• Welcome to Orpington Astronomical Society.
 

News:

New version SMF 2.1.4 installed. You may need to clear cookies and login again...

Main Menu

A bad attack of the pox

Started by MarkS, Jun 27, 2008, 07:08:40

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

MarkS

Here is Wednesday night's attempt at the veil with the newly modded Canon EOS 300D (scaled down by factor of 3)  21 x 5 min exposures:





I'm not at all happy with it - the stars look absolutely terrible.  They have a bad attack of the pox - a white pimple surrounded by angry red.  Here is a region of the image at full size:





The problem is chromatic aberration on my ancient Nikon 300mm lens - the one I normally use for guiding.  It will not focus the H-alpha in the same plane as the blue.  I guess it was never designed to do so.

Looking back at my test focus image the problem becomes quite apparent:





The telescope focus knob is adjusted as the star slews from left to right.  At the left hand end (2nd segment in) the blue is clearly focused but is surrounded by a red haze. At the right hand end the red is clearly focused but is surrounded by a blue haze.  The veil was imaged with the focus knob near the left hand end of the scale.  Hence every star is surrounded by red haze.

In retrospect, the best compromise is around the 3rd segment where nothing is completely in focus but at least all the colours are more or less registered.  The other alternative is to stop the lens down further (say F8 instead of the F5.6 which I used) but this would reduce the light too much.






Carole

Your talk was very interesting last night Mark, and I can now understand where you are coming from when you talk about your imaging on the forum as I tend to forget you are using a DSLR.  The only subject you did not get to explain was modding.

Any chance of any explanation here, or is it too long/complicated to put in a post. 

I thought you method of focusing was rather clever. 

Carole

MarkS

Carole,

It's easy to explain.  A CCD is sensitive to a much greater range of wavelengths than the human eye is e.g. infra-red.  So the manufactureres of consumer digital cameras put a filter in front of the CCD that allows only the visible wavelengths to reach the CCD.  If they had no filter then your digital camera photographs would look very strange - they would include features that are invisible to the eye!

The problem is that the IR filter put in by the manufacturer also cuts out most of the Hydrogen-alpha wavelength.  The H-alpha wavelength is very useful (and pretty!) in astrophotography even though it is invisible to us.  Modding the camera means replacing the manufacturer's filter with an alternative that lets the H-alpha reach the CCD.  In the case of the Canon, this improves the brightness of H-alpha by a factor of 4.  Depending on the alternative filter you use, the camera can also be made sensitive to infra-red.

A modded camera is great for astrophotography but it means everyday photos now come out extremely pink and show features invisible to the human eye.  A special additional filter can be put on the lens to return the camera back to its "normal" sensitivity range (i.e. only the visible wavelengths) to make everyday photographs appear with the correct colour balance.

Mark

Rick

Quote from: MarkS on Jun 27, 2008, 07:08:40
I'm not at all happy with it - the stars look absolutely terrible.  They have a bad attack of the pox - a white pimple surrounded by angry red.
...
The problem is chromatic aberration on my ancient Nikon 300mm lens - the one I normally use for guiding.  It will not focus the H-alpha in the same plane as the blue.  I guess it was never designed to do so.
Is any of that red the result of infra-red appearing red in the modded camera? That would definitely focus a way off from the main visible components...

Quote from: MarkS on Jun 27, 2008, 11:16:06The H-alpha wavelength is very useful (and pretty!) in astrophotography even though it is invisible to us.
Ummm... I think the point is that it is visible (or at least in the visible part of the spectrum)... ;)

Ian

but only just, it's right at the edge of the eye's red response. If it wasn't visible, a Ha filter would appear opaque, rather than a very deep red. I've never seen a reasonable explanation why DLSRs have such a poor response to Ha by default...

Rick

Yeah, but it's visible enough for instruments like the Coronado solar telescopes to show a nice image, admittedly from a high intensity source... ;)

MarkS

Rick,

Your suggestion about the infra-red focusing to a different plane is very plausible.  However I was using an Astronomik CLS filter and I believe it cuts the IR but I'm not absolutely certain.

