Telescope filter mapping for VPHOT

Mike,

I don't think that VPhot recognizes TG, TR, TB filter names, so you will have to change them for use by VPhot to V,R, B one way or another.  If you can get the Filter Mapping tool to do this, fine.  (I say this with some hesitation since I haven't had much success using the mapping tool.)

Another way to to change the filter name is when you upload the image.  Then you also have the option to specify the filter name.  I believe either method will rewrite the FITS header value for keyword FILTER in the VPhot version of your image file.

One issue: 

When you submit your measurements using the (very convenient) VPhot to WebObs pathway I think your report will give the edited VPhot header value for FILTER, i.e V instead of TG.  This would not be correct unless you have transformed  your TG measurements to V. 

Untransformed TG observations should be reported as TG, so you will probably need to edit the report to show TG rather than V before submitting your magnitude measurements.

Transforming TG to V is not required.  This is something you can do later once you have more experience.

Phil

Phil, you wrote:

"I think your report will give the edited VPhot header value for FILTER, i.e V instead of TG.  This would not be correct unless you have transformed  your TG measurements to V."

Whether ot not DSLR green channel measurements are transformed, the WebObs extended file format instructions are to report the filter as TG. There is a separate field (TRANS) in which the content is either YES or NO, for transformed or not transformed.

Roy

Roy,

I was commenting on reporting data from monochrome CCD or CMOS cameras using tricolor imaging filters in place of standard photometric filters.  I'm not familiar with how DSLR observers report their data, but I think it should conform with the standard practice.

Could you please point me to document and page where you got this idea.  i think this needs to be corrected, or at least clarified.

Phil

Agree 100%, Phil. Transforming DSLR TG magnitudes to V and then reporting them as TG is technically just wrong. If they're properly transformed, they are then on a V magnitude scale, full stop.

I don't know where the disconnect is, but yes it sounds like some manual or file format doc somewhere needs to be sharply corrected or at least well clarified.

Phil and Eric,

Go to WebObs. In the first paragraph there is a link to AAVSO File Format Specifications. The paragraph containing the link entreats observers to ensure that they adhere to the specifications.

The sequence of links is AAVSO Home Page>WebObs>AAVSO File Format Specifications>AAVSO Extended File Format.

I’ve copied and pasted part of the specifications below – only those that deal with the issue under discussion. I don’t normally like to appeal to authority, but when this issue was discussed elsewhere in the Forum, after I pasted an identical comment about TG etc., Arne commented that I’d given a good description of the requirements.

• FILTER: The filter used for the observation. This can be one of the following letters (in bold):
  • U: Johnson U
  • B: Johnson B
  • V: Johnson V
  • R: Cousins R
  • I: Cousins I
  • J: NIR 1.2 micron
  • H: NIR 1.6 micron
  • K: NIR 2.2 micron
  • TG: Green Filter (or Tri-color green). This is commonly the "green-channel" in a DSLR or color CCD camera. 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. 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. These observations use R-band comp star magnitudes.
• TRANS: YES if transformed using the Landolt Standards or those fields that contain secondary standards, or NO if not. Document the method used to transform in the "NOTES" section.

The important point in the above is that the FILTER and TRANS fields are independent. Thus, a TG filter observation can be transformed or not, just as a Johnson V filter observation can be transformed or not.

To me it is not logical to label a transformed DSLR green channel observation as V. The passbands of the filters as everyone knows are not the same, and different spectral characteristics of stars will result in differences between transformed Johnson V observations and transformed DSLR green channel observations. Transformed DSLR green channel data may be identifical to a Johnson V filter result for some stars, but the point is it will not be the identical for all stars.

Roy

I probably should give a little clarification. I was referring to a post under "Using Pixinsight for Photometry" which mentions the following  :  

VPhot and DSLR images 

VPhot can handle DSLR images and it will get better in the future.

I believe the steps now are:
- You need to calibrate and de-bayer the images first and upload the single channel fits files.
    These images can be big. Convert them to 16bit integer instead of leaving them as floats.

- The filter names are TG, TB and TR.

- In the telescope profile establish the mapping of TB->B, TG->V and TR->R. This tells VPhot what reference data should be used. Yes, I know,  it should know automatically that TG is defined as DSLR green referenced to V; working on it.

- You can report to webobs TB, TG and TR values. But better you establish transform coefficients for your camera so that you can transform the data and be able to report it as B, V. R is a little problematic to transform.

