This is my first HaRGB shot. It’s a combination of grayscale luminance data that I captured while camping at Coyote Lake, CA over the weekend, and color data that I captured from the backyard on Tuesday night. Looking back, I probably should have done it the other way around. Oh well.

Also, while camping, I was using the Canon 300D DSLR, but the AC adapter for the 350D showed up Tuesday, so I switched to the 350D for the RGB data.

Because the 300D and 350D have different pixel counts (6mpx vs 8mpx) but the same size CCD (APS-size), the 350D magnifies the image slightly given the same optics. All this is to say that I had to do some scaling and rotating in Photoshop to get the two images to line up. It’s enough to make a person want a filter wheel.

The Ha data is 5 full hours, fifteen 20m subframes. That’s close to a record for my deepest exposure yet.
The RGB data is a little over 2 hours, twenty-six 5m subframes.

I’m not sure how to calculate a “total” exposure time off of that. shrug.

This is a full-frame crop of the original (as much of the full frame as was still useful after rotating)

It was cloudy on Saturday night, but Sunday night I decided to try out another run at the Cygnus Milky Way.

I have a decently large target list of nebulae that are in the area, but actually a lot of them are in Cepheus (which is not really up far enough to shoot right now), so I thrashed around a little before settling on the nebular complex that includes Sharpless 115, 116, and 112. These 3 nebulae lie near Deneb, on the other side of the star from NGC7000 (The North America nebula).

Like I’ve done for the past few photos, I started the images as soon as I could after dark, then left the mount running all night. By sunup, I had captured 16 frames of 20 minutes each (that’s 5h 20m). Here is the result:

A reminder that the reason these latest images are in B&W is because I’m shooting through a narrowband (13nm passband) Hydrogen-alpha filter. So my full-color images coming out of the DSLR only have data in the “red” channel of the image (the “green” and “blue” channels are essentially empty of useful data). I can go back and re-shoot these objects unfiltered, then use that RGB data to turn these into color images. But for now, all I have is a red channel, so the images appear as levels of grey representing the amount of red data in the image.

This time I tried 20m subframes (polar alignment not good enough — much field rotation) and settled on 10m subframes (lots of detail but no star trails). I forgot to change the number of subframes when I switched, so it came out to 18 frames total, 3 hours of data.

I could have easily done 2 more hours of data and chased the sunrise. Oh well.

3 hours is a lot of data for one object. The result has a lot of depth of nebulosity and a lot of intricate detail, considering the huge image scale. I am going to have a lot of fun with narrowband this summer.

As well as the DSLR is doing so far, there’s something to be said for a dedicated astrocam for narrowband. With the same scope and exposure, I am getting a noisier image than people who are shooting with astrocams. This will not stop me from continuing, of course.

Here is NGC7000, The North America Nebula. 3 hours of data (18×10m).

While I was testing out the Ha filter last night, I decided to test out a theory.

Like most astrophotographers, I take lots of short subframes of my subject, then stack them together into a longer “integration time” image.

I hardly ever get to do side-by-side comparisons of these image stacks against single subframes of longer length.

But, I’d done a run of fifteen 5-minute images, and was finishing up, so I decided to run a single 20-minute image to compare them.

Here’s what I got:

Note that you can see lots of hot pixels in the single frame. But, barring those, the longer subframe captured at least 1 magnitude more stars, but the nebulosity is noisier and less detailed. The stacked image has fewer stars (due to the shorter subframes), but the nebulosity looks smoother.

Not a definitive test, but certainly food for thought.

Also, 20min subframes of this object are definitely desirable.

http://docs.sdaa.org/AISIG/CCD_Imaging_Primer.pdf

lots of basic information about equipment and software selection, and a really nice section on image processing.

Here’s last night’s result. Not too bad.

I really like NGC 4725. It’s one of those galaxies that I never hear about, but it’s very photogenic and has a really cute companion nearby. There is another “buddy” in the upper left corner, too.

I need to ply my trade in Leo some more (4725 is up near the Dipper), but a lot of the galaxies in Leo/Virgo are small. I don’t think I’ve done a good M65/66 yet, though, so maybe that’s next.

I’m also noticing that galaxies which are dimmer than mag 9 are a little disappointing.

Processing notes:
I did remember to use Median stacking this time (which tightens up the stars).

I also tried out the “quality” filter (a new thing in Max 5 I think), using “roundness” of “0.2” (the default for roundness), which ate about 5 of the later images, which is consistent with I thought I noticed later ones with tracking problems.

There’s something wrong with the flats. I turned off boxcar this time, but there is still a bad color gradient from center to edge. I’ll reshoot the flats and see if I can fix it. The good thing is, the dust bunnies seem to have stayed put; all of these calibration frames were from almost a month ago, and in particular the flats are still flattening the image very well, if throwing the color off a bit. I could probably do some tweaking to the master flat (ie throw it into PS and grayscale it) to make it work a little better. But I think I’m just going to shoot more flats and see what happens.

