Astrophotography and MFD

[theguywitha645d]

Ray had asked to see a results from the Pentax 645D for astrophotography. I am relatively new to astrophotography and I really enjoy wide-field work. This is a single 3-minute exposure at ISO 1600 and, if I remember correctly, 35mm lens at f/6.7. This was on an AstroTrac mount.

[ondebanks]

Thanks! I had asked you via PM, but it's even better that you shared it publically.

That's a great result for such a relatively small amount of signal integration!

Let's see...3 minutes at f6.7 is equivalent to 2 minutes at f5.6; 1 minute at f4; 30 seconds at f2.8.

I'm doing that exposure conversion because all of our digital backs/cameras can shoot 30 sec at f2.8, and in 30 seconds the sky turns little enough that with a wideangle or normal lens, you hardly need a tracking mount; a fixed tripod will do. Anyone can take such an image. I'd love to see more people share an image?

Ray

[theguywitha645d]

Ray, sorry. I was not sure I could send an image via PM. But I also thought there might be some other crazy folks doing astro with MDF. I can't believe there are only two of us trying this kind of work.

I have a question which you may know the answer. I occasionally get a sensor split with astro photos, but not always--a green/magenta split. And it is only apparent with a great contrast shift applied with a histogram or a curve. Usually, it is a star field with one side closer to sky glow from a neighboring city. Is the sensor just looking at each half differently and trying to WB balance? I am not sure this is something to worry about. I never see this in normal daylight conditions. Thoughts?

[ondebanks]

Is it a straight-line split, down or across the middle? That would indicate slightly different gain/LCC calibrations for different sides of a stitched chip (as in, the manufacturing process used stitching). There have been threads on this phenomenon, I believe, for other backs. IIRC, a fresh LCC image taken when shooting would correct it.

Or is it more wavy, like my attachment below? If I really level-torture a file with a fairly strong vignetting gradient from centre to edge (wide open lens) and low background counts, I can get colour banding from my DCS645M back. My educated guess is that this is because it produces a 12-bit output RAW file; 12 bits could be undersampling the noise and thus quantizing the gradient. 14 bits would prevent that (ha! see the "Why are only MF 16 bit?" thread!). That's given me an idea - I think I can simulate this pretty easily.

Ray

[theguywitha645d]

Is it a straight-line split, down or across the middle? That would indicate slightly different gain/LCC calibrations for different sides of a stitched chip (as in, the manufacturing process used stitching). There have been threads on this phenomenon, I believe, for other backs. IIRC, a fresh LCC image taken when shooting would correct it.

That is the one. The interesting thing is it is not consistent--I can take several different astro photos in one session, but it only shows in a few. I will have to look into that more.

[theguywitha645d]

Hey, your turn to post an image (without the rings).

[yaya]

That is the one. The interesting thing is it is not consistent--I can take several different astro photos in one session, but it only shows in a few. I will have to look into that more.

I believe the Pentax uses dual readout so each side of the sensor is being read separately and either the camera or the software puts the data together and normalises the exposure readings to get a uniform image.

It could be that consecutive long exposures and the accumulation of noise causes some slight errors in the process

[theguywitha645d]

I believe the Pentax uses dual readout so each side of the sensor is being read separately and either the camera or the software puts the data together and normalises the exposure readings to get a uniform image.

It could be that consecutive long exposures and the accumulation of noise causes some slight errors in the process

Yaya, thanks. That would seem to make sense.

[Brian Hirschfeld]

I'll try and throw a IQ180 image on, if I am up and about tonight..

[ondebanks]

Hey, your turn to post an image (without the rings).

Here we go. Using the wonderful removable IR filter feature of the DCS645M, to boost signal sensitivity (it now picks up more IR than visible light, so the colours tend to white).
Mamiya 645AFD, 200mm APO at f2.8, 30 seconds, untracked.
Comet Hartley is the greenish fuzzball left of centre (green from its Cyanogen emission).

