Band shift vs more integration time - Forums

Topic starter

New contributor

1 day ago

What would you choose?

I have shot an image I’m quite happy with at f/2 and my 7nm Optolong Ha filter. However, I am aware I might have gotten some band shift. I was shooting other things and I didn’t pay attention to the f-stop at the time. Since then I have wondered something.

Is it better to get some band shift but having less imaging time or have double imaging time and zero shift?

For context

I shot at f/2, 135mm at 7nm for a total of 30 minutes. If I did shoot at f/2.8 for 1 hour, would I get more/better signal?

My uneducated guess is that it depends how much light you are losing. If light loss is 50% or more due to the shift, I guess it IS better to stop down. But, if you only get e.g. 70% of the maximum transmission, is it truly better to get your band shifted and halve your integration time?

I guess this would need testing, and not all filters are the same. For example, Antila and Baader filters are notorious for still working properly for fast systems. However, Optolong released an L-eXtreme F2 because the original LeX lost about 80% of that light.

For what it’s worth, I think I lost more signal than I gained from the wider aperture, but I’d love to hear someone’s opinion on this.

The image mentioned: The North America Nebula in H-Alpha - AstroBin

Clear skies.

Simone

Helpful Insightful Respectful Engaging

Médéric Hébert

1 day ago

With 7nm, I don’t think you’re losing all that much data.

Might be a problem with 3nm…

andrea tasselli

Frequent contributor

1 day ago

I wouldn't be too difficult to measure the net effect by sheer photometry. I bet that the change by going to f/2 from f/2.8 isn't as stark as loosing half of the aperture.

Simone Scarpa

Topic starter

New contributor

1 day ago

I also thought it was not going to be a big deal but then I read the report on the L-eXtreme at f/2 and I actually think f/2 is too much for this Optolong filter specifically. I will try to test it out the next clear night.

andrea tasselli

Frequent contributor

1 day ago

The L-Extreme isn't the same as the 7nm Ha.

Armin Lukas

1 day ago

I don’t have an answer, but what about a quick test? Take a few subs each with 2.0 and 2.8 and see what looks better to you? Which verison has more detail, etc?

Simone Scarpa

Topic starter

New contributor

1 day ago

Indeed the next clear night I fully intend to test this. I will make a quick report when I do and how the data differs

Armin Lukas

1 day ago

https://www.researchgate.net/publication/344786253_Optolong_L-eXtreme_Filter_Comparison

I found this test, conducted by Jim Thompson, where different filters got tested at different light angles. Maybe this gives you a better understanding. I wonder how much it really effects the transmission.

I use the Sv220 filter on my Sharpstar 13028HNT, and the results are not bad. But never tested a F2 filter (yet).

shenmesaodongxia

1 day ago

I ran my ALP-T Highspeed with the Samyang 135mm f/2 at around f/3.3.
You can check how much blue shift there is by taking a flat frame. You just need to auto-stretch the flat frame and apply auto white balance. If the result is completely white, like a wideband flat frame, that means it is not good.

A proper flat frame should show some green or red color in the middle.
Yes, you can still get some results at f/2, but you will not get as much detail as you would with a slower focal ratio, even with a long exposure time.

Helpful Concise

Simone Scarpa

Topic starter

New contributor

1 day ago

Funnily enough I went back to the Autosave and I actually found some actual signal in the other channels. It usually isn't very visible as 95% of it resides in Red but here I'd say 60% the signal was in R and the other 40% in G,B in this case.

I may have a clear night on Thursday to test this further.

andrea tasselli

Frequent contributor

1 day ago

That's to be expected from an OSC as both G and B filters let in some amount of R.

Well Written

Alex Nicholas

1 day ago

Bandshift is certainly an issue - but I honestly do not think it’s what its made out to be… As with so many things, on paper, it’s horrific and you would never want it..

