UV/IR cut filter or not with an OSC camera in a newtonian?

[Pitch Black Skies]

4 minutes ago, ONIKKINEN said:

I suppose the camera doesn't care whether the IR is in the coma corrector or not, as long as its not getting all the way to the sensor so not important how far the filter is. Assuming no vignetting of course, but it is a 2 inch filter so not an issue.

Thanks, just curious for myself too as I am in the process of ordering one. I have the 533MC and luckily a 1.25" filter completely covers it's small sensor.

[vlaiv]

40 minutes ago, Pitch Black Skies said:

I think I'm confusing chromatic aberration with star bloating and halo's, as seen in John's image.

Ok, I see it now.

Pink circles are due to reflections. Anti reflection coatings are usually optimized for visual part of spectrum and will now work well on say IR part of spectrum.

These are however not chromatic aberration - they are simply filter reflections (or maybe other optical element) - one that has coatings for visual part of spectrum and these reflections are in IR part.

[Pitch Black Skies]

25 minutes ago, vlaiv said:

Not sure what image are you taking about.

Star bloat can be due to many things - one of those is chromatic aberration.

Chromatic aberration is due to refractor lens not bringing all wavelengths of light to same focus - lens has shorter focal length for some wavelengths then for other. Effect is very recognizable as it presents most times as blue (sometimes red) halo around stars (especially white/bluish ones).

Take for example this image of M45:

There is no mistake - stars have blue (and a bit red closer to core) halo.

This does not happen with mirror telescopes.

It does not happen with good APO scopes neither unless you let in IR and UV part of the spectrum. ED doublet and Triplet scopes are optimized to bring to common focus two or three wavelengths in visible part of spectrum. Further away from visible part of spectrum - bigger deviation of focus.

If you don't remove these (essentially not important wavelengths as far as color and visual representation of target are concerned) - they can cause star bloat. Depending on camera used and its QE - star bloat can have color or be white (if camera is equally sensitive for each color in far IR).

This does not happen with mirror telescopes. Even using corrective refractive elements in optical path don't usually lead to this. Most corrective elements don't disperse light very much in order to introduce significant amount of chromatic aberration.

 

 

Thanks vlaiv and apologies, was referring to CCD-Freaks first pic. It's probably somehow connected his fast focal ratio? Probably nothing to worry about at F4/F5?

Edit: just seeing your above post now. That is very interesting and we'll spotted.

Edited December 16, 2021 by Pitch Black Skies

[ONIKKINEN]

18 minutes ago, vlaiv said:

Ok, I see it now.

Pink circles are due to reflections. Anti reflection coatings are usually optimized for visual part of spectrum and will now work well on say IR part of spectrum.

These are however not chromatic aberration - they are simply filter reflections (or maybe other optical element) - one that has coatings for visual part of spectrum and these reflections are in IR part.

How can you tell between the 2? Looks a lot like CA but just red instead of the usual blue. Also, if it is just reflections this would mean that some correctors might not have this issue if they are better coated with infrared in mind. I doubt the latter is true for me, since the Maxfield isn't that great of a corrector anyway.

[vlaiv]

13 minutes ago, ONIKKINEN said:

How can you tell between the 2? Looks a lot like CA but just red instead of the usual blue. Also, if it is just reflections this would mean that some correctors might not have this issue if they are better coated with infrared in mind. I doubt the latter is true for me, since the Maxfield isn't that great of a corrector anyway.

These are probably due to filter or sensor cover glass (multiple bouncing close to sensor).

They are very common on fast mirror systems like Tak Epsilon 130 and alike - look here:

It is the same type reflection - but this time being "pink" - due to IR light involved. Image discussed was taken with ASI533 and SharpStar F/2.8 hyperbolic newtonian (similar to Tak epsilon).

If you look at ASI533 QE curve and in particular IR part of spectrum (700nm and above) - you'll see that red is still a bit dominant over blue and green:

Same amount of IR light is likely to produce pinkish color.

 

[ONIKKINEN]

On 15/12/2021 at 14:42, ONIKKINEN said:

IF i were able to free the filter from its prison.

I got the filter out, with some out of the box thinking and trickery. I screwed another adapter on to the threads of the stuck filter holder to act as a better gripping surface and then put my 60mm guidescopes holder rings around that adapter. Tightened as tight as i could with pliers and it cracked open after a few tries. The nylon tipped screwheads also did not mark or noticeably deform either of the adapters or threads, turns out this is a pretty effective method of removing stuck things and much better than strap wrenches that i have tried! I am never screwing a single piece of kit together without some sort of lubricant in the treads again...

Now just to wait for a gap in the clouds to test a filter/no filter difference with my corrector. Think ill shoot the double cluster, which doesn't take a long integration, has M type red/IR stars well as very bright B type blue/UV stars. Should bring out any internal reflections or possible CA out nicely.

