observations for these long term light curves? 



Visual versus CCD

While the timing of visual observations may be less precise than that of CCD observations, it is still useful (in a statistical sense at least) to have visual observations.  I have been using the AAVSO Zapper app to identify discrepant observations on various stars, and have recently been intrigued to see that in some cases there are more discrepant new CCD observations than visual ones!  So digital does not guarantee perfection, nor even superiority over carefully done visual observations.  I had a brief discussion with a conscientious PEPper about that, and he immediately agreed that CCD observations are not always accurate.

Log in to post comments
AAVSO 49 Bay State Rd. Cambridge, MA 02138 617-354-0484