6 replies to this topic

[mgermani]

Hi folks:

 

My apologies if this has been covered already, but I've been having some issues stacking images taken with my recently-acquired Svbony CLS filter clip-in filter. Briefly, it's installed in my unmodified Canon T3i and I've noticed the problem with both my 80mm achromat and my 50mm canon lens.

 

While I understand that red is greatly reduced when using a CLS filter, I seem to have a modest amount of red captured in my light frames:

 

Canon RAW file: https://www.dropbox....tf_cls.CR2?dl=0

JPG file: https://www.dropbox....tf_cls.jpg?dl=0

 

But by the time it is stacked (I'm using SiriL) almost all the red data is gone, and it's reduced to just pin-point stars, even when stretched. I have good data in my blue and green channels, but I can't balance the colour without the red data. Photometric colour calibration sometimes helps create usable images, but I'm wondering why my red data has gone? Is it all noise?

 

Stacked FIT file, no processing: https://www.dropbox....etched.fit?dl=0

 

The CLS filter seems to be increasing the chromatic fringing around stars (on my achromat) and I'm often having to deal with plenty of colourful gradients. Is my cheap CLS filter to blame? Is the processing/SiriL to blame? Is this compression of red data totally normal? I'm kind of baffled. I'm including a link to one of my flats, below, taken with the filter in:

 

Canon RAW file: https://www.dropbox....at_cls.CR2?dl=0

JPG file: https://www.dropbox....at_cls.jpg?dl=0

 

Should I stop using the filter, and if so, should I switch to another filter? Modifying my camera is not really an option now, but perhaps down the road.

 

I'd appreciate any advice/input. Thanks!

 

 

[klaussius]

While I understand that red is greatly reduced when using a CLS filter, I seem to have a modest amount of red captured in my light frames:
 
Canon RAW file: https://www.dropbox....tf_cls.CR2?dl=0
JPG file: https://www.dropbox....tf_cls.jpg?dl=0
 
But by the time it is stacked (I'm using SiriL) almost all the red data is gone, and it's reduced to just pin-point stars, even when stretched. I have good data in my blue and green channels, but I can't balance the colour without the red data. Photometric colour calibration sometimes helps create usable images, but I'm wondering why my red data has gone? Is it all noise?

I don't know about SiriL, but if I run that single CR2 file through my own workflow it looks alright:



I see a healthy amount of red there.

What may be happening is that, during demosaicing, the weak red signal in comparison to the green and blue can get clipped on the low end. The CLS filter has a very narrow bandpass in red, and the histogram tends to lean heavily to the left. Part of the demosaicing process involves a color matrix to recover some color saturation, since the bayer filter's channels overlap significantly in their passband. With a heavily filtered red channel, that tends to clip the red signal on the low end and make it all black.

But that only happens when there's a strong signal in the other channels. Background extraction should be done on the raw linear data before applying the color matrix, and will remove that blue cast (light pollution). Then the red data will be safe.

That's what I do, and what I did to get the above image. I'm not sure how one does that in SiriL. But your stacked FITS looks alright as well so I'd say you're doing something wrong to the data during postprocessing:


 

The CLS filter seems to be increasing the chromatic fringing around stars (on my achromat) and I'm often having to deal with plenty of colourful gradients. Is my cheap CLS filter to blame?

Definitely not. About fringing, Achromats have that drawback. It may only be more noticeable since the CLS has a wider passband in blue, where the achromat has the strongest CA.

About the color gradients, that's probably light pollution. A CLS filter will decrease it but by no means fully remove it.
 

Is this compression of red data totally normal?

Yes, it's normal. The CLS passband in the red is very narrow, almost like a NB Ha filter. So your red signal is very very clean.

The signal seems to disappears because the signal is clipped on the low end when applying color correction on the raw data before background extraction as I mentioned earlier, but the data is there and can be extracted with proper processing.
 

Should I stop using the filter, and if so, should I switch to another filter?

For Andromeda, I think you shouldn't use it. But not due to any of this, but rather because a CLS filter is designed for emission nebulae. It won't really help that much for broadband targets like galaxies and reflection nebulae, as it's removing as much signal as it is removing light pollution. Keep using the filter on emission nebulae, but I would not recommend you use it on Andromeda.
 

Modifying my camera is not really an option now, but perhaps down the road.

Modifying your camera would give you much stronger red signal. It would be worth it.

[mgermani]

Thanks Klaussius!

 

I took some of your advice and experimented with applying background extraction BEFORE photometric calibration, and this helped reduce gradients in my images with or without the CLS filter, so that's helpful.

