AstroChris - Processing Tutorial using M57 with the MX7C


Step By Step Processing of M57
For this M57 image, I took 12 images 600 seconds hires self guide mode, using guide exposures of 2seconds binned 2x2. It was a windy night, and after looking through the images, only 6 were usable as the others suffered from the dreaded LX200 backlash and had elongated stars. I took 12 dark frames from a couple of nights before that were 600 second exposures, hires self guided (without the telescope connected) with 2.1 second guide exposures in 2x2 binning. I was trying an experiment to have slightly shorter guiding exposures on the actual images so the dark frames would match the amp glow, but as you see further on it didn’t work, so I am now back to taking identical exposures. They can still be used by manually adjusting them (scaling them). I took 24 flat frames at dusk, with the LX200 in Land mode using exposures from 2 seconds to 20 seconds (keeping the background fairly constant by increasing the exposure time as I went) using hires interlaced mode. I then also took 24 dark frames for the flats in the same mode, with exposures of 6 seconds. Scaling the darks to remove amp glow First I took all these images and went to Tools / Preprocessing and put all the images, dark frames, flats and flat darks into their boxes. On the options tab, I had everything on average, automatic alignment unchecked and also confirm each image unchecked. This creates an image from which you can check out your amp glow. Split the created image into its Luminance part using whichever plugin you use (I use AA2 with ‘ MX5/7 Colour Synthesis v1.2 Steve Hill’). Then with that luminance image look at the amp glow on the left hand side using the black and white sliders, if it is darker than the rest of the background in the image, then the dark frames need reducing slightly, and if it is lighter, then the darks need increasing. Our background for this M57 image is about 100 counts darker on the left hand side, than it is for the rest of the image, use the magnifying tool and it gives a background reading on the status bar of the local background value. Save the dark frame that was created, then click Arithmetic / Coefficient / Multiply and enter 0.95 – this will take scale the image to 95% and save it as something else. Close all your images. Goto Tools / Preprocessing and remove all the dark frames from the Dark Frames box and replace them with your scaled dark, now run the process again and you will now have an image without amp glow. You may have to do this process a few times to get it right by trial and error, that is why we kept the original saved. Now we have a nice averaged dark frame to use. Calibration Create a master dark for you flat frames – Tools / Preprocessing and put all the flat darks into the image box, save this as your darkflatmaster. Next subtract your masterdark from the images and same them in a different folder. (I do this in AIP4WIN, but in AstroArt you can go to Tools / Preprocessing to do it and select keep each image, the you have to save each one individually). Colour synthesise your darked images using the plugin that you use, and batch convert all the frames into their LRG & B channels. Take all your flat frames and combine them, if you can by normalised median. This mean you have to make all the backgrounds equal. I use either AIP4WIN or if I need more control to get rid of stars in the flats, then I use Sigma-Beta 10. In AstroArt you need to open all the flats and then click on each one and normalise to the first one, then save each and then combine on median (Tools / Preprocessing). Now subtract the flatmasterdark from this. Split this into its RGB & L components and save. On the Luminance channel, run an AntiVBE filter to minimise the vertical blind effect. Close all images so you have a clean workspace. Open any luminance image, put a square around a medium brightness star, one where no other stars are nearby. This is our alignment star. Close the image. Open Tools / Preprocessing and put all you L images into the image box and your L flat into the flat box. On the options tab select auto alignment an it will automatically put in the box you set earlier. Set the images to average and deselect confirm each image. Click OK and this will create your luminance frame, apply the AntiVBE filter to this and save. Do the same with the RG & B’s using their own flat frames. You will now have a set of calibrated channels. Take each channel and put them in the images box (Tools / Proeprocessing) on autoalignment, but selecting ‘Keep Each Image’. This will align the images together, save them. Flatten Each Channel Depending, on the image and object, you may want to flatten it by using the artificial flat plugin. I do this with each channel image, gaussian blurring the resulting flat sigma 15, before dividing the channel image with it. This now creates a set of flat aligned images. These are the masters. Process the Luminance When I present this M57 image, I want it to show the