I finally got around to using my QHY8 last night and took 10x5min shots of M20 before the moon came up. I also used my new light box to do flats and took a series of 2 sec and 0.2 sec shots of flats to make master flats. I was pleased with the raw shots from the QHY8 but I totally messed up my flats as they didnt correct the dust or gradient in the image. This stopped me being able to process the image further.
Can someone give me a DETAILED explanation of how to use Nebulosity V2 to get a proper exposure for the flats using the QHY8 and then to then use them in DSS to correct the light image.
Thanks for your time. And Ive included the flat and processed image to see my problem.
All comments welcome as I sure need them
Allan
I don't do flat fielding in Neb 2. I do all my calibration in CCDStack but the process would be similar in Neb or DSS I'd assume. I target 9000ADU for my flats in the brightest part. I stack 40 of them to make the master. You'll also need to shoot 40 bias frames to make a master bias. You need the bias frame when you divide your master flat into each of your subs. Otherwise your flat will over/under correct and you may get inverse vignetting or darker or brighter dust bunnies. HTH.
PS: Forgot to mention - all calibration is to be carried on prior to debayering. There is not gain debayering then flat fielding each channels. You get the same result for 2 times the amount of work.
Last edited by multiweb; 10-09-2009 at 12:27 PM.
Reason: Debayering
Hi Allan
I have not used a QHY8, and advice from those that own the camera is probably best, but I do see that its a 16-bit camera. You should aim for an exposure length that results in an ADU count of between 35-50% of the SATURATION level of your camera. On an SBIG ST10, full well depth is 77,000, but saturation occurs at around 56,000. So I shoot for an ADU count of between 18,000 – and 28,000.
That said, if you want to test the effectiveness of a flat field you have created, do not apply it to an image containing nebulosity. Test your flat exposures and techniques on just a plain star field. If your resultant background ADU values at the 4 corners and centre of the image are within just a few counts of one another, then you've cracked it and you know what exposure length fits. The fact that you are using a light box is great..as long as you are happy it is supplying a uniform light source.
You can also play with a flat, just for test purposes. Take a master flat you have created, and inrease/reduce its value using pixel math. Apply the resultant image to your light frame to see what happens.
By the way, a perfect flat will not remove a gradient.
Thanks for that Martin and Marc. Ive read Eddie T's article on Flats and it makes sense to me with your information and some other from the web. I will have to pay more attention to the flats, flat darks and bias than I had before - thats obvious. I will have to experiment with the QHY8 and see exactly what well depth that I need - however that begs the question, where in nebulosity can I determine the ADU count for a particular exposure. I was trying to use the histogram but it really didn't give me the information I needed.
where in nebulosity can I determine the ADU count for a particular exposure. I was trying to use the histogram but it really didn't give me the information I needed.
haaa.... I got caught on that one too. Beware of the bayer matrix... it's everywhere Seriously don't look at the histogram. Take your flat in preview mode then center it and rollover your mouse in the brightest part of your flat (usually the center) and look at the pixel ADU value in Nebulosity status bar bottom right. Now here comes the tricky part. I told you by trial and error with my QHY8 I go for a max of 9000-10000 ADU right? (might experiment with your QHY*, may need different levels) Well, the ADU values are radically different in the 1xR, 2xG or 1xB pixels. You may get only 2000ADU or so in the blue when the red is already well over 10000ADU. So I check the flat level in the brightest pixel of the bayer matrix regardless of the color and base my exposure time on this. I got caught the first time by doing Ha that falls only in the red, then checked the levels on an area of pixels rather than individual so although my average ADU was well within 9000 ADU I still had pixels well over 30000 ADU in the reds. Talk about inverted vignetting as soon as I debayered
I just loaded the 2 sec flats I did and ran the cursor over this file and found that the peak intensity given by nebulosity was 20,000 while the 0.2 second flat frame gave an intesity of 2-3000. So I figure that to get the intensity I need, Marc, I have to expose between these levels -maybe somewhere between 0.8-1 sec for the light intensity I was using. When I increase the magnification of the preview field I assume that at 400 times, you see individual pixels and that the level you are after is that of the green pixels as they are the ones giving the highest values in a normal OSC shot?
This has been quite informative. I suppose there is no way to decrease the intensity of the flat ie calibrate it to the lights so that it becomes an effective flat. I saw somewhere where it was stated that no matter the exposure as long as there was no exceeding the well capacity then any flat could be used. But from my shots that doesnt seem to be the case.
I suppose there is no way to decrease the intensity of the flat ie calibrate it to the lights so that it becomes an effective flat.
Actually yes there is a way. That's what Martin mentioned before. If you do a pixel math on the raw flat you can raise or lower the global ADU value by a given value. Intensity will depend on your optical train. Without filters I usually do up to 0.5s exposures, maybe 4s if I have a Ha filter in line. Regardless I try to keep everything centered around 9000ADU - that's what works best for me.
Allan , Hi Mate, you can down load the tutorial for nebulosity from Craig Starks web site . It gives a good explanation on the method of taking flat frames and bias frame images in this document.