 

I also tried stacking a series of images I took with my 50mm camera lens and the CLS filter without using flats, and I was able to remove the gradients. Flats with CLS filter + 50mm lens = bad stacks, it seems. Sometimes I read people saying clip-in CLS filters don't work with wide-angle lenses, but I guess I didn't consider 50mm wide-angle. Maybe it's borderline.

 

I recently popped-out the CLS filter and managed an hour and a half of Andromeda (under a near-full moon) https://www.astrobin.com/fkof88 - which I'm thinking is heading in the right direction. I used your suggestions for processing order, which definitely helped. Some of the feedback I received on this image included the suggestion of the Baader Semi-APO filter, which is based on a CLS substrate. A lot of people seem to like this filter a lot for visual, but I don't see many people imaging with it, and it wouldn't be compatible with future lens & apochromat purchases. I had been eyeing the Optolong L-Pro, as it seems to be less dramatic in terms of colour balance than my Svbony. Thoughts?

 

Thanks again, I really appreciate all the advice!

 

Cheers,

Mark

[Fivel]

Hi mgermani.

Two things I can add to this discussion.

1. I have the same Svbony filter and an OSC ZWO camera on my 80mm scope. I do not seem to lose any reds, in fact, I get better that without it.

2. As to the star color fringing, I believe that PS has a defringing tool where you  can select the colors to minimize.

I use Affinity Photo and it has that tool. I get good results using that. Of course I start with a decent image since my AT 80EDT is a triplet Apochromat. I also use Siril for pre and post processing. Sometimes I do not care for the colors I get with the Photometric tool, but that is just how it works for me.

My basic Siril PP steps starts with Background extraction then Background Neutralization followed by (if needed) Green reduction.

3.Your last image (astrobin) seems a bit red in the core. Are you looking at the Histogram colors when post processing? 

4. One more thought-does your camera have an IR blocking filter on it? The imaging experts in my club told me not to use both at the same time, since the CLS does block some of that range. If you get the IR filter removed, it gives you better flexibility for when you image with and without the CLS filter.

As for a better (more expensive) light pollution filter, the Optolong L-Pro is a popular one within my club.

For me, I am pleased with the results with the Svbony.

 

Keep trying the previous suggestions and use the CLS as recommended.

 

Stay healthy

 

Fivel

[klaussius]

Glad you got everything working.

 

The L-Pro IMHO is probably the best fit for galaxies. I have no direct experience with it, but I'm actually planning to acquire one soon. I know a guy that did Sculptor almost at the same time I did Sculptor, with similar equipment and sky, the only difference being the filter (he used an L-Pro), and the results were actually quite improved.

 

Still, broadband filters aren't very aggressive and their benefit will be mild, if they provide a benefit. Whether they do or not depends a lot on your skies.

 

Take some time to inspect the origin of the light pollution in your area, which kind of lighting produces it, and their spectrum. Compare the spectrum of the light pollution to the bandpass of the filter to get an idea of whether it would help or not. If your light pollution is concentrated on the blocked bands, it may help. If you have broadband light pollution, however, it won't help as it will not increase contrast.

 

So YMMV with that filter.

[mgermani]

Thanks guys!

 

Glad to hear from somebody with the same filter, Fivel. I seem to be getting some healthy red levels, but they get crunched down during stacking so I wasn't sure what was going on there. Still not sure how I can avoid that, unless I stop using SiriL and switch to another stacking program.

 

I've been using Photoshop to remove some of the fringing, but it's not perfect, and it will remove colour from areas where I'd like to keep it. The fringing is not so bad without the CLS, though, especially if I stop the scope down (when imaging very bright stars).

 

Regarding the reddish Andromeda core on Astrobin, I noticed that on the majority of photos of Andromeda there is pale yellow or reddish hue to the core, so I left it in. I suppose I could remove it.

 

My camera is not yet modified, so it has the IR filter intact. It was my understanding that filters such as the CLS-CCD had the IR cut filter added so you could use it with a modified camera, but that the Svbony filter was just a plain CLS. At some point I do plan to modify my camera, but not just yet. That's why I'm figuring I'll stick with broadband filters until that point.

 

From my reading, my city (Vancouver, Canada) uses mostly high-pressure sodium lighting. I've seen the new LED lights go up around town, but only in a few areas, and not near me. I'll see about comparing emission graphs, but I'm thinking less aggressive might make processing easier for me, for now.

 

Thanks again for all the advice!

[mgermani]

Fivel - just a quick follow up on your mention of "background neutralization". I came home and had a hard time finding the function (I'm using 0.99.6) but managed to track it down under "color calibration" in the image processing menu. What a lifesaver! I'd been relying on platesolving and photometric calibration to tame the colours in my image, and this helps so much when that's not possible, or when I'm not happy with the results. Thanks!