Step by Step Instructions for generating Transformation Coefficients using VPhot / Transform Generator (TG) are included in this pinned post.
Vphot/TG Transformation Coefficient Generation Process – Step by Step
Task 1 – Generate Ensemble Sequence in Vphot for Use in Vphot/TG.
It is assumed that you have collected images of an AAVSO Standard Field (e.g., M67) in multiple filters (e.g., BVRI) and uploaded them to VPhot.
In Vphot, select/open one of your V images of an AAVSO Standard Field (e.g. M67).
Under Catalog, click ‘Load AAVSO Standard Stars’. Many stars (e.g., 165 comps) will now have blue apertures/names identified on the image.
On left side window in the standard star/comp list, note the list of star names all have AAVSO Unique Identifiers (AUID) (e.g., 000-BLG-879).
Select one of these standard comps (e.g., 000-BLH-002) on the image (zoom in to see comps more easily), left click the comp, box opens, click radio button for ‘Check star’ rather than ‘Comp star’ to make this the check star. (Note that the check star is not used for generation of coefficients but is needed to meet normal analysis report protocol). Click ‘Update’ and Close.
Select one of these standard comps (e.g., 000-BLG-956) on the image (zoom in to see comps more easily), left click the comp, box opens, click radio button for ‘Fixed Target star’ rather than ‘Comp star’ to make this the target star. (Note that the target is not used for generation of coefficients but is needed to meet normal analysis report protocol). Click ‘Update’ and Close.
Check star will appear with orange aperture on image and in left comp list. Target star will appear with green aperture and in left comp list.
Click blue ‘Save’ link at bottom of star list and give your saved sequence a name like ‘M67 BVRI All Std Comps 240801’.
Do not worry about what the Image Window and Photometry Report window looks like at this point. TG will take care of removal of poor comps due to overlap, saturation, and other problems by visual observation/deactivation of poor (> 2 sigma) data points on the subsequent TG plots!
Note: It is possible to initially/carefully delete/de-select/remove std stars from the image/sequence by zooming into the image and removing any stars from the sequence that appear overlapped, saturated, or too faint. However, in ‘most’ cases, this tedious process is not necessary since deselection within TG plots is more efficient and statistically valid. Just use ALL the standard field stars in the saved sequence.
Task 2 - Generate Instrumental Magnitude Report in Vphot for Use in TG.
It is assumed that you now have all your AAVSO Standard Field images (e.g., M67 BVRI) listed in your VPhot ‘Available Images’ list. Check boxes to include one set of all these standard M67 images (e.g., 4 filter images) in your Available Images List (e.g., BVRI x 1 night).
Select/click Time Series.
Select your saved Std Field Sequence from Task 1 in the sequence pull-down box. Leave other settings as defaults.
Click ‘Start the analysis’.
Click ‘Continue to the result page’.
Click ‘General Export’ on Time Series page.
At bottom of ‘Export Time Series Results’ page, click ‘Save to AIP fmt for TG’. Save file on your computer with name like ‘M67 BVRI TG Report 240801’ and save in some folder as desired).
Repeat Task 2 for other (e.g., 2) day’s BVRI images. This will permit the calculation of average Transformation Coefficients.
Task 3 – Generate Transformation Coefficients for use in Vphot/Transform Applier.
Note: Look under the following link to find how to install/set up Transform Generator (TG) on your computer: Transform Generator (TG) -- (Version 6.9a) | aavso (I recommend using the Single File Installation version).
Open Transform Generator (TG) on your computer. May take a while, be patient.
Click in ‘Add Scope’ pull-down box, click ‘Add Scope’ again, type a ‘System Name?’ that you understand. Click ‘Enter’ to save.
Select the M67 (or other as appropriate) radio button in ‘Select Standards Field’.
Select TG/AIP4WIN radio button. Should be checked as default already.
Click “Select File(s) button. Find/Select the Saved Vphot file from earlier step 17. File name will appear in ‘Current file’ box.
Click ‘Calculate Transform Set’ button.
List of Coefficients will appear in bottom of window.
Click first coefficient line (i.e., Tbv). Plot opens. Expand to full screen.
