Affiliation
American Association of Variable Star Observers (AAVSO)
Sun, 10/18/2015 - 14:54
Hello all.
I have a bucket full with data from XX Cyg. B filter and V filter.
I can transform this data and create a AAVSO report and uploading it to the database, great.
But VPHOT transform allows me only select 1 V image and 1 B image , the Vphot doing the analysis, make a report , save, webobs , and put it in to the database.
is there a way to select more images at one time? this avoid a lot off work.
With time series I can select all images from say b filter, processing and make a report off it , again webobs, upload ..., whats the point for this ? i carefully processed my coefficients, witch are not added to the "single" B filter data..same for V filter.
with best regards.
Willie.
Willie:
Yes, as has been mentioned before, VPhot currently only transforms one filter pair at a time.
Time series transformation is on the list of future features that we are considering.
However, have you looked at George Silvis' Transformation Applier (TA) that is found under Data-Data Analysis-Transformation on the main AAVSO web page. George designed this software to deal with this issue. Check it out.
Ken
PS: Many observers just run single filter time series for temporal analysis where transformation to the standard filter system is not critically needed (e.g., CBA).
Hello Ken.
Thanks for your answer.
I did look at george TA application, and try to use it, in this way... time series, I created a V filter report from all images, check, filled in my transform coefficients, load the txt file= AAVSO report into the application, got a big error , something with floatingpoint data... tried on a differend way, no succes, skip it.
not so handy if only 1 filer set is possible, no drama, just a lot off work...
PS: Many observers just run single filter time series for temporal analysis where transformation to the standard filter system is not critically needed (e.g., CBA).
Could you explain that some more, how is this untransformed data usefull for AAVSO?
met vriendelijke groet.
Willie.
Willie:
90+ percent of the CCD data submitted to AAVSO is untransformed. Transformation corrects your system to the standard system BUT this is normally only a correction of a few percent (1-10%) since most filters used by AAVSO members are not that different than "standard" filters used by Johnson-Cousins in their original work.
NOTE, that transformed data is preferred/desired by AAVSO since it removes this small but not insignificant bias compared to professional systems. The AAVSO has long discussed this issue and tried repeatedly to get members/observers to generate their transformation coefficients, regularly run two filter photometry and transform their data before submission. This was one of the reasons for running the recent Transformation Campaign! It remains a constant battle for truth and justice!
SO, transformation is desired but we are not about to ignore all the data collected by observers with only one filter (V). If the observer is only interested in measuring the variation of magnitude with respect to time (e.g., time of maximum or minimum) to determine the period, an extremely accurate magnitude is not critical. Other light curves can be adjusted to match magnitudes and confirm the time of max/min.
BTW, you seem to indicate that you are trying to transform single filter data? This is not normal and is subject to error for stars that change color over their normal light curve. Am I correct in assuming that is what you are trying to do? You did mention collecting images in both B and V?
TA takes only a little effort to learn but if you need help ask George or others. Transforming many pairs of images in a long time series is too tedious for me!
Ken
Hello, Ken.
Thanks for your reply,90 % NOT?, I,m new in this photometry work, liked it, but its not that easy, so did a lot off reading mainly from the AAVSO website, ccd guidlines, that kind off stuff, so i bought 2 filters B&V and would like AAVSO guidelines use them to, first attemps where hopefull, with the problems as we have them all in the beginning, but I think i have to follow the transform path (B-V imaging).
you wrote:
BTW, you seem to indicate that you are trying to transform single filter data? This is not normal and is subject to error for stars that change color over their normal light curve. Am I correct in assuming that is what you are trying to do? You did mention collecting images in both B and V?
answer:no no what i mean with single filter data is the dat calculated from the B-V & transform coefficient, as i did read in the guideline.
Yes I have B and V image dataset, from old faithfull XX Cyg. with the transform generator I calculate the transform coefficient with NGC 7790, witch i wasn't that easy at all, the B Image had to faint stars, so I needed some longer intergation time, using the CCD guideline I calculate the transform manually with the values off the know stars in the xx Cyg chart, these values are different as calculated with the transform generator, I think the transform generator does it knowing better then this beginner , so I use those values.
Ok I use the Vphot transform function, save as AAVSO report, did this many times, I uploaded these via webobs to the AAVSO database. at the website now i,m choosing (pick a star) XX CYG, check recent observations, and yes there is data, how ever I see a lot off data from me but with B & V Filter, is this good? I don't know, hope you can help me.
I will contact George.
met vriendelijke groet.
Willie.
Hello Wille
Did you load both the B and V time series into TA? What version fo TA are you using. YOu should use 2.45. Prior versions will not properly access the AID for mags.
