Jump to content

SkySurveyBanner.jpg.21855908fce40597655603b6c9af720d.jpg

STC Duo-Narrowband filter unboxing


Miguel1983

Recommended Posts

3 hours ago, Demonperformer said:

(Yes, I know ... spherical chickens in a vaccuum!).

That's a good one! :D

3 hours ago, Demonperformer said:

I have been thinking about this and it would be interesting to see some sort of comparison figures for how much time would be saved. So, for example, if I have a target that, with my individual Ha, Oiii & Sii filters, requires 1 hour with each filter to achieve a given result, how much time would I need with each of the now 4 individual filters to get an identical result, assuming identical scope, camera, etc and identical conditions (Yes, I know ... spherical chickens in a vaccuum!).

That is actually not as easy as it might seem, because it involves human response to brightness and color (which are different - we seem to be much more sensitive to variations in brightness than in color).

It would also depend on the way you are processing your data. I guess we can do couple of hypothetical cases and see what sort of data we end up, but I'm not sure how to do processing phase calculations - as this depends on processing method used. Just to explain what I mean by that - if you capture 3 sets of NB data - with Ha, OIII and SII filters - you don't have luminance component there, so you can produce one by adding these three sets, but you don't have to do this - you can assign each data set to particular channel - R, G or B.

Now it also depends on what channel did you assign which data set. One way of calculating luminance from RGB data is L =  0.299*R + 0.587*G + 0.114*B (often recommended online for sRGB color space, but I'm not sure if that is luminance that we would perceive). From this it clearly shows that resulting L SNR (one that we would use for comparison to 3band filter luminance) - depends on how you assign channels - HST or some other palette.

Other option would be to assign sRGB channels and convert that to CIELab and extract L from that - this is yet more complicated, since sRGB assumes gamma of 2.2, so there is non linear transform to linear RGB prior to conversion to Lab, just to get L channel there.

We could go the other way around and leave single filter data as is and compare on basis of that (per channel SNR), but we would need a way of incorporating tri-band filter luminance with single filters to arrive at final color - and again that depends on workflow.

Let's do a simple case where we don't want to compare actual result, but just examine SNR of individual stacks in relatively simple case. I'm going to "assume" quite a lot of things here to create hypothetical scenario.

Conditions: 80mm F/4.8 scope , about 2"/pixel sampling, mag22 target (per "^2), mag24 sky (this would be the case or ~mag19 light polluted sky and narrow band filter), 4h total time (so in case of nb filters alone, each filter would get 1:20, but in case of triband + nb, each would get 1h total), low read noise camera, 10 minute individual subs, and QE of camera, signal strength and FWHM of filters combine in such way that signal strengths have ratio of 5:2:1 (so Ha is x5 as strong as SII, while OIII is x2 as strong as SII - I would say reasonable assumption).

SNR for 1:20h per filter (3 NB case):

Ha - 23.68
OIII - 13.38
SII - 8.83

SNR for 1h per filter (triband + NB case):

Ha - 20.5
OIII - 11.59
SII - 7.12
Triband - 26.64

Results are as one would expect - each Ha, OIII and SII have less SNR for 1h exposure than for 1:20, and of course triband filter has the most SNR of all, although it is only 1h.

Above results will obviously be different under different conditions (like different target and sky brightness and such ...).

We can do one more step of calculation and that is to add Ha, OIII and SII from first case (solo NB filters) to produce output similar to triband filter and see how they compare.

Result of summing first three stacks would have SNR of 28.00 - that is better than triband filter (26.64) - but remember this is 4h (three times 1:20h) vs 1h with triband filter

Link to comment
Share on other sites

  • Replies 53
  • Created
  • Last Reply

OK, I think I must be missing something ... if I am reading your figures correctly, in each of the 4-hr integrated-time examples, all 4 channels have a higher SNR when only 3 filters are used [23.68 vs 20.5; 13.38 vs 11.59; 8.83 vs 7.12; 28.00 vs 26.64]. So doesn't this argue against using the Tri-Band, as it would require a greater integrated time to get the same SNR?

Link to comment
Share on other sites

2 minutes ago, Demonperformer said:

OK, I think I must be missing something ... if I am reading your figures correctly, in each of the 4-hr integrated-time examples, all 4 channels have a higher SNR when only 3 filters are used [23.68 vs 20.5; 13.38 vs 11.59; 8.83 vs 7.12; 28.00 vs 26.64]. So doesn't this argue against using the Tri-Band, as it would require a greater integrated time to get the same SNR?

That just shows that if you take Ha filter and integrate for 1 hour and 20 minutes - you will get higher SNR than taking the same filter and integrating for only 1h - perfectly normal thing. It also happens to OIII and SII filters.

We just spent 3 hours on Ha, OIII and SII in second case and remaining 1h on triband filter.

What is interesting is that 1h of triband filter has higher SNR than any of other filters used for either 1h or 1:20h, and that it has just a bit lower SNR than sum of Ha, OIII and SII totaling 4h (each 1:20h), but taken over 1/4 of time - for just 1h.

Now there are other combinations what we could have used here, like instead of using just tri-band filter for luminance in 1h - we could have created mix of (Ha 1h + OIII 1h + SII 1h)*weight + tri-band 1h - and that thing would produce the best SNR in luminance by far (same signal will be captured if we add Ha, OIII and SII as would by use of triband - simply because such filter includes these three NB filters in one) - but that would be "advanced" processing.

We could have gone for 2h, 0:40h, 0:40h, 0:40h - triband, Ha, OIII, SII in second case - that would also total for 4h. In that case triband luminance would have by far the greatest SNR of any particular filter or even sum of 1:20h single NB filters.

Above analysis is the same as RGB vs LRGB case. If we were to take 1:20h of each in R, G and B versus 1h of L, R, G and B - we would end up with lower SNR in one hour exposures in color, but added luminance creates better looking / smoother image (higher SNR in luminance which human eye / brain system perceives as more important for smoothness than color part of image).

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...

Important Information

We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue. By using this site, you agree to our Terms of Use.