Camera Labs

Inspired by Gordon's post I thought I would have a go. Guess what, it's more difficult than I'd hoped.

My biggest difficulty continues to be getting a really crisp focus. As Gordon says, the lens needs to be set to manual focus with image stabilisation (VR) off. It really is a case of trial and error even when starting from a known good setting.

My set-up for the shots to follow is a Canon EOS 400D/XTi with an EF 24-105mm f/4L IS USM. A Hutech LPS-P2-FF Front Filter was used to block scatter from the local sodium street lighting. The camera was set to manual, ISO 800, f/4 and daylight white balance. The lens was set to its widest zoom of 24mm. The Bulb exposure was controlled by a remote switch to stop unwanted camera vibrations. All exposures were 180 seconds for this sequence. Colour information was discarded. Three minute exposures demand that the camera counters the rotation of the Earth so the whole shooting match was mounted on top of my telescope to take advantage of its motor driven mounting like so

The telescope's optics weren't used.

IRIS software was used to process the RAW images. Here's how the shot looks like with virtually no processing except re-sizing

This and all subsequent images are click-able to allow examination of the full size originals.

I took a second three minute exposure of the Milky Way and then covered the lens and took a three minute dark frame

This shows a few stuck pixels and the effect of sensor noise.

The next shot is of this dark frame subtracted from the first image

As before, minimal processing apart from a small adjustment to the image thresholds.

The last image was produced by taking the second three minute exposure of the Milky Way and subtracting the dark frame as above. The intention was to add this to the image above to increase the signal to noise ratio. While I succeeded the result could have been better as, unknown to me at the time, the camera had rotated slightly between shots. To its great credit, the IRIS software did a pretty good job but rotating a digital picture by just a few degrees is impossible without introducing subtle artefacts and those artefacts severely limit the degree of unsharp masking and other tweaks that can be accomplished before they become visible. Here is the result


The Milky Way here in the Northern Hemisphere can't rival the show it puts on at this time of year in the South when one is looking towards the centre of the galaxy. Even so, I am disappointed that it isn't a bit more obvious in these pictures.

That said, I am really pleased with the performance of the EF 24-105mm f/4L IS USM lens for both its sharpness and lack of chromatic aberration. I have used this lens at other zoom settings (35mm, 70mm and 105mm) as well and so far have detected no zoom creep at all despite near vertical orientation. The same can also be said for my EF 70-200mm f/4L IS USM at various zooms.

Even so, an f/4 zoom of any focal length is hardly ideal for this sort of work. In my opinion, big sky shots on a cropped frame sensor call for 20mm focal lengths or less to get the angular coverage and apertures of f/2 or faster to overcome some of the noise issues with such sensors.

That sort of glass ain't cheap and if you are that serious about astrophotography then a DSLR sensor with its Bayer pattern of colour sensors is pretty hopeless. Good glass will focus a star onto a single pixel and that could be red, green or blue

However, within these limitations there is still fun to be had. While the camera was busy doing its stuff I had time to enjoy beautiful clear skies and the occasional Perseid meteor. Without the camera I would probably have been rattling the windows with my snoring.

You don't have to put your camera on a motor driven set-up as Gordon's post shows. If you want to try longer exposures with the camera on a fixed tripod then you can get star trails which can be quite artistic if set against the right backdrop. Give it a go.

Bob.

_________________
OM-D E-M5 + ED 12-50mm, H-PS14042E, H-F007014E, L-RS014150E, M.ZUIKO DIGITAL 45mm 1:1.8.
Gitzo GT1541T, Arca-Swiss Z1 DP ball-head.
Astrophotography: TEC 140 'scope, FLI ML16803 camera, ASA DDM60 Pro mount.

Hi Bob, great posting! Yep, when it comes to astrophotography you can't beat a dark sky and a bright aperture. In my example I was shooting at f2.8, which as you know gathers twice as much light as a lens at f4 - and I was also shooting at 1600 ISO, so effectively doubling what you were working at again...

I was playing with a 50mm f1.4 yesterday and the jump of having two more stops (4 times brighter) than an f2.8 for this kind of thing is amazing!

Go on, buy yourself a 50mm f1.8 or 50mm f1.4 for your Canon!