As much as I love MFD, it really is not that well suited for astrophotography...I discovered that too late!  But now I also use my Mamiya lenses on a 5DII as well .
The MFD sensors are huge, but noisy...but at least they're slowly getting better. That's why I'd love to see some examples, and I'm encouraged by your 645D shot.

Ray

[theguywitha645d]

Ray, thanks. I knew of a guy working in telescope instrumentation that had the same conclusion about medium-format sensor. But I am encouraged by the 645D.

I will try to post some more. But like I said, I am just starting and I am also trying to figure out my 645D--and why do the stars come out so late at night.

I would like to play some more with stacking to see if I can reduce noise--the noise at 1600 is not really bad, but when you push the contrast with things like the Milky Way, the noise obviously get pushed as well. I have also been playing with moon-lit landscapes, but I am really waiting for this winter to try some snow scenes at night. But all in all, I have been impressed with what Pentax has done with the sensor. I just came back from Japan with a 300mm lens I would like to play more with.

I will see if I can post a few more tonight--jet lag is still with me.

[WIFoto]

As much as I love MFD, it really is not that well suited for astrophotography...I discovered that too late!  But now I also use my Mamiya lenses on a 5DII as well .

Ray

Perhaps not!

Hi,
I am not an expert, but over the last few years I have learned some very interesting things about digital astrophotography. There are some wonderful techniques that can be used to tame the noise generated in sensors during longer exposures. Much astrophotography is done with CCDs that can be pretty noisy. With the right techniques, you can get rid of the noise!   

If you have not looked into this topic I can recommend the book:

The New CCD Astronomy by Ron Wodaski.

It is amazing what stacking a few image frames can do to reduce noise. Four stacked will make a significant difference in reducing noise. After 16 frames, the gains diminish quickly.  Also, if you take the time and create a series of calibration frames (dark, bias, and flats) and apply them your image, quality will improve even more.

There are some fantastic software tools for both Mac and PC users that are free, or very inexpensive, that make processing digital images easy. One favorite is Nebulosity by Craig Stark of Stark Labs.

As an overview: you shoot your image files and put them in a folder. Shoot your dark frames, put them in a folder. And so on with your flat and bias frames (if you elect to use them). Then fire up a program like Nebulosity. The software will ask you to point to your dark frames. It will then stack them into a master dark frame (with low noise) and save it. You do the same thing with your flat frames and your bias frames. You end up with a master frame for dark, flat field, and bias.

Next, you tell the program the folder for your image files. Typically you will shoot 4 or more frames of a single subject on rig tracking the sky accurately like the AstoTrack mentioned in the original post. The program will work with you to align and stack the image frames. At this point the stacked photos will be nice. But then you tell the software to apply the master calibration frames for dark, flat, and bias, and everything gets even better. Noise should be all gone. Your image should be free of any natural vignetting. You will still have work to in order to stretch the image, correct color, sharpen, and so on, but your starting point should be massively better than if you had made single frame of the night sky.

One final point. It is important to keep the sensor as cool / cold as possible. Heat equals more noise recorded. Dedicated astro imaging cameras use active cooling systems to manage the heat. Big observatories will use liquid Nitrogen cooling. I have found shooting on cooler / colder nights vs on a warm summer night makes a difference in image noise.

Ray, I bet you can do some terrific asto work with your MFDB. With accurate tracking, good focus, and some free / inexpensive software I think you will be as surprised.

