Friday, November 24, 2017

#119 - Wednesday, November 22, 2017 - Filter Phenomena

Happy Thanksgiving!
I got an early start out at the observatory on "Thanksgiving Eve," arriving out there around 5:45 PM, about a half hour after sunset.  I ate a Jimmy John's sandwich for dinner on the drive, and after turning on the heater in the warm room, I hustled over to the mini-dome where the memorial scope is kept to start getting set up.  I needed to test whether I had finally properly cleaned by blue filter since I finally found my multi-coated optics & filters cleaning solution (exactly where I thought it was, not sure why I couldn't find it before).  However, I realized on the drive there that I had forgotten the USB cable for the camera, since I transferred it from its normal camera bag to my DSLR camera bag, since my DSLR is currently on its way back to Nikon for repair (see log entry #118 for what happened).  Fortunately, I found a spare in the warm room's computer desk.  I started unpacking...and realized I had left the filters at home in my other camera bag too!  So yes, I drove the 25 minutes back to my house to grab them.  (The club doesn't have any imaging filters, at least that I can find.)  So it was nearly 7 PM by the time I finally started imaging.

On my last trip out, I learned that the menu button on the memorial scope's hand controller (a Losmandy Gemini II mount) does't like the cold (I learned over the summer that it doesn't like the humidity either), but I also learned that I have full control over the scope from my tablet thanks to the Gemini Telescope app that comes with its ASCOM driver.  It even has a "perform meridian flip" button that I am super excited to try!  It also lets me use a wider catalog of stars for the initial sync step of the alignment process.  (I don't actually align the mount besides the one-star sync; I haven't had much luck with the alignment working properly, and since it's very well polar aligned, it's not terribly necessary.  I just pick a new sync star if I'm changing areas of the sky between targets. and the gotos are pretty close).  I synced it to Mirach, since my first target of the evening was getting blue data on M33 so that I can finally process it (I have L, R, and G data already from previous trips).  I couldn't see the diffraction spikes from my Bahtinov mask very well during focusing, however, which wasn't a good sign.  I got it close, and then slewed over to M33 to take some test images.  The stars looked better than they have previously (see this post for some background on the blue filter issues), but only in part of the image - they still had weird diffraction-y mess on stars on the other half, it looks like.
Ignore the high noise level; it's a screen grab of an unprocessed FITS file, with the gain adjusted arbitrarily.

So I think the blue filter is shot.  Luckily for me, I have a $125 gift certificate to Oceanside Photo and Telescope for the Astronomical League imaging award I won earlier this year, which I will spend on a set of Astronomik parfocal LRGB filters that I have already picked out.  Having parfocal filters will be nice anyway - with the ones I currently have, I have to slew to a nearby bright star, re-focus, slew back and re-center the target every time I change color channels.  The L filter cuts IR and UV, which don't focus perfectly and can leave your stars looking de-focused.  However, I only have a 3-position filter wheel at the moment, so the additional ease that that brings will have to wait.

Aside: Filters

As I was writing this and researching which LRGB filters I wanted, I decided to find out exactly what I already had - and it turns out the filters I have are not RGB at all, but UBVRI filters used for photometry!  I've been using three, the BVR ones - blue, visible, and red.  The visible filter appears green.  I have the I filter too, which is IR-pass, but haven't used it - I have read that using an IR-pass filter is your luminance channel for planetary imaging can make really sharp images, though, so I may have to try that sometime.  The U is ultraviolet, which I don't have.  I couldn't find much information out there on what exactly the difference is between the two types, but I think the main take-home is that UBVRI uses a standard, well-defined spectrum and is used for scientific purposes, while RGB filter spectra tend to vary by manufacturer.  Below are the two spectra.
Johnson-Cousins UBVRI filter spectrum

Astronomik LRGB Type 2c filter spectrum (the yellow line is the L, or luminance, filter)

A few observations are that the blue and green filters have greater overlap in the RGB spectrum than the UBVRI one; the green and red filters have greater overlap in the UBVRI spectrum than the RGB; and the red UBVRI filter extends much farther into the IR than the RGB set does.  I will note, however, that my red UBVRI filter still provided some very sharp images (sharper than the green and blue, in my opinion) despite passing a decent amount of IR, which in refractors usually has a slightly different focal point than the visible part of the spectrum because they are made for the visible spectrum.