On 16/12/2021 at 23:53, vlaiv said:

This does not happen with mirror telescopes. Even using corrective refractive elements in optical path don't usually lead to this. Most corrective elements don't disperse light very much in order to introduce significant amount of chromatic aberration.

Usually, yes, but maybe sometimes? Attached is the spot diagram of my corrector on an F5 system, so my F4.5 primary should be slightly worse still. What do you make of this?

It looks like not all colours are treated the same out of axis, so a bit like CA? Although CA would be on axis as well of course. I am not educated enough to guess how much refraction must take place before CA becomes noticeable, but you might be... The first lens element on the coma corrector is actually very curved, not unlike maybe a finderscope lens? It is about 130mm from the focal point and doesn't advertise ED elements of any kind, so i would assume CA is there, even if very small.

[vlaiv]

21 minutes ago, ONIKKINEN said:

Usually, yes, but maybe sometimes? Attached is the spot diagram of my corrector on an F5 system, so my F4.5 primary should be slightly worse still. What do you make of this?

I'd use UV/IR cut filter with OSC camera simply because of getting color info correct. For mono + filters it makes sense not to use filter for luminance if system is well behaved in IR simply because many sources out there give of significant signal in IR part of spectrum and one improves SNR for luminance that way.

RGB filters should cut off IR anyway so that is not concern.

As far as above diagram - it shows that said coma corrector does not do very good job. It does correct for coma up to about 4/3 sensor size (11mm away from optical axis, on 14mm we still see coma but to much lesser extent than system without coma), but it makes scope not diffraction limited very quickly. We don't have closer data, but spot diameter 7mm of axis is ~10.3µm (geometric radius is ~5.16µm) and airy disk diameter is 6.2µm for 200mm F/5 system (5.9µm for F/4.75 system). One of definitions of diffraction limited system is that all rays hit within airy disk.

I would be more worried by star bloat from coma corrector even with IR/UV cut filter than from IR/UV itself.

Take for example this spot diagram:

Yes, indeed, it has slightly larger spot in the center 2µm RMS vs 1.2µm RMS - but keeps that spot all the way to the edge of the full frame sensor at 22mm. Btw, this last diagram does not say which scope is it for, so that is a problem, but let's assume it is indeed F/5 of moderate aperture.

[ONIKKINEN]

2 minutes ago, vlaiv said:

I'd use UV/IR cut filter with OSC camera simply because of getting color info correct. For mono + filters it makes sense not to use filter for luminance if system is well behaved in IR simply because many sources out there give of significant signal in IR part of spectrum and one improves SNR for luminance that way.

RGB filters should cut off IR anyway so that is not concern.

As far as above diagram - it shows that said coma corrector does not do very good job. It does correct for coma up to about 4/3 sensor size (11mm away from optical axis, on 14mm we still see coma but to much lesser extent than system without coma), but it makes scope not diffraction limited very quickly. We don't have closer data, but spot diameter 7mm of axis is ~10.3µm (geometric radius is ~5.16µm) and airy disk diameter is 6.2µm for 200mm F/5 system (5.9µm for F/4.75 system). One of definitions of diffraction limited system is that all rays hit within airy disk.

I would be more worried by star bloat from coma corrector even with IR/UV cut filter than from IR/UV itself.

Take for example this spot diagram:

Yes, indeed, it has slightly larger spot in the center 2µm RMS vs 1.2µm RMS - but keeps that spot all the way to the edge of the full frame sensor at 22mm. Btw, this last diagram does not say which scope is it for, so that is a problem, but let's assume it is indeed F/5 of moderate aperture.

Yes, the Maxfield is a bit of a useless corrector. Only 2/3rds of the price of the GPU and does a significantly worse job than it. The only 2 reasons i bought it were: it was available and i could afford it 😃. Also didn't do research and just assumed that a coma corrector corrects coma and that's it.

I assume your graph is from the GPU? I am getting one whenever i get fed up with fat stars.

[vlaiv]

In either case - I would recommend UV/IR cut filter to be used with OSC / newton combination - because of color accuracy.

[ONIKKINEN]

1 minute ago, vlaiv said:

In either case - I would recommend UV/IR cut filter to be used with OSC / newton combination - because of color accuracy.

I trust this, and i plan on doing that. Still going to test what the difference is, if for no other reason than curiosity.

[vlaiv]

5 minutes ago, ONIKKINEN said:

I trust this, and i plan on doing that. Still going to test what the difference is, if for no other reason than curiosity.

There are several "projects" where I think IR is essential part. For example M42 - there are number of young stars that you can't really see due to dust in optical part - but will show up in IR part of spectrum.

 

[ONIKKINEN]

Just now, vlaiv said:

There are several "projects" where I think IR is essential part. For example M42 - there are number of young stars that you can't really see due to dust in optical part - but will show up in IR part of spectrum.

 

One such star in the earlier example from CCD-FREAK in the lower part of the dust cloud. Very bright pink star in the unfiltered version and almost undetectable in the filtered one, very easy to tell the difference. Embedded in dust so IR passes this better, i assume.

