Good morning dear members.
Is known that since 2012, photometric measurements have been collected on the RGB channels, of the various types of color cameras, which are included in the Label DSLR photometric. However, these cameras offer more possibilities. When we see the transmission curves of these sensors without the infrared cut filter, we appreciate that the R channel draw the original drop of the monochrome sensor (without micro-filter array), while the blue channel recovers sensitivity beyond 750 nm (Towards the red side of the infrared spectrum), this is normal in all models and surely we have seen it at some opportunity, in figure 1 annex 4 different models as an example. Canon 350D, Olympus E500 Nikon D200, and two models CDD1240C / DCC 3240C
Some astrophotographers remove the infrared cut filter to enlarge the photographic spectrum, and even more rarer to see, some observers incorporate an infrared pass filter from 700 nm, such as the Kodak 89B, Hoya R70, Hoya R72, or Opteka R72, this Last very cheap on Amazon, you can find it if you look good, even in ($ 17). When incorporating this filter we can count on three infrared channels, of which the R and the B can be transformed to the bands Ic and z' Sloan respectively (although the latter more remote from the original filter z´sloan). This has been my case. Note graph 2.
In my case, I use this pseudo I Cousins filter, to follow nebular variables, because these are very red presenting greater brightness in the IR, which makes them easy to follow in these wavelengths. On the other hand, as it is enveloped in nebular regions, its light in the optical channels, V and its DSLR (TG) equivalent, is greatly weakened by the absorption of dust and gas from the nebula, and where Rc or R DSLR channel Would present contamination of the emission of the Hydrogen Alpha Nebular "Ha", which difficult to correct the background brightness in the measurement, so we avoid using a band I.
Figure 3 shows a filter Ic with respect to the R channel DSLR (infrared), as the original filter Ic has a cut towards 900 nm, is logical is produced a red leak in the R channel DSLR with respect to the original filter (Shown in light blue), even so the photometric measurements are very precise and more if they are transformed, this is seen in the graph 4, which shows their correlation with the magnitudes Ic using the stars of the Landolt SA98 field for two different nights (21-22/01/2015). Can be seen that this channel coincides less with the band i' Sloan.
Some time ago 3 or 4 years, I have been doing photometry using this channel experimentally, but this year 2017, I started to report to the AAVSO database with this, however, I had to write in NOTAS, which I am using the R channel of a DSLR using an infrared pass filter, as I show in image 1, but this is very long and only 100 characters are allowed (see figure 1), this eliminates the possibility of writing another note referring to the photometry.
It would be a good idea to incorporate a DSLR reporting label, it can be named "TI" to designate this filter. This would open up more possibilities in the AAVSO database, although I do not know other people who are doing so, having the possibility in the future will appear.
I attach several curves of light in which I have reported with this channel, my measurements I point them with Red lines, there are TG and TI measurements as you can see.
Oliver Christopher Lòpez, (Andrés Bello Astronomical Complex)
The AAVSO Extended Format record (AEF) note field has plenty of room for detailed comments. Take a look at https://www.aavso.org/aavso-extended-file-format . So, don't be shy about documenting the special circumstances of your observations.
Also note that if you develop transform coefficients between one of your configurations to a standard filter, then you can report that transformed data as the standard filter. For example, if you can develop a good set of coefficients to the filter 'I', then you can report your data as transformed 'I' data. Just document what you are doing in the extended Notes fields. The Transform Generator and Transform Applier tools would work well for this case.
Normally transforms are for standard filters for which there is reference data available. For example, you are observing with an 'I' filter and are using reference 'I' data. The transform is then adjusting for uniqueness of your optical train. In the case you describe above the response curves might be more varied then the case of your 'I' filter vs the reference 'I'. So you would want to use the transforms for objects that would not have peculiarities in the wavelengths where your system and the reference filter are different. So, avoid novae or CV's that might have emmission or absorbtion features.
I hope wiser heads can chime in here with more detailed guidance on transforming your setup.
I knew that when the measures are transformed, we can report them as the original band.
On the one hand, the difference between this band and the Ic band is not a big problem, the OGLE A.Udalski 1998 Project uses a band I, whose glass is a Schott RG9 of 4-mm, whose right fall presents the same scenario shows, also the band I (ASAS, Pojmánski 1997, 1998, 2000), and both are commonly used as Ic to be transformed using: I = i Ogle + 0.029 * (V-I).
But, my main intention, to designate the channel as TI, is that the users of the AAVSO database know the origin of the measure.
Besides, now that you touch the subject, I take the opportunity to ask something that I have never completely understood.