Gordon

**Hugs 50mm f/1.8**

_________________
Nikon D80 | 18-135mm Kit Lens | 50mm f/1.8 | 4GB 150x Card | SB-600
http://flickr.com/photos/antmanisfubar

Hi Gordon,

A month or two ago I did have an order in with an Amazon reseller for a 50mm f1.8 as it is relatively cheap and optically very good value but it never arrived so I cancelled. It's currently available for less than £70 so I may try again.

Here's a pretty subversive question for you. To set the scene, star images are essentially of a digital nature rather than the analogue nature of planetary or conventional subjects. Given the geometry of a DSLR sensor is it a good thing in stellar astrophotography to have a star focussed down to a spot much smaller than a pixel or (and here's the subversive bit) is it better to have a star image the same size or even a tiny bit larger than the pixel area?

If the latter is true then as long as that performance is maintained to the sensor edge at the widest aperture of the lens it may be pointless (oh dear, another bad pun!) to invest in expensive glass purely for stellar astrophotography.

Bob.

_________________
OM-D E-M5 + ED 12-50mm, H-PS14042E, H-F007014E, L-RS014150E, M.ZUIKO DIGITAL 45mm 1:1.8.
Gitzo GT1541T, Arca-Swiss Z1 DP ball-head.
Astrophotography: TEC 140 'scope, FLI ML16803 camera, ASA DDM60 Pro mount.

Hi Bob, that's a very good point. Literally!

I have found my digital astro photos do have some nasty artefacts which weren't there on film, and I suspect much of it is to do with a point source of light striking a pixel or two. I'm not sure there's a solution to it either...

But there's still value to be had with bright aperture lenses and those which are also well-corrected to the corners of the frame. There's nothing like stars to reveal nasty coma effects...

If your sky isn't too polluted, I'd try increasing your ISO to 1600...

Gordon

Thanks for the advice, I'll give it a try. The Hutech filter does a great job in cutting out the backscatter from the street lights and with any luck any increase in sensor noise can be countered by producing an extra frame or two for IRIS to stack. A very powerful program but unfortunately poorly documented and some of its best tricks aren't available via the GUI.

Bob.

_________________
OM-D E-M5 + ED 12-50mm, H-PS14042E, H-F007014E, L-RS014150E, M.ZUIKO DIGITAL 45mm 1:1.8.
Gitzo GT1541T, Arca-Swiss Z1 DP ball-head.
Astrophotography: TEC 140 'scope, FLI ML16803 camera, ASA DDM60 Pro mount.

.
I tried out my new EF 50mm f/1.8 lens this evening. This is the cheapest lens Canon make so I wasn't expecting too much.
That said, reviews of this lens are reasonably complimentary about the optical quality but are rightly, in my view,
disparaging about the build quality and the awful auto-focus motor.

The set-up was as described in the first post of this thread with the camera at ISO 800 and the lens wide open at f/1.8.
Four normal exposures were taken, each lasting two minutes, and a final two minute exposure with the lens cap on
completed the session.

I might add that most of the session was devoted to trying to obtain a sharp manual focus by means of around a dozen
20 second photos, examining the magnified result on the cameras LCD each time. The new 40D with its ability to control
and display the whole process in real time on a notebook computer connected via an USB cable sounds incredibly attractive
though maybe not with the EF 50mm f/1.8 lens' focus motor!

Anyway, here's the result of subtracting the dark frame from each of the four exposures, aligning and stacking the four resulting
frames, tweaking the dark threshold and applying a light unsharp mask:


Click on the image to see it at full size (2.5MB download)

The photograph is of the Milky Way in the general area of Cassiopeia which is almost overhead the UK around midnight at this time
of year. To help orientation I have drawn lines connecting the five major stars.


Apologies if the lines aren't quite in the right place.

Considering the lens only cost £60/€90/$120 it's not a bad result although examining the corners of the full size image tells a tale.
It certainly shows how fast optics can bring out the nebulosity of the Milky Way.

To borrow a famous quote in a film from my youth "My God, it's full of stars".

Bob.

_________________
OM-D E-M5 + ED 12-50mm, H-PS14042E, H-F007014E, L-RS014150E, M.ZUIKO DIGITAL 45mm 1:1.8.
Gitzo GT1541T, Arca-Swiss Z1 DP ball-head.
Astrophotography: TEC 140 'scope, FLI ML16803 camera, ASA DDM60 Pro mount.

