Would appreciate some advice...

[ejp1684]

I recently bought an Altair Astro colour GPCAM, to do a bit of planetary imaging and start some basic spectroscopy. The planetary bit is going ok. I took a .ser video of Jupiter, ran it through PIPP and Registrax and produced what I think isn't a bad image for a first go.

I want to start getting some images using Star Analyzer 100. I've been using SharpCap to capture images, and on my first try I used the RAW8 option and saved the image of the star and its spectrum as a fits file. (The other options are RAW12, RGB24 and Mono08). I then tried to follow a tutorial for BASS to calibrate the spectrum, but when I imported the image of the spectrum it came up grey with lots of speckles on it. A search through these forums enlightened me on the subject of debayering. I found a link to IRIS which looked promising, so downloaded that and tried to do things with the image, but got a bit lost. I found from a text capture file that I need GRBG when I come to debayering. I tried that in DeepSkyStacker but the image stayed grey with the speckles.

I'm a bit tired of using all these different pieces of software. I just want to get the spectrum into BASS (or RSpec demo version) and move on a bit.

Am I doing this correctly, or have I missed something?

Eric.

[Pete Presland]

I cant help you with your Spectroscopy, but that is a fine Jupiter image!

[Oddsocks]

Hi, Eric,

I hoped someone with more experience would have replied by now but I can pass on what little I know, I have been taking spectrums for just a couple of years now though I have the advantage of coming from an astro imaging background going back to the days of film using hypered emulsions.

I have BAS but only little used, nearly all my spectrums are processed with RSPEC so I will just stick to an explanation with that.

The colour GPCAM is basically a video webcam that can also take long-time exposures but it is uncooled and therefore each single exposure will contain a lot of noise so to take a single frame exposure of a spectrum will produce a noisy image where the individual emission and absorption lines will be difficult to differentiate unless additional processing is carried out, and this usually requires extra software and processing steps.

The easiest way to get started is to capture a colour video file in AVI format, just the same as you did for the planetary video but select AVI and not SER. 

RSPEC does not need a colour file, it processes and calibrates only in mono and then produces a synthesised colour image once you calibrate the mono waveform, but it can not use a video file that is not debayered first so it is easiest to capture the video as debayered colour AVI, import it into RSPEC where it will strip out the colour automatically and show only a monochrome image to begin with and then produce the full colour synthesised spectrum once you calibrate the waveform against a known reference.

In SharpCap make sure you capture video in deBayered AVI mode, not SER, orientate the camera and the telescope so that as near as possible the zero order (star) and full spectrum is spread out horizontally or vertically and the zero order (star) and full spectrum is contained within the frame, you may need to add or remove spacers between the SA100 and the camera in order to achieve this and the best spectrums will be those that use the maximum number of pixels available, if the zero order and spectrum only occupy half the frame then the resolution of the spectrum will be poor. Adding spacers widens the spectrum, removing spacers shortens the spectrum.

You do not need super fast frame rates so set the camera gain quite low and adjust the exposure time to suit the target star and spectrum, a frame rate of less than ten per second is more than enough, some of my best spectrums in video mode have been with just 2.5 frames per second, then capture the AVI file with a run long enough for at least 250 frames. Obviously, if you are imaging a bright target then you can up the frame rate and capture a shorter run but in any case try to keep the gain low to minimise the noise.

Now you have your video file which is already debayered (in colour) and in AVI format, just open this in RSPEC as a video file, use the rotate image slider to position the zero order star on the left and the spectrum on the right and as horizontal as possible, then drag the two measuring frame lines up the image and place one just above and the other just below the wanted spectrum avoiding other nearby stars or hot pixels.

RSPEC will then sample each frame in the video file and average the spectrum to calculate the waveform, you can select how many frames to average, between 50 and 150 is a good place to start, this part of the program does the same sort of thing that you do in planetary imaging where multiple frames are combined with each other to produce a high signal-to-noise ratio stacked and averaged image, in RSPEC instead of stacking and averaging a whole two dimensional image it stacks and averages a sample through the spectrum of each video frame as defined by the user positioned upper and lower frame lines to produce an averaged waveform with a high SNR.

All that is left is to calibrate the image using the calibrate tool, the zero order, wavelength = 0, is the star on the left and this correspond to the first high peak on the left side of the waveform, then choose one of the easily identified and known absorption lines in the spectrum, click on it and then enter the wavelength, now RSPEC will display the full colour synthesised spectrum for you to analyse, export to photoshop or whatever.

It is worth taking the time to watch the tutorial videos on the RSPEC website, there are a whole bunch of short tutorials covering each step in the process and one longer workshop video which goes over all the same ground again but with added interruptions from the audience! There are also samples of captured video and captured single image spectrums that you can download and practice using the RSPEC software.

When you run out of bright stars that can be captured with a video file then you need to begin looking at single or multiple long-time exposures, capturing in RAW FIT format, but for these you need to add dark and bias frames, these need to be debayered and stacked in separate software such as DSS, the output image is in colour TIF format and you can import this into RSPEC or BAS as a single image, once again, this will appear as a mono image until you have calibrated the zero order (star) and first reference line in the spectrum, only then will the image display as a full colour synthesised spectrum. But this is a whole separate topic and is part of normal astro imaging which I will not attempt to describe here.

So far my most difficult spectrum involved a mag 14 star and for this I was using a cooled mono camera and needed forty exposures at twenty minutes each to get a good result, added to this were the darks and bias frames, overall this took three consecutive night in the observatory to capture the data. The first bright star spectrums that I took with a mono video camera took just ten minutes and required no calibration frames so this is where I would suggest you start.