Take for example, the difference between a 60mm F/7 refractor with a 420mm focal length, vs a 150mm f/2.8 reflector with a focal length of 420mm. On paper - you would NEVER EVER buy something 60mm f/7, when its theoretical max resoultion is significantly worse, its light gathering power is significantly worse and its focal ratio is almost 8 times slower… But again - theoretical, on paper specs.

In reality, the F/2.8 vs F/7 yields maybe a 2x imrovement in speed, resolution is likely more limited by your local seeing than the max theoretical resolution of the optics.

When I bought my F/2.8 Newtonian, I was 100% positive I’d have to replace my 3nm SHO filters, and sure, They definitely do not get the same throughput as they did in my F/7 APO, but, I can say with confidence that 1h of data with the 3nm filter in the f/2.8 newt is giving stronger SNR than 1h of data with the same camera and filter than the F/7 APO.

Is it optimal to be running 3nm filters with an f/2.8 optical system? no.
Would it be better if I bought F/3 optomised 3.5nm filters? yes.

Is it undeniably necessary to get the F/3 optomised filters? Not immediately I don’t think… I will get them, but they are certainly not top of my priority list at the moment.

There is undeniable science behind bandpass shift.. It happens. without doubt, but I don’t think without first assessing your results on a full image, should you be too concerned.

Helpful Insightful

GalacticRAVE

22 hours ago

I don’t think your considerations quite lead you to your intended goals….

Let’s first look what happens with band shift: the near axis light (which falls in perpendicular to the filter surface) get’s through unaffected, the very off axis light rays (which are inclined to the optical surface) get shifted, so eventually no Halpha light gets through (to be precise, Halpha gets blocked but slightly shifted light gets through, i.e sky).

What happens if you step down your lens: the near axis light get’s through unaffected, the very off axis light rays get blocked (as it hits bare metal).

Looks very similar, doesn’t it? So in other words: stepping down with the filter has little effect, no change for the near axis light, off axis you in the best case may block light rays that are blocked by the filter anyway, in the worst case you may block light which otherwise would (partially) go through.

Note, the situation is somewhat different for RASA, fast Newtons etc as in that case the central light is blocked by the secondary, so the effect of band shifting is considerably larger as all light is off axis.

I strongly recommend running the test suggested by others here to see what the net effect of the filter is - or you do some detailed analysis (some info for the mathematically/physically inclined you can find on the web, including a couple of well researched youtube videos, Brent Mantooth has a number of excellent clips). If the effect is more than f2.8, you may also consider to put the lextreme in front of the lens (get a 77mm to 48mm step down ring and screw the filter into the M48 opening) as the light incoming light is (alsmost) perfectly perpendicular to the optical surface - ie no bandshift here. For the lens-astrophotographers it would indeed be nice the get somewhat larger filters - but these would also be quite expensive …

Matthias

PS: just saw that Brent Mantooth has a video specifically on the Samyang: the net effect of a 6nm filter is basically making your lens a f2.5. There is another reason tough to consider stepping down (or better to put a filter in front): you get a vignetting-type of feature (the on axis rays pass to near 100%, but off-axis only 65% of the light gets through). But you cannot correct that easily by flats unless you have a (hypothetical) monochromatic flat panel …

Helpful Insightful

Simone Scarpa

Topic starter

New contributor

21 hours ago

andrea tasselli · Nov 24, 2025, 10:53 PM
That's to be expected from an OSC as both G and B filters let in some amount of R.

Of course, the Bayer matrix isn't perfect and there is bleed onto other pixels but I was just saying it looked like G,B signal was more than the usual amount.