Anyways, moving on.

Moving on: Back to imaging

Since my blue filter still wasn't working, I decided to go ahead and take some luminance data on a few things while I was out there and the skies were nice.  I used my Astronomik CLS filter (a nifty light pollution-blocking filter that does a very nice job, particularly with my DSLR images).  I've been wanting to image the Heart Nebula, IC 1805, so I slewed over to it.  However, it is a bit too large for the FOV (field-of-view) I have in the memorial scope; it's a good 1.8x2.6 degrees to get the whole thing, but I only have about 1.3x1 degree FOV with the SBIG ST-8300 on the memorial scope, which is a Vixen 140mm neo-achromat, f/5.7.  I decided to image just the heart of it (tee hee).  It's in Cassiopeia, so it will be up for quite a while from my latitude.  

Unfortunately, it didn't turn out very well - I had some guiding problems, and it looks like I was ever-so-slightly out of focus.  For some reason, I couldn't get nice diffraction spikes to form on Mirach with the Bahtinov mask.
Date: 22 November 2017
Object: IC 1805 Heart Nebula (central region)
Camera: SBIG ST-8300M
Telescope: Vixen NA140ssf
Accessories: Astronomik CLS filter (2-inch)
Mount: Losmandy Gemini II
Guide scope: Celestron 102mm
Guide camera: QHY5
Subframes: 5x600s (50m) | Luminance Only
Darks: 10
Biases: 20
Flats: 0
Temperature: -25C (sensor), 25-29F (ambient)

Our local professional astrophotographer tells me this one is hard to get even under dark skies.  I'll give it another try with my new LRGB filters on my Borg when it warms up a bit, since it requires more setup (I have to bring my mount and put it together).  The Borg has a shorter focal length (500mm), and thus a wider FOV - 2x1.5 degrees with the SBIG.  We'll see.

After having all the guiding problems, I homed the telescope, power cycled it, and tried a different target - the Flame & Horsehead Nebulae, since Orion was finally up.  Alnitak formed very nice diffraction spikes with my Bahtinov mask, so I was able to get much better focus.  Bonus, I could see the Flame Nebula in 1-second long exposures in the Focus mode of the CCDOps software, which made framing the image much simpler.  (The CCD chip is by far and away more sensitive than my DSLR chip).  I did have to rotate the camera though to put the Flame and the Horsehead side-by-side on the longer axis of the sensor than the shorter one, otherwise you would have a tall and narrow image of this wide scene.  

Now, Alnitak, the left-mode star in Orion's Belt, is a very bright star - magnitude +4.2.  Actually, it's a double star, with its partner being +9.55.  They are only 2.2 arcseconds apart, which at its distance, corresponds with a physical separation of 574 AU (astronomical units - the distance between the Sun and the Earth).  It heavily saturates camera chips, particularly the sensitive CCD chip.  Now, if I was smarter, I would have taken some shorter-exposure frames too, but it wouldn't have mattered because some high clouds rolled in anyway, and there may have also been frost on the scope too (I didn't feel like getting a stepstool from inside to check).  But the shorter-exposure frame would have let me done a better job at processing out the super-saturation of that star.  

Let's make a long post even longer, and I'll walk you through the processing process.

Stretching the Histogram

I haven't really addressed this yet, but I have recently made the switch from doing stretching and initial adjustments in DeepSkyStacker to stretching the histogram in Photoshop instead.  Everywhere I read online, astrophotographers say to avoid doing this in DSS and do it in Photoshop instead.  I finally gave it a try, and I did notice that the noise was much lower by doing it in Photoshop.  Everyone also says that everyone has a different way of doing it, so here's a quick-and-dirty I've seen several people do.

