Digital Spectroheliograph (SHG)

[Merlin66]

I'm currently testing a digital SHG. This is based on a Carton 100mm f13 objective and a 60mm f13 Littrow spectrograph fitted with a 1200 l/mm blazed grating.
An ASI 174MM USB3 fast frame camera (laptop=MSI C, 16Gb memory, 500Gb SSD) collects the spectral data. Tests so far indicate a >400fps with a ROI 1930 x 48 pixel.
The basic design was evaluated using the SimSpec SHG spreadsheet
This SHG should have the ability to record solar images at any visible wavelength with a bandwidth of 0.24A.:eyepop:

My recent book "Imaging Sunlight" gives all the necessary info.

[algwat]

Ooohh , that looks like a beauty.

Please post some test image when you get then.
And one of all the insides in position , I think I just found my ten year project...? I better get the book then ...8)

On slightly different tack, i volunteer at science works regularly during the days, and on some of the Discovery the night sky. This would make a really good technical demonstration during the day time.

kind regards, Alan

[Merlin66]

Alan et al,
Hmmm
There seems to be an "issue" with the ASI 174 working in this application.
There appears to be significant banding/ spectral ripples generated which have the potential to screw up the image extraction process.
The attached image shows the dark Ha absorption band (across the image) and the resonance bands above and below. The vertical lines are artefacts from the slit plate.
Could be a show stopper.

[Derek Klepp]

Looks good Ken.

[Merlin66]

Managed to extract a frame from the SHG Ha AVI file and calibrate it.
The dispersion is 0.7A/pixel (Exactly as predicted by the SimSpec SHG spreadsheet!)
The ripples are almost exactly 2A separation and seem to present a +/-2.6% variation.

[Maurice]

Hi Ken

Are the ripples repeatable spatially from frame to frame?
If yes, then a simple flat frame could solve your problem.
If no, then is the ripple frequency constant?
If yes, then you could run a FFT over the data & then 'mask' the offending frequency in the real part of the transform, then perform an inverse FFT to rebuild your image without the ripple.
If no.... well then it gets interesting....

cheers
Maurice

[Merlin66]

Maurice,
Not sure how to prepare video AVI flats and then apply them to an AVI file????

[Maurice]

Take a short AVI of an evenly illuminated area in the frequency that you are viewing with your scope. Try to obtain illumination values that are somewhere in the middle of your range.
Process this AVI in AS2! without aligning on anything - the resulting image will be your 'flat' frame.

Inspect the frame by stretching the low to high range to see if the ripple that you observe is exactly repeatable when comparing your flat to your light frame. If it is, then you can define this frame as your 'flat' for AS2! to apply to the AVI of your data.

It will be able to do this on a frame-by-frame basis so that you end up with LOTS of single frames, or I believe that you can apply the flat to the AVI and generate a new one that has been corrected.

Cheers
Maurice

[Maurice]

...looks like its even easier now, there is a 'create master flat' button in the 'Image calibration' drop down menu.....
Be sure to set the image stabilization control to 'surface' & not ' planet COG' when you generate the flat... then you can 'save master flat' from the drop down menu.

[Maurice]

...have a look here for more detailed (yet simple) instructions about what to do ...

http://www.schursastrophotography.com/solar/Articles/solararticle5.html

cheers
Maurice

[Maurice]

If the ripple frequency is constant, but the effect is displaced from frame to frame, then fft filtering can work.
Attached is a mars image from a few years ago.

'Original mars' is the frame as stacked & processed.

'Noisy mars' is the same frame with a fixed pattern noise applied to simulate your problem

'fft noisy mars' is the frequency map (real component) of the noisy image. You can clearly see the frequencies where the pattern resides (hot spots).

'masked fft frequencies' is the same map with the offending frequencies masked.

'fft filtered' is the inverse fft of the frequency map showing the reconstructed image with the masked frequencies. Its not perfect, but it is a big improvement on the 'noisy' image. With a little care in the masking applied, a better result can easily be obtained.

It is fairly close to the original, pre-noise image.
All processing done in IRIS.

cheers
Maurice

[Merlin66]

Maurice,
Thanks for the input/ comments.
I'm familiar with Chris S solar work but didn't think of AS2! for the possible processing.
The flat can't be obtained from a bright sky - still contains the spectrum.
A Quart Halogen is used in conventional spectroscopy, I'll give it a go.
Thanks.

[Merlin66]

Maurice, et al,
OK managed to obtain a "flat" AVI (I still need to spend time on analysis)
BUT I find AS!, FireCapture, SharpCap all need to stack the target AVI to apply the flats!!
I don't actually "stack" the spectral images from the SHG, just extract the data from each consecutive frame in the AVI and then sequentially lay them together to form the image.

[Maurice]

Hi Ken

If I recall correctly, it might be Registax or PIPP that allow you to save the filtered AVI.
I thought that AS2! did allow you to save the individual corrected frames. You could then make an AVI out of those?

cheers
Maurice

[Merlin66]

Maybe, but playing with 9000 individual frames is not something I look forward to.
(Checked with Torsten - re flats in FireCapture for SGH - no lucck.)

[Maurice]

Batch processing is easy for a sequence of images with many programs - IRIS is a good example; (no need to process an image at a time) and Virtual Dub is another.

You can apply a flat to each image then save the result consecutively and automatically. You can then use a series of images to make an AVI in Virtual Dub very easily.

Lots of ways to do this.

[Merlin66]

Alan,
The attached images show the layout of the components in the housing.
Any questions let me know.

[Merlin66]

Still testing...
Using the 6mm long Surplus Shed slit and double 2" diagonals.
First light Ha spectroheliogram...
This shows the AR2599 and the prom on the opposite edge...
Working at only 70fps with the ASI 174
Image lacks "depth"....
Still a long way to go.
Onwards and Upwards

[Merlin66]

Robin, the author of SharpCap is currently working on a version which will incorporate the processing of flats in the target AVI.....
Look forward to trying it soon.

[Merlin66]

Hmm
Besides the ripple/ banding issue with the ASI 174 the project success may be in doubt...
To achieve my "fast" requirement I need to operate the camera at around 400 fps. This in turn means that every individual frame must have an exposure of <1/400 sec.....
To achieve a "good to great" signal from the Ha absorption line I find the exposure needs to be probably more like 1/20 sec.

The "fast" SHG option therefore appears to be restricted to very large aperture systems.....
Onwards and Upwards

[Merlin66]

Just an update....
I think the faster frame rate may be solvable - using the horizontal ROI (1930 x 100) definitely seems to be capable of >400 fps. (Note: a vertical ROI doesn't seem to have the same improvement - only achieving +/- 100 fps)
But getting a workable exposure to match is a bit more problematic.
With single frames from a 8 bit AVI the density range (full well) is 0 -256 ADU.
With the local continuum over exposed (>256) the core of the Ha is only registering at 70 ADU.

Fred Veio made a nice illustration in his book "The Spectrohelioscope" which shows the depth of exposure needed for the various Ha features - it also gives an indication of the intensity variation which can be achieved...

A profile of the Ha line shows the issue.

I'll do further testing with various exposures at the higher frame rate.
If things improve I'll move onto the nodding mirror option to give scans of 4 secs.