Testing Baader’s solar filters, in the ersatz-obsi

As you know, due to various reasons, lately I’ve been doing solar imaging almost exclusively, focusing on full disk observations. Of the many filters I have turned towards the Sun — some dedicated solar filters, some more in a hacker spirit –, the Calcium K wavelength is one of my favorites, obviously. Now that Baader started selling its generation ii CaK filter, I figured I’d test it against itself, the first generation, and against Lunt’s Calcium module.

UPDATE 2024: in the meanwhile I tested Altair Astro’s early bird sample of Calcium K 3nm, against Antlia’s 3nm CaK, and got a B1800 Lunt CaK (btw fuck Bresser for wasting my time, for the second time in a row regarding Lunt products), to deploy in a long, f/11 tube and protected by a full aperture ERF (from Altair), to do full disk solar with chromosphere and all.

As you also know, I record images as a hobby, without any specific scientific goal, I just like the rigor and if some aesthetics can arise from it, in the form of nice solar disk images in this case.

TLDR0: not much difference, but still.

TLDR1: if doing full disk, if you have the old CaK and like it — keep it and your setup the way it is, the new filter is nice, the profile is standard, but overall, doesn’t show anything new nor better than the old one. Both are that good. Obviously not as good as Lunt’s CaK, but they are a fair competition, given the price range.

TLDR2: if you have the Venus U filter, do use it. It’s not as good for the Sun’s active regions as the CaK, but still pretty good. Also, give the dark blue a chance. Or basically any filter you own.

TLDR3: for Venus, the new filter may yield better images, due to it letting through about two times more light (half the expo time). I like very much the CaK images I obtained of Venus, in the 2020 epic season.

Venus, on 2020-04-05, using Baader’s CaK (gen1), as a false bi-color image with a green filter, mapped to cyan and orange respectively, with a N150/750 tube.

Setup, conditions

In the ersatz-obsi, on 2023-03-19, I went with my usual setup, with the 1.25″ filterwheel containing: Baader’s Solar Continuum 7.5nm, Baader’s Calcium K generation II, Baader’s old Calcium K generation I (the double stacked, high cell version), a UV/IR cut filter, and Baader’s Venus U filter. On other days, since they were there, to see against the Continuum, I also had a ZWO Ha 7nm cell, ZWO’s DouBand, and even Optolong’s L-eXtreme (out of curiosity, since they were in the wheel at the time) — I included some pictures from earlier too.

EQ3-mod, mountpusher, Baader Astrosolar ND3.8 Film (brand new), TS-Optics 76/342 TS76EDPH, automated filter wheel, ASI 178MM (cooled)

For the Lunt CaK module, a different tube was used, EQ3-mod, mountpusher, Sharpstar 60/330 ED-APO, Lunt CaK B600 LS6CAKMDS2, a barlow’s lens part, ASI 178MM (cooled)

I wanted to give a fair chance to all filters, while checking that the atmospheric conditions remain fairly the same — this is why I dropped 2023-03-18’s data from the comparison. Even though there were some cirrus clouds, resulting in some of the raw material’s discarding, there was plenty of the 2.5 TB obtained in about 3 hours, including some Lunt Ha images, not exposed here.

The ersatz-obsi’s setup, a good half of the equipment I built and programmed myself

Goals, methods

To bring out the best from Baader’s two CaK filters, compare them against each other, against other filters, and see how they fare against Lunt’s Calcium Module. Even though Lunt’s CaK, even the B600, is more expensive than all the other filters put together, it’s still meaningful to see how all fare.

To compare Baader’s filters, I maintained the same tube, the same solar film, and only adjusted the focus. To honor John Cage, I record 4’33” long ser videos, maximum framerate possible. The exposure time resulting from the setup is usually on the order of milliseconds, as shown in the camera CSVs posted below, trying to maintain about 90% histogram coverage. Focusing is always a bit tricky, but the electronic microfocuser I built allows a perfectly steady focusing, a slight displacement of a few pixels happens when the direction of the focusing changes.

Due to the ersatz nature, the polar alignment is approximate at best, so an autoguider is keeping the Sun in the middle of the field of view of the camera, but also allowing for some dithering.

I didn’t check for reflections. I know the old one had a ghost image, also the glass layers are tilted against each other, but I also knew the ghost image can be ignored for most intents and purposes. I once caught a fairly bright solar prominence with the old CaK filter, so it is possible — that’s when the ghost image became important, but the Lunt also suffers from this.

I also didn’t check longer focal ratios. I have some DSLR pictures from quite a few years back, taken with the old filter, obviously, through a MC102/1300.

