Askar C2 Filter Halo Tests with NIR sensitive OSC cameras Generic equipment discussions · astrovic47 · ... · 15 · 548 · 12

Discussion

As can be seen, the SV220 filter is doing a good job preventing halos with the exception of the bright star Alderamin, but I think that is acceptable given the cost of the filter. I should also note that the type of halo it shows (a thin ring) to me, is kind of pleasant? It almost adds the appearance of a lens flare which I personally don’t mind.

Now compare the background stars in the Askar C2 image. While there isn’t the same thin ring halo on Alderamin, there are definitely some unique halos around some, but not all, of the stars.
Unfortunately, I find these donut looking halos unappealing, and they also provide a challenge when using star removal with Starnet. 


Is it an NIR leak?

When I first ran into this problem a few months ago, one person had suggested to me that the filter may be ‘leaking’ NIR light. I unfortunately don’t own or know anyone that has a spectrometer that can verify this, but by drawing some conclusions from the results, it does seem like it’s the culprit.When qualitatively looking at the stars with the most severe halos (See number 1 above) I notice that the star appears orange in colour. This would suggest that it is a cooler star, and possibly emitting more light in the NIR.

If I review the stars in Stellarium, I can use the Colour Index (B-V) value as reference, where negative numbers indicate hotter stars, and positive numbers indicating cooler stars. Star 1 has a value of 2.30 indicating a cooler star emitting light at longer wavelengths. Star 2 has a value of 0.48, still cooler but not as cool as 1. Finally Star 3 has a value of 0.29. At least for this small dataset, a trend can be seen where stronger halos are proportional to the (B-V) value. The higher the (positive) number, the stronger the halo appears. Take this with a grain of salt as I am by no means 100% of my assumptions, and the data set is extremely small. 


NIR response of the Camera.

Below is the published spectral curve of the Touptek ATR585C using the Sony IMX585 sensor. Beyond 850nm the bayer matrix appears irrelevant and the response in all 3 colour channels is nearly the same.
If I review the halos in each colour channel separately, I don’t discern any noticeable difference which further pushes me in the direction of NIR being responsible for the halos.


Optical Density.

A filter specification that I had never heard about until I started looking into the difference between the SV220 and the C2 (aside from the band passes and band widths) was that the SV220 has an OD value of ≥5. While the C2 has an OD ≥3.Again, I’m by no means an expert on this property, but fundamentally it is the ability for the filter to reduce or attenuate the incoming wavelengths it’s not designed to pass.
In the case of the two tested filters between roughly 300-1000nm. From some quick googling, the difference in OD3 to OD5 is 1,000x and 100,000x respectively. 


Is it just a bad copy of the filter?

When I was originally looking at this issue with a different camera (Player One Neptune CII) that also has increased NIR response, I reached out to Askar to see if they had any suggestions. They simply offered to replace the filter if I sent it back to them. Luckily my local astronomy store was kind enough to also reach out to them and agreed to do the exchange locally, which I appreciated.Upon receiving the new filter, I noticed the same issues but decided to keep it anyways because…


Solution
A simple solution one person had suggested to me (assuming the problem was NIR) was to simply add an additional UV/IR cut filter and see what happens. And as the results show, it fixes the problem. While not ideal if you can only add a single filter to your imaging train, I luckily can put it on the nosepiece of my field flattener and enjoy using the filter in my drawer without halos.[b] [/b]

Zoomed in on the bubble nebula, despite the poor focus on my part of the image on the right, the donut halos are eliminated by the addition of the UV/IR filter. Lower image pair shows the dramatic difference adding a UV/IR filter has made on bright stars above the star cluster.:




Conclusion

So for those who made it this far, thank you. Hopefully most of what I said makes sense. Based on my tests it seems to me that Optical Density (OD) may be a critical specification when looking at narrowband filters for OSC cameras that are sensitive to NIR and don’t have a built-in UV/IR glass over the sensor. I’m unsure what OD value some of the higher end filters have, but admittedly didn’t spend a lot of time looking.

My best guess is that the Askar C2 being one of the cheapest dual narrowband filters on the market, corners have to be cut somewhere. One of them I suspect is using the lower OD value. Assuming that their published spectrum of the filter is 100% accurate. Their D1/D2 line of filters indicate that they are OD5. It would be interesting to see if one of these filters shows the same halos or not but at nearly 3 times the price it’s not within my budget.

But as I said at the very beginning, the addition of a cheap UV/IR filter is an economical way to ‘enhance’ the filter making it at least as good as a more expensive budget filter when it comes to rejecting these supposed NIR halos.

Tony Gondola:
Thanks for taking such a detailed look at this. I'm wondering if this is a problem with reflecting optical systems as well? I would suspect that it might still be in the far IR, just not to the same extent.

I'm glad you found it insightful, I forgot to add a few additional photos showing the effect of adding the UV/IR with the C2. 

To your question, I believe the halo problem is, or at least majority due to the filter and not the telescope design, assuming the issue is NIR leak. This is shown by adding the UV/IR filter in front of the C2 and the elimination of the halos. According to the published graph of the filter, it should be cutting out everything after sulfur out to 1000nm like the SVbony filter.

But can also see that the halos may be worse because of the refractor as I don't think it's possible to simultaneously have NIR and VIS light at the same focus? Would be interesting to see the results with the same setup but with a reflector instead.

@Phil Eng you can definitely 'stack' filters as I have to do when using the C2. For my setup the narrowband filters get swapped in my filter drawer that is attached the camera. Then on the nosepiece of my field flattener is where I put the UV/IR cut filter.

A thing to keep in mind is that while my method works, adding additional elements in the imaging train can also lead to some other problems like reflections between components and a reduction in the overall signal reaching the sensor (although this may be negligible?). 