I hope to improve VPhots handling of color imagery this year.

George

I am just starting to learn VPHOT  and also dslr photometry, however it looks like when you map the filters as mentioned above , the report will show the filter as"V".  If I do not map , and if I have 'filter=TG" in the fits header than that will show up in the report.  From what I have read we want to report it as TG, not as V unless we have performed the transformatios.  So it looks like, from my trials, that I do not want to do the telescope mapping of the filter channels so that the report indicates "TG" from a dslr.  

I just want to make sure that when I finally submit a observation that I have done it correctly.  I hope I have explained myself.  

Thanks,

Mike

Mike,

Re: data submission.  I think you've got it right, but below are the instructions from the DSLR manual. 

For VPhot, if you don't (somehow) change TG to V in the header, I believe (but I could be wrong) that VPhot won't know what to do with it, and you'll get some kind of error message.  When you change the header FILTER value from TG to V, you just have to remember to report TG for untransformed magnitude measurements.

As I said previously, I recommend that you don't bother with tranforming TG to V until you have more experience.  Trying to do this from the get-go will just slow you down.  Get used to submitting TG measurements first.  If you observe non-red targets your results should be good. 

AAVSO DSLR Observing Manual V 1.4, page 79  regarding WebObs submissions for Tri-color observations:

"If you are submitting standardized magnitudes (i.e. non-transformed magnitudes) leave the “transformed” check box under the Magnitude field unchecked and select the appropriate “Tri-Color...” option from the Filter drop down list. For example select “Tri-Color Green” if the standardized magnitude was derived from green channel instrumental magnitudes.

For transformed observations (with or without extinction correction) be sure to check the “transformed” check box and select either Johnson B, Johnson V or Cousins R from the Filter drop down list, depending on which DSLR color channel was used."

Phil

Phil,

I can see what's in the DSLR manual. The problem I have with this policy is that someone using a colour CMOS astronomical camera (not a DSLR) and submitting a transformed green channel observation would be submitting "CCD" as the Obstype and "V" for the filter.

Which is the same entry as for someone with a monochrome CCD camera and a V filter submitting a transformed observation. I believe that's not good.. Such differences should be transparent when analysing data in the future.

At least with DSLR cameras the Obstype is listed as "DSLR".

Roy

Phil,

When you use the submission form for individual DSLR observations, as described on page 79 of the DSLR observing manual, there is a drop-down list in the Filter field. In my opinion, when this form was designed, the filter field was meant to represent the physical filter. The choices include "TG" etc, and "Johnson V" etc.

I submit that it is patenly wrong (even though in the DSLR manual) to list your filter as "Johnson V" when in fact you are NOT using a Johnson V filter.

However, the people who designed the form and who wrote the DSLR manual would need to comment.

Roy

Again with the perennial confusion between filters and passbands.

V filter is not V passband, in fact they have nothing to do with each other outside of convenience (and past practices for better or worse). Johnson V passband is a well-recognized, worldwide standard--you can look up the transmission spectrum. But "TG passband" doesn't even exist; for magnitudes reported as TG filter without transformation, at least we know what the observer did.

The specific confusion here probably arises from the Extended Format doc's specifying Filters for results when it definitely should specify Passbands. Transformation merely serves to get magnitudes from whatever-filter-you-used to a proper, standard target passband, by way of using a color index from a catalog (or failing that, from backcalculation from 2 magnitudes--tricky math but it works).

So whether you take images with a DSLR green channel, or through a V filter, or through Saran Wrap^®, or through your brand-new Martian ♦^Œ§¥¥¥ filter--after doing a proper transform (if possible) to V passband, that magnitude is now cast into V passband. That's what transform means. And so of course it would best be reported as V passband, even if the Extended Format doc still calls it (sigh) "Filter". Or perhaps best: if transformed, both filter and passband should be specified.

It seems that the DSLR manual gets this right, so yay! And yay Phil for finding the relevant text.

"Or perhaps best: if transformed, both filter and passband should be specified."

How do you define a passband easily (one field in a database)?

But otherwise, your statement supports what I am suggesting: that the actual physical filter used and whether or not the data was transformed both be specified.

I haven't seen any actual discussion on the problems I have raised with the official DSLR manual instructions, or the fact that a transformed DSLR/one-shot-colour camera green filter observation is NOT the same as a Johnson V filter observation for all stars.