Cheers,
Paul

[jsiva]

My twin does a lot of astro stuff using a CCD and a 5DII.  I think there is a contact link on his site if you have any questions.  I'm sure he'd be happy to help with any questions.

http://sanjeevsivarulrasa.com/

[Brian Hirschfeld]

As promised here is a set of 8 images I took with my IQ180 in my backyard tonight. They were taken using the PhaseOne 645DF with my IQ180 and the Mamiya 55mm f/2.8 lens. I took images starting at ISO800 @ f/2.8 and 60.0s like recommended above. However this was over exposed since their was light spill from Bridgeport (closest major city) near by. Exposure was adjusted and images were taken full-frame 80mp at ISO400 and ISO800. Then for comparison I flipped into Sensor+ (20mp Images) and took images from ISO400, ISO800, ISO1600 for comparisons. I was generally pleased with the results surprisingly well handled noise, this was my first time trying out the camera with long exposures at night. Methodology is better explained in the descriptions.

http://www.flickr.com/photos/brianhirschfeldphotography/sets/72157628255284141/

Some notes: A few planes etc flew by because apparently in this distance is part of JFK's flight path for its 9:00 PM international flights heading up towards Greenland before cutting over. In most of the images Orion's Belt is visible. If anyone wants to see the full-res images PM or e-mail me montblanc100years@mac.com and I'll put them in a dropbox or something.

Enjoy
~BH

[theguywitha645d]

My twin does a lot of astro stuff using a CCD and a 5DII.  I think there is a contact link on his site if you have any questions.  I'm sure he'd be happy to help with any questions.

http://sanjeevsivarulrasa.com/

WOW! Those are really excellent. I don't want to post now.

Actually, if you could ask your twin, the landscapes with stars, are those fixed tripod shots? At the image scale it is hard to see if the stars are trailing and I would really be interested in focal length, ISO, and exposure information. And please pass along my complements.

[theguywitha645d]

Thanks Brian.

Here are some other tests and mucking about at night. The painting with light image is the longest at five plus minutes. All of these are stopped down to f/5.6 or f/6.7. I think all were at ISO 1600. Focal length are a combination of 35mm, 55mm, and 120mm (Big Dipper).

[Brian Hirschfeld]

I was checking out that Astotrac mount, how does that work?

also edited one of my Astrophotography Images, so you can only see the stars, just Levels and B&W

http://www.flickr.com/photos/brianhirschfeldphotography/6448480871/in/photostream

[theguywitha645d]

Brian, I mount mine on a Gitzo tripod with a Manfrotto geared head--the geared head make polar alignment easier. Basically, there is a polar scope (which unfortunately is optional) which aligns the mount to the North star. I simply attach the camera to the mount with a ball head. You turn on the drive and it slowly rotates the camera to counter the Earth's rotation. You have about an hour before you need to reset the drive. This mount will also works in the Southern hemisphere. It is very light and compact and runs off AA batteries--the mount comes with a cigarette lighter adapter; I went to an electronic parts store and fashioned two battery packs for about $6; the AstroTrac battery pack is about $15.

Getting used to polar alignment takes a bit of doing--it is amazing how stars all look the same. You actually use the north star and one or two others whose position is marked in the reticle of the polar scope. The better aligned, the longer the tracking and/or the longer the focal length. Polar alignment for a 35mm lens is easy, tricky with a 120mm.

The AstroTrac is one of the most compact equatorial mounts available, but there are others including mounts for telescopes.

[jsiva]

WOW! Those are really excellent. I don't want to post now.

Actually, if you could ask your twin, the landscapes with stars, are those fixed tripod shots? At the image scale it is hard to see if the stars are trailing and I would really be interested in focal length, ISO, and exposure information. And please pass along my complements.

Brian,

I have no idea.  All I know is that he goes off in the middle of the night for hours on end and spends quality time with his telescope!  I mostly do landscapes and compared to my efforts, his are monumental investments in time and patience. What you are looking at is likely multiple stacks on an automated mount.  Drop him a note, he's quite geeky about explaining what he does

[larkis]

This might not be astro photography but I did a 10 minute+ long exposure test with the 645D, the image is here on my blog: http://blog.dominik.ca/2012/01/20/salvation-mountain/

The raw is still under exposed so it got a boost in Lightroom which brings out more noise. I will try to open the lens more or increase the ISO to let more light in quicker for my next test.