(One of the reasons) for using the TB, TG, and TR channels of the DSLRs, is to enrich the photometry in the standard filters, for example by combining transformed TG magnitudes with V Johnson magnitudes. The transformed TG magnitudes can be combined with transformed and untransformed V Johnson magnitudes, because the difference between transformed and untransformed V Johnson are smaller than untransformed TG magnitudes, but when we report TG magnitudes, we have the option of activating a square If the magnitudes were transformed. Then how to separate the transformed TG magnitudes to visualize them combined with V Johnson, if TG transformed or not, are enclosed in the same package ?. It would then be more convenient to report TG transformed measures as V Johnson, instead of TG transformed. Of course, I assume that it is to differentiate them from V Johnson, but for this to make more sense, then there should also be an option in the light curves visualization tool, which allows to choose to visualize, transformed measurements TG separated so that we can choose how to combine the measurements V Johnson and TG in the light curve.
The optical R channel dslr, can also be combined with r` Sloan, rather than with Rc, attached its curves and relationship using the landolt field SA98, the dispersion is due to the use of stars of all brightness, which causes the same, But this is another different story.
Hi Oliver, DSLR G channel can be transformed to Johnson V with a very good accuracy and I do not see why to separate it from CCD camera measurments that often need similar correction and are no more accurate under similar condition.
There is also another way to make the color correction: my VSF technique, this is a correction at fluxes level and it's even more reliable than the classical transformation. I would disagree to see DSLR photometry considered inferior to other techniques, this not my experience from near 10 years of DSLR photometry.
Another very effective capability that is specific to DSLR is the direct measurment of B-V. It is possible as the B and V channels fluxes measurments are simultaneous and very little affected by the sky condition. I would like to have a possibility of reporting "pure B-V" in addition to V to keep the very good stability of the direct B-V unaffected by the sky instead of B that is then affected by the V sky scatter.
This is an excerpt from the AAVSO DSLR observation manual ( https://www.aavso.org/dslr-observing-manual ):
"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
I do not see any practical use for reporting something as TG and transformed.
I absolutely agree with Roger that a DSLR TG transformed to V measurements in general are very accurate and useful.
And think about it: Many comparison stars we are using got their V magnitude from the Tycho2 catalog (so from the "green" channel of one of the photometers onboard the Hipparcos satellite), and that filter was more similar to the RGB green filter in a DSLR than to Johnson V, see Fig 1 in https://www.aavso.org/media/jaavso/2879.pdf . So if we trust the transformation from Tycho2 enough for comparison star data, we should not be too sceptical about using a DSLRs green channel transformed to V either.
Many thanks Roger and Bikeman
In the precision of the G DSLR measurements, I completely agree with you, DSLR is the only photometry that I have been working on for years, and its accuracy is fully equaled to V Johnson (and R to r` Sloan). Now I think I understand better the aspect of the report, DSLR channel G-transformed measurements, must be reported necessarily as V Johnson, and not optional as I thought (to see the "transformated" box for TG), then I do not know if this Box, this extra for this band ?. But I still maintain the original idea, an option for TI, because not every time I have been able to transform R infra-red DSLR to Ic (and even so, I have had to report it as Ic), like no every time I have been able to transform G-channel to V-Johnson.
With respect to VSF technique. I'm not sure if this is: Stellar Photometry With DSLR: Benchmark of Two Color Correction Techniques Toward Johnson’s VJ and Tycho VT, https://www.aavso.org/media/jaavso/2880.pdf I read about it in a superdicial way, today look for the document in my PC, and now I understand why I do not study it deeply , the translation came out with some codification and I do not read it in detail (I speak Spanish). But if I understand that the idea is to perform photometry to air masses very close to the horizon, I had developed a simple concept for this subject, which incorporates a software that develops, and that I attach some of its windows of operation, my concept is Take coefficients of a standard field, which we capture from the zenith to the horizon, then we establish the relation of the slope in relation to the air mass, and the software automatically determines a coefficient for any air mass, the zero point is calibrated with The star (or stars) of comparison. At this moment I am incorporating another method that uses extinction coefficients for each channel and a linear regression relation of these vs the color index, to give a different treatment to each star according to its color. This is already shown Mark Blackford in the new manual DSLR, attached my result for my observation site and the star alkaid. I incorporate the iterative method of Bruc L. Gary and optionally activate it. All this we can apply it to a series temporat and the operations are performed sequentially, I attach a light curve with the falling brightness of the star XY Leo using a webcam and RGB Fotocalc. I am going to study the method carefully, because I can also incorporate VSF technique at this software. I would like to know your opinion.
Also consider reporting B-V RAW in notes, which would be B-G.
Oliver Christopher Lòpez
Yes, the paper 2880 is my original publication on VSF. That technique was part of a software I developped to enable easy wide field DSLR photometry, at that time current softwares being not well matching those needs. Anything is free, no problem.
