Photometry Without Photometric Filters

American Association of Variable Star Observers (AAVSO)
Sat, 12/31/2022 - 23:46


I am just getting started with photometry, and trying to test out my current equipment and images to get a feel for how to correctly use VPhot.  I'm looking for more detailed guidance about how to use astrophotography LRGB filters for photometry in VPhot.  To be clear, I am NOT referring to DSLR filters/bayer mask.  I am referring to a setup with a monochrome camera, filter wheel, and LRGB filters (in my case from ZWO).

I've seen this question asked a couple times over the years, but it seems the question was never really addressed in full, just with pieces of advice:

  • Tri-color (DSLR) is OK in some cases.  Avoid emission nebulae and novae 
  • Stick to cepheid and rr lyrae types
  • Use clear filter and V Magnitudes 

I've read through the photometry with CCD/CMOS cameras guide, searched the AAVSO forums, and watched a few presentations on YouTube.  I watched the 7 or 8 "Photometrica" tutorials, too.  I've used VPhot to upload some existing data just to get a feel for the process.  It's been raining for the past two weeks, so I can't collect new data right now.

I understand how to calibrate (dark, bias, flat) and I understand the linearity and saturation limits of my camera.  I understand the basic theory of how to select an appropriate comparison star and check star.  I understand how to set the aperture around the stars in VPhot.  What I don't understand is at least the following:

1. Which of my LRGB filters would be best to use for photometry, given that magnitudes in VPhot and the AAVSO database are given w.r.t. photometric filters?  I can compare the wavelengths of each filter to those of the photometric filters, and see that some are a better fit than others, but none of them really match up very well.

2. What is the right way to tell VPhot that I am using a non-photometric filter?  I see that I can choose the filter during upload, and I could override the FITS header to force the filter to be V, for example on a Green filtered image.  I can get very good error readings on the target and comp star (0.01), but the check star magnitude differs by 0.5 magnitudes from the expected: presumably because I'm using a Green filter and comparing against a Johnson-V filter magnitude.

3. Do transformations help here?  I couldn't quite understand if this made sense to do with non-photometric filters.  Will that help translate my Green filter magnitudes to reasonable V filter magnitudes?  I understand that you need at least two different filters to create the transformation coefficients, but given that I have zero photometric filters, I wasn't sure if this is still useful to do with my LRGB filters.  If so, which coefficients should I compute, and how should I enter them in VPhot and/or the AAVSO report format?

4. How would I format the report to make it clear that the filter I used was not photometric, but something else?

I appreciate any more details you all can provide.


Variable Stars South (VSS)
Carl,Out of curiosity, I…


Out of curiosity, I did the same thing with my ZWO mono camera and RGB astrophotography filters, so here are my thoughts on your questions.

Q1. Which of my LRGB filters would be best to use for photometry?

The G filter would be best to start with, because the majority of photometric observations in the AAVSO database are reported for V or TG.

Q2. What is the right way to tell VPhot that I am using a non-photometric filter?

I think VPhot picks up the filter from the FITS header from the equipment profile in your capture software, but perhaps it would be best if George Silvis or Ken Menzies could answer this for you.

Q3. Do transformations help here?

Yes. Your filters may differ from mine, but my Tv_bv is about -0.145 whereas Tv_bv for Johnson V filters is typically smaller (absolute value), often but not always less than 0.05, and occasionally less than 0.01. The issue for large (absolute) Tv_bv values is that they may yield large differences between transformed and non-transformed magnitudes, estimated as Tv_bv multiplied by the B-V difference between the variable and a comp star. Errors in non-transformed magnitudes are therefore minimized by selecting comp stars with B-V values as close as possible to the B-V of the variable.

Determination of transformation coefficients involves using your own filters: for example B and G if you determine Tbv and Tv_bv.

Q4. How would I format the report to make it clear that the filter I used was not photometric, but something else?

Correct entries in the relevant fields in the AAVSO Extended File Format will take care of this. Depends on whether you submit transformed or non-transformed magnitudes. If transformed, the FILTER should be identified in your submission as (for example) V, the band to which you have transformed the observation, and enter ‘Yes’ In the TRANS (Transformed) field. If non-transformed, the FILTER should be identified as TG (or TB or TR), and enter ‘No’ in the TRANS field.


American Association of Variable Star Observers (AAVSO)
Thank you so much, Roy! …

Thank you so much, Roy!  This is helping me understand a bit more.  If I understand correctly, I think you are suggesting using the G filter, and comparing against the V magnitudes.  With respect to Q2, I found that I can override the filter during upload, so even if my FITS file reports G as the filter, I could override it to V.  This seems like a reasonable thing to do, as it makes VPhot happy, and shows me the V magnitudes that I want to compare against.  Plus, as you answered in Q4, I could report as V if I do transformations, and TG/No otherwise.

I appreciate your explanation regarding selecting comp stars with similar B-V values w.r.t. the transformation.

I think that pretty much answers all my questions... at least until I can get some clear skies and take a crack at calculating the transformation coefficients.