Simone Scarpa

Topic starter

New contributor

21 hours ago

GalacticRAVE · Nov 25, 2025, 05:23 AM
I don’t think your considerations quite lead you to your intended goals….
Let’s first look what happens with band shift: the near axis light (which falls in perpendicular to the filter surface) get’s through unaffected, the very off axis light rays (which are inclined to the optical surface) get shifted, so eventually no Halpha light gets through (to be precise, Halpha gets blocked but slightly shifted light gets through, i.e sky).
What happens if you step down your lens: the near axis light get’s through unaffected, the very off axis light rays get blocked (as it hits bare metal).
Looks very similar, doesn’t it? So in other words: stepping down with the filter has little effect, no change for the near axis light, off axis you in the best case may block light rays that are blocked by the filter anyway, in the worst case you may block light which otherwise would (partially) go through.
Note, the situation is somewhat different for RASA, fast Newtons etc as in that case the central light is blocked by the secondary, so the effect of band shifting is considerably larger as all light is off axis.
I strongly recommend running the test suggested by others here to see what the net effect of the filter is - or you do some detailed analysis (some info for the mathematically/physically inclined you can find on the web, including a couple of well researched youtube videos, Brent Mantooth has a number of excellent clips). If the effect is more than f2.8, you may also consider to put the lextreme in front of the lens (get a 77mm to 48mm step down ring and screw the filter into the M48 opening) as the light incoming light is (alsmost) perfectly perpendicular to the optical surface - ie no bandshift here. For the lens-astrophotographers it would indeed be nice the get somewhat larger filters - but these would also be quite expensive …
CS
Matthias
PS: just saw that Brent Mantooth has a video specifically on the Samyang: the net effect of a 6nm filter is basically making your lens a f2.5. There is another reason tough to consider stepping down (or better to put a filter in front): you get a vignetting-type of feature (the on axis rays pass to near 100%, but off-axis only 65% of the light gets through). But you cannot correct that easily by flats unless you have a (hypothetical) monochromatic flat panel …

Very interesting. However, even if the filters are supposedly durable I don't feel comfortable enough keeping an L-eXtreme in front of the lens. I already don't like the heavier vignetting produced, but I feel like these things are meant to be in the closed up in imaging train near the sensor. People have used it as such and there are many examples, but a 10nm STC Duo-NB costs similarly, has (even if a bit wider) similar FWHM and is a clip version..

Also, I read up something on Roger Clark’s website stating that even with high quality filters, putting one at the end of the lens will inevitably degrade quality. I haven't seen much info on that. I suppose it's more to do with UV filters and such. For narrowband it's a harder comparison since stars will shrink inevitably.

Thanks for the YouTube video recommendation. I'll check it out.

Clear skies

Helpful Insightful

Simone Scarpa

Topic starter

New contributor

21 hours ago

Alex Nicholas · Nov 25, 2025, 12:48 AM
Bandshift is certainly an issue - but I honestly do not think it’s what its made out to be… As with so many things, on paper, it’s horrific and you would never want it..

Take for example, the difference between a 60mm F/7 refractor with a 420mm focal length, vs a 150mm f/2.8 reflector with a focal length of 420mm. On paper - you would NEVER EVER buy something 60mm f/7, when its theoretical max resoultion is significantly worse, its light gathering power is significantly worse and its focal ratio is almost 8 times slower… But again - theoretical, on paper specs.

In reality, the F/2.8 vs F/7 yields maybe a 2x imrovement in speed, resolution is likely more limited by your local seeing than the max theoretical resolution of the optics.

When I bought my F/2.8 Newtonian, I was 100% positive I’d have to replace my 3nm SHO filters, and sure, They definitely do not get the same throughput as they did in my F/7 APO, but, I can say with confidence that 1h of data with the 3nm filter in the f/2.8 newt is giving stronger SNR than 1h of data with the same camera and filter than the F/7 APO.

Is it optimal to be running 3nm filters with an f/2.8 optical system? no.
Would it be better if I bought F/3 optomised 3.5nm filters? yes.

Is it undeniably necessary to get the F/3 optomised filters? Not immediately I don’t think… I will get them, but they are certainly not top of my priority list at the moment.

There is undeniable science behind bandpass shift.. It happens. without doubt, but I don’t think without first assessing your results on a full image, should you be too concerned.

Thanks for the reply, very insightful. You're right, probably blowing the issue out of proportion. As GalacticRAVE said, it would probably be worse on RASA or fast newt systems.