Hubble palette on OSC, Hb or OIII?

Ha+Oiii plus UV/IR for stars.

Hello, Willlem!

As far as I know, you can create a synthetic SII channel if you want to use the Ha-OIII filter only. But usually, you would use the SII-OIII filter to capture SII.

Hope this helps. Clear Skies!

Willem Jan Drijfhout · Jun 2, 2026, 03:46 PM

Is there a general consensus about what is the best dual band filter to use on an OSC camera for a Hubble palette, besides the Ha/OIII?

Would that be an SII/OIII or an SII/Hb?

I would go quadro-chromic and get an Antlia SII/Hb DNB. Not cheap though…

I don’t think they are comparable, one is Hbeta-SII and the other is OIII-Halpha. If you mean the ALP-T then it is much better than the L-Extreme but then one is 5nm and the other is 7nm (in fact even more). Performance with fast systems is the real discriminant here and there is no real alternative to Antlia for that.

If you don’t want halos Antlia is your only option really. Others are “virtually” halo-free (real YT influencer-speak) - and you will be unpleasantly surprised.

Rafał Szwejkowski · Jun 2, 2026, 11:07 PM

If you don’t want halos Antlia is your only option really. Others are “virtually” halo-free (real YT influencer-speak) - and you will be unpleasantly surprised.

Askar’s D1/D2 are stunningly good and 100% halo free in my testing… That said, not sure how great they are on a fast optical system.. I only used them as fast as f/4.8.

Furthermore, remembering that if you go Ha/OIII + SII/Hb, you’re pidgeon-holing your self into a deeply annoying truth about the night sky…

OIII and SII are incredibly sparse by comparison to Ha. Ha rules the night sky… If you’re only capturing OIII data while also capturing Ha, you will either end up considerably weak OIII data, or you’ll end up sacrificing SII or OIII in favor of the other…

If you get Ha/OIII and SII/OIII, you can shoot 4~5h with Ha/OIII, then 10h with SII/OIII, you’ll have 5h Ha, 10h SII and 15h OIII, which will typically mean your SII and OIII won’t get swamped by Ha data…

YMMV, but, As nice as getting the Hb is, I would prefer to double down on OIII… or in a perfect world where the sky was always clear and you have no other responsibilities, get all 3…. 5h Ha/OIII, 10h SII/OIII, 5h SII/Hb resulting in 5h Ha, 15h OIII, 15h SII, 5h Hb… then go to town with your data…

They have IR leak, small but it is there.

Mikołaj Wadowski · Jun 3, 2026, 05:48 PM

I’d go for Sii/Oiii, Oiii often requires the most signal out of SHO. As far as I’m aware Hb looks nearly identical to Ha, just a fair bit fainter, so I don’t think it’s useful at all.

You’ll be wrong on both counts, generally speaking…

I have been happy to combine the ALP-T HA/OIII with the D2 SII/OIII data, such as this version of Fleming’s Triangle (AKA Pickering’s Triangle} : https://app.astrobin.com/i/7lwijc

There are several other images in my gallery using the same combination.

Mikołaj Wadowski · Jun 3, 2026, 09:57 PM

andrea tasselli · Jun 3, 2026, 05:50 PM

Mikołaj Wadowski · Jun 3, 2026, 05:48 PM

I’d go for Sii/Oiii, Oiii often requires the most signal out of SHO. As far as I’m aware Hb looks nearly identical to Ha, just a fair bit fainter, so I don’t think it’s useful at all.

You’ll be wrong on both counts, generally speaking…

Huh, interesting. Here’s an example of Ha vs Hb, with the only difference being I added noise to the Ha data as Hb was far noisier. This is the first dataset I found with Hb data included. Can you tell which is which?

If you know why Hb can differ from Ha you can probably tell them apart by looking at specific clues. This is a rather extreme example as it’s an area with very thick dust, which would show more differences. Despite that, I think you’d agree the difference is barely there.

As to Oiii requiring more data, check out https://www.simg.de/nebulae3/dr0_2/. The linear narrowband FITS maps are presented in Rayleighs, so you can quite easily see what each filter would look like on a CMOS sensor by setting the max cutoff to similar values, accounting for QE.

I present my own data as proof that you can’t generally assume that Hb and Ha are scaled copy of each other. Starts with M27, see here: M27 Ha-Hb-OIII-RGB Composite - The Dumbbell Nebula - AstroBin

Here is another one:

📷 image.pngI can’t see a way to transform one into the other

Now for an emission nebula:

📷 image.pngHb: 78×180s

📷 image.pngHa: 70×180s

And as soon I have got a chance let’s see how cygnus fares. I’ll be there’ll be surprises there too.

As for OIII vs. SII, again I took both advice from people I know have more experience than me in doing SHO and did my own testing and checking. While there are obvious zones where SII is dominant compared to OIII by dint on not having it at all or being obscured by dust or what else, the fact is that Oxygen is the third most common element in the Universe and Sulphur is just down in 8th position. That must have a bearing on how much we see of it, hasn’t it?

This are few shots from areas of the skies that are particularly rich in SII and for the life of me I can’t see the OIII emission being particularly weaker than the SII, quite the opposite if we look at their distribution. Granted one is more concentrated than the other.

📷 image.png20 min each channel

📷 image.png10 min each channel

📷 image.png20 min SII 15 min OIII

In all cases the starless SII has been linear-fitted to the starless OIII so that they have the same background level. Auto-stretched in PI.

Finally, two raw unprocessed frames of a zoomed in image comprising the Thor’s Helmet. No price to guess which one is SII and which is OIII.

📷 image.png5min each.

So I re-state my original assertion : you can’t, generally speaking, assume your original statements to be true.

I quite like your IC2944. That would be my aim too.