nebula in all its beauty, I also want to show the very faint stars (down to mag20) and I would also like to show the little spiral galaxy as best I can. To do this I want to run a deconvolution on the nebula, a star bloat filter on the stars and because the signal to noise of the tiny galaxy is so small, I don’t want any filters applied to that. So I will make 3 images and combine them later. We already have the luminance with nothing applied to it, so we just need to manually fix any dark spots, raise the white point using the slider so you can see all of the culprits, click on the image to put a small cross on it where there is a dark spot and hit Ctrl x, you can then manually paint in the offending pixels using copying a nearby pixel, do this until you have no more dark spots, save this as you lum image. Now run a star bloat filter aggressiveness 3, background 500 and save the image lumsb. Back to your lum image, run a Lucy-Richardson deconvolution sigma 0.9, 20 iterations. Normalise the Colour Take each colour channel and normalise them to the one with the lowest background that way we have a good place to start when colour combing, save the images. Note at this point we go over to Photoshop, so if you can’t open .fits images in Photoshop, then save the 3 colours and all 3 luminances as 16 bit TIFF’s. Photoshop 7 In Photoshop there is a very powerful processing step called levels and curves. These are histogram adjustments. To understand a histogram, open one of your luminance images. Press Ctrl L, and the levels dialogue box will open, there is a graph which shows all the brightness levels in your image, the graph shows the number of pixels in the image with same brightness level. On the left you can see the highest point of the graph, this means there are more pixels with this value than any other in the image – this must be the background. We can ‘squeeze’ the image and show what levels does, use the black slider just under the graph and bring it up to just before the background, this sets all pixels below that level to black, then using the white slider at the other end of the histogram, pull it right down towards the black slider – you can see the image changing, everything above the white slider becomes white, and all the levels in between are adjusted to fit in between theses level linearly. Cancel the levels dialogue box and press Ctrl M, this opens the curves dialogue box. You will see a box with a diagonal line going down through it, make sure that the black is on the left and white on the right, if it is not then click the little arrows on the black / white bar at the bottom. What curves do is a non linear stretch, this means we can brighten a particular part of the image without brightening the rest. The left hand side of the curve (diagonal line) are your darker areas in the image and the right is the brightest areas, just like a histogram. If you click above the curve towards the left to create a curve sloping upwards then you will brighten the darker parts of the image without over brightening the bright parts. You can even increase the background to show the dim areas and lesser the bright areas so as not to saturate nebulae etc. We want to apply curve to brighten the dim parts of the image while not saturating the stars. Create a curve like this and save it as M57 Tutorial. So now we know how levels and curves work. Open your image lumsb (the luminance image that went through the star boat filter). This is going to be our base luminance image. Adjust the levels by moving the black slider about 15 before the background (numbers are next to ‘Input Levels’, with or M57 image the input levels should read about 20. Now adjust the curves by applying our saved curve, then apply another curve the same. The background will be getting brighter now, so again adjust the levels. Keep doing this procedure until the noise in the background of the image becomes objectionable. At the end the background should have a value of no less than15 on the levels graph. IMPORTANT never adjust the back point over the background level of the images, by doing so you are deleting all the dim areas of the image and creating an unrealistic image with a jet black background – the sky is not black, it is just very dark. When you have applied all your levels and curves, you should have an image that is just starting to get grainy (noisy) with a background that is a little too bright. Change the image to 8 bit – (Image / Mode / 8 Bit/Channel). Now open your lum image (the image without any processing). Again apply levels and curves as before, but this time, don’t be as aggressive, what we are looking for is a nice smooth background. Make sure that the background is much lighter than the previous image, if you background is 15 on the previous image, then leave this one at 20. Convert the image to 8 bit – (Image / Mode / 8 Bit/Channel). Now you should have a noisy luminance showing the very faint stuff, and a smooth image that doesn’t go as deep. Click on the noisy image and click Ctrl