Click/select each green data point on plot that is ‘outside’ the blue 2 sigma lines. They will turn red. Continue the process as blue 2 sigma lines shrink inwards. This process WILL get to the point of diminishing returns. Do NOT overdo it! You may leave points that are on the 2 sigma lines but do not remove data points inside the 2-sigma lines. Data points near the extreme color ends of each plot are the most important in defining the Fit line.
At the end of selection, just close the window as normal in windows. BTW, there is a tool (‘save the figure’) at bottom of screen to save the plot.
Continue this process for ALL other listed coefficients.
Then, click ‘Save Transform Set’. The Coefficient File is automatically saved as an *.ini file in a hidden location. Click “OK’.
Repeat the TG process for any other day’s magnitude reports (usually ~3).
After you have saved all the days results, click on ‘Review/Average Transform Sets’ to open all available coefficients result sets.
Note: Report any problems with these instructions by replying in this forum topic.
Select the results from all the image days (~3) and then click ’Retrieve transform sets’. All saved days should be listed. Check each day’s results that look reasonably similar (coeff and error).
Click ‘Compute Average of Checked Transform Sets’. The ‘Avg Transform’ set will appear in Yellow highlight.
Compare the day’s results to confirm reasonable consistency!
Click ‘Save File of Average Transforms, Enter/Select File Name’. A prompt appears to input an appropriate file name to identify the target (e.g. M67 BVRI System Date of analysis). Don’t worry about the extension type. Just click Save. The file will be automatically saved as an *.ini file. This file can be opened in a text editor to view.
This ‘average’ *.ini file can be opened/used in Vphot/TransformApplier (TA) to transform applicable AAVSO Report files!
HTH, Ken
PS: Post a reply if you have any problems with this process.
I congratulate Ken Menzies for posting this ‘Step by Step Instructions for Generating Transformation Coefficients Using VPhot / Transform Generator (TG)’.
One question which occurred to me is, what exposure should I use when imaging the standard fields. Do I need to do it for all the exposures I use?
Thanks
Great instructions Ken. Just renewed mine. One thing, those little green data point circles. They can hide in amongst the graphics.
I like these kind of workflows. I am still not sure about what exposures to use so I used my most common ones, and the coefficients do vary a bit. I’ll post the results and ask for comments.
As long as I look at this VPhot/TG tandem, one question puzzles me.
We are calculating transformation coefficients for a standard field at the given night and for given the airmass. How can we apply them to the photometry taken at a different night and with different airmass? From what I learned from photometry books (Henden & Kaitchuk, Warner) we should find extinction values and daily zero points.
The approach explained in AAVSO CCD photometry guide should work well if we have a big enough ensemble of standard stars in the same filed with the target, not for the scenario discussed above.
Transformation Coefficients are related to the ‘slope’ of your plots. It doesn’t matter what the ‘intercept’ is. If you move your plot up or down, the slope is still the same.
You are conducting ‘differential aperture photometry’ in a small field of view for each image where extinction differences across this field may generally be ignored and the zero point is the same. (This is not true for all sky photometry or in a very large field of view where such corrections should be made.)
Assume we repeat this measurement at different altitude. Shouldn’t second order extinction affect the slope?
Assume I calculated my transforms yesterday for Melotte 111 and used it to transform my today’s observations of SZ Lyn. Isn’t it a case of all-sky photometry?
The detailed procedure is much appreciated! Very clear. Thank you.
One follow-up question: When “outlier” measurements are deactivated in any of the coefficient plots it does not seem to be deactivated in for the other coefficients. I would have thought that the correlations between the coefficients would require that the same measurements go into the calculation of all of the coefficients. Do I have that wrong?
I’ve just caught up with this thread which started a few months ago.
On March 7 Dmitry wrote:
“We are calculating transformation coefficients for a standard field at the given night and for the given airmass. How can we apply them to the photometry taken at a different night and with different airmass?”
On March 8 Ken replied:
“Transformation Coefficients are related to the ‘slope’ of your plots. It doesn’t matter what the ‘intercept’ is. If you move your plot up or down, the slope is still the same.
You are conducting ‘differential aperture photometry’ in a small field of view for each image where extinction differences across this field may generally be ignored and the zero point is the same.”
Ken’s reply emphasizes that TCs relate to the slopes of the plots, not the intercepts. Of course that is correct, but I think Dmitry’s issue remains valid. Instrumental magnitudes of a given star will differ at different airmasses. Extinction will affect blue wavelengths at high airmass more than it affects red wavelengths. Thus b-v (for example) is sensitive to airmass.
