A common problem I am having with photo's from the camera is green fringing on stars. I can only presume it's a combination of false colour from the Megrez 90 refractor and incorrect colour balance ??? I is a little strange, as I don't get any false colour at all when using my Canon 6D on the Megrez 90 any thoughts?
38 x 300s
William Optics Megrez 90 + Borg DGL reducer (~500mm fl, cropped)
Still though, i'm reasonably happy with he amount of colour variation I have in the galaxy. More than I often achieve with my ST8 and lousy LRGB skills
I'm interested to see your results from the OSC as I've been experimenting with a Trius-C SX694. I too see a greenish fringe on stars even when the colour balance looks OK on the main object.
I wonder if it might have anything to do with the Bayer matrix where there are two greens for each R/B in the array of 4 pix? I know in CCDStack there is an option to use only one of the 2 greens when Debayering. I've not experimented enough with it to see if it might make a difference.
The other issue I'm curious about is your method of extracting the RGB frames. In CCDStack there is the choice that needs to be made to extract to half resolution, or to interpolate to full frame. Adam Block recommends to go the half size route however I found with my data that I did not get nearly as good a result compared to interpolating immediately to full frame.
Finally, I was surprised to see that you used relatively short subframe times. With my Trius mono camera 5 min is about all I can do before LP steps in too heavily. When I shot NGC253 with the Trius-C I went up to 12-15 minutes based on Terry Platt's recommendation. I found that these much longer times gave a better result on the colour camera given the reduced sensitivity of the CCD due to the matrix and the lower inherent resolution. I used the Trius-c colour data and combined it with hi resolution luminance. I'm pretty happy with the effort. The picture is posted on these pages recently.
Hi Roger,
I spent 5 minutes on it in Photoshop -
I hope you don't mind?
Does this look better?
cheers
Allan
Thanks Alan - I appreciate your attempt. The colour is certainly different and less green in general, however I still see the green fringing on one side of the stars. So, different colour balance but not necessarily solving the colouration I was hoping to.
Quote:
Originally Posted by PRejto
Hi Roger,
I'm interested to see your results from the OSC as I've been experimenting with a Trius-C SX694. I too see a greenish fringe on stars even when the colour balance looks OK on the main object.
I wonder if it might have anything to do with the Bayer matrix where there are two greens for each R/B in the array of 4 pix? I know in CCDStack there is an option to use only one of the 2 greens when Debayering. I've not experimented enough with it to see if it might make a difference.
Very interesting! I haven't played with the different de-bayering settings in CCDStack either. I will have to do that. I am using CCDStack for my processing, typical workflow is:
Reduce images in MaximDL
De-bayer in CCDStack
Register in CCDStack
Normalise in CCDStack
Data Rejection & Mean combine in CCDStack each of R, G, B
Create RGB colour image from previously stacked RGB files.
Post-processing in PixInsight and Photoshop
I find it frustrating that unlike DSLR images the CCD OSC images need to be split in to their component RGB channels, it introduces a significant amount of additional processing which makes it not very beneficial over a filtered mono system.
Quote:
Originally Posted by PRejto
The other issue I'm curious about is your method of extracting the RGB frames. In CCDStack there is the choice that needs to be made to extract to half resolution, or to interpolate to full frame. Adam Block recommends to go the half size route however I found with my data that I did not get nearly as good a result compared to interpolating immediately to full frame.
Curious. I have never noticed the half/full frame option at all! I can only presume I am using the full frame option, but can't check right now as my PC's HDD died last night
Quote:
Originally Posted by PRejto
Finally, I was surprised to see that you used relatively short subframe times. With my Trius mono camera 5 min is about all I can do before LP steps in too heavily. When I shot NGC253 with the Trius-C I went up to 12-15 minutes based on Terry Platt's recommendation. I found that these much longer times gave a better result on the colour camera given the reduced sensitivity of the CCD due to the matrix and the lower inherent resolution. I used the Trius-c colour data and combined it with hi resolution luminance. I'm pretty happy with the effort. The picture is posted on these pages recently.
Perhaps just comes down to laziness I don't know why I chose 180s in this case, but I do tend to stick to 300s or less so I don't need to bother with any autoguiding .. hence laziness 300s is more common as my choice.
I don't find that at 300s I am missing significant detail. I do need more subs than my mono and more sensitive ST8-XME to smooth out the noise, I find that the noise in the ST8300c can easily become a problem in processing the images, and lots of subs is required.
I wouldn't mind 600s exposures, but wouldn't bother going beyond that personally.
Thanks for your answers to my various questions. I've written to Terry Platt at SX about the green colour stars. I'll be interested to hear his response particularly to learn if this is just an anomaly we just happen to see or an actual "issue' with OSC.
Personally I don't think you absolutely need to treat the images any differently than a DSLR. In CCDStack you can play with the settings and immediately produce a colour image from each subframe. Just untick R/G/B but leave color ticked. Then select the correct Bayer pattern. Out comes the colour image. I think you can do that and stack the results. Whether that is the best way to treat the data is open to question. On the other hand it ought to be possible to take DSLR images and extract each color in CCDStack and do the individual stacking before generating colour.* The trick there is to type the word "Bayer" in the filter column of the image manager. Once you do that all the color controls become available. Adam Block said in doing individual stacks of each colour that it is essential that the same subframe be used for registration of each colour. That has worked well for me.
