DSLR - the next step... light pollution filters, HARGB, LRGB...

luka · #1 29-03-2016, 12:15 AM

Hi

I just finished doing the full IR mod of 1200D and have to order Baader UV/IR cut filter (or similar) filter to avoid bloated stars with my ED80. This got me thinking about the future improvements, in particular about the light pollution (LP).

I am guessing that there are quite a few people like me that use a DSLR and battle the LP close to the CBD. How "far" is the next step? I have found the following methods and I would like to hear thoughts/recommendations from the more experienced:

1. Dark sky is of course the best but rarely possible. Actually not possible for me ATM.

2. Narrowband/CCD is unfortunately too expensive.
Cost: few $1000s

3. LP filters - I read mixed comments about those, ranging from great to waste of money with some having horrible colour cast which was impossible to fix in post processing.
Cost: $200-$300
(I am about to order DGM NPB filter for visual but I don't think it is suitable for imaging.)

4. HaRGB imaging with a DSLR. The narrowband hydrogen alpha filter is used in front of DSLR as luminance while RGB component comes from the standard DSLR image. The main idea is that the Ha is narrowband, avoiding LP (imaging can even be done during the full moon). Efficiency is 1/4 for Ha on DSLR as only 1/4 of pixels is used unless the camera is debayered.
Cost: $300-$500

5. LRGB with a DSLR where UV/IR filter is used for luminance.
I could be completely wrong here, I saw this mentioned on few places but without a clear description.
Where does the RGB component come from then? The standard DSLR image would give RGB but that would be identical data to the luminance channel? Do they just extract the L channel from the DSLR image and then mix it back with the RGB image.

Anything I missed? How do these methods compare and how much improvement is to be seen over the DSLR images? Any other thoughts?

I know that nothing compares to the dark skies but sometimes that is not an option

Thank for your thoughts
Luka

P.S. As a side question, do I need 2" filters for the DSLR? 1.25" are much cheaper and I believe can be used if I ever move to CCD imaging.

Atmos (Colin) · #2 29-03-2016, 07:33 AM

Dark skies if possibly are always the way to go.

CCD's can be on the expensive side especially compared to your current setup.

Although I have never used a LP filter, I wouldn't from what I have heard which appears to be the same as you.

With the HaRGB, you do only get 1/4 of the signal so it can take a lot longer to get the data but it is better than nothing right

My understanding is that you are going to need 2" filters with a DSLR. Although you can get away with 1.25" filters with some of the similarly sized CCD chips (KAF-8300 comes to mind), the filter is considerably closer to sensor and even then it is usually recommended to get 36mm to avoid vignetting.

With your DSLR, just take your normal RGB and then extract the luminance channel and blend that into the RGB data. Being a OSC there is no way to capture true luminance data.

glend (Glen) · #3 29-03-2016, 07:43 AM

If you have removed the stock camera filters you will be able to gather in about five times the amount of Ha light than previously. I suggest you use the camera awhile before you think about 'more mods'. Shoot the main Ha targets (Eta Carinae, M42, Lagoon, etc) and you will be amazed at the difference.
The bayer matrix makes shooting narrowband a real torture because of the limited number of pixels available to collect light. Going to the mono mod of the sensor makes narrow band possible as every pixel is working, however it is a difficult mod to do and damage to the sensor is a real possibility; however your resolution improves significantly but so does your image capture and processing workload. As far as I know I have the only mono cooled Canon in Australia. at this time. You will need to consider filter investment and changing them and the need to refocus. Be aware that the longer sub times required of narrow band make cooling almost mandatory, as heat build up in multi long subs. The cold finger mod would be my suggestion as a mod to do before you go to filter shooting and it will improve your OSC on the way. Read Rowland's cooling thread.

rmuhlack (Richard) · #4 29-03-2016, 09:22 AM

now that you have removed the stock filter, my suggestion would be to use an Astronomik CLS-CCD clip-in filter, and then if you're desperate to mod further, to look at a cold finger mod. Both of these options will help to improve signal-to-noise, which is what these mods (and the progression to cooled mono high QE) help to achieve.

I ran two cold-finger modded canons for a while (albeit from dark skies), see here for some deep-sky examples (note the lunar pics in that link were taken with another camera). Dark noise is reduced considerably with the cold finger mod

rustigsmed (Russell) · #5 29-03-2016, 10:05 AM

Hi Luka,

Definitely can recommend a light pollution filter. My setup only allowed 70 second subs before the background light pollution destroyed subs, after buying an astronomik ccd cls, my sub length went up to 3 mins and dramatically improved my images and made processing way easier too. I can't recommend a LPF highly enough.

with regards to NB, I did quite a lot of experimenting with my modified DSLR and NB filters. my plan was to eventually have a ccd but I had to stagger the expense out and bought my NB filters before the CCD (couldn't think of something more frustrating than having the camera but not having filters!). I wouldn't recommend using the Ha as luminance as the reconstructed/debayered image is a quarter of the resolution of the colour image. As Glen has pointed out noise is the killer for dslrs and NB, long exposure and uncooled is a bad recipe. still you can get some alright details on the brighter objects.
I tried very hard on some of the dimmer ones and just couldn't get over the noise issue. I used 2" filters as I was going CCD later on, you are able to use 'clip in' filters on a dslr. http://www.bintel.com.au/Astrophotog...oductview.aspx

Here are some examples: including NGC 6164 normal, Ha and Oiii blends - probably my best NB/RGB blend with a dslr. I think that is the way forward rather than using it as a NB only option.