[vlaiv]

3 minutes ago, ONIKKINEN said:

Embedded in dust so IR passes this better, i assume.

Yes, amount of scatter depends on wavelength - that is why sky is blue - because blue light (shorter wavelength) scatters more than red (longer wavelength) - IR is longer still so it scatters the least.

[iantaylor2uk]

I understand vlaiv's argument that he suggests using the UV/IR for colour accuracy but surely that doesn't matter too much since we are shooting RAW files (fits files) so we can get whatever colour we want in post-processing, can't we?

[ONIKKINEN]

5 minutes ago, iantaylor2uk said:

I understand vlaiv's argument that he suggests using the UV/IR for colour accuracy but surely that doesn't matter too much since we are shooting RAW files (fits files) so we can get whatever colour we want in post-processing, can't we?

Try to colour balance an image taken entirely without blue for example. You'll find that no matter what you do you cannot get a true colour result in the end, i think that's the point. With the IR part of the spectrum the light is unevenly spread to the R-G-B pixels so colour balance is skewed or mixed in capture, mostly towards reds since thats where the IR response is strongest. I dont think you can balance this out if its too much out in the first place. Take the image CCD-Freak posted as an example, the colours are just not right and wont be although i think that might be an extreme example.

[vlaiv]

26 minutes ago, iantaylor2uk said:

I understand vlaiv's argument that he suggests using the UV/IR for colour accuracy but surely that doesn't matter too much since we are shooting RAW files (fits files) so we can get whatever colour we want in post-processing, can't we?

Not really.

It is about ratio of colors - or if something is blue - you want it to remain blue.

Our eyes don't see IR photons so their contribution to our color vision is always 0. Camera sees those photons unless we put UV/IR cut filter.

Say you have "perfect blue" and "perfect green" both with some IR photons. Our eyes will see them as perfect blue and perfect green as we won't see those IR photons.

But camera will record RGB ratios that are not perfect blue and perfect green - because cameras have some sensitivity for every component in IR part of spectrum.

Blue wont be 0:0:1 any more and green won't be 0:1:0 any more - they will be something like 0.1, 0.1, 1.1 and 0.1, 1.1, 0.1 (scaled down or whatever).

Problem is - that you can't undo this unless you know IR light intensity and exact QE for each R, G and B in that part of spectrum - and you usually don't unless you take same image with IR pass filter to subtract from original one.

[StarryEyed]

Comparing the two charts its not possible to say given the first does not go past 700nm. There must be a better chart than this about for the IMX571. 

Edited December 20, 2021 by StarryEyed

[vlaiv]

23 minutes ago, StarryEyed said:

Comparing the two charts its not possible to say given the first does not go past 700nm. There must be a better chart than this about for the IMX571. 

Even if there is no published chart for IMX571 that goes up to 1000 - we can sort of guess from other OSC sensors made by Sony.

IMX462 is really exception in its sensitivity, but we can look at others and they all seem to have some sensitivity above 700nm

We can also look at sensitivity of mono version and do some extrapolation (again by comparing other model OSC vs mono graphs).

[StarryEyed]

3 hours ago, vlaiv said:

Even if there is no published chart for IMX571 that goes up to 1000 - we can sort of guess from other OSC sensors made by Sony.

IMX462 is really exception in its sensitivity, but we can look at others and they all seem to have some sensitivity above 700nm

We can also look at sensitivity of mono version and do some extrapolation (again by comparing other model OSC vs mono graphs).

Odd that. but there are differences in the red filter I think on the later chips they are all basically transparent at 800nm. From QHY's web site for the QHY268M with the IMX571. A different presentation but shown none the less

 

 

 

Edited December 20, 2021 by StarryEyed

[ONIKKINEN]

4 minutes ago, StarryEyed said:

From QHY's web site for the QHY268M with the IMX571. A different presentation but shown none the less.

 

From FLO, on the ZWO 533MC product page:

Not the same sensor at all, but this smaller sensor is suspiciously similar in specs to the bigger 2600MC (which has the IMX571) so it might not be such a wild guess to say it would be similar. Could very well be that the 533 is just a smaller size version of the 571 with similar tech behind the curtains.

[ONIKKINEN]

Just now, ONIKKINEN said:

From FLO, on the ZWO 533MC product page:

Not the same sensor at all, but this smaller sensor is suspiciously similar in specs to the bigger 2600MC (which has the IMX571) so it might not be such a wild guess to say it would be similar. Could very well be that the 533 is just a smaller size version of the 571 with similar tech behind the curtains.

Actually here is the 2600MC, with the IMX571 from ZWOs website (cant believe i missed this somehow):

[StarryEyed]

With respect to the location or the UV/IR filter I did get reflections with the IR/UV filter behind a Baader skyglow filter in doors with room lights but not when I put them the other way around. It was odd. My Baader sky glow is an old filter which I am sure doesn't have IR/UV on it. Never really bothered to figure it out once it went away.

Edited December 20, 2021 by StarryEyed