Beginner questions : calibration, demosaicing and other.
Stacking program, astrometric (plate) solver, fits file viewer
Brought to you by:
han59
Menu
▾
▴
Hello Han, hi all,
First, thanks for working on and sharing this piece of astronomical soft, which seems very promising, and full of interesting options.
I'm new here, and in the (somewhat heavy) process of learning how to use Astap in many fields. Image calibration, plate solving, photometry.
I do imaging with a canon 650D DSLR, and using CR2 raw files.
I suppose I will have a few questions before being able to go out alone from the sandbox, if you don't mind.
My first question will be about the calibration – Stack (Calibration only No-demosaic)
22 lights, 14 darks, 19 flats and 12 biases.
There was a warning flag : exposure times differences.
Indeed, for a theoretical exposure time asked for 120 sec, the automatic timer did exposure times ranging from 120.4 to 120.6 secs. I don't thing this is very important.
Also, there was a second warning flag : sensor temperature differences between darks and lights.
All temperatures ranging between 18 and 22 degs.
I even didn't know that there was a temperature sensor in those DSLR. Yet, the lights and darks where taken outside with an external temperature of a few deg above 0 celsius (say 4-5 deg)
Is this of any importance, and may that affect the data, in the purpose of extracting photometry ?
Thanks in advance,
Christophe
Hi Christophe,
For photometry these practical problem are not of importance. Even for photography a few degrees have a very minor influence.
Other topic. Note you can calibrate your files directly from the popup menu in tab photometry as long the darks, flats and flatdarks are loaded in the corresponding tabs.
Feel free to ask any other question.
Han
Oh ! So soon !
Thank you Han for this soon reply.
It is currently snowing, so I have plenty of time to understand and test Astap ;-)
Ok, I have noted that I could directly do calibration from the photometry tab. That's good.
The main issue I get now is about debayering the light images.
I tested every available pattern, and never get a perfectly colored image.
The best result I got is, in auto mode, or RGGB, with libraw, which seems usual for a canon cr2 file. But as you will see, the colors are not vivid, something like extinct, or reddish.
Is this normal ?
Will separate channels be correctly extracted anyway ?
Another question :
is there a way to extract R, G and B separate channels in just one “click” ? Exactly as one could extract G, or R, or B in one click ?
I did not find this option. Sometimes we have to perform photometry in each available channel.
Thank you for further help,
Merci !
Christophe
I forgot ... Here is a copy of the original photo used to test de debayring pattern.
It should looks more like this.
Auto mode for demosaic is in most cases correct.
ASTAP doesn't apply a supplier recommended colour correction matrix like a DSLR camera would do. However it tries to balance the colour such that the average star colour is white and also the background is gray. This will work well for celestial images. but not so good for terrestrial images. The same way a DSLR camera will work fine for terrestial image but could produce greenish or reddish astronomical images.
I see. That's fine.
I did a few more tests and it works very well.
Thank you for that.
Christophe
No there is not. For the moment I suppose most people are just using the green pixels since it is 50% of the pixels. Blue or red are only 25% of the pixels.
You're right, Han. There is no more than 25% of the light in red and 25% in the blue.
Despite this, it is possible to observe some phenomenon and differential LC, when just comparing the 3 different RGB bands. This is particularly interesting when observing eclipsing binaries, and novae, when there are two or more involved stars that are not of the same colour, EG : a blue star orbiting a red star.
The case of T CrB is particularly interesting, as there are large discrepancies between the 3 bands in this case.
Most observers don't know this about DSLR photometry, or did not try.
Maybe some day, if the option became available and automatic, with RGB extract and simultaneous photometry in the three colors band, with automated AAVSO report in three band TR, TG and TB, it should be great, and probably more observers will use it.
More researchers in AAVSO alerts are requesting/encouraging those type of observations with DSLR.
Currently, this is feasible, but it requires some times.
Cheers,
Christophe
I have considered doing the three colours in one step. You could do it directly on the raw. There is no need to separate in three files. It would require code changes. Also the users should have the image out of focus to get a reasonable flux distribution. Maybe in future.
Cheers, Han
Han,
You mean, doing photometry directly on the CFA no-debayerized images, right after the calibration ?
Well, a good solution, probably.
Defocusing is not a problem for most of observations. We use to do that almost every night in order getting a longer exposure time and a better SNR, without saturating the stars. And sometimes, the images are largely defocused. What about the plate solving, in this case ? I understood that the images should be well focused for beeing plate solved.
By the way, if you add such an option, would it be possible to add another option ?
In some cases (eg = exoplanet transit on a star of low brighness), we have to perform photometry on the "full" light, independently from R,G or B filtration, in order to get better SNR, becouase we have no other choice. So the images are transformed to monochrome (grey levels), and then, the photometry is extracted from the monochrome images.
Food for thought ;-)
Christophe
The worry with using the blue or red channel is that the peak of the star shape lies at a a green sensitive pixel. So defocus would fix that partly. Too far out of focus the image will not solve. I estimate it will be fine at a HFD of maybe 8 or 10.
It would require a modification of the star detection to keep record of the red, green, blue flux of the raw image. Technical not a problem. But at the moment I assume ASTAP is only used by a niche of the users active in photometry. So at the moment I'm little reluctant to develop it. A large group of users would help. It would definitely require feedback from the users to be a success.
First step would be a field test of the red and blue channel extraction and show it works well.
You could just bin 2x2 the image for that.
Han
Indeed, photometry is a niche. I understand. Don't worry about that.
Anyway, I will perform some test this weekend, in R G and B, and report here.
The fact is that defocusing is a very effective method in astro photometry,but very few software are able to align or plate solve such images. AstroimageJ can do that, but it is somewhat complexe to use, and definitely not for the newbies.
The plate solving is effective even when stars are as big as donuts or annulus (eg 60 pix in diameter or more). I would prefere a more friendly approach, and more intuitive like in astap.
Christophe
Listen ...
So you want to solve an image ? Ok. You just need to click on the "Solve" button, as written in the tutorial.
Then you click on this famous button, and wait for a few seconds, but nothing happens.
You click one more time to the said button "Solve". But nothing happens, again.
Suddenly, you realize that the image was already solve twice. In less than a mouse click.
Seriously. 0.1 sec to plate solve an anonymous image ?
This software is a killer !
Last edit: ChrisCiel 2024-03-12
If the initial position of the mount is in the header it takes no time. Blind solve takes longer. :)
Han