Unfortunately, where image processing is concerned, unless you buy a commercial software package you will not get all the process steps you need in a single freeware package, you need several.

For simple colour conversion of a single image in RAW FIT format then I would recommend the free ESA/ESO/NASA FITS liberator 3 software package, it's only function is to import, debayer if colour, and display FIT files, you can save the image as a colour TIF file to import into RSPEC or BAS or Photoshop etc and it can be found on the Hubble website here:

https://www.spacetelescope.org/projects/fits_liberator/

Hope the above may be of use until someone else with more experience can help, also when discussing image problems it is better if you can post the image with your question, it makes it so much easier to form a diagnosis since textural descriptions can be open to varying interpretation.

William.

[Merlin66]

Eric,

I think you should be able to process your spectral image along the same lines as the process you used for Jupiter. Stack the frames in Registax and save the final image as a .bmp or TIFF.

This can then be opened directly inBASS Project and yo can follow the BASS tutorial to remove the background and work on calibration.

Hope this helps

 

[ejp1684]

Thank you both so much for your advice. William, your detailed explanation has helped greatly, and has answered some other questions I hadn't got around to asking!

Eric.

[ejp1684]

At last a few hours of clear skies the other evening, well, apart from the clouds appearing every now and then. By 11.30pm they had decided to stop proceedings. But I did manage to get a few spectra even though the moon was up, and I'm surrounded by houses and flats giving off heat into the cold night.

I tried to process avi files using RSpec, but I was disappointed with the results. The curves were very 'curved' if you know what I mean. No nice dips, more a sort of gentle undulation. So I tried BASS and obtained the following result for Merak, an A1V star which was the closest I could get to an A0V. I found the instructions for BASS quite straightforward.

The camera isn't very good below 400nm or above 650nm, but as this is just to try and get everything to work I'm just pleased something's come out reasonably close. It took about six goes at doing the Instrument Calibration curve to get it to look like this. I'd appreciate any comments and advice, and if I'm on the right track then I'll have a go at some more spectra I took that night.

Eric.

[Merlin66]

Eric,

Glad you managed to get a reasonable outcome from BASS project - as you say it's pretty straightforward to use....

I think you're off to a good start, focus on the spectrum not the zero order star image and get a bit more practise, practise....

What's the spacing from the SA100 to the camera chip?

Ken

 

[ejp1684]

Thanks for your reply. Yes, there seems to be as much 'art' to this as 'science'! I've tried to focus on the spectrum, but find it a bit tricky as some parts of the spectrum look clear while others don't. For example, I can get the right hand side looking good, but then the left hand side isn't as clear. In particular, some stars give a sort of fish tail in the blue. Hope that shows in the picture.

I'm wondering if I remove the UV-IR filter if that will give me any improvement in the IR region so I can get the H alpha line in. At the moment it doesn't register with the filter cutting out at 650nm.

The SA100 is 66mm from the chip. This produces the star near the left, and the right hand side of the spectrum near the right of the image, so both the star and the spectrum take up about 90% of the image. This gives a nm/pixel of 0.58 which I think is a bit low. I should say I'm not expecting great results as the camera is a general one and I use it for planetary imaging as well. I realise a mono camera would be better. Here are the response curves for this one.

 

Hoping for some clear skies tonight without the moon and practice some more.

Eric.

[Merlin66]

Eric,

The "fish-tail" is not unusual..... this is caused by chromatic aberrations in your system. By re-focusing on various regions of the spectrum you'll see the "necking" shift position within the spectrum.

I would have thought the UV-IR filter would have cut closer to 700nm rather than 650nm.

Practise.....

Onwards and Upwards.

 

[ejp1684]

Thanks for the reassurance about the "fish tail". And as soon as we get some more clear skies I will certainly put in some more practice.

Eric.

[robin_astro]

Hi Eric,

What scope are you using?  One which gives a small star image will give the sharpest spectrum but too low focal ratio gives aberrations with the simple Star Analyser setup which will ultimately limit the resolution so there is a bit of a compromise there, particularly with larger apertures. As you have seen, setups with chromatic aberrations (achromatic refractors in particular but also some focal reducers) cause focus issues, particularly at the blue end.  

Concerning the optimum dispersion, you can use the calculator to give you an idea of what should work best

http://www.patonhawksley.co.uk/calculator/

because it is a colour camera the effective pixel size is larger than the actual pixel size so within the "OK" band of spacings the calculator suggests, try to aim for the  higher dispersion end (larger spacing,lower A/pixel), provided you can still fit the spectrum and zero order in the camera field. (The calculator will warn you about this)

Cheers

Robin

[ejp1684]

Hi Robin, thanks for your advice.

The scope is a Celestron Evo 8 with F6.3 reducer. I tried it without the reducer at f10 but still got the focus issue at the blue end, so the reducer didn't seem to be contributing to the effect. It seems worse on some stars than others.

I've used the calculator you mentioned, but hadn't realised the impact of the colour part on the effective pixel size. When using the grating at about 66mm from the chip I get 5.8 A/pixel with both the zero order and the spectrum in view. That's with the IR-UV filter in place, gives from 380-650nm.

Eric.

[robin_astro]

Hi Eric,

That's interesting. I would not normally expect to see such significant chromatic aberrations in an SCT.  Here is an example from my C11 with focal reducer for comparison.  Perhaps it was atmospheric conditions.   I suggest moving the focus a bit towards the blue end which might bring the Balmer lines blueward of H beta into focus.

Cheers

Robin