Affiliation
American Association of Variable Star Observers (AAVSO)
Mon, 03/30/2015 - 03:53
Hi all,
I am a novice to DSLR photometry. I have been monitoring recent Nova in Sagittarious (although it is clouded out here in Mumbai for last four days).
After de-bayering, I computed Green magnitude using Citizen Sky analysis Spreadsheet. However, if I mean to report Blue magnitude (which is sensitive to H-beta), how do I calibrate instrument magnitude so as to compute Blue magnitude? Also to compute Red magnitude?
Thanks in advance,
-Shishir
Hi Shishir,
good to see you're still monitoring the nova in Sagittarius. I've only managed a couple of visual estimates so far. Been too busy with the DSLR course, but it is wrapping up now so I should be able to start observing again soon.
As I mentioned in my March 17th email, novae are NOT suitable targets for transforming DSLR instrumental magnitudes to Johnson-Cousins magnitudes.
DSLR Bayer filters have different spectral response curves to Johnson B and V filters and Cousins R filter. The green filter is close the the V filter, but the blue and red filters are quite different to the standard filters. For stars with spectra that closely approximate black body radiators (i.e. insignificant emission lines or absorption bands) we can transform all three DSLR instrumental magnitudes successfully.
But novae are not black-body emitters. At certain phases in the evolution of novae their spectra contain very bright emission lines which makes transformation invalid.
The best you can do for novae is measure the DSLR green channel magnitude, and report the untransformed value as TG filter (tricolor green). Forget about the blue and red filters for novae.
The Citizen Sky spreadsheets perform transformation automatically but I think if you set the variable's B-V colour index to zero the calculated magnitude will be effectively untransformed. [If Brian Kloppenborg, Roger Pieri or Heinz-Bernd Eggenstein are reading this can you please confirm this?]
I will contact you off-line about transforming DSLR blue and red magnitudes of normal stars. Cheers,
Mark
Hi Shishir,
I have to agree with Mark's comments above. First, it is worth reviewing the mathematics behind the calibration spreadsheet in our DSLR photometry paper (in JAAVSO so its free to read). This is really the (extremely shortened) version of the DSLR photometry manual from the AAVSO which I really suggest anyone who is new to the hobby read prior to getting started. In the paper, equation 3 shows the airmass corrected formula that is implemented in the spreadsheet. There you see that the calibrated V-band magnitude is linearly dependent on the (B-V) color of the star. Thus one must know this value (or be able to approximate it somehow) in order to calibrate the photometry. This is quite troublesome for DSLR photometry as the cameras's other channels (B and R) are not ideal for determining these quantities. In fact, in some DSLR cameras the B channel can have a fairly bad leak of near-infrared light!
Because novae undergo fairly dramatic color changes, have copious emission lines, and DSLR's G channel does not correspond perfectly to V-band, novae are not well suited targets for DSLR photometry. I was originally going to suggest that you could calibrate the data using (B-V) colors from other's photometry, but I realized that the differences in filter response between G and V might cause a systematic bias between the two systems. For example, imagine there is an emission line in the nova that is just shortward (i.e. at a smaller wavelength) than the standard V-filter would measure. As shown in Figure 4a of the paper, the DSLR camera's green filter would measure this whereas (Johnson) V-band would not. Unless extreme care were taken to correct for this, the G measurement might have more flux than the corresponding V measurement, thus the nova would appear brighter in transformed G than it should.
Thus I think Mark's approach of submitting untransformed tricolor G is probably the best solution for now. From these data astronomers will still be able to estimate the time decay coefficients thereby making the data quite useful although they are not in a standard filter.
I hope my comments were useful.
Brian
Hi all,
I often see messages from Mark, but this is long time we have not exchanged Brian, I am very pleased you keep interested in DSLR !
The first thing I would say is that photometry of nova is a problem for anyone including CCD with Johnson filters. We have had such experience during the Nova Delphini 2013 campaign. You could review the long discussions with Arne about that on the forum. There are important differences between commercial BVR Johnson filters, and few of them have really integrated the change of detector types from the time of Johnson (move from intensity to photon count). When you look at the filters response from various pro observatory you would find increadible differences !
During the Nova Delphini 2013 campagne my own results using the 450D (the one of the paper cited by Brian) were not ridiculous at all and much more coherent than number of observations done with CCD and BVR filters. In addition the B-V and V-R direct measurments that are possible with a DSLR are very meaningfull as all made at same time, same conditions, same image. What I would like to see is a specific way to report that direct B-V and V-R. I would note that the coherence between such B-V and V-R is a very good indicator of specific spectra: when B-V and V-R are coherent we have classical star spectral types , when they diverge this is a special case like a nova.
Now, for DSLR, there is another technique than the classical transformation to take care of the filters response. I called it VSF, this is a technique that simulate the V response by combining the B, G and R channels. This is very simple to apply, you don't need catalogue reference after calibration is done, and it is much less sensitive to the spectral type of stars than pure V itself. My paper on the subject is here:
http://www.aavso.org/media/jaavso/2880.pdf
Brian you said some DSLR could have a huge near infrared leak in the blue channel. This is not my experience and measurments. The normal photographic application of DSLR needs a very strong rejection of IR, otherwise the color space of the photo will be strongly affected. All digital cameras have a strong stack of rejection filters in front of the overall sensor surface. This is made from one IR dye filter PLUS one interferential IR cut (strong cut at 700 nm). Maybe you refer to the response of the pigments that are directly deposited on the CCD or CMOS pixels. Yes, they are all transparent to IR, but the filter stack that is in front of the sensor totaly rejects that IR. Even much better than the classical Johnson filter do ! This is true for normal DSLR, but it's clear we should be carrefull with modified DSLR from which the filter stack has been removed, they should be more or less considered like unfiltered CCD and external filters shall be applied, at least an IR+UV cut (specific sets exist that can be inserted within Canon cameras).