On the other point, yes, the eye is not completely insenitive to H-alpha.

Mark

Ian

the other thing to bear in mind when modding a camera in this way is the bayer pattern filters on the CCD are not completely opaque to IR, so you need some sort of IR filtering for anything approaching a reasonable colour rendition.

Rick

Quote from: Ian on Jun 27, 2008, 12:28:22the bayer pattern filters
Yeah, that could make things complicated...

While we're on the matter of colour representation of IR images, how did you take your "DSC in IR" image? (Link to gallery via thumbnail below...)


Ian

standard Nikon D70 with a Cokin IR pass filter. In fact, that's given me an idea for IR pass photography, need to adapt the filter holder to fit my refractor...

Carole

QuoteModding the camera means replacing the manufacturer's filter with an alternative that lets the H-alpha reach the CCD

Ah, is this what you were telling us about when you took the camera apart and changed the filter from a red one to a white one, in which case you DID tell us about modding?  So modding simply means modifying.

QuoteA CCD is sensitive to a much greater range of wavelengths than the human eye is e.g. infra-red.  So the manufactureres of consumer digital cameras put a filter in front of the CCD that allows only the visible wavelengths to reach the CCD.

Getting in a muddle now as one minute you are talking about a digitial camera and the next CCD.  I thought CCD was when you use a laptop instead of a DSLR and yet on CCD we put ON an IR filter.

Maybe I don't really understand the true meaning of a charged coupled device.
Can some-one elucidate all this.

Does this go part way to explaining why the naked eye does not see astronomical things in colour?   

Thanks
Carole


Carole

Mac

getting back to the focus problem, red blue.



Band C is the Ha line, so would be in focus when your blue was also infocus, unless you have severe chromatic aberation.

Which even on your old nikon lens would probably be nill.

The more plausable answer is the out of focus red is in fact it is the IR, as this does focus at a different plane, compared to visible light.

and now that you have the full range of your CCD which goes right in to the IR spectrum

Some of my old pentax lenses still have the IR focus point on them.

Now that you have removed your IR filter your chip will be recording the IR out of focus, whilst the rest is in focus.

You probably find that the IR filter fitted to most modern cameras probably starts from 650nm upwards.

Just as a thought have you tried photographing UV objects.

http://www.naturfotograf.com/uvstart.html.

Ian

Carole,

a CCD is the actual sensor chip. You can see a small one when you unscrew the lens off your webcam, or if you take the lens off a DSLR and lock the mirror up, you'll see a big one. The reason we tend to associate CCD with astro cams, is to distinguish between film and CCD, DLSRs didn't exist when that phraseology came into being. So, all digital cameras are CCD devices (actually there's a couple of types but the distinction is technical, the other is known as CMOS) and we're just a bit sloppy with names. Bit like Hoover really...

Mac, yup alot of old lenses have a IR mark on them (often in red or orange) and that can really show you how far away the focus point of IR is from the visible (bearing in mind the wavelength for IR is twice that of blue light, and refraction angle is a function of wavelength).

However, when we're talking critical focus the next question is how well corrected is the lens. In my experience, not particularly. Camera lenses rarely see the sort of exacting requirements we have on our equipment. Point parallel light sources are as hard as it gets for an optical system and camera lenses are not designed to perform perfectly with them. However, ED and APO camera lenses should do what the name suggests, in the case of APO should be corrected to three wavelengths across the visible spectrum. None of my camera lenses are APO and I don't think your Nikon one is either Mark. So actually, camera lenses can have chromatic aberration but generally get away with it. They can also suffer from significant coma and spherical aberration too. Do a google search for "Boke", there is a view that spherical aberration is good for a camera lens, as long as it's the right way...

MarkS

Carole,

Digital cameras, specialised astro cameras and webcams all have a sensor chip.  Most sensor chips use CCD technology (but some are CMOS - that's a different story).  I just used the word CCD as shorthand for sensor chip.