VSF is a pure color correction fully independent of the extinction gradient calculation, by the way not specific to low elevation. The major reason for is to avoid the risk of cross-contamination between the transformation coefficient and the extinction gradient coefficent that exists in classical techniques. In the last it's often impossible to get comparison stars that provides the good color and spatial distribution that avoids the cross-contamination. The VSF technique is more resilient than the transformation in case of M spectral types. In addition it is very easy to apply when the software delivers simultaneous R,G and B fluxes. I use it for many years without problem.
There are aspects that have not been covered in the paper. In fact the coefficients "a" and "b" that work are not a single couple of values but a familly providing similar results. Within that familly there is a couple of it (a, b) that is better self-adaptive to the air-mass reddening. By the way it's possible to just use a single coefficient set for a significant air-mass range and make things very simple.
In narrow fields there is no need for extinction gradient correction. In wide field I just calculate the gradient into the image related to pixel position, in fact there is no need to calculate the extinction itself. In addition this takes care of non-regular extinction condition and other residual flat errors (both so common !).
The B-V is just a direct log relation of the FB/FG fluxes ratio.
By the way it is few affected by the sky condition. This is the reason why it's so interesting. Then the calculation of the true B-V and related coefficients is based on a minus square derivation, function of the comparisons B-V and the air-mass, simple.
Clear Skies !
Many thanks Roger.
Precisely the aspect of the transformation to large masses of air (that originates the crossing of which you talk between the transformation coefficient and the extinction gradient coefficent), is what gave me more work in my software, for this I went by different canines.
I have implemented 5 different forms in my software, this is also free. I would like to see your software, where can I download it ?.
The 5 forms are as follows:
1) the classical method, where with a single transformation coefficient applied to the different masses of air using the standard atmospheric model.
2) a single coefficient of transformation, but the extinction is corrected from our own determination for the channels of our camera and stars of different color to a range of masses of air.
3) just like the previous, a single coefficient of transformation, and the extinction is corrected from our own determination for the channels of our camera, and stars of different color, to two different air mass ranges.
4) the application of two coefficients taken to 2 different air masses, and the intermediate air mass coefficients, are derived with a linear regression fit, where the slope of the relation is used as the Y axis and the mass of air as X axis, is very good at zenith distances no higher than 60ª
5) the application of 4 transformation coefficients, two for zenith distances no higher than 60º and two for zenith distances greater than 60º, the intermediate values are derived the same as the previous method (this method has a lot of calibration work).
6) VST Method (Roger Pieri), it will take me a while to incorporate it. After making the new modifications.
If I have any doubt in your method, I will let you know.
Undoubtedly, B and R can not be compared with G, but the R channel with IR pass filter, if it fits to Ic. Almost as much as G, I would even put it above G for not affecting so much the atmospheric extinction. I have measured with this infrared channel. Then with a combination like G and this pseudo Ic, if you can do much science with a simple DSLR, although as I said, I do not know anyone else who is using this pseudo Ic channel.
Hi Oliver, my software is not a Windows application (or any classical OS). It's a set of functions under a Dyalog APL workspace, like a tool box. Dyalog is a commercial developpment system that is also the engine needed to execute any APL function (with license issue). I use it for my own experiments only, it permits very quick programming and easy manipulation/analysis of large data sets (like our images..). The APL language is an array processing algebra far from classical language like C or similar. I think the best is we talk about the possible solution off-line.
For the extinction I prefer to use the local gradient measured into the image. Measuring the extinction epsilon is a lot of work and is valid only a given day in a well photometric sky, here in urban western Europe it's not often the case !
Hi Gary, it's clear the B and R channels of the DSLR are made to generate the right colors to human vision and this is somewhat different than the Johnson's B and R. In addition the sensitivity of the red is low compared to the green (4~5 times ! ) and at less extend the blue. By the way the SNR is not as good, then we could also have sampling problems with some optics. The B and R responses curves are very different than Johnson ones but the transformation remains possible for number of spectral types. The B-V extraction works well for most except certain MxIII types which the B-V doesn't follow the Teff.
The attached table is a simulation of the photometry of the DSLR filters for the full Pickles spectra library. V-Rc and B3-V are the reference values provided by Pickles (mag), Bm-Gm and Gm-Rm are the DSLR results (mmag). Vjm is the Vmag error from a Johnson filter at one air-mass, Gm the uncorrected DSLR error, Gtm the classic transformation, Gcm the VSF, all in mmag. At end, OTxxx are results from several spectra of Nova Delphini.
You could see there are many cases for which the DSLR B-V works.
Clear Skies !
I do not yet see any appreciation for the accuracy of the R channel using the pass infrared filter (pseudo-Ic) when its accuracy in relation to Ic is comparable to that of the optical G-channel in relation to V johnson.