Do I need a UV/IR cut filter?

[billhinge]

I have been using my GH2 for astro imaging, attached to the telescope for planetary video and standalone with a lens for dso

(being a x2 crop a 300mm gives an equivalent fov to a 35mm 600mm tele) and it has very good video (in terrestrial terms anyway) with 1x1 bining

I was considering getting it modified for astro or getting a canon but I'm not sure now?

Looking at some images taken with a 720nm and 850nm filter I'm getting handheld shots at 720nm @iso800 and I can image red diodes, tv remotes and uv lamps no problem. This leads me to think that my cameras UV/IR filter is quite weak, i.e. not an abrupt cut at 400nm and 700nm

Here is a an image I took of the pleiades as an experiment at iso12,800 13x60s subs. I notice I get quite a bit of halo, even on original frames, is this a symptom of unfocused light in the blue end (I toned down the blue a bit in this image)

Also when imaging planets the blue channel is more blurred than the red or green (the saturn images are blue, red, green in that order with the combined at bottom, I tweaked the blue channel in pixinsight)

When I did some shots of orion there was quite a lot of red in the image that I had to filter out. The two red images are outside handheld at 720nm 800iso and inside in shade illuminated with a window at iso3200. I realise I wont get the same response as a full spectrum but I assume I would get a decent response at Ha?

Anyway I have a set of LED's arriving from maplin tomorrow ranging from uv to 700nm to allow me to plot the frequency response

So the question , is it possible that the GH2 filter is not sufficiently cutting deep blue and deep red, hence a 67mm UV/IR cut filter on the 300m lens would be advisable for dso shots

If so this may be easier than getting a modified DSLR?

here is hopefully some evidence

[billhinge]

Preliminary testing with leds seems to indicate yes, I think Ha should be fine out of camera unmodified.

I guess I shouldn't be surprised as it uses a deep red auto assist lamp for autofocus in dark ! (uses a contrast based focusing method called CDAF unlike DSLR's)

I'll do some more detailed tests after dark

@$&% spell checker!

[nytecam]

I understand normal [terrestial] lenses are not designed to focus beyond a restricted range ~ 400-700nm hence haze/fuzz outside this waveband. Pure reflectors [no lenses/correctors] bring all wavelengths to focus but limited by detector spectral sensitivity:cool:

[billhinge]

yes that's correct, you can see in the second red image of the cat, I used an AF lens and it was fooled by the IR .

The issue for me is that I use a C11 SCT which is one reason why I'm asking this question, plus sometimes I use a telephoto on the GH2 and attach it directly to a GEM, the result as you say is haze/fuzz.

The other question being is it possible to find an off the shelf camera that will record Ha un-modified without having to replace the IR cut filter - I suspect it maybe

An aside - I'm also investigating a technology solution that will allow me to take ~raw (actually apple prores) from the HDMI port

http://atomos.activehosted.com/kb/article/ninja-tests-with-popular-dslr-cameras

It already has 1:1 binning so if everything works as I hope then I should have a 1 stop camera for terrestrial, planetary, lunar and dso - well thats the theory plus I don't need to buy any additional lenses

[billhinge]

Having done a quick and dirty test in the dark I came to the conclusion that spectral response is pretty even from deep blue (dipping into uv) up to at least 700nm therefore it should image hydrogen alpha out of the box with no mods

I have ordered a couple of super bright 660nm leds for a fun test just to be absolutely sure (should be here by Thursday)

So my belief is that the fuzziness I see in some images is due to deep blue and red and I need to add a stronger UV/IR cut filter (which is easier than taking one off

Here are some pics, taken from above to get maximum beam (they are very directional). Note the hazing around the UV diode, in real life the UV was exceedingly bright - even in the viewfinder (which is electronic not optical).

The LED's have different strengths 500mcd for UV down to 5 for the 700nm so I normalised the values based on data sheets, exposure times were all similar except of course for the uv

I'll try to plot this on a graph by the weekend when I get all the data sorted

[nytecam]

Interesting use of multicolour leds to test spectral response My own tests done with a prism or grating on starlight eg spectroscopy which is the inverse of your solution. There's a spectro forum on SGL of course.

Beyond early disappointing DSLR result via my 300D I don't use this cam for astro eg severe LP and too big a fov for my DSOs preferring a small sensor unfiltered OSC CCD for max photon collection getting down to mag 16/17 in a few seconds exp:hello2: I refuse to spend time on long exposures on a single object - preferring to capture many objects in a session with very modest exposures - it works for me!

[billhinge]

I'd love to claim the LED idea, someone on here mentioned that red led's have similar colour to Ha so I just extended the idea a bit!

I live in a town and the light pollution from the docks can be pretty bad some nights, the first image was 13 subs which were virtually yellow being shot at iso 12800 (no typo) before subtracting darks and bias (not even the correct ones but leftover iso 800's). But I know what you mean

[iwatkins]

Also bear in mind that LEDs have a bell curve output and most filters have a bell curve block/pass plus their leakage, sometimes at harmonics.

To test a filter, I would look to use a quality laser with a certificated frequency.

But yes, the GH2 won't be different from most other DSLRs, block/pass will be weak/inconsistent. They are designed for visible light, sub second exposures firstly.

If you want to do anything differently, you'll need to go dedicated camera for serious work.

Red LEDS have similar output to Ha, true. But Ha is actually a very specific and very narrow frequency and you'll find LEDs will swamp anything useful.

Cheers

Ian

[billhinge]

yes I took note of the diode spectral curves, some of the diodes I used were supposedly pure, others not. But I take your point, if I had the right kit then I would use it but I just wanted a rough expt to test my photographic experiences which to my eyes it seems to

I believe the correct kit is called a monochromator?

[iwatkins]

I'm not a spectral expert, but I understand that most LEDs are pretty pure, but the packaging can affect output.

Yes, a monochromator basically allows you to mechanically tune the output of a source to a specific waveband, usually using a narrow exit slit from what I remember.

But keep experimenting, if what you find gives you images you want, then its a win:)

I think you would enjoy spectroscopy, there is a section here for that.

Cheers

Ian

[billhinge]

The answer is definitely yes. I ordered a B+W UV/IR cut filter for my zoom yesterday and it arrived this morning. Makes a clear difference when shooting through haze in daylight, noticeably better colour and detail

The glare around UV diode also disappears. My normal and super bright Ha (660nm) diodes arrive tomorrow for conclusive proof

[billhinge]

Yes!

Here is a bright 660nm clear diode taken with an unmodified GH2 fitted with a screw on UV/IR cut filter

easy!