DSLR Tutorial - Calibration: Beginner spreadsheet questions/comments

Hi Paul, To answer your questions: 1) lambda Aur as the Check Star: lam Aur is used by several ameteaur astronomers as check star for the International Epsilon Aurigae Campaign. All stars are variable at some level and I was concerned that lam Aur may not be entirely stable. I had a lengthy discussion with Jeff Hopkins about it's long-term stability (he stated it was stable and offered proof) and have inspected it's light curve from the SMEI spacecraft. In both cases, it is stable below the detectable limit of ground-based photometers so it serves as a good check star. It's magnitude also appears stable over the long-term (at least 30+ years now). 2a) About standard deviation: This was more for the motions of entering the equations rather than for statistical accuracy (I should not that someplace). I would really suggest that one take 10 star field images and use those for statistics. We should also be using the standard deviation of the population, stdevp, rather than stedev as we are using the entire population instead of just a subset of the data. The difference between the two functions is that stdev returns a number divided by N - 1 and stdevp returns a number divided by N, the entire population size. Stdevp should be more accurate (i.e. smaller), but in small number statistics I'd rather report a larger error. 2b) Using stacked images: The jury is still out on this one. Stacking the images gives you better SNR, but unless your photometry program reports errors on the counts, then you also loose the statistical information about the distribution. I'm in favor of analyzing each image independently then averaging and stdev'ing the numbers to give you a proxy of the uncertainty in your measurement. The big down side is that you have to measure each image which takes more time than taking the images! So if your photometry program includes errors in the counts, stack, otherwise, measure independently. We'll include a discussion of this information in the advanced tutorial. 3) Modifications to spreadsheet We probably won't make many more modifications to the beginner spreadsheet as it doesn't include airmass corrections and is therefore useful fairly infrequently. We also were hoping to get people typing stuff in so they can learn a little bit about spreadsheets in the process. I'll pass your recommendations along though. Have a good day, Brian

Hi! If it were just for AIP4Win use, I guess it would be easier now to use the MMT Magnitude measuring tool where you do photometry on a series of frames (AIP4Win does the tracking). Anyway, Richard berry considers the Single Image Photometry tool as a legacy tool only and is focused on many MMTimprovements in the most recent version of AIP4Win (2.4.0 as I write this). BUT, I think we also want to have a tutorial for a free software like IRIS, and there it's probabyl easier to stack first, measure then. Maybe a compromise would be to re-write or amend the AIP4Win tutorial (to cover the MMTand other improvements that went into 2.4.0) and leave the stack-and-measure approach in the IRIStutorial? As for my personal experience, now that eps aur can be observed at low airmass & cold temperatures, I get comparison star magintude fit errors that are really low (<< 0.01 usually) and very consistent Eps Aur magnitude estimates. If I measure 4 stacks of ca. 40 seconds total each, I now get a standard deviation that is well below 0.005 mag (with an F/2 50 mm lens). So I'm really not so worried about losing the distribution of the mag in the individual frames, I assume it's not the limiting factor in the measurement error. CS HB

Hi Heinz, I too have seen great errors in the fit parameters, but what I was concerned about were the errors in the count values because each measurement of instrumental magnitudes also has some uncertainty with it. Thus far we have ignored these factors, but if we want to take DSLR photometry to the next level (i.e. the "advanced" tutorials), we should fully propagate the errors. I'm fairly sure MaximDL includes the uncertainty values along side the count values. Do you know if AIP4WIN or IRIS do the same? If they don't, I can write Christian and Richard to see if they would do this (so far they have been very good about updating things to meet our needs). See ya, Brian

Hi Brian, Just advising you that AIP4Win 2.3.1 shows a +/- uncertainty value in the Single Image Photometry tool for every star chosen. How it determines this uncertainty I haven't worked out as yet, but it's probably somewhere in the accompanying Handbook. Version 2.3.1 is the one for which the Beginner tutorial was written. I've now downloaded the latest version (2.4.0) but haven't installed it yet - I'll hold off until I am happy that I am processing the images correctly and getting sensible results. Regards, Paul.