You put an IR filter on your webcam to prevent the infra-red which would otherwise appear out-of-focus on your image.  Out of focus IR is possibly the problem I am seeing with my "pox" picture.

Mark

Carole

Thanks Mark and Ian for your explanations.

Just one more while we are on the subject.  I tried explaining to the doctors at work why the eye does not see in colour, I know it has something to do with less sensitivity and the fact that the eye does not see in long exposure, but I felt I was not sufficiently able to explain. Obviously the eye does not see in IR and other wave lengths is there anything else I can add to this?

Carole

Ian

I'm surprised you'd need to explain it to some doctors ;)

The eye has two sorts of receptors, known as rods and cones. The cones have a colour capability, the rods intensity only. However, the cones need a minimum amount of illumination to work, and are *much* less sensitive than the rods. In low light, the cones are completely inactive and so we have no perception of colour.

Interestingly, the one colour that can sometimes be visible is green. In a big scope, like the society lightbridge, bright nebulae like the orion nebula can look green. This may well be due to the green sensitive cones coming "on-line" as it were first making it seem green. Most attempts at true colour images of that nebula are rarely green, although that might be in part due to convention. Of all the colours, green response is the greatest...

Rick

For a photographer back in the days of film there's an advantage to having the IR out-of-focus at the film plane. Consider the effects of taking a setting-sun photograph...

Carole

Thanks Ian,

I had been showing the doctor's some of the members imaging and mentioned that I had actually seen things like the Dumbell and Ring nebula but they hadn't appeared in colour as they did in the images. 

These are gynaecologists, so although they get a general training, the eyes are not their best specialty.

Carole


mickw

Biting tongue.......................

Beer coming out of nose.......................

:lipsrsealed:
Growing Old is mandatory - Growing Up is optional

Mac

Puts a new slant on the ring nebula......................... :o

Ian

and I was thinking things are looking up...

MarkS


Rick,  it looks like you were correct.  The CLS filter does allow infra-red through.  So a lot of the red around the stars will be long wavelength red and maybe infra-red as well not coming to focus.  I need to get an IR filter to prevent this.

When I imaged the Dumbbell, the main optical elements in the scope were mirrors so there no differential focusing.

Ian, you're right about IR passing through the RGB elements of the Bayer matrix - Christain Buil has some interesting response curves here: http://www.astrosurf.com/~buil/350d/350d.htm   It explains why your DSC IR image is almost monochrome.   

Ian

Quote from: MarkS on Jun 27, 2008, 23:32:31
It explains why your DSC IR image is almost monochrome.   

Actually, it doesn't entirely. My D70 has the OEM IR block in place. The interesting palette is mostly down to the fact nearly zero response in blue and green and you really have to be creative with the colour balance (the in-camera colour balance doesn't stand a chance)

I'll probably take the filters down to DSC in August, bright sunlight is best, not least as the IR pass filter cuts more than 4 stops. If you get enough signal in the blue and the green then the most effective colour balance is to make green foliage white (apparently it's as close to a complete reflector of IR you're going to come across in a sheep field in Kent). This PBase photo album is a good example.

What would be fun would be to try my IR pass filter on one of your modded Canons. It would have a much wider response across the who sensor.

MarkS

Quote from: Ian
Do a google search for "Boke", there is a view that spherical aberration is good for a camera lens, as long as it's the right way...

Boke - now you risk getting me onto one of my other favorite subjects - point spread functions and deconvolution of de-focused images.  I've written some (buggy) software to do this. 

An example: here's a random picture I found on the internet.  Yes, that's Beckham in the foreground, but who are those England supporters in the background?



Deconvolution provides the answer:



The same technique works remarkably well on planetary images  ;) 


Ian

that's impressive Mark. So how does your program differ from other methods of deconvolution? (I'm prepared to not understand the answer, but nothing ventured... ;) )