As I said in the PM Check the intesity levels in Pixel Information in the image menu in Nebulosity but use the levels in the bottom half which gives Max for the entire image... Keep it around the 9000 mark and you must use Bias frames or the flats are a waste.
1. Keep your camera clean and blow off dust from the chip and filters so its clean to start with so you're not relying on flats to correct housekeeping. Camera needs to be at the temperature you are going to be imaging at. Wrong temperature flats may be useless.
2. I take dusk flats with a white T-shirt over the end of the scope.
make sure the camera is in focus and is in the position you will be using it without any change. I point my scope about 15 -20 degrees angle and towards an even part of the sky towarsd teh west (not around clouds that are moving around).
3. I go for 3 seconds or longer. I find below 3 seconds you can get a picture of the shutter (Sbig cameras don't have this effect as they have a different type of shutter but cameras with a normal type shutter need longer exposures otherwise you see the shutter in the image.
4. I go for 20,000 ADU more or less.
5. I take 3 of each filter. You have to be kind of fast as the light is dimming and you've got maybe 15 minutes to get them.
I take 3 bias. Not sure what your 0.2 second shot was for, was that for bias? CCDsoft has a setting for bias so I set it to that. I set it to flat field when doing flats.
6. Create a master flat for each filter or if colour just one. I subtract the bias from each flat so I make a master bias first. Then a master flat subtracting the bias. I do this in CCDstack and it is quick and easy.
The above procedure seems to work just fine.
If your camera and filters are very clean and your scope produces a very even flat then flats aren't so important. If there is vignetting or your camera and filters are dirty then it becomes quite important.
I take 3 bias. Not sure what your 0.2 second shot was for, was that for bias? CCDsoft has a setting for bias so I set it to that. I set it to flat field when doing flats.
Greg.
Greg
Thanks for your detailed reply. Ive taken it all on board. The 0.2 sec shot was for my flat with the light box. It gave me an adu of ~2000 and 2 sec gave about 20,000. So I have to experiment to get the sweet spot for DSS. Bias taken at 0.001 sec for the QHY8.
Greg
Thanks for your detailed reply. Ive taken it all on board. The 0.2 sec shot was for my flat with the light box. It gave me an adu of ~2000 and 2 sec gave about 20,000. So I have to experiment to get the sweet spot for DSS. Bias taken at 0.001 sec for the QHY8.
Not sure how the shutter is on the QHY but with Apogee and FLI cameras they use a 3rd party shutter which opens and closes like a normal camera lens. If you took a .2 sceond flat with one of those cameras you can clearly see the shutter leaves in it.
Go for ADU of 20-25,000. 2000 is too dim and may not show up the dust donuts very well. The camera handles ADU up to 65577 so go for 1/3rd = approx 22,000.
Curious as to why the master flat didn't work as it looks ok on it's own.
It seems to have picked up the gradient/vignetting and dust specks.
What settings did you use in DSS to apply the flats?
Not sure how the shutter is on the QHY but with Apogee and FLI cameras they use a 3rd party shutter which opens and closes like a normal camera lens. ......
Greg
Well I went out again last night and had a heart-attack as for an hour I couldn't figure out why my QHY8 was not working. Worked OK last night but nothing for about an hour of messing around checking cables. I missed the S video cable into the power box was loose. What freeked me out was that Nebulosity recognised the camera and it functioned alright with the program and the little green light was on but no picture. Finally when I found the loose cable all was OK. Anyway I took a whole lot of shots to determine the best spacing for my focal reducer and for my flats. To cut a long night short, I took Marc's advice and settled for a flat with 9200 average ADU, took bias frames and dark flats etc - and the frames came out perfect from DSS. No vignetting or minor dust specks (after cleaning the chip beforehand).
Well thats a relief as now I can go away next weekend and spend all my time with just the QHY8.
Thanks to everyone that came to my aid - its really appreciated and shows what a great site IIS is.
Forgot to add photo of M20 with flats and bias removed in DSS - no vignetting or dust so Im happy. Was 2x5min shots stacked.
Last edited by allan gould; 12-09-2009 at 04:39 PM.
The USB cable powers the USB interface. The SVIDEO port is for the rest of the cameras electronics.
Theo
Now you tell me!!!!
Theo,
Thanks for all you assistance on the phone - your feedback and help was much appreciated and really was of great information as usual. Service above and beyond.
Regards, Allan
Seriously don't look at the histogram. Take your flat in preview mode then center it and rollover your mouse in the brightest part of your flat (usually the center) and look at the pixel ADU value in Nebulosity status bar bottom right. Now here comes the tricky part. I told you by trial and error with my QHY8 I go for a max of 9000-10000 ADU right? (might experiment with your QHY*, may need different levels) Well, the ADU values are radically different in the 1xR, 2xG or 1xB pixels. You may get only 2000ADU or so in the blue when the red is already well over 10000ADU. So I check the flat level in the brightest pixel of the bayer matrix regardless of the color and base my exposure time on this. I got caught the first time by doing Ha that falls only in the red, then checked the levels on an area of pixels rather than individual so although my average ADU was well within 9000 ADU I still had pixels well over 30000 ADU in the reds. 
Talk about inverted vignetting as soon as I debayered 