Nice work Bob!

Thanks Gordon.

While sitting here listening to a concert by the late and much lamented Luciano Pavarotti, who I had the great privilege of seeing once at the Met, I spent a little time counting the stars in a small section of the image in my last post.

Very approximately, I think the whole image must contain around 75000 distinct stars, not counting those stars too faint to show up individually but which contribute to the glow. We now know that planetary formation is likely to be the norm as part of star birth though many planetary systems look very different to our own.

I look at those stars and consider the great beauty we can create here on Earth and can't help thinking of the line used in Carl Sagan's film Contact. When asked if there are people on other planets the reply was "If it is just us, it seems like an awful waste of space."

_________________
OM-D E-M5 + ED 12-50mm, H-PS14042E, H-F007014E, L-RS014150E, M.ZUIKO DIGITAL 45mm 1:1.8.
Gitzo GT1541T, Arca-Swiss Z1 DP ball-head.
Astrophotography: TEC 140 'scope, FLI ML16803 camera, ASA DDM60 Pro mount.

Hi Bob, it is mind-blowing to think about this stuff, which is why it's often easier for us humans to look past the scale and focus on optical aberations instead! I'm certainly guilty as charged, although I've never lost the wonder of gazing skywards on a clear night. I was also sad to hear about Pavarotti and regreat not having the chance to have seen him perform in person.

So did you count the stars in a small portion and extrapolate?

Hey, I'm not so sad that I counted all 75,000.

Yes, I inverted the image (stars as black dots) and copied out a 300x200 pixel section which seemed to have an average density. In the end I only counted the stars in a 150x200 pixel section (used "Paint" to add a small dob of red over each star counted) and then extrapolated. Thus the 75,000 figure is very approximate but at an average of one star per 133 pixels it doesn't seem wildly out.

As you know, our galaxy is thought to contain around 100,000,000,000 (100 billion) stars so even with that delicious 85mm f/1.2 lens I think any big sky pictures would come up a little short!

Bob.

P.S. I saw Pavarotti in I Lombardi: not an opera I knew. Pavarotti eventually shuffled on stage, opened his mouth and the most wonderful sound washed over the audience even though he was near the end of his career. I've never heard a recording which really did that moment justice.

_________________
OM-D E-M5 + ED 12-50mm, H-PS14042E, H-F007014E, L-RS014150E, M.ZUIKO DIGITAL 45mm 1:1.8.
Gitzo GT1541T, Arca-Swiss Z1 DP ball-head.
Astrophotography: TEC 140 'scope, FLI ML16803 camera, ASA DDM60 Pro mount.

.
In an earlier post in this thread I shared an image of the Milky Way in the region of the constellation Cassiopeia. That image was obtained
by reading the RAW files directly and, by choice, discarding the colour information.

As the camera was saving large fine JPEG files as well as RAW files for each exposure I thought it would be interesting to try the same
stacking technique using the JPEG files as the source and also retaining the colour information. The four original shots were processed
in the same way as earlier in order to produce the stacked image. I then adjusted the thresholds and applied a little sharpening. Finally,
I increased the colour saturation by about 50% and then removed the slight blueish tint introduced by the Hutech filter I use to counter
the local sodium street lights.


Click on the image to see it at full size (1.4MB download)

Retaining the colour information seems to increase the background noise and in removing this I have lost some of the nebulosity of the
Milky Way. The final result is certainly very sensitive to the post-processing choices made.

Despite the Bayer pattern of red, green and blue sensors the camera wasn't fooled into producing green stars! If you are unsure why
green stars in this picture are not a good idea then have a look at this article.

Bob.

_________________
OM-D E-M5 + ED 12-50mm, H-PS14042E, H-F007014E, L-RS014150E, M.ZUIKO DIGITAL 45mm 1:1.8.
Gitzo GT1541T, Arca-Swiss Z1 DP ball-head.
Astrophotography: TEC 140 'scope, FLI ML16803 camera, ASA DDM60 Pro mount.

Interesting technique Bob... having played around with your images, what do you feel you're missing from your astro shots, in terms of equipment, conditions or technique?

