We are excited to announce the launch of our new forums! You can access it forums.aavso.org. For questions, please see our blog post. The forums at aavso.org/forum have become read-only.
Announcement: New Applications
We are excited to announce the launch of our new applications! We're opening up early access to our new applications for searching, downloading, and submitting photometric observations. You can now access these applications through these links:
We ask for your feedback in order to help us improve these applications. Please send feedback for the applications above to feedback@aavso.org. Note: please avoid duplicating submissions across the two submit applications.
My experience has been, especially as I grow older and my eyes "yellow", that the comps should be of the same colour, (ie B-V) as themselves, and if possible, close to that of the target star.
I agree with MDP--its true for even PT. I always try to do this, unless there are other considerations. Usually other systematic errors will be larger that this one.
It's depending what we are doing. For observation, in particular in small FOV, high enough in the sky, we prefer to have the same color for the comp and the target. But ok, this is not always possible ! But for the determination of transformation coefficient of the DSLR, at opposite, we need a large range of color of the ensemble being used (but late M stars shall be avoided).
But the CS "intermediate" spread sheet was dedicated to observation at relatively high air-mass and very large FOV (Eps AUR was very low at a critical phase of the campaign, and the comp far, an unusual case). It includes the determination of the extinction across the FOV - AND - the color correction coefficient for each image stack (DSLR+atmosphere). By the way a significantly large range of color of the ensemble shall be used in this case. In fact that method is not perfect, the choice of color and position of the stars of the ensemble is critical in this process ! Then if stars of similar color are used the equation system would deliver very wrong coefficients ! In fact I would not recommend to use that technique.
This is one of the reason for which I did develop my own technique, the said VSF. Both have been described in JAAVSO papers (2012).
The "beginner spead" sheet doesn't include that extinction calculation, only the color coefficient calculation, it should be used only for small FOV and/or high in the sky. It's clear it needs a large enough color range of the ensemble.
I am interested to read your article. I have just downloaded it!
I have a related question.
In the Citizen Sky spreadsheets (beginner and intermediate) there are best fit plots (for example "Residual Mag. vs. Color Correction"). In those plot there is the R-squared value. I imagine that this value have to be near 1 for have best results. Is it right?
I have noted that, by choosing a different set of comparison stars, I have different values for R-squared (in the range 0.6 - 0.9)
Can this value give information about the correct choose of the comparison stars?
I have not been involved in that spread sheets, it was Brian and HB if I remember well, I rarely used it at that time, more than 5 years ago. I prefer my own software ! The explanation is after the "step 7" in:
The paper says the squared residual should be below 0.7 in case of good data. I don't think it's a good indicator of the specific issue of the "intermediate" process. The issue is the fact there are distributions of color and positions in the ensemble that generate a cross-contamination of the color transform coefficient and the extinction gradient. This could happen even if the overall fitting is perfect.
The "intermediate" technique should work properly if the number of stars of the ensemble is huge and at well random position. In case of a few stars, let say ~6... the variations in function of the position due to color contaminate such of the extinction gradient. That even if the equation system is perfectly satisfied... To avoid that the position distribution of colors should be neutral (not obvious ! ).
There are other general issues, one is the catalog mag accuracy, this is not obvious to find stars known at the accuracy needed by such process. The best source for this range of mag is Tycho 2. But the VT and BT magnitudes are different than Johnson one, the transformation is often not easy. One good point is that the VT spectral response curve is very similar to the current response of the DSLR G channel: no transformation needed ! But ok we are not in the Johnson's system...
In the case of the "beginner" spread sheet that problem doesn't exist as the only variable is the color. Then the residual is a god indicator of the data quality.
My experience has been, especially as I grow older and my eyes "yellow", that the comps should be of the same colour, (ie B-V) as themselves, and if possible, close to that of the target star.
Hi Patrick,
I was referring to the photometry and not to the visual estimates.
Regards,
Luigi
Hello Luigi
I agree with MDP--its true for even PT. I always try to do this, unless there are other considerations. Usually other systematic errors will be larger that this one.
Gary
WGR
Hi Gary, Luigi,
It's depending what we are doing. For observation, in particular in small FOV, high enough in the sky, we prefer to have the same color for the comp and the target. But ok, this is not always possible ! But for the determination of transformation coefficient of the DSLR, at opposite, we need a large range of color of the ensemble being used (but late M stars shall be avoided).
But the CS "intermediate" spread sheet was dedicated to observation at relatively high air-mass and very large FOV (Eps AUR was very low at a critical phase of the campaign, and the comp far, an unusual case). It includes the determination of the extinction across the FOV - AND - the color correction coefficient for each image stack (DSLR+atmosphere). By the way a significantly large range of color of the ensemble shall be used in this case. In fact that method is not perfect, the choice of color and position of the stars of the ensemble is critical in this process ! Then if stars of similar color are used the equation system would deliver very wrong coefficients ! In fact I would not recommend to use that technique.
This is one of the reason for which I did develop my own technique, the said VSF. Both have been described in JAAVSO papers (2012).
The "beginner spead" sheet doesn't include that extinction calculation, only the color coefficient calculation, it should be used only for small FOV and/or high in the sky. It's clear it needs a large enough color range of the ensemble.
Clear Skies !
Hi Roger,
thanks for answer.
I am interested to read your article. I have just downloaded it!
I have a related question.
In the Citizen Sky spreadsheets (beginner and intermediate) there are best fit plots (for example "Residual Mag. vs. Color Correction"). In those plot there is the R-squared value. I imagine that this value have to be near 1 for have best results. Is it right?
I have noted that, by choosing a different set of comparison stars, I have different values for R-squared (in the range 0.6 - 0.9)
Can this value give information about the correct choose of the comparison stars?
Regards,
Luigi
Hi Luigi,
I have not been involved in that spread sheets, it was Brian and HB if I remember well, I rarely used it at that time, more than 5 years ago. I prefer my own software ! The explanation is after the "step 7" in:
http://www.citizensky.org/content/calibration-intermediate
The paper says the squared residual should be below 0.7 in case of good data. I don't think it's a good indicator of the specific issue of the "intermediate" process. The issue is the fact there are distributions of color and positions in the ensemble that generate a cross-contamination of the color transform coefficient and the extinction gradient. This could happen even if the overall fitting is perfect.
The "intermediate" technique should work properly if the number of stars of the ensemble is huge and at well random position. In case of a few stars, let say ~6... the variations in function of the position due to color contaminate such of the extinction gradient. That even if the equation system is perfectly satisfied... To avoid that the position distribution of colors should be neutral (not obvious ! ).
There are other general issues, one is the catalog mag accuracy, this is not obvious to find stars known at the accuracy needed by such process. The best source for this range of mag is Tycho 2. But the VT and BT magnitudes are different than Johnson one, the transformation is often not easy. One good point is that the VT spectral response curve is very similar to the current response of the DSLR G channel: no transformation needed ! But ok we are not in the Johnson's system...
In the case of the "beginner" spread sheet that problem doesn't exist as the only variable is the color. Then the residual is a god indicator of the data quality.
Clear Skies !
Roger
Thanks Roger,
your answers are always very enlightening for me!
Regards,
Luigi