Peter
* Having not used a DSLR I'm now unsure about that remark! Does the camera immediately give you a colour image? If so it still might be possible to split the colours and then stack the individual subs and finally recombine. I have a feeling that this might give superior data rejection compared to rejecting RGB and then stacking.
Guys, I don’t know why you are doing these complicated operations of extracting colours and then re-combining them again. I really don’t see any advantage of processing already colour image that way.
Roger, you use PixInsight, so, calibrate frames, register frames and integrate them. Then equalise RGB values to 1 and extract L and work separately on L and colour. Then re-combine L+colour(RGB).
Personally I don't think you absolutely need to treat the images any differently than a DSLR. In CCDStack you can play with the settings and immediately produce a colour image from each subframe. Just untick R/G/B but leave color ticked. Then select the correct Bayer pattern. Out comes the colour image.
Interesting, I'll have to find what you mean there, look around the options.
Quote:
Originally Posted by IanP
Guys, I don’t know why you are doing these complicated operations of extracting colours and then re-combining them again. I really don’t see any advantage of processing already colour image that way.
Roger, you use PixInsight, so, calibrate frames, register frames and integrate them. Then equalise RGB values to 1 and extract L and work separately on L and colour. Then re-combine L+colour(RGB).
I'd love to not need to split the RGB channels. It seems silly to do so, but I found if I didn't I ended up with grayscale images, colour data lost I'll try in PI again, I'm not sure if I tried there originally or just CCDStack which is what I prefer using for .FIT
For sure if you just stack all the grey scale images from the camera the resulting image will not give colour. If you do the suggestion I said earlier in CCDStack you can easily convert all the Bayer masked images directly to colour and then stack those.
I'm betting that a guy like Adam Block would not say to separate the RGB frames and stack individually if there were not some significant advantage. It's a lot of extra work so I'm thinking there must be something gained. He does a tutorial as part of the CCDSoft tutorials just on this subject so I may watch it again and see if he spells out the reason.
When I used to use a one shot colour CCD (STL11 OSC) I used Images Plus but it did a lot of that stuff automatically.
It split the image into its component RGB and then processed them from there. It wasn't a lot of work at all. I did notice the stacking process that split the channels seemed to give the best results so I think there is a gain in doing all that extra processing. Whether CCDstack is the best software is probably more the question.
I found at that time CCDstack was not OSC friendly software. Perhaps that has changed but CCDstack seems to not handle colour images well in my experience. Try opening one of your Photoshop processed CCD colour images in CCDstack - yuck!
OK, I went back and had a listen to Adam's tutorial again. I'm now completely convinced that he is correct and that a better outcome is possible if RGB is separated and processed and only recombined to colour at the end. Here is a brief summary:
Processes like data rejection, registration, and normalization, take information from luminance data and then interpolate neighboring pixels. So consider these two situations. 1 Colour image with a red hot pix. How can one accurately interpolate from a neighboring pixel when in fact that pixel may have a different colour space? 2. Consider the same situation of a hot red pixel on just the red channel. One can now be 100% sure that the interpolation is correct because the neighbor is in fact red. The same argument can be made for registration and really the question must be asked about normalization too. What does it mean to normalize a colour image? The brightness values from image to image may very well be different, but also show different colours. By keeping the data separate and "pure" these difficulties go away.
All it does is remove the green elements from the image. You can read more about it via the link above and they have some example on their website. I have no affiliation with HLVG etc. I just love the tool and use it all the time with my SCT and Canon 600D
I can upload my result if you'd like? I tested it and it only took 10sec. The outcome was identical to your initial image, just without the green tinge
All it does is remove the green elements from the image. You can read more about it via the link above and they have some example on their website. I have no affiliation with HLVG etc. I just love the tool and use it all the time with my SCT and Canon 600D
I can upload my result if you'd like? I tested it and it only took 10sec. The outcome was identical to your initial image, just without the green tinge
PS. Your image looks great by the way. Well done!
Cheers, Evan
Feel free to share
Reading the plugin page it looks very suitable to my current problem .... I have never noticed CA on my Megrez 90 before but perhaps with a broader spectrum of sensitivity (compared to the Canon 6D) I am now detecting some.
Hi Roger - just wondering if you're seeing thin, high level cloud in there, as the focus looks a bit soft, and there's been an awful lot about recently.
I'd be surprised if it was chromatic aberration, as that would normally give you a blue halo - I think it would be tricky to focus red and blue, but not green?
The other theory about the extra amount of green due to the Bayer matrix configuration might be worth following up too. I would assume if the images were debayered then registered (which is how I do osc processing) that it's not subframe misalignment either.
Interesting.
Cheers,
Andrew.
I had a quick answer from Terry Platt at SX about the green star issue. What he said doesn't entirely make sense to me and I've written back to him for clarification. He essentially wrote that the CCD is most sensitive to red (I question this as I'm pretty sure it is most sensitive to green). He said that red is saturating first due to increased sensitivity and that one sees blue or green as they are not saturated. I cannot figure this out unless what he possibly meant to say is that green is saturating first and the R and B are not. To me this would create a green tinge. Am I missing something obvious?
Feel free to share
Reading the plugin page it looks very suitable to my current problem ....
Hi Roger, please see attached files (the HLVG processed image is on the right). In addition to the stars, you can see that it's also removed a slight green tinge from the galaxy.
I is a little strange, as I don't get any false colour at all when using my Canon 6D on the Megrez 90 
any thoughts? 
