Filter suggestions for lunar?

[StuartT]

Hi Lunar Types

Just getting into this and wondered if you have any recommendations for a good filter to use for imaging the moon? I hear good things about the Astronomik Pro Planet 642 IR pass filter

Thanks

[johninderby]

The Baader Filter seems to be the most popular ir pass filter

https://www.firstlightoptics.com/uv-ir-filters/baader-ir-pass-filter.html

 

[StuartT]

Thanks.

• The Baader passes all wavelengths longer than 685nm
• The Astronomik 807 allows all wavelengths longer than 807nm
• The Astronomik 742 allows all wavelengths longer than 742nm
• The Astronomik passes only between 642nm and 840nm (so blocks the longer IR) 

Not really sure how to choose between them. My camera is a ZWO 2600MC and it has an IR cut filter which blocks some IR, but it's not entirely clear to me which wavelengths as the manual appears to show two conflicting graphs

[johninderby]

You might try the Svbony IR pass 685 filter. Very cheap but surprisingly good and a way to try a 685 filter before spending more money.

https://www.amazon.co.uk/Svbony-Filter-Effects-Planetary-Photography/dp/B08P453951/ref=sr_1_1?crid=3AO63NGBK6WVR&keywords=svbony%2Bir%2Bpass&qid=1642864519&sprefix=svbony%2Bir%2Bpass%2Caps%2C103&sr=8-1&th=1

Edited January 22, 2022 by johninderby

[StuartT]

Thanks. I think I first need ZWO to answer the question of their conflicting graphs and what the IR cut window actually cuts in the 2600.

[neil phillips]

IR 685 is good. The top graph seems to show everything after around 750nm getting cut

Edited January 22, 2022 by neil phillips

[StuartT]

6 minutes ago, neil phillips said:

IR 685 is good. The the top graph seems to show everything after around 750nm getting cut

If the top graph is indeed the correct one, the 685 filter would mean I am imaging the moon only in a window of about 15nm. Is that ok? I guess light is not in short supply where the moon is concerned

[neil phillips]

Just now, StuartT said:

If the top graph is indeed the correct one, the 685 filter would mean I am imaging the moon only in a window of about 15nm. Is that ok? I guess light is not in short supply where the moon is concerned

To be honest i am not familiar with that camera. Think you should talk with some who are familiar with it. Hopefully others on here use it. And can give a more definitive answer. 

[Elp]

Haven't imaged it often but do you need one?

[vlaiv]

5 hours ago, StuartT said:

Thanks. I think I first need ZWO to answer the question of their conflicting graphs and what the IR cut window actually cuts in the 2600.

Top is IR/UV cut

Bottom one is simple AR (anti-reflex) coated window - or "plain glass" type.

You don't really need filter for lunar imaging, but if you want one to optimize for seeing conditions - in your case (and otherwise) I would recommend Baader solar continuum filter.

It is sort of narrowband filter although it is not really centered on any particular wavelength that is prominent in narrowband imaging.

Second one that I would recommend is Ha filter (not too narrow - if you can find one that is about 10-15nm FWHM that would be ideal).

Why not narrow? There is no particular wavelength of light you are trying to catch here. These filters serve different purpose in lunar imaging - they serve to reduce seeing effects.

This image contains all the clues:

Seeing is the same thing as prism above - it is bending of the light and there is important feature of this bending that we can see in the image above - red is bent less than green which is in turn bent less than blue (that is why we have dispersion as each wavelength is bent slightly differently).

Because of this - there are two important thing that we achieve with NB filter - first, we can select red part of spectrum as it gets bent less so seeing affects longer wavelengths less than short ones. Second is - wavelengths that are close together are bent about the same - so there is no "smear" due to different amount of bending of light.

Why not very narrow? Because effect on close wavelengths is small and we still need enough signal to get good SNR in short exposure. We should not block too much light - that way we will get very noisy image in exposures that are of interest (less than 5ms).

Why green over red?

I recommended solar continuum over Ha - although green is bent more. Because it is a trade off.

Longer the wavelength - less detail will be accessible with same aperture.

You'll get sharper detail at ~540nm than at 656nm without influence of atmosphere. Atmosphere will bend green light a bit more - but we are doing lucky imaging and trying to beat the seeing. Both filters are narrow enough to provide benefits of reduced wavelength "smear".

There are two more benefits - one of which is applicable to your case and other might not be (but is still valid).