If you have the time and are able to post an image with the C2 (or C1) I'd be interested to see if you also see halos. I looked at the spectral graph for the 533MC sensor (below) and at 850nm the response is around 35% whereas on the 585 sensor (second image below) it's around 75%. Again assuming the problem with the C1/C2 filters is NIR leak, I wonder if the lower NIR sensitivity of the 533 sensor is enough to make the halos much less severe. 

IMX533

IMX585

Hopefully the link to the Google drive will allow folks to down load the masters, if not let me know... Also if I need to convert these to fit files... Let me know that as well

I did download both files but I'm not seeing a way to open them in Siril unfortunately, looks like they may have disabled that feature in their more recent releases? 

If possible, if you can export in either .fit or .tif would work for this purpose.

@Phil Eng thank you for the images, they are definitely adding some interesting and useful information to these tests.

I ran both your C1 and C2 images through the same simple processing as described in my original post to be consistent. I also searched astrobin for an amazing image of the Wizard nebula that contained more or less the same field of view, shot with a 533MC pro and a doublet refractor which led me to the following image from Anthony MICHEL.
https://www.astrobin.com/mbgf1j/0/

I only really found one star exhibiting the same extreme donut halos as in my tests, and even more interesting, it was really only noticeable on the C2 image. 

This group of stars is located above the hat of the wizard and the brightest star in the images is at the centre of a dust cloud. There is one star in the area (A) that I've annotated in the below image that when compared to the reference image, should not be that big especially in the C2 image.

Unfortunately Stellarium doesn't have this star as clickable therefore I can't determine it's (B-V) index. But if we simply assume based on star colour, then we can see that it's subjectively one of the more orange-red stars in the field. And following the original thought that it's being caused by NIR, it certainly adds support.

So why is the halo either not present, or much less noticeable on the C1? I can't say for certain maybe it's more difficult to produce the C2, or maybe  it just wasn't manufactured as precise as the C1? 

Another interesting difference between the C1 and C2 is that when comparing the halo in the individual RGB channels, on the C2 there isn't much difference (same as I observed) but on the C1, only the red channel really shows a halo. As to why that is, that might be reaching the limit of my speculation 

In any case, I would still say do some tests by adding a UV/IR filter and see how much of a difference it makes in the stars. And I suspect that if I image this region with my setup I'll see an even stronger halo around 'star A'. Perhaps I'll try on my next available clear night.

Great write up of your observations.  It has halted me procuring the C1/C2 filters pending further research.  The vendors often post filter bandpass characteristics, but frequently leave off the NIR.  Were you able to find bandpass curves that included filter NIR characteristics?

I own and use an IDAS GNB filter which does have known NIR characteristics.  I do not notice the “ringed stars” with it that you observe with the C2 filter, but I shall go back and check.  I have an inage and subs of the M52, Bubble, and Lobster Claw Region that may contain the same stars.

Thanks again for reporting your findings.

Being short on time and only having very quickly read through this thread, I think you are on the right track as to the cause of these tight halos.  Please see my experience in which I report in this post: https://astrob.in/cd9mc1/C/.  Note that my description also relates to a Forum post I made on this issue.  But all that the information that I see you provided here is consistent with what you are demonstrating so well in this thread.

Aside from the IR signal, likely the reason that all stars or most bright stars do not show these bright donut halos is that many are not strong emitters of NIR/IR.  Note the uncanny correlation of stars with these distinct donuts, regardless of brightness, to Long Period Variable (LPV) stars.  In fact, when I first made the correlation in my images from short f.l. lenses, I thought I had found a great way to identify LPV!  Alas, it may be, but for all the wrong reasons!  As you will see, the contribution to this artifact from some filters seems to be limited to filters manufactured with interference methods.  It is known that interference filters are very dependent on the angle of incidence of the light onto the filter surface.  So it is just that these filters were not designed to work at short focal lenghts and it seems to be the IR light that escapes the filtration effect easiest.

For me this was even with cameras with poor NIR sensitivity.  How can this be?  Well I urge you to get on Simbad, call up images of your fields choosing one of the NIR cameras that are available and you will see that uniquely for these stars, that the NIR/IR emissions from these stars is orders of magnitude greater than in the visible bands!  Some stars, essentially invisible in the visible can become the brightest stars in NIR/IR!  So, even with a camera not designed to see NIR, if you give it 100x, or more, the amount of NIR light, even non-IR optimized cameras will see it and make those halos.  Also, consider the NIR halos occur because NIR is either not well focussed by the optic, basically because the optic is not designed to do so.  My solution with those optics that give me the optic is to ditch the L-Pro (which I assumed was serving as my IR/UV cut filter) and bought a dedicated IR/UV cut.  As you state, that was the solution and worked very well.  On any other optic, my L-pro is still my go to.  

Also be aware that IR leaking through any filter will affect star color when using a OSC camera.  (It may also do so with RGB, but don't have the time to think through that problem, though the color filters used during RGB capture are probably much better at IR/UV cutting than the Bayer layer.)  The reason is that the Bayer filter layer on a OSC camera is designed to work on a camera that has the IR cut filter as part of the system.  The R, G, B, filters in the Bayer layer are perfectly transparent to IR.  I use a camera for IR work that has the extended NIR range, but also has a Bayer layer.  When I image only in IR (using an edge filter that removes all vis light), my images are colorless.  With my defective fast f.l. setups that lead me to investigate the issue, I noted that stars with high IR signals appeared off-color.  Depending on the amount of IR light leak, the result will be a color-skewed image of many stars.  

In any case, thanks for bringing this up again, since I think it could help those who may be struggling with this issue with this sort of filter and fast optics, which is more popular than ever.  As long as this thread is visible....   Unfortunately, all these tidbits of wisdom on a form such as this seem to fall away with time and get lost much too soon!