It's OK to quote an instruction manual. The issue is whether or not the manual should be questioned.

Roy

Roy,

I do understand your points, but if we are submitting data to the AID we must all follow the same rules for how we characterize our data.  This means that if proper procedures are followed a V magnitude is a V magnitude regardless of which filter or camera was used to derive it.

To your points, we are encouraged to include with our submissions comments on how the measurements were made, especially if there was something different from the normal CCD camera with a V filter used to derive a V magnitude.  I think techniques worthy of comments would include standard magnitudes derived from OSC cameras, standard magnitudes derived from tri-color imaging filter observations, and especially, Saran Wrap observations.

Phil

I've messaged Mark Blackford asking for clarification.

Roy

Hi Roy and All,

I can see how there can be some confusion but in my mind it is pretty clear.

When submitting observations using the extended format you need to include the parameter OBSTYPE = DSLR which indicates the detector was a DSLR camera.

If you measure the green channel instrumental magnitudes then use catalog V magnitude of the comp star to get the standardized magnitude of the target you report that as FILTER = TG and TRANS = NO

If you then transform that TG standardized magnitude to the V passband you report that as FILTER = V and TRANS = YES.

The current DSLR manual (version 1.4) is very clear on that I believe. This was discussed at the DSLR Manual Workshop way back in 2013 (?) and I’m pretty sure Arne Henden clarified that in email correspondence around the same time.

I agree with Eric in comment #12 above, it would be better if the FILTER parameter was call PASSBAND.

Cheers,

Mark

This has been a very interesting discussion - and civil!  The comments that follow are mine alone and not necessarily that of HQ.

I agree with Mark's description, and also agree with Mark and Eric regarding FILTER vs PASSBAND.

There are lots of free parameters here, and we simplified things when setting up the Extended Format so that the complexity wasn't overwhelming for most observers.  You have monochrome CCD, one-shot-color CCD, monochrome CMOS, one-shot-color CMOS, DSLR (both CCD and CMOS).  Some sensors are blue sensitive; some are back-illuminated; some are deep-depletion red-sensitive.  You have TG filters from different sources, with or without red blocking, and V filters that mimic the Johnson passband or are flat-topped.  At some point, you have to condense down to a reasonable subset.

I think the DSLR manual is correct in how it tells observers to submit their data.  One-shot-color (OSC) cameras are a hybrid between DSLR and "CCD", having the DSLR tricolor filters but without the consumer camera baggage.  My feeling is that OSC cameras are closer to DSLR than CCD, and observers might consider using DSLR for their obstype.  CCD and CMOS sensors are very similar in their response, and differ primarily in the supporting electronics.  Because CCD and CMOS are similar, we grouped them together as "CCD", primarily to indicate that they are 2-dimensional imaging sensors as opposed to single-channel photoelectric photometers.

Once you've transformed your measures onto the standard Johnson/Cousins system, then using V as the passband is correct.  I agree that transforming a measurement taken through a filter whose passband is very different from a standard passband may have systematics.  Any transformation error should show up in the measurement error, if you want to do things properly.

If you are worried that someone might misconstrue your measures, include an expanded note for the measure, where you might include information about your camera or filter.

I hope this helps!

Arne

Thank you Phil, Eric, Mark and Arne for the discussion and clarification. I’ve got some editing to do in my previous DSLR WebObs submissions!

It seems to me that a practical consequence of the discussion is that the WebObs AAVSO Extended File Format definitions need clarification, because they are not specific enough to be read in isolation (without reference to page 79 of the DSLR Observing Manual).

I suggest something like the following, with the additions in bold text at the end of each paragraph.

• TG: Green Filter (or Tri-color green). This is commonly the "green-channel" in a DSLR or color CCD camera. These observations use V-band comp star magnitudes. Use only if the measures are non-transformed. If transformed, use the appropriate passband instead of TG, typically V.
• TB: Blue Filter (or Tri-color blue). This is commonly the "blue-channel" in a DSLR or color CCD camera. These observations use B-band comp star magnitudes. Use only if the measures are non-transformed. If transformed, use the appropriate passband instead of TB, typically B.
• TR: Red Filter (or Tri-color red). This is commonly the "red-channel" in a DSLR or color CCD camera. These observations use R-band comp star magnitudes. Use only if the measures are non-transformed. If transformed, use the appropriate passband instead of TR, typically R.

Roy