Hi Paul, Thank you for clarifying that. I think Richard implemented it correctly (he had a note about this in his book from what I recall). Could we try a quick test? Could you measure the counts and uncertainty (for one star) on five frames and then do the same for those five frames stacked and then post those 12 numbers here? I think we'll find the noise will go down as ~1/sqrt(N) where N is the number of frames. Thanks! Brian

Hi Brian, Paul, I'm pretty sure there is nothing in IRIS that would help here. As Paul mentioned, AIP4Win does display uncertainties for each single magnitude it computes, but of course it depends on some constants for the sensor characterization that you have to provide first: gain, read out noise and dark current). This is not-so-constant for uncooled DSLRs I would guess, but AIP4Win even has a tool that automagically does the sensor characterization for you given a set of test frames (bias, flats, darks). CS HB

Hi Brian and Heinz, Just a few general comments on your replies and discussion: 1. What you seem to need to do, from my perspective, is to decide how far each of the different tutorials should go wrt improving the results. I would suggest that the Beginner level should not worry about StdDev but use just the E of E the different programs provide, with the proviso to the newcomer that this is only a rough estimate based on a single value - the Intermediate and/or Advanced levels take it further. This is based on the idea that some/many newcomers are not technically trained, and the concepts of StdDev and E of E are not easy to grasp in the first instance, particularly while you are struggling with all the other issues associated with photometry. If you adopt this approach, then I would suggest that you a) advise the user not to report the results of their own star measurements using this method, and b) adjust the spreadsheet by removing this standard deviation calculation (but retain it in Intermediate &Advanced spreadsheets). A Beginner's tutorial, after all, is simply an introduction (which most people would understand and accept) designed to introduce the newcomer to the main processes involved. I haven't yet examined the Intermediate tutorial so don't know what it covers. 2. It seems, from reading your comments, that you are faced with the same problem that most authors of the computer textbooks face: a shifting and developing subject. Here we now have a ver. 2.4.0 of AIP4Win with a different processing protocol available. I'm sure that during this year we are likely also to see changes to Iris and MaximDL that will impact on your tutorial notes. This is going to become onerous over time, and I sympathize with you. Good luck with your efforts. I have found the Beginner tutorial quite useful and has saved me hours trying to learn packages on top of all the issues associated with photometry measurement. If that is what you have been trying to achieve, then you have succeeded with me (despite my critiques of the tutorial in another posting). Great work, chaps, Paul.

Hi Brian, Many thanks for replying to my post - I appreciate your giving time to help out a newcomer. 1. Lambda Aur: Since this is a Beginner's spreadsheet and tutorial, then the fewer questions one has the easier the path is for the newcomer, like myself. I would advise strongly that you note why lam Aur is chosen as the Check Star in the tutorial to stop the user from chosing another star. For example, when I examined the star chart I wondered why rho Aur wasn't chosen as it is the nearest to eps Aur - lower Air Mass effects - and have actually gone ahead and done this. I'll change back to lam Aur as you indicate it has an impecable pedigree. 2. Stacked image vs individual frames: I'll leave that for the experts to decide - I'll just add that the simpler the better for a Beginner's tutorial. The aim of a Beginner's tutorial, surely, is to introduce the newcomer to the main features of the discipline without bogging them down on ifs, buts and maybes. Now, if the software actually calculates a StdDev from the individual frames used to make up the stacked image when you click on a star in that stacked image, then this would be a wonderful time saver. Richard Berry - are you monitoring this discussion? 3. I've made comment on what I think you need to do to the Beginner's spreadsheet elsewhere - remove the StdDev calculation section entirely with the tutorial note that a) the software may provide you with an Error of Estimate which is second best to a true StdDev, and b) don't use this spreadsheet and Beginner processing method to report results of your own star measurements. I hope these comments help make a more useful tutorial for everyone. Paul.

Hi Brian, OK for the test. Of course, the tutorial files supplied are only 3 in number (Beginner and Intermediate) so I will have to use my own files. This caused a hiccup - vers. 2.3.1 couldn't read my Canon 60D files, so I had to install vers. 2.4.0. I hope this is not going to complicate matters for us. Please note that I will be away and out of Internet access for the next 3 days, but I will progress this during my absence. Now, Iseek your advice on one thing: Should I make master dark, flat and flat-dark frames and apply these to each individual image prior to making the measurement? My thought is that I should, but I would like some confirmation of this. Thanks for your help and best wishes, Paul.

Thanks for explaining what those extra parameters do, Heinz. I haven't gone chasing their meaning in the help file. I'll have to find out how AIP4Win does this automatically - does it occur when you select your series of flats, darks, etc., or do you have to go through another process to get these? Cheers and best wishes, Paul.