I am still pretty wet behind the ears in this subject so my views should be taken with a pinch of salt! I don't think there is a single Eureka! improvement which will allow me to step up a gear or two.

• A really fast lens (f/1.2) would allow me to capture more nebulosity. As you know, aperture determines the faintest stars you can capture but f-number determines the faintest nebula.
• A fast lens requires dark skies with neither the Moon or street lights spoiling the party. I'd either have to use the car or vandalise the local street lights. The environmentalists would no doubt prefer the latter!
• A better sensor. DSLR sensors are pretty good considering they are optimised for well lit subjects but for deep sky work they fall well short of the capabilities of a cooled CCD with its low thermal noise and long integration times.
• Increase the exposure time. Two minute exposures do exhibit some sensor noise but I can probably go further and/or go from ISO 800 to 1600.
• Better software. The IRIS software I have been using is powerful but the user interface is a little quirky and some of its best tricks can only be accessed from the command line. I am sure it is capable of more if only I could find out how!
• More patience and dedication. The most difficult area to improve and the one that maybe needs the most attention!

Not everything on the list costs money but the big improvements certainly do.

There is one more thing I would like to try with the Cassiopeia images. I would like to take the colour image and clean up the background so that it is completely black. Then register it exactly with the monochrome image.

The next step can't be done with IRIS (or if it can I can't find the command to accomplish it). The idea is to load the colour image in Photoshop and extract the chroma information and discard the luminance. Then take the monochrome image and overlay the chroma information. That way I should get the best of both worlds.

Bob.

_________________
OM-D E-M5 + ED 12-50mm, H-PS14042E, H-F007014E, L-RS014150E, M.ZUIKO DIGITAL 45mm 1:1.8.
Gitzo GT1541T, Arca-Swiss Z1 DP ball-head.
Astrophotography: TEC 140 'scope, FLI ML16803 camera, ASA DDM60 Pro mount.

Hi folks,

I finally got to play with my new EOS 40D in the dark last night after the rain showers cleared away. The set up was similar to that
of the first post in this thread (the photo shows my old 400D/XTi plus EF 24-105L lens) but this time I used an EF 50mm f/1.8 lens
and, of course, the 40D. The camera was skewed around slightly to avoid the main telescope getting in the field of view. The only
reason for putting it on the telescope was so that I could use the telescope mounting to track the stars.

The other big difference is that I had the 40D coupled to an ancient (and very slow) notebook via a USB cable. That meant that I was
able to control focussing, aperture, ISO and white balance as well as the Bulb exposure from the notebook. This ability to use Live View
with the picture displayed on the notebook screen is a real boon. It needs a bright star but you can see exactly how well the lens is
focussed on the sensor. No more sequences of manually tweaking the focus and examining trial shots by craning around to see the
camera's LCD screen. Best of all, the whole process can be controlled from a comfy chair. Bliss!

Following this advice at Christian Buil's web-site I restricted myself to ISO 400. The following shot was taken with the 50mm lens wide
open at f/1.8 with a 120 second exposure. Click the image if you want to see it at full resolution - much more detail but a 2.5MB file.



In the centre of the image is the North American Nebula (geographic north would be to the right), so named because of a resemblance
in shape to that continent (WiKi). The image has been post-processed using IRIS.

I was running a sequence of test exposures and, unfortunately, after the fifth the dear old steam age ThinkPad ran out of boiler pressure
and locked up on me so not only was I unable to expose a proper dark frame but I had to cut the session short. It would have been nice
to run some more exposures at the settings above with a view to stacking some frames. That will have to wait for another night when,
hopefully, I can correct some of the other technical defects in this image.

For all that, I am dead chuffed. The 40D has already proved that for astrophotography it is a significant step up from my 400D/XTi. The
North American Nebula is an object I have known from pictures since my childhood but I was unprepared for the thrill of discovering it
for myself using a DSLR and a cheap lens.

Bob.

_________________
OM-D E-M5 + ED 12-50mm, H-PS14042E, H-F007014E, L-RS014150E, M.ZUIKO DIGITAL 45mm 1:1.8.
Gitzo GT1541T, Arca-Swiss Z1 DP ball-head.
Astrophotography: TEC 140 'scope, FLI ML16803 camera, ASA DDM60 Pro mount.