The wheel got the new filter’s data before the delivery

First notes

(0) On the previous day I concluded that the old one performs better. But I also concluded that one observation is no observation and wanted to repeat, accounting for the changes in seeing. And yes, it was a drastic change in the seeing.

(1) The new CaK cell needs only half of the exposure time, give or take.

(2) It is really hard to see any difference, at all, on the live image. Like really. I thought it was my memory playing games, then decided to put in the old one, next to the new filter, to see with my own eyes. There really is no difference.

(3) I still like it. I already have my filter wheel adapted, with a T2 washer, to make sure the higher-than-normal filter frame fits into the wheel, and it worked like a charm, so no advantage here with the normal filter frame profile.

Processing

My workflow includes the several gigabytes, usually around 5-10k frames, of which AutoStakkert 3 selects the 500 best frames, and stacks them. I usually go with around 200-1000 alignment points, as the seeing indicates the resolution vs pixel noise, mostly by gut feelings. Since I take many 4’33” recordings, and the Sun itself produces some white noise like „pixelnoise”, also visible in the SDO frames at 1600 and 1700 Å, stacking together the individual result-frames of individual videos, usually results in a smoother image, sometimes desirable, sometimes not (as it looks really unnatural), from an artistic point of view. Wavelets are adjusted in Registax 6, then photo editing — omitted here for today’s Baader CaK, not even rotated into the Sun’s equatorial plane. As they popped out of the camera, mostly Earth’s horizontal grid, with zenith flipping.

To emphasize the active regions, I obtain the „brighter than their context” areas by subtracting a large radius median from the original frame, and adding the resulting „cracks” to the original frame. This is different than sharpening, as the lights don’t rob light from the surrounding pixels, so they don’t get darker.

The end result

I honestly don’t see much difference.

Maybe there is, or would be, at higher f ratios, maybe. Since both filters perform in the very same scope, safety film etc, and I couldn’t find a high focal ratio requirement in either’s description (compare that to the Sundancer Hα). Baader’s site does mention that mirrors are preferred, OK… What I can tell is that this APO tube blows my Canon tele lens out of the water, runs circles around it, and eats it for breakfast, as far as blue and UV are concerned. Been there, done that. I wouldn’t worry about the quality of my APO refractor. I may repeat the test with my N 150/750 tube, same filterwheel, but I am simply not interested in disk details at this point — it’s clumsy to set up the newton in the ersatz-obsi, and to obtain a full disk without coma errors and distortions, onto a really large sensor, the question of the coma corrector’s (also Baader, I mean…) UV performance, it just leads too far into areas I don’t really care about.

Also, the Venus U perform similarly, though much harder to image due to both camera sensitivity, atmospheric seeing and … the glass is glass, after all.

gen1, stack of stacks, pre-wavelet

gen1, stack-of-stacks, wavelet (same for all baader CaK pictures here)

gen1, stack of 500 frame out of the 4min33 video, pre-wavelet

gen1, stack of 500 frame out of the 4min33 video, wavelets shaperning

gen2, stack of 500, pre-wavelet

gen2, stack of 500, with wavelets

gen2, stack of stacks, pre-wavelet

gen2, stack of stacks, with wavelets

Lunt stack of 500, pre-wavelet

Lunt stack of 500, with wavelets

Lunt, stack of stacks, with wavelets

Also, with careful processing, even deep blue and the Solar Continuum shows the active regions, and though really hard, bordering on PITA, the continuum-Hα (meant for night time, nebuale) also shows the same active regions (and even prominences, their presence, not beauty).

ZWO’s 7nm Hydrogen alpha Sun from 2023-02-13. The active regions are visible, but it was a PITA to emphasize them in an aesthetic way.

The very first steps of processing for the above, 7nm width H alpha image

2023-03-19, Solar Continuum, the Emerald version

2023-03-19, Solar Continuum, the same raw picture as the above, but in grayscale

2023-02-21, The Sun through the Venus U filter, with post processing, emphasizing the active regions

And below a series from 2022-06-18, depicting the Sun through various filters, all processed for aesthetic purposes.

2022-06-18, Baader CaK, old

2022-06-18, Baader Venus U

2022-06-18, Baader Dark Blue

2022-06-18, Lunt Ha

2022-06-18, Baader Solar Continuum, old version + UV/IR cut

2022-06-18, Lunt CaK, 6×5 min (hence the lack of high frequency details)

2022-06-18, ZWO Duoband

The Sun on 2015-07-03, Baader’s CaK (old version), MC 102/1300 and Canon 1100D, image reprocessed on 2021-12-03

The Lunt filter and my processing technique allows me to show: the Black Sun

The epic prominence, in Lunt’s CaK on 2022-05-10 — Baader’s filter is not suitable to do this, and digging this picture out of the Lunt raws was several days’ work. It got a prize though…