You are using OSC sensor. It has RGGB bayer matrix - which means that half of pixels in your sensor are sensitive to green light while only 1/4 of them is sensitive to Ha. Also - QE in green is higher then in Ha. This means that you'll have higher SNR per recording with Solar continuum filter versus Ha filter.

Second benefit is related to refractors. Refractors are usually optimized so that best color is green around 510nm. This is where optics is the sharpest and has the least spherical aberration (spherochromatism). If you are using refractor for lunar imaging - this will give you edge in sharpness of recording (and subsequently less need for wavelets).

Just remember - optimum F/ratio is not the same for NB filters and in particular not the same for Solar continuum and Ha filters. ASI2600 has 3.76µm pixel size and therefore optimum F/ratio for solar continuum is F/13.9 while for Ha it is F/11.4

 

[StuartT]

12 hours ago, vlaiv said:

Top is IR/UV cut

Bottom one is simple AR (anti-reflex) coated window - or "plain glass" type.

You don't really need filter for lunar imaging, but if you want one to optimize for seeing conditions - in your case (and otherwise) I would recommend Baader solar continuum filter.

It is sort of narrowband filter although it is not really centered on any particular wavelength that is prominent in narrowband imaging.

@vlaiv Thanks once again for a full and expert explanation. This is great news as I already own the Baader solar continuum filter 😃

[StuartT]

On 22/01/2022 at 22:33, vlaiv said:

Just remember - optimum F/ratio is not the same for NB filters and in particular not the same for Solar continuum and Ha filters. ASI2600 has 3.76µm pixel size and therefore optimum F/ratio for solar continuum is F/13.9 while for Ha it is F/11.4

Sorry @vlaiv but I don't quite understand this part. Why would f ratio affect filter performance?

[vlaiv]

2 minutes ago, StuartT said:

Sorry @vlaiv but I don't quite understand this part. Why would f ratio affect filter performance?

I did not say f/ratio will impact filter performance (although filter performance is affected by F/ratio - but that is another topic altogether), I said that optimum F/ratio will change based on filter used.

There is optimum F/ratio for any given pixel size. This is because pixel size in combination with focal length defines sampling rate of the image (arc seconds per pixel), and aperture of telescope defines maximum detail that this telescope can resolve (larger telescopes are able to resolve more).

There is no point in sampling more than telescope can resolve (this is called oversampling and is generally not a good thing for number of reasons) - and if you want to record all the detail available - you don't want to undersample as well - so there is this optimum sampling that will be just right for level of detail available. In long exposure this mostly depends on seeing and mount performance, but in planetary where we want to circumvent seeing effects by using lucky imaging - detail is solely defined by aperture size (and wavelength of light).

There is simple formula for calculating optimum F/ratio where this happens, and for ASI2600 with 3.76µm pixel size - it is F/13.9 for solar continuum filter (~540nm) and F/11.4 for Ha filter (656nm).

On a separate topic - filter performance is affected by F/ratio used if you use interference filters. These filters depend on thickness of different dielectric layers and these layers have different width depending on the angle light hits the filter. For this reason - there is shift in filter curve for different angles of incidence. With slow F/ratios - most of light hits at very steep angle and "sees" layer thickness close to designed. With fast F/ratios - portion of light hits at shallower angles and see layers much thicker.

For this reason - you can see some NB filters "optimized" for - fast F/ratios - like down to F/2 or similar (where this effect was taken into account).

 

[StuartT]

4 hours ago, vlaiv said:

I did not say f/ratio will impact filter performance (although filter performance is affected by F/ratio - but that is another topic altogether), I said that optimum F/ratio will change based on filter used.

There is optimum F/ratio for any given pixel size. This is because pixel size in combination with focal length defines sampling rate of the image (arc seconds per pixel), and aperture of telescope defines maximum detail that this telescope can resolve (larger telescopes are able to resolve more).

There is no point in sampling more than telescope can resolve (this is called oversampling and is generally not a good thing for number of reasons) - and if you want to record all the detail available - you don't want to undersample as well - so there is this optimum sampling that will be just right for level of detail available. In long exposure this mostly depends on seeing and mount performance, but in planetary where we want to circumvent seeing effects by using lucky imaging - detail is solely defined by aperture size (and wavelength of light).

There is simple formula for calculating optimum F/ratio where this happens, and for ASI2600 with 3.76µm pixel size - it is F/13.9 for solar continuum filter (~540nm) and F/11.4 for Ha filter (656nm).

