M51 WIP Odd Colours from OSC and Noise Levels

[ianaiken]

Hey all,

I've been stacking some data I captured this week of M51. It was clear but I think there was a high level of haze.

I have two questions where I am looking for some input.

1. Is the colour on this image normal for an OSC, I seem to have blue to the left, green in the middle and red to the right. If so, what is the best way to process it out? Or, is it caused by stretching the image too much?

2. There is a lot of noise in the image. Is this caused by the haze or the fact that I put my gain up to 14. I've compared to an image done at gain 10 and at 14 there is a lot more noise, but I am not sure if this is due to the sky conditions or not.

Camera is a QHY8L, 27 subs at 600s. Full-size jpg here http://dl.dropbox.com/u/1595099/M51_Master_DBE.jpg

Thanks for your input. If anyone wants to have a look at the FIT file then here ya go http://dl.dropbox.com/u/1595099/M51_Master_DBE_SGL.fit (should be there after 21:30 tonight)

Many thanks for any input.

[samtheeagle]

I'm afraid I cannot help you Ian, but your issue looks a lot like the one I've just posted a thread about. I've been testing the AstroArt 5 demo, and I'm getting this same kind of strange colour problem with DSLR data. Hopefully someone will be along to enlighten us.

What software are you using?

[ianaiken]

I tried with AA5 and Pixinsight to calibrate and stack and same result on both. Did you use flats? I'm wondering if its flat related, may have another go without flats.

[samtheeagle]

I've tried all combinations of stacking with and without the various calibration frames, but no joy. I'm getting the same with nothing but the lights. There must be something I/we are doing that's not quite right, a lot of people use these apps, so it must be possible!

[harry page]

Hi

I get worse gradients than that A little DBE will sort it ouy

Harry

[ianaiken]

I have a lot to learn my master

I tried DBE and saw no difference. I will revisit I probably forgot to subtract the background or something equally as stupid.

Many many thanks

[ollypenrice]

My own OSC camera, an Atik 4000, also produces a slight colour gradient with one side a bit red and the other a bit green, the division between the two being a diagonal. It isn't all that strong but matters in deep stretches and, as Harry says, DBE is the answer.

If the gradient is strong you may need to raise the 'Tolerance' from 0.5 to 1.0 or more. If you don't do this, many of the markers you put on won't 'take' and will turn red (I think it is) to show that they are above the chosen tolerance and will be ignored. I find that a higher Tolerance does not seem to damage much of what is real faint signal, though that is the danger. Also beware of markers generated at the very edge since edges are often not typical of the rest of the image. I find it best to crop them off before using DBE.

Olly

[ianaiken]

Olly thank you. I see two improvements I can make, first is to crop the image and then secondly adjust the tolerances. Looking forward to trying this later today.

As for the noise do you think this is typical of high gain or haze? Or maybe both?

[ianaiken]

Right, well after restacking in AA5 in a two part process and some processing in PI here is the image now. SO far only 4.5 hours of data, and to be honest the FWHM is quite high due to a high cloud haze at time of imaging. I think this is reflected in the noise in the image, or my gain is too high (which I don't think it is).

Image Details: 27x600s @ f7.65

Full size viewable at AstroBin

Thanks for looking and any feedback you have.

[samtheeagle]

Looking good ian I must return to this target myself, I've had one crack at it with my old achro, but I've yet to turn the ED80 to it.

[ianaiken]

Thanks I will be spending a lot more time with it, and maybe will also get some Ha before I'm done.

[ollypenrice]

You've got the background much flatter but clipped the black point, I think. I wouldn't feel you have to get a black sky and a bit of colour and noise there looks natural enough.

Olly

[ianaiken]

Olly - thanks for the feedback. I always seem to clip the black point, not intentional, it just seems to happen somehow. I'll revisit, I probably clipped it on Hist Transform. I was racing to get an end result after stacking so easily in AA5.

I'll also get some more data for my next version.

[ollypenrice]

Olly - thanks for the feedback. I always seem to clip the black point, not intentional, it just seems to happen somehow. I'll revisit, I probably clipped it on Hist Transform. I was racing to get an end result after stacking so easily in AA5.

I'll also get some more data for my next version.

I still do just the same, every time, and have to go back! Beware that switching from 16 bit Adobe RGB (best for processing) to sRGB 8 bit can take a bite out of the black point and my web gallery site often takes a second bite!!

Olly

[ianaiken]

Olly,

How do you see that the black point is clipped? Are you checking out the histogram?

I saved this as a TIF in PI to open in Adobe PS to then save as JPG per dave_galera saving advice. Is this the best way to do that?

Here is another reprocess. I've increased saturation and reprocessed histogram to avoid clipping on the shadows and I have also ran through HDR transform again. I seem to have my red gradient back in the background.

[ianaiken]

And now I think I have removed the red with a mask and SCNR. Maybe it needs background neutralization instead.

Full size is here M51 WIP

[ollypenrice]

I follow Dave's advice as well!

I didn't check the histo, I just looked, so it may not be gospel by any means.

That slight background gradient ought to go in DBE if you raise the 'Tolerance' value a little. It is quick to try a value, increase it, try again, etc. If, in DBE, you place a marker on a point and it won't 'take' (I think they turn red when you try the next one?) then you need to up the tolerance.

Olly

[ianaiken]

Thanks Olly, greatly appreciated.