I guess one nit-pick question remains: I thought I read that TG is "Tri-Color Green," aka the green channel of a One-Shot-Color or DSLR camera.  It seems that the "DSLR TG" is actually a closer fit to Johnson-V than my ZWO G filter.  The DSLR Green has an exponential decay as frequecny increases, whereas my ZWO G has a steeper drop.  Is there some other, more accurate way to report that I'm using a G filter that's not from a DSLR (i.e. not technically TG)?

Variable Stars South (VSS)
TG, TB and TR in the Extended File Format

The following are in the official instructions for use of the Extended File Format. I've emphasized the relevant words in bold font.

"TG: Green Filter (or Tri-color green). This is commonly the "green-channel" in a DSLR or color CCD camera, or Astroimaging G filter. These observations use V-band comp star magnitudes.

TB: Blue Filter (or Tri-color blue). This is commonly the "blue-channel" in a DSLR or color CCD camera, or Astroimaging B filter. These observations use B-band comp star magnitudes.

TR: Red Filter (or Tri-color red). This is commonly the "red-channel" in a DSLR or color CCD camera, or Astroimaging R filter. These observations use R-band comp star magnitudes."


American Association of Variable Star Observers (AAVSO)
Filter definitions

Yes, the official filter list does include Red, Green, Blue and Yellow. Orange too I believe.

But these have never been used. Not one record in the AID uses code 22 for 'Green'

So, we should ask for HQ instructions on what should be done.


Variable Stars South (VSS)
Red, Green, Blue, Yellow and Orange Filters


Yes, those filters appear, but they are specifically labelled "Blue Visual", "Green Visual", etc. except for the Orange, which somewhere I found as "Orange (Liller)". I wonder if the latter was used by Liller when he was doing his Schmidt camera imaging on film for galactic novae, from South America I think.

They appear in the list at, and are annotated "historical data only; do not use".


American Association of Variable Star Observers (AAVSO)
For two-plus years, I used…

For two-plus years, I used nothing but an RGB-based system for photometry. I have a ZWO ASI533MC color camera used in conjunction with a UV/IR cut filter to suppress the near-IR response of the camera. When I started photometry (coming from an EAA background), I did not want to spend the extra money on a mono-camera and photometric filters without knowing if this branch of the hobby was for me. Now that I mentor for the AAVSO on photometry, many of my students have RGB-based systems since they come from either astrophotography or EAA. Thus, it makes sense I continue to use such a setup, although I'm now beginning to think about purchasing photometric filters. That being said, you can do lots of good science with an RGB-based arrangement. In fact, for my current project (flare monitoring of m-class stars) an RGB-based system has the advantage of taking simultaneous measurements in three bands without losing measurement cadence or over-exercising a filterwheel. 

I agree with most of Roy's comments. The G filter is close to the Johnson V filter and is a good place to start. Transforms can be used to convert results to BVR, but you should indicate in the WebObs file that a transform from RGB has been undertaken. I report most of my data as TB, TG, and TR. I'm not a big user of VPhot; I prefer AIJ. I do, however, use VPhot for transform determination, used in conjunction with the Transform Generator. When using VPhot, I select the appropriate BVR filter to represent TB, TG, and TR, respectively, as VPhot does not have these filter designations. The transform coefficients you derive will have larger coefficients than if using BVR-derived data - see the DSLR observing manual for further information.

Hope this helps.


American Association of Variable Star Observers (AAVSO)
Thank you, Gary. I will…

Thank you, Gary. I will check out the DSLR guide for more information about computing the transformation coefficients with non-photometric filters.

I do want to understand if my monochrome camera with a filter wheel should be treated any differently from a One-Shot-Color camera, like yours, or a DSLR. Do you know if that should be reported or treated differently in any measurements?

I have tried using AIJ, but it seems rather difficult to A) Identify and analyze the characteristics of potential comparison stars and B) Compute absolute magnitudes (it seems, more geared towards relative magnitudes) Do you have any good resources for doing AAVSO-quality photometry with AIJ? In particular, it seems that the latest version somehow lacks or changed the way to enter magnitudes for comparison stars.

American Association of Variable Star Observers (AAVSO)
Relating to AAVSO reporting,…

Relating to AAVSO reporting, a monochrome camera with a filter wheel using RGB filters should be treated the same as an OSC camera.  That is, results should be reported as TG, TR, or TB or transformed and reported as BRV, indicating in the report that these are transformed from tri-color measurements. Ideally, the DSLR Observing Manual should be updated and renamed to cover all tri-color systems, but that comes down to having the resources to do it. There are, of course, differences between OSC and DSLR-based systems.  For example, a DSLR system typically has a much wider FOV, which requires consideration during processing.

AIJ can seem a bit overwhelming at first, but it is an excellent tool. I love that I can do all my analysis locally; it's fast and can be used for the real-time display of photometric data - I've used it since I started photometry.  Many new features have been recently added, including a great macro for creating AAVSO-formatted WebOps reports. I would be happy to discuss AIJ with you if you like.  Also, I'm a mentor for the AAVSO.  So, if you want, I'd be happy to support you under this program as well.  Send in a request for a mentor and ask for me specifically.  Either way, I'm happy to provide further assistance.