I've only just started to undertake full image calibration so not got a great deal of experience. The question I have relates to getting the best possible flat frame from my scope/lightbox combination. What I've noticed is that flats are taken with the lightbox in different rotational positions, the cancellation against the flat frame taken in the original position is not as good (a spread of several histogram bins) compared to virtually 0 if calibrated against itself. I'm guessing what I'm seeing is the effect of imperfections in the light box itself. This would suggest it may be better to average the flats over say four rotational positions of the lightbox to average out any illumination errors? Wondered if anyone else using lightboxes had come across this and also how they'd dealt with the problem?

Thanks, Jon.

---

8" Mead LX-90 + DSI III
Isle of Wight

Hi, Jon

I don't use a lightbox, but I can tell you that if the illumination is uneven, you will likely introduce as many problems as you solve by using flats. By all means, try rotating the box. If that fails, work on modifying the design to correct the illumination problem at the source.

Good luck!

Chuck

---

Chuck Reese
CDIP Certified: Nebula, Galaxy, Solar System, Stars & Clusters Imager
CDIP Mentor: Nebulae & Solar System Categories
For information on the Certified DSI Imager Program, see:
http://autostarsuite.net/forums/1427/ShowForum.aspx

Hi Chuck,

Thanks for the reply... I did these tests just to see how even the illumination was. I might have a look and see if there is anything I can do to improve the light distribution, but this is not easy as the differences are only just discernable by blinking the flat images... However the worst difference I've seen is about 400 ADUs edge to edge in an approximately gaussian shaped histogram. I've got an example posted from re-processed earlier work on M-57:

http://autostarsuite.net/photos/jonw/category1343/picture16963.aspx

If you overstretch the histogram, you can see it's not perfect, but then what is perfection? Still, no dark skies, so nothing better to do at present!

Thanks again for your help,

Jon

---

8" Mead LX-90 + DSI III
Isle of Wight

In the end, only you can judge if the result is "better". There really is no such thing as "perfect", but it appears to me that your flats are over-correcting in the center of the image. For some targets, this will be a problem. I do believe you can do better if you can sort out the issues.

good luck!

Cuhck

---

Chuck Reese
CDIP Certified: Nebula, Galaxy, Solar System, Stars & Clusters Imager
CDIP Mentor: Nebulae & Solar System Categories
For information on the Certified DSI Imager Program, see:
http://autostarsuite.net/forums/1427/ShowForum.aspx

Thanks for the advice Chuck, this is greatly appreciated. The picture itself is not a particularly "tidy one" as the scope was not very well polar aligned in the first place (scope operated straight out of the garage with no additional alignment), but also had the same concern that the correction was not truely "flat" (hence this posting). The vignetting and coma with this particular arrangement are also too extreme for my liking, so intend to get a modified nosepiece for the DSI in order to move the FR closer to the chip and get something closer to F4/F5 rather than F2.2 that appears to be the case from measuring the star separations.

With no clear skies on the horizon, I've a got few weeks to sort out the light box. I think I'll endeavor to get some flats by other means and compare these to the light box too.

Thanks again Chuck!

Jon.

---

8" Mead LX-90 + DSI III
Isle of Wight

I've thought of a further problem... I've taken the flats with the DSI IIIc in monchrome mode, and processed the LRGB image files with the same flat. Thinking about the way the CCD is put together, the pixel sites are different for the RGB image components, so there will be a slight difference between each colour channel of the flat. Not had a chance to test this yet, and the scope has gone untouched for the last couple of weeks due to the weather, so will have to wait for the next imaging session to test this as I expect things will have changed significantly since the last session.

Again, any thoughts/suggestions gratefully received.

Thanks,

Jon.

---

8" Mead LX-90 + DSI III
Isle of Wight

Hi, Jon

Never tried that myself...I would not recommend it though :-)

If you want to minimize the workload for image calibration, you can save to raw format and use that to apply darks, flats, flatdarks, align, and combine. When done, you "split" the colors and work in RGB just as you would with a mono camera from there. (I used to do this with AIP4WIN)

Otherwise, you probably want to process each color channel as its own separate stack (I've done a lot of this too!). One thing to be aware of is that most artificial light sources are very deficient in blue. I think the Envisage histogram displays the green channel...your red flats will usually be somewhat more exposed than indicated, and your blue flats may be very underexposed than indicated. Just check the histograms in photoshop to find out.

Chuck

---

Chuck Reese
CDIP Certified: Nebula, Galaxy, Solar System, Stars & Clusters Imager
CDIP Mentor: Nebulae & Solar System Categories
For information on the Certified DSI Imager Program, see:
http://autostarsuite.net/forums/1427/ShowForum.aspx

Hi Chuck!

Well, I've just done a series of quick comparisons, and sure enough there is a significant difference between flats for the LRGB files, just as you would expect to see with a mono camera & filters. So, processing each colour channel separately for flats seems to be the best way to go. The colour channel difference is definately the most significant error in the previous image calibration (100's of ADUs), and not the lighting uniformity. Should have been obvious really, different colours= different focus and scattered field, let alone the pixel site differences. Got me thinking the "C" has probably not saved me that much effort processing wise in the long run...

Picking up on the colour front, I've used a white LED in the lightbox, so have the opposite problem, loads of blue and green and noticeably lower red output, but enough to get a good flat frame in each channel.

Thanks again for all your advice here, great help in getting to the root of the problem! Just could do with some clear skies here!

Jon

---

8" Mead LX-90 + DSI III
Isle of Wight

For info, I managed to get some information together following some clear skies (unfortunately with a near full moon). I've posted the final output here:

http://autostarsuite.net/photos/jonw/category1343/picture17043.aspx

Broke my usual rule of not changing more than one thing at time... changing the focal reducer position in the optical system, it's my first autoguided image, the stretch and sharpening processing (used DDP) and the flat frame processing. Anyway, the point of the post is that post processing as LRGB, with flats derived for each channel seems to work really well. Taking lightbox flats with the average of two lightbox orientations (0 and 180 degrees) results in a gaussian spread of about 20 ADUs when flat processing an average in either of the orientations. I tried to get the processing with the raw image data, but have found that neither AIP4Win or FITZ can read the DSI IIIc raw data in successfully.

Chuck, thanks especially to you for your help and advice... Can't wait for that elusive clear sky (with no moon!)

---

8" Mead LX-90 + DSI III
Isle of Wight

Fearless - I love it! :-)

Congratulations on a fine image, Jon. Perserverance does pay dividends in this hobby!

Chuck

---

Chuck Reese
CDIP Certified: Nebula, Galaxy, Solar System, Stars & Clusters Imager
CDIP Mentor: Nebulae & Solar System Categories
For information on the Certified DSI Imager Program, see:
http://autostarsuite.net/forums/1427/ShowForum.aspx