Perhaps post a little more detail for George to help. TA is sooooooo easy. Its really worth perservering. I transform my data every clear night, and it literally takes me less than 10 seconds per object.
Gary
Hello Gary, thanks for the reply.
I think it,s not clear to me, how to put this data into that TA field.
What I did is follows, in Time series I produced a AAVSO report, saved as text file, open it in wordpad,
copy everthing , then open TA and paste it into that TA field, is this the right way ? hopefully not, So there will be a change that I can work with it.
I,m very curious how it works.
met vriendelijke groet.
Willie.
Willie:
What you stated is correct BUT you did not indicate that you generated a separate AAVSO Report for both the V and B time series?
I assume your time series images were collected in the following order: BVBVBVBVBV........?
Then did you run a separate times series calculation in VPhot for the B images and then the V images?
Subsequently, did you cut and paste the data from both the B time series AAVSO report and ALSO the data from the V time series AAVSO report into TA's upper box?
Is this correct?
Ken
Ken,
******What you
Ken,
******What you stated is correct BUT you did not indicate that you generated a separate AAVSO Report for both the V and B time series?
Yes, I did.
*****I assume your time series images were collected in the following order: BVBVBVBVBV........?
Yes, you are right
*****Then did you run a separate times series calculation in VPhot for the B images and then the V images?
Yes
*****Subsequently, did you cut and paste the data from both the B time series AAVSO report and ALSO the data from the V time series AAVSO report into TA's upper box?
I guess I did not, will try that
Is this correct? allmost.
thanks.
Willie.
Hi Willie,
I know how you feel :-) If you want to transform data from one night and you have only two filters and have equal number of exposure in both of them - you could use just a spreadsheet. Having a compstar in the same FOV and close to variable translates transformation to trivial calculus, basically (Delta V = Delta v + coef*(Delta (B-V)). Simple to implement, simple to use. For better uncertainty estimation, use multiple comps, average found V-s, use scatter measure as an indication of uncertainty etc.
More trickier is when you have data from e.g. 50 nights. Sometimes a frame or two are not usable in the series (because of anything) - e.g. 2 random B-frames and 3 random V-frames are missing: how to automatically select potentially some thousands of pairs "scientifically correctly"?. What if you have same object but with very different exposure times (e.g. 10 sec and 180 sec) because of some reason? What if you have many variable stars in the FOV? What if you want to measure every single star in the FOV and from night to night that list changes a bit. What about quality checking? Etc ;-)
Best wishes,
Tõnis
Hello Tonis, thanks for your reply.
Yes I also did use a spreadsheet, calculating wtc, no its only one night data set from a HADS star, i,m a beginner and maybe HADS is a good way to begin with. I use MAxim photometry as well and yes more comp stars i preffer, the VPHOT application , they say, seems to be a good application, but it isn't all that easy, att least for me.
but i,m struggel on and with help from this forum, ill get the right answers, push buttons is monkey work, I need to understand why and whats happening with the data, so I can understand the data.
met vriendleijke groet.
Willie.
Greetings Willie,
TA offers mechanisims for transforming timeseries data. Normally it collects observations into groups (eg BV combinations) But it can also work with estimates of colors so that many B observations can be transformed with a single V obs.
Easiest if you can send me or post the observations you are trying to transform and I can provide you with a worked example.
Cheers,
George
Hello George, thanks for the reply.
you wrote..
TA offers mechanisims for transforming timeseries data. Normally it collects observations into groups (eg BV combinations) But it can also work with estimates of colors so that many B observations can be transformed with a single V obs.
answer.. I think I don't understand this, but I did shoot data with B and with V filter, so i like to use these data.
I send you a mail with the data in it.
in between I discover where the first steps goes wrong.. look at here.
Cyg,2457292.48902,12.134,0.035,V,NO,STD,ENSEMBLE,na,Star 1,10.360,1.24641,0,X15386HRE,XX cyg V filter
the decimal point occurs a error, something with floating point. make this any sense?
met vriendelijke groet.
Willie.
Willie:
If you look at the TA help tab, you will note that TA does not use ensemble comps. There is a workaround but you should not start there.
Also, note that your check star is Star 1 not a comp like 103. So this will not work either. You have not set up the check in VPhot properly.
Also note that the target name at the beginning of the data line is Cyg not XX Cyg. This is another problem.
Ken
PS: did you want to skype to talk about this?
Willie,
A couple notes on the data you sent to me:
- TA cannot transform ENSEMBLE observations. That process, transforming ensemble records, is very involved and the AAVSO Extended Format records are not reallly set up to accept that data with its required backup information. If its a question of submitting ensemble records or transformed records, the AAVSO would prefer transformed records. So pick one of the stars from the ensemble to be your comp star.