The 900×600 image is my normal treatment; I select a 1500×1000 section of the image that contains my object, do all the processing on that, and then resize down to 900×600 to tighten everything up once I’m finished.

The FOV in this image is around 40′x27′.

Doug (author of MaxIM DL) suggests shooting everything RAW Monochrome. This has never worked correctly for me, but whatever.

IMPORTANT POINT: If (and only if) flats are shot in RAW Monochrome, then “boxcar” filter needs to be turned on.

Since I shoot in RAW Color, *no* boxcar filter is needed for flats.

I finally knuckled under to the pressure of the dust bunnies and decided to shoot “flat field” frames with my astrophotos. A “flat” is a photo of an evenly illuminated surface. Since there’s no “subject” data in the frame, what the flat shows is any darkening aberrations in the optics, like vignetting (darkening at the corners), dust (all those lovely dark donuts), etc.

The problem is, how do you produce an evenly-illuminated surface? This can be a pain in the neck, so most people, myself included, skip flat calibration in their photography. But my photography is starting to get to a point where I think I could get better results if I started with cleaner data. So, in search of flats I went.

There are many methods to produce flats (shooting the sky at twilight, shooting the inside of the observatory dome, creating a lightbox), but the concept is the same: find a way to back-light some kind of diffuser.

So here’s my version. I made it oversized (the box is 22” square by 8” deep, with a 16” aperture) so that I can mount it on the wall of the observatory and in theory use it to do flats for any of my telescopes, including Trixie (the 13”).

The project was a fun build; it was equal parts construction, papercraft, and electronics. Cost about $15 at OfficeMax and RadioShack. The resulting box is very light; I imagine it weighs less than 2 pounds.

I used 6 bright “white” LEDs for the light source (and I added a dimmer switch, ooo…), which look yellow when I shot a photo in a darkened room, but look very very blue when shot through the telescope. I’ll have to work on the blueness; either I’m going to end up adding red and green LEDs to the mix, or I’m going to figure out how to deal with the blue mathematically (it’s all about the grayscale (: ).

The photos are fairly self-explanatory, except for the blueness and for the very uneven lighting in the telescopic image. From my extensive study of lunar eclipses, I think that the upper LH corner is dark
because of actual vignetting in the optics, but that the upper RH corner is dark because the lightbox was not centered over the telescope. It was a quick test, after all.

Anyway, now shooting lights should be as dead-simple as shooting darks and bias frames — the lightbox is (or will be) set on the wall of the observatory such that when the scope is parked, it points at the

lightbox, and voila! Easy flats.

I’m not surprised that it took this long to “get around to” building a lightbox. But I am definitely ready for those dust bunnies to disappear…

Keep looking, um, up?
Jimbo

PS: in case you’re really interested, the way that image calibration works is that you do this:

calibrated image = (original – dark) / flat

The dark (an exposure the same length as the original, but with the lens cap on) is heat-based data that increases linearly over time, so subtracting it is easy.
The flat (vignetting and dust bunnies) needs to be divided out; think of it as “pixel without a dust bunny has stronger response, so you’re averaging the response of all the pixels”.
Once you’ve calibrated the image with flats and darks, all that should be left in the image is what was actually in the sky.

Art Morton wrote:
> You hit the cruxes of the matter, the secondary size. The other thing
> I remember battling was f/ and being able to see the secondary when
> focusing at low f/s.
>
> Sounds to me you have that all work out, and with a little more
> testing you will have it solved.

Either solved, or unsolvable with a workaround. Either way is OK with
me. I like shooting flats with the lightbox. Rain, rain, go away. Or not. (:

J
> On Feb 24, 2009, at 11:24 PM, Jimbo S. Harris wrote:
>
>>
>> On Feb 24, 2009, at 4:16 PM, Art Morton wrote:
>>
>>

>>> Refectories and Maks are notorious for hot spots as the f/ goes down
>>> to 3 or so…..
>>
>> My particular Newt came with a 2.1” secondary (or was it .1.8”?) and
>> I replaced it with a larger secondary (the size usually found in 12”

>> scopes), which was the size that NEWT suggested would be large enough
>> to evenly illuminate a field at least 30mm across.
>>
>> So I have it more or less narrowed down to 3 culprits:
>> – The focuser drawtube (2” is sufficient, but a longer drawtube means
>> that the light cone is chopped to 2” further from the focal plane…)

>> – The MPCC (my most likely culprit, but my biggest fear)
>> – the T-ring (only 1.45” in diameter, but very close (within 2”) of
>> the focal plane, so possible but unlikely)

A,

Just ordered a 3/4” travel drawtube from Moonlite. I have to wait for
their next CNC run to get it, but he’s only charging me $50 for it. I
should be able to eliminate the focuser as a culprit in vignetting.

I’m pretty sure it’s the MPCC. But I’d rather do flats than take the

MPCC out of the image train. (:

J