On a separate topic - filter performance is affected by F/ratio used if you use interference filters. These filters depend on thickness of different dielectric layers and these layers have different width depending on the angle light hits the filter. For this reason - there is shift in filter curve for different angles of incidence. With slow F/ratios - most of light hits at very steep angle and "sees" layer thickness close to designed. With fast F/ratios - portion of light hits at shallower angles and see layers much thicker.

For this reason - you can see some NB filters "optimized" for - fast F/ratios - like down to F/2 or similar (where this effect was taken into account).

 

Sure, I know about image scale, but I thought the formula was

(206*pixel size)/focal length

Where does frequency come in?

[vlaiv]

36 minutes ago, StuartT said:

Sure, I know about image scale, but I thought the formula was

(206*pixel size)/focal length

Where does frequency come in?

That is image scale or pixel scale - arc seconds per pixel.

Cut off frequency is something related to Nyquist sampling theorem and aperture and all of that.

Here is formula for that:

https://en.wikipedia.org/wiki/Spatial_cutoff_frequency

If you apply Nyquist sampling theorem to that (need to sample at twice highest frequency component) and rearrange for F# (F/ratio) you get:

F/ratio = pixel size * 2 / wavelength

(where pixel size and wavelength are in same units)

So for example - if you want to see what is optimum F/ratio for 540nm wavelength and 3.76µm pixel size then you calculate:

F/ratio = 2 * 3.76µm / 0.54µm (as 540nm = 0.540µm)

F/ratio = ~13.925925926 = F/13.9

similarly you get for Ha, but substitute 656nm as wavelength.

When you want to calculate critical sampling frequency for full spectrum / RGB - it would be sensible to take 400nm, but blue part of spectrum is affected the worst by seeing and this leads to slight over sampling in other wavelengths. Peak of human luminance sensitivity is around 500nm - so that is where we perceive the most sharpness - for that reason, I advocate use of 500nm wavelength to calculate cut off for full spectrum / color images

This gives handy formula:

F/ratio = pixel_size * 2 / 0 .5 = 4 * pixel_size - easy to remember.

[StuartT]

On 22/01/2022 at 22:33, vlaiv said:

 

You don't really need filter for lunar imaging, but if you want one to optimize for seeing conditions - in your case (and otherwise) I would recommend Baader solar continuum filter.It is sort of narrowband filter although it is not really centered on any particular wavelength that is prominent in narrowband imaging.

@vlaiv can I just check with you that this filter would be good with either a mono camera or an OSC? (I assume they both have similar frequency responses). Sorry if this is a dumb question

[vlaiv]

34 minutes ago, StuartT said:

@vlaiv can I just check with you that this filter would be good with either a mono camera or an OSC? (I assume they both have similar frequency responses). Sorry if this is a dumb question

Filter will do it's thing regardless of sensor used, so yes.

[StuartT]

14 hours ago, vlaiv said:

Filter will do it's thing regardless of sensor used, so yes.

thank you. One thing the ZWO don't seem to provide in their documentation is the frequency response of their cameras. Do you know if this can be found anywhere?

[The Admiral]

14 minutes ago, StuartT said:

thank you. One thing the ZWO don't seem to provide in their documentation is the frequency response of their cameras. Do you know if this can be found anywhere?

Unless I'm barking up the wrong tree, isn't this what you want (from the ZWO website)?

It is relative response, but should that matter?

Ian

Edited February 10, 2022 by The Admiral

[StuartT]

1 hour ago, The Admiral said:

Unless I'm barking up the wrong tree, isn't this what you want (from the ZWO website)?

It is relative response, but should that matter?

Ian

Ah. Is it? Thanks. I didn't know quantum efficiency was the same as sensitivity. Much appreciated!

[Mikey0368]

I was just about to come on and ask a similar question. I have an image of January's full moon taken using the ZWO IR 850nm Pass Filter, will I be able to process colour into the picture or does the filter stop that?

[vlaiv]

1 minute ago, Mikey0368 said:

I was just about to come on and ask a similar question. I have an image of January's full moon taken using the ZWO IR 850nm Pass Filter, will I be able to process colour into the picture or does the filter stop that?

You might be able to get some sort of color - but it will certainly not going to be true color or anything near it.

There is no color in IR part of spectrum - but pixels can have slightly different sensitivity to same part of spectrum and that will give "impression" of color.

[Mikey0368]

Cheers for that. Saves me wasting any more time on it!