TCs determined at significantly different airmasses will therefore differ, and TCs determined at low airmass may not provide accurate transformations for measurements of stars at a significantly higher airmass if extinction is ignored. The practical issue is what the word ‘significantly’ implies. For differential photometry at ‘low’ airmasses, the effect is small enough to be ignored, but at ‘high’ airmasses this may not be the case.
The basic color terms in the transformations are determined largely by the telescope hardware, and should be stable if there is no change in the set-up. The ‘TG’ program does only this part of the process. The atmospheric extinction is a separate term in the transformation, and will vary night-to-night (and even during a night), and is not dealt with by ‘TG’. So it is not a complete photometric reduction program. The assumption is that comp stars in any field are established and one uses those to set the magnitude/color zero-points, and assumes the extinction effect is absorbed in that zero-point. At least under cloud-free conditions, one can assume a typical extinction value of something like 0.2 magnitudes/airmass in V, with no color term, and as long as you don’t to extreme airmass (say sec z > 2.5) you will be OK at the few-percent level.
Starting from exactly the same question as Dmitry from the same appendix D, I was also wondering why determination of a TC doesn’t simply include an airmass term.
Not sure about the second order magnitude effect (negligible or not?). Isn’t it logical to expect shifting to red at lower altitudes?
Therefore, is it a bad idea to try TC calculation including airmass effect, to try to combine multiple standard fields at different sky declinations, because many standard fields only have very few reference stars, which doesn’t allow accurate TC calculation statistics.
There is a lot of airmass between Landolt +50° and 0° standard fields, and I’m not sure how to apply a TC to intermediate declinations/airmasses.
Could you also comment if ‘True Color Index’ is assumed to be at airmass = 0 ?
Hi Ken:
Thanks for this excellent description of the process. One point I found a little confusing was that we need to process a full set of filters, one each, then repeat that for however many images we have in each filter.
I am posting because TG6.9 is frequently unwilling to save the plots. It seems to happen regardless of whether or not I have removed any stars from the plot. The save button appears to depress as it should but immediately returns. Repeated clicking on the button has no effect. Sometimes I can open a different plot and it works fine, sometimes after 10-20s the button works. I have tried restarting TG without luck. I am saving to a remote drive, if that makes any difference. Any ideas?
Thanks, Richard
I’m not sure I understand your confusion about filter sets? There are a few ways to analyze different sets of filters depending on how many sets you collected. If I collect a BVRI set and then another BVRI set, I might choose to run a time series of the first set and calculate the coefficients for that BVRI set. Then I might run a time series of the second BVRI set and calculate those coefficients. This would allow me to average the coefficients for that night. Alternatively, I might stack the BVRI pairs and get one set of coeffs. Another option is to run a time series of the BVRIBVRI together and calculate one set of coefficients. So, there a few options. What exactly do you do most often?
I don’t save the plots too often and when I have, I don’t recall having a problem with clicking the save image button. I don’t think anyone has reported that problem? Have you had any problems with your mouse? BTW, I think TG6.9a is the most recent version?
Perhaps someone else could make a comment if they have had the same problem? I hope you are still able to generate your coefficients even when you don’t get a plot image?
HTH, Ken
PS: I guess saving to a remote drive could be the cause (delay?) but you could simply try to save to a local drive to see what happens?
Ken:
I did figure out the filter set confusion, I just didn’t find it clear from the way it was written so it took me a couple of tries. And I didn’t fully appreciate the flexibility (process a whole set at once, stack by filter and process results, process by BVRI filter group) so thanks for that clarification. I will try them all out.
I wanted to save the plots so I can refer back to them for quality of fit etc, also as a record of the individual coefficients with error estimates. I didn’t see any way to save the coefficient sets in a human/spreadsheet readable way other than ‘averaging’ a single set and saving. Which I will now do.
I tried running TG 6.9a on the local machine with local VPhot files and saving the plots locally. I still get random failures to save the plots. The problem is certainly not the mouse/trackpad since the button does respond to clicks (i.e. it pushes in and out) but no save dialog shows up.
Honestly I am mainly using VPhot and TG this time to get values against which I can compare those from my own programs. I have slow internet and uploading any significant number of images to VPhot is totally impractical.
Richard