As a conclusion I would say the Johnson system itself is not adapted to the observation of nova, Arne recently wrote about that:
http://www.aavso.org/nova-photometry
I think this proposal should also be considered for DSLR. They are not very sensitive to Ha but Ha is usually so strong in nova it should be no issue. OIII is ok. V no problem, in particular when using VSF.
Brian, Mark, if you are interested we could discuss such subject off-line.
Clear Skies !
Roger
Hi Roger and Brian,
good to hear from you both. We all seem to be on the same page regarding DSLR observations of novae.
After feedback from participants in the two DSLR Photometry courses I plan to rewrite much of the manual. I'll include a discussion of novae photometry problems - with both DSLR and conventional CCD/filter systems.
In the second DSLR course we used an updated version of the Reduction-Intermediate spreadsheet (mainly cosmetic changes). Further improvements are planned such as blue and red channel processing and time series analysis.
Roger, I'll look at the possibility of including your VSF technique as an option. Right now I'm doing the VPhot course so when that is over I'll re-read your JAAVSO paper and start thinking about how to implement it in Excel. I'll probably have some questions so I'll contact you off line then. Cheers,
Mark
Hi Mark,
Processing the B and R channels is ok, but I don't think it is very safe to transform them to standard photometric filters since their profile are so different.
By the way, there was a pretty nice paper about DSLR photometry that appeared on ArXiv last week. It is well worth a read: http://adsabs.harvard.edu/abs/2015arXiv150603097Z
Regards,
Brian
Hi Brian,
yes I saw that paper - very interesting. I understand why they chose to have tightly focused star images but the additional processing required to overcome undersampling issues seems very complicated.
Also, I'm sure twilight sky flats would be compromised by intensity gradients across their 9.6 x 6.4 field of view.
Regarding transformation of blue and red channel magnitudes, I've had good success with many E Region standard stars which I've used for determining transformation coefficients. But a thorough study is required to identify which types of stars are suitable, and which are not. Cheers,
Mark
Hi Mark,
Sorry if this post is outdated, but has there been sufficient acceptance of means to use all three channels of the DSLR as a one-stop-shop for rgb photometry?
Any help is appreciated.
Best regards,
Steven
Hi Steven,
the blue and red channel transformation remains problematic for stars that are not approximately blackbody emitters, e.g. novae. So the non-transformed TG magnitude is still the recommended way to report observations of these types of targets.
You could also report the TB and TR magnitudes where the catalog B and R magnitudes are used to standardise the DSLR blue and red channels instrumental magnitudes, respectively. Catalog B magnitudes (or more commonly B-V color index) are readily available, however, catalog R magnitudes (or V-R color index) are rarely available.
I still believe DSLR blue and red instrumental magnitudes of normal (i.e. blackbody-like) can be transformed successfully. Cheers,
Mark
I think if you
I think if you made observations of a nova over long timespans, it would be a shame not to completely ignore the B channel. You will see a color change, you can report the data as TB (tri color blue), and if this serves to quantify how much off this is from standard filters, this will be useful as well.
I guess this thread will also be found by people looking for DSLR color transformation for objects that are more or less black-body-emitters, and in this context the dicussion here: http://www.aavso.org/bv-index should be useful on how to get transfomration coeffecicents for transforming to standard B filters.
CS
HBE
Hi Heinz-Bernd,
Thanks for reminding me of that forum thread, I never did get around to generating transformation coefficients from the M67 (not M46 as I wrote in my post) images I recorded back in April. I'll do that after the VPhot course finishes in a couple of weeks. Cheers,
Mark
Many apologies for
Hi All,
Many apologies for replying a bit late. My sincere thanks to Mark and Brian for guiding and Roger and 'Bikeman' to given invaluable inputs. Yes, Brian as I was monitoring nova, I was collecting information about DSLR photometry, transformation coefficients and related issues and it was at that time that I came across your papar a couple of months ago apart from papers of Hoot (2007), Loughney (2010) and a few articles on nova photometry.
My motivation to monitor Nova Sagittarii 2015 No.2 was that it was accessible from my urban sky (Unfortunately, with the arrival of monsoon, now I am forced to stop my observations and will be missing decline phase). Since it isn't seem a classical nova with gradual decline, I am rechecking and refining my observations so as to manifest the fluctuations as accurate as possible but its highly time consuming task :-(. My initial idea was to provide researcher varied channels of observations by transforming RGB into BVR using transformations such as Park et al.(2015) available at: http://adsabs.harvard.edu/abs/2015arXiv150104778P OR to in addition to RGB, compute V mag using assumed brightness of nova vide van den Bergh & Younger (1987) or Miroshnichenko (1988) or Hachisu & Kato (2006) or Hachisu & Kato's universal decline law (2014).
In any case, I hope my 60 days observations should be useful, isn't it?
Regards,
-Shishir Deshmukh