Eta C Ha
https://www.flickr.com/photos/803366...posted-public/

No NB

https://www.flickr.com/photos/803366...posted-public/

Ha RGB
https://www.flickr.com/photos/803366...posted-public/

Ha Oiii RGB
https://www.flickr.com/photos/803366...posted-public/

Russ

glend (Glen) · #6 29-03-2016, 10:44 AM

Well since Russell has shared an image as an example I will offer a couple of my own, of the same object Eta Carinae. The purpoe of putting these up here is just to demonstrate the difference that some types of mods can make.

Both of the images linked to below, are taken with Canon 450Ds, the difference lies in the mods. The first image was taken with my cold finger cooled colour 450D, this is a full spectrum modified camera thus the presence of much more Ha (Red) in the image colour. As stated before going full spectrum will give you access to up to five times as much Ha as a stock filtered camera will provide.
The second image was taken with my mono cold finger cooled 450D, this is also full spectrum (no filters in front of the sensor), but I run a EOS filter drawer on the mono camera and had an Ha filter in for this image. Thus the image is entirely Ha light, and only Ha.
Two things to things to point out: the mono mod of the sensor opens up a while new world of resolution as all the pixels are gathering light all the time (not like in the colour matrix covered sensor). There is a full 12.2 megapixel resolution capability. The level of detail is thus substantially better than the colour camera can produce. Secondly, the image content seems very different because the colour camera cannot 'see' what the mono camera with the Ha narrowband filter can reveal.

Eta Carinae, Full Spectrum Colour, cooled 450D on a Skywatcher MN190:

http://www.astrobin.com/full/185615/0/

Eta Carinae, Full Spectrum Mono, cooled 450D (with Ha filter) on Skywatcher MN190:

http://www.astrobin.com/full/243048/0/

Just a final word on cooling and dark noise. Canon 450Ds have been shown in testing (by rcheshire (Rowland), and myself) to best perform when the sensor has been cooled close to 0C. There is not alot more to be gained by going sub zero, and I have tested down to -15C. At 0C the Darks and Bias Frames become almost identical. A Dark (or Master Dark) library can then be easily created and negate the need to shoot them each time or track the temperature they were shot at. Obviously you still stack with settings to eliminate hot pixels, etc. I should add that most Canon DSLRs return similiar results when cooled to these temperatures. Cooling really works, but brings you into the world of condensation management - another topic for later.

rmuhlack (Richard) · #7 29-03-2016, 01:15 PM

Quote:

Originally Posted by glend

Just a final word on cooling and dark noise. Canon 450Ds have been shown in testing (by rcheshire (Rowland), and myself) to best perform when the sensor has been cooled close to 0C. There is not alot more to be gained by going sub zero, and I have tested down to -15C. At 0C the Darks and Bias Frames become almost identical. A Dark (or Master Dark) library can then be easily created and negate the need to shoot them each time or track the temperature they were shot at. Obviously you still stack with settings to eliminate hot pixels, etc. I should add that most Canon DSLRs return similiar results when cooled to these temperatures. Cooling really works, but brings you into the world of condensation management - another topic for later.

My own testing (with both a cooled 1000D and a cooled 450D) concur with this. To manage condensation I eventually settled on a regulated sensor temp of 10C. This reduced dark noise considerably compared with uncooled operation and substantially improved SNR, with regulated cooling allowing the use of dark and bias masters libraries that were temperature matched.

glend (Glen) · #8 29-03-2016, 01:31 PM

Yes absolutely you need to be able to regulate the TEC to achieve your setpoint temperature, and that means some sort of controller is required as well. Rowland's thread gives a couple of examples that work well. I went with the external box that attaches to my pier, which contains a PWM to power the TEC and a temperature controller with a relay to switch it on and off. Cost almost nothing to build. I generally keep my setpoint at 2C, as TEC overrun will take it just beyond that before it switches back on. Condensation is managed by enclosing the camera in an argon filled ziplock bag (although any dry gas that does not transport moisture would work); sealing is a chore when installing the bag, but it only has to hold the gas inside not allow positive pressure, and argon is heavier than air anyway so it tends to stay in vessels. It only has to stay in there for overnight runs and I can top it up the next night. I run dessicant packs inside the bag but they never need changing. I get condensation on the outside of the bag but never inside. I've tried other approaches but keep coming back to the argon bag.