First, open Levels (Ctrl + L).  Move the left slider (black) to the foot of the histogram (but don't cut into the main part of the histogram itself), and the middle slider (gray) to the right end of the main chunk of the histogram.  Don't adjust the right (white) slider, as you will lose star data.  Hit OK, and repeat the process a few times.
For 16-bit images (Photoshop apparently can't do 32-bit), the brightness scale of a given color channel (in this case, since I have monochrome images, there's only one channel) ranges from 0 to 255.  With a raw image fresh from DSS (choose "Embed" instead of "Apply" when you save out the image), all of the data is contained in the narrow spike you see in the above histogram.  You want to "stretch" this data out so it covers the whole 0-255 scale, basically.
Stretch #2.

Stretch #3.

Once you've done a few rounds of this, open the Curves dialog (Ctrl + M).  

Now, if you're used to editing the curves of a color image, you will notice that this one looks a bit different - it's backwards!  I don't entirely understand why Photoshop does this for monochrome images, and it feels like I'm editing the image while looking in a mirror, but you can accomplish the same objectives.  Adjust the curve to your liking.  I find that I get a nice result when I select a black point (the leftmost dropper) in a dark area of the image, and then pull the middle-left end of the curve down to increase the brightness of the dim areas of the DSO.

Dealing with Blown-Out Stars

I followed Dave Rankin's video on how to bring out the Trapezium region of M42, the Orion Nebula, for how to do this.  Basically, what you're going to do is stretch the histograms for both the long exposure (bright) image and the short exposure (dim) image, make them both layers of the same image, and use a layer mask to combine the two.  It works surprisingly well for how easy it is.  I tried it once on some Orion Nebula data I had as well.  I need to use it more often to deal with bright stars.

In my case, since I don't have short exposure data, duplicated the layer (in the Layers box in the lower right-hand corner, I right-clicked the Background layer, clicked Duplicate Layer, and named it "star"), and then used Camera Raw Filter on the "star" layer to reduce the exposure to make Alnitak dimmer.  Normally, instead, you could copy & paste your stretched and adjusted short-exposure image on top of your long-exposure image.  You can change a layer's name by double-clicking on the name in the Layers section.

Next, re-arrange the layers so that the short-exposure one is on the bottom.  (Just click and drag the layer - I had to unlock the background layer by double-clicking the lock symbol in order to do it).  After that, click on the Background layer (your long-exposure image) and click the "add layer mask" button, which looks like a white box with a dark circle in the middle (it's at the bottom of the Layers section).
Now, with the "Background" (long-exposure) layer highlighted, click the Brush tool over on the left-hand side, and make sure the color is black.  "Paint" the area around the blown-out star to expose the short-exposure, less-blown-out image underneath.  Don't go too far into your DSO, or you'll lose data there.
Obviously, that looks ridiculous.  But we're not done yet!  Go to Filters->Blur->Gaussian Blur, and turn it up pretty high - mine was 182 pixels radius.  Click OK.
Date: 22 November 2017
Object: Flame & Horsehead Nebulae
Camera: SBIG ST-8300M
Telescope: Vixen NA140ssf
Accessories: Astronomik CLS filter (2-inch)
Mount: Losmandy Gemini II
Guide scope: Celestron 102mm
Guide camera: QHY5
Subframes: 4x300s (20m) | Luminance Only
Darks: 23
Biases: 20
Flats: 0
Temperature: -25C (sensor), 23-25F (ambient)

All right, so it kind of helped.  I think there was additional haze from frost or something on the telescope objective - I should have checked for it.  Oh well.  I'll bring my dew heaters next time out.  Now flatten the layers (Layer->Flatten Image), and continue editing - noise reduction, Carboni tools, whatever else you need.

I'll try it again soon!

Now, for the rest of the post

It was cold!  As low as 23F.  Guiding worked better after I re-started PHD2, however, and I didn't have to go check on things that often, so I was able to warm my fingers and toes in the warm room.  A bit of a disappointing night, but I'm starting to get my workflow down on imaging with the CCD camera.  Woot!

No comments:

Post a Comment