A (select all) and then Ctrl C (copy), then select the smooth image and click Ctrl V (paste). You have now layered one image on top of the other. Set the blending to lighten (show the lightest pixels from both images). Now select the Background from the layers palette and adjust levels - the black point slider until your image shows the faint portions of the image from the noisy image, but the smooth background from the smooth image – enlarge the image to check out the results. When you are happy, merge the layers – Layer / Flatten Image. Enhancing the Nebula Open the deconvolved luminance image, this is going to be used as the nebula. Apply levels and curves until the nebulas has the desired brightness and contrast – don’t worry about the stars or the background, when you are there, duplicate the image. Now apply an aggressive unsharp mask filter to the image – Filter / Sharpen / Unsharp Mask. Use 500%, Radius 1.0, Threshold 0. This is very aggressive and will leave dark rings around the stars in the nebula. Convert the image to 8 bit. Also convert the duplicated image to 8 bit. Copy the sharpened image and paste it into the unsharpened image (Ctrl A select all, Ctrl C copy, then in the unsharpened image paste Ctrl V). Set the blend mode to lighten and this will have got rid of the dark halos around the stars. Flatten the image. Blending the Nebula into the Image Now we need to blend the sharp nebula image into our background image. Copy the nebula image and paste it on top of the background image. Enlarge the image so the nebula almost fills your screen, then carefully with the polygonal lasso, trace an outline around the nebula. With the nebula nicely selected, feather the edge of the section by 6 pixels – Select / Feather. Now we want to delete the rest of the layer so the underneath layer can show through so Select / Inverse, then hit delete. If you are happy with the result, flatten the image. Selective Blurring Photoshop has a tool that no other processing software I have coma across has. It allows you to select and then blur or sharpen parts of an image without affecting the rest. We don’t need to sharpen the image further, but the background may benefit from a little blurring to reduce any noise present. But we don’t want to blur any stars or nebula (or even the little galaxy). Select / Colour Range – this open a box, set the fuzziness to 20 and click with the cursor on the darkest point of your image, now holding shift down click on other parts of your background being careful not to click on anything bright. As you do this you will see in the box the selection that you are making, and once you have all the background and no stars selected click OK. The selection will be right at the edge of the stars so in order to not blur the edge when we perform the blur, we want to contract and feather the selection so it blends nicely. Select / Modify / Contract by 5 pixels and then Select / Feather by 3 pixels. Now we have just the background selected we can perform a gaussian blur, Filter / Blur / Gaussian Blur and set it to about 0.8 to perform a mild blur. Control D deselects everything and you can now save your final image. Note that the background is still a little bright, but we will deal with that in the final touch ups. Colour Combining Open the three colour channels. Change each one to 8 bit. Now on the layers palette select the channels tab. Click the little arrow on the right hand side of the palette and select from the menu Merge Channels. Select RGB Colour and 3 channels, then specify the RG & B channels and click OK. A colour image will be created. Open (if it is not already open) your final luminance image. Change the mode to RGB Colour – Image / Mode / RGB Colour. Now copy the colour image and paste it on top of your luminance image. Set the blending option to Colour. A colour image will be created, but it will be very pale. Select the colour layer and increase the saturation – Image / Adjustments / Hue/Saturation Set it to about +85. This will show a nice colour on your nebula, but the stars will probably not show much colour. Now using the polygonal lasso, select around the nebula being careful not to actually select any part of the nebula. Now Select / Inverse so we have everything in the image selected except for the nebula and increase the saturation to +75. This will now show the star colour well. For final touch ups we want to darken the background (or increase the contrast), resize the image and crop the edges. Select the background layer, and either perform an ‘S’ shaped curve with the background lowered slightly and the bright portions increased slightly. Or select the background and set the black point to darken the background. Whichever gives you the most pleasing result. Now resize the image to eliminate the rectangular pixels – resize to 779 x 580. I then crop 5 pixels from the edges because of processing artefacts. Image / Canvas Size 769 x 570. And there we have the final image, flatten it and save.

Step By Step Processing of M57