- the Check star name, "Star 1", will not be recognized. You need to use an AUID or star label from the star chart referenced in the record.
- Your earlier note mentioned a problem with a floating point number; I believe that is because the record you submitted had an embedded carriage return. Each record must be continuous, with out embedded breaks.
Once these problems are cleared you will be ready to transform
- The data you sent me had 181 B observations and 151 V observations.
If you just transform these you will get 151 BV pairs transformed, and 30 untransformed.
Since this is a fast moving variable, I would suggest doing just that. Sorting the records by date show that they were collected BVBV...
with the last 30 B's you can transform them with a B-V color estimate. Examine the transforms for the first 151 BV pairs. TA will give you B-V color values and error estimate in its report section. If the color is stable over the stars cycle, use this in in separate TA run, using just those 30 B obs and follow the instructions for adding a #BVCOLOR record to the input.
George
Hello!
[quote=SGEO]
- TA cannot transform ENSEMBLE observations. That process, transforming ensemble records, is very involved and the AAVSO Extended Format records are not reallly set up to accept that data with its required backup information. If its a question of submitting ensemble records or transformed records, the AAVSO would prefer transformed records. So pick one of the stars from the ensemble to be your comp star.
[/quote]
George, I'm a bit confused. VPhot can create transformed magnitudes using an ensemble of compstars. E.g.:
#TYPE=EXTENDED
#OBSCODE=ETOA
#SOFTWARE=VPhot 3.1
#DELIM=,
#DATE=JD
#OBSTYPE=CCD
#NAME,DATE,MAG,MERR,FILT,TRANS,MTYPE,CNAME,CMAG,KNAME,KMAG,AMASS,GROUP,CHART,NOTES
ASAS J075512+2203.9,2457006.29314,13.290,0.020,B,YES,STD,ENSEMBLE,na,143,15.098,2.718,na,13722HE,na
U Gem,2457006.29314,14.411,0.023,B,YES,STD,ENSEMBLE,na,143,15.098,2.718,na,13722HE,na
I think that AAVSO Extended format itself is not limiting the use of ensembles. Well, it indeed doesn't contain any information about compstars, just a single checkstar label, but seems that this is OK for AAVSO database.
What do you think if it would be possible to implement similar scheme in TA?
Best wishes,
Tõnis
Hello George, forgive me the delay, had some other priors at home.
at your mail..
- TA cannot transform ENSEMBLE observations.
actuelly what does dis mean?. I saw it in the data telegram, I maybe did use the wrong conversion, how to fix this?
I just followed the Vphot instructions..
Ok its clear that star1 thing, have to look how to rename that to AUID...
with best regards.
Willie
TA is not able to transform ENSEMBLE records because the record doesn't contain enough information. TA would need to know the ID's of the comp stars, their raw mags and the methodology that combined the ensemble (ie simple average, weighted average, what ever).
One of the goals is that records submitted to AID include enough information that the observation can be refined if COMP or CHECK star magnitudes are refined. TA does this by putting the raw information into the comment section of the record along with the transform coefficients.
Yes, VPhot will transform ensembles. But its methodology is not well documented. And it does not include the details of the transform in the comment section of the record. This might be fixed in the future.
For now, the AAVSO is recommending that you transform your data and that means do not use ensembles, use a single comp. VPhot will do 2 filter transforms, but if you want the more than that and better documentation, use TA.
George
Hi Willie:
My comments are provided at each step you described below, in parentheses:
In VPHOT:
selecting images B filter (CORRECT)
selecting Time series (CORRECT)
selecting sequence, start analysis, done, continue. (CORRECT but did you have one comp in your sequence or several comps (ensemble means more than one comp)?
in analysis screen,click ensembles , hit all(3 ), create AAVSO report, create report file,save. (THIS STEP IS CORRECT BUT the fact that you clicked the ensemble button and selected all(3) means you did have multiple comps (three comps) in your sequence. This is normally good ensemble procedure BUT you seem not to understand what George meant. TA will NOT work unless you use one comp and only one comp in your sequence. It is just due to how TA works right now. Nothing else. It just doesn't !
Try another step: Check only one comp not all 3 comps in the ensemble button. Save your AAVSO report. Do the same for the V images. Try this data in TA and report what happens!
Hope this helps, Ken
Willie:
It looks like I edited rather than replied to your post. I know I'm a moderator on the VPhot forum but didn't know I could edit someone else's post. Sorry about that!
Ken