3.94
#...
·
·
2
likes
|
|---|
|
As I had quite good results with the normal Antlia ALP-T Dualband Filter I also bought the High Speed Version of that filter as soon as I realized it exists. So now I have a set of Baader High Speed Filters and the Antlia, so why not try to compare the results.... Some basic Data: The ALP-THS filter is designed for telescopes with an aperture of F3.6 to F2.2 The Baader Filters are designed for systems with an aperture of F2.3 or faster, so my [email protected] (F2.2) should work with both products. I had very few clear nights in the last weeks, but in the last days I had 3 short nights to both finetune my Starizona and to record a few subs. My priority was to record some data so please forgive me for sharing data that has stars that are a bit affected by tilt and a little by backfocus........ So far I have collected: 16x5min subs ALP-THS 5nm 16x5min subs Baader Ha 3.5nm Highspeed 15x5min subs Baader Oiii 4nm Highspeed and 19x5min subs Baader Sii 4nm Highspeed to be able to create a final image some day soon..... My target for this evaluation is the Lagoon Nebula because it has rather strong signal on all three wavelengths and I am shooting from my balcony and cannot reach all those nice targets that are high up in the sky at the moment. For me, currently the nights are just to short to justify driving to my preferred mountain site for better seeing. But I live in a Bortle-4 area so things are not too bad on the balcony. 'This test will be easy' I thought, but I start realizing that the word 'easy' is not something that you can use when your sentence is about astrophotography. So I guess to do those filters justice I will also need your input on how to best look at the data to do a fair and meaningful comparison... To enable that I have shared my data, you can find the stacked images and full size comparisons here if you feel you can improve the way I compared: https://www.mycloud.ch/s/S00E3CA97C39D95B2B3E6CD246FEC0503462967892A This data is shared as Creative Commons, mention my name, non-commercial,share alike (CC-BY-NC-SA) That out of the way, let's continue: The main plus point for using a Dualband Filter together with an OSC camera is that you get two filter runs for the price of one. Both Ha and Oiii will be recorded and the Red Channel will include the Ha signal and the Green Channel has the Oiii (Blue is usually too noisy to effectively use the Oiii in that channel) But the OSC camera is not limited to be used with a Dual-Narrowband filter (although it feels as if many people think that this is the case) it can also be very successfully used with single HA, Oiii and Sii filters. Resolution will be not as good as with a mono camera, but the difference is not night and day and with a Sony IMX571 sensor there are lots of pixels to make up for the loss of resolution of the bayer matrix. The first test I did was to compare SNR in PixInsight with Herbert Walter's SNR script. I did look at single subs and at some integration runs and the results were rather surprising. I debayered and then extracted the red channel for a comparison of Ha and for a single Frame I got an SNR value of 32dB for Antlia and 38db(!!) for Baader, so a whole lot more SNR for the Baader filter. First surprise.... The two subs were done 10 minutes apart. I did the same thing for Oiii, the recordings time of the two subs was also close-ish, in this case 40min but overall the sky quality changed little during those 40 minutes. For Oiii the results were much closer, 31.4 dB for Baader and 30.6 dB for Antlia, 2nd suprise, I would have expected better SNR for the Baader Filter. When comparing the full stacks, the tendency of the single frame was also visible in the stack: Baader HA SNR: 47.3dB Antlia HA SNR: 41.6dB Baader Oiii SNR: 39.4dB Antlia Oiii SNR: 39.7dB and just for the fun of it: Baader Sii SNR: 43.0 dB So based on this single value there is an advantage for the Baader Ha filter while on Oiii the filters are pretty similar. I also used the Noise Evaluation script and in general it gave the same tendency as the SNR script with a small advantage for the Antlia Oiii compared to the Baader Oiii (4.559e-01 vs 5.267e-01), in Ha the numbers were in favour for the Baader filter. Now for visual comparison, and that part is the part where I am not 100% sure if I am properly comparing apples to apples. The reason is that I simply used ScreenTransferFunction to default-Stretch the images. I also tried to do a GHS stretch with identical parameters but a slight change in Symetry Point changes output a lot so I gave up on that approach. If you know that you can do better than me with the ScreenTransferFunction then please add to this thread, your insight is apprechiated! First Comparison is HA, Baader on the left and Antlia on the right. There was very little done to the image, only stretching, no DBE, no crop, no nothing. What puzzles me is that Antlia looks as if it has a tiny little more contrast although it is a little bit darker, but besides that the amount of details and nebulosity looks pretty similar with a very, very slight advantage for the Antlia. The little loss of contrast for the Baader can be caused by the fact that both filters received most of their data on different sides of the meridian. When you look at the full scale pics you can see that on the light pollution for Baader Ha seems mostly to be on the left and for the Antlia Ha the light pollution is more on the right side of the frame. When I look at the frames at 1:1 then I can see that noise is a little better with the Baader filter, that matches well the theoretical numbers from above, but honestly the difference is not huge and nothing that could not be fixed with a little bit of NoiseXTerminator.  Looking at Oiii things are more like I would expect them. The background of the Baader is darker, it also has a smaller bandpass so less light pollution can pass. But there is more......  When I look at 1:1 then I start to wonder if the Baader does kill meaningfull data. When you look at the whispy nebulosity then I'd say that there is more of it in the Antlia than in the Baader..... Noise also seems better under control with the Antlia, but again, not too much of a difference in noise to me.  One aspect that I have not yet touched is Halo's and when you look at the 1:1 in Oiii then my fear is that some halo is building up with the Antlia filter. I have some more subs from the Butterfly Nebula, there I have more bright stars in the frame... from what I can see Halo's are not an issue with the Antlia, but I will have to create a picture with a much brighter star to be sure, I will add new data to this thread when I have the data. So for now, which filter is better? When you are about absolute quality then the Baader Ha has a small edge for fast scopes. My guess is that for really fast scopes in the F2.0 range the Antlia will not perform good in Ha, I already now get less signal than with the Baader at F2.2 so perhaps the Antlia is shifted less to support more of the not so superfast telescopes out there. Baader has recently created a new line of filters that is also not shifted that much for Scopes >F2.3. So when you own a fast RASA I'd only buy the Antlia when you reserve the right to return it in the case that my 'fear' is right, that it is not shifted enough and will fail miserably in Ha. For Oiii I do not want to decide yet, my tendency goes to Antlia but first we'll have to see how it performs on a really bright star. When you have limited budget or you have limited clear Skies and have a fast scope in my >F2.2 range: I think there the answer is easy, the Antlia is cheaper compared to buying 2 seperate filters and you also get a lot more integration time when the moon is not to bright and you can use the benefit of the 'free' Oiii channel. The filters are not that far away in quality when the Halo's are a non-issue with the Antlia, and even when there are Halo's, they are likely under better control than with what we were used to with the Optolong L-eXtreme... More to come, I hope somebody else will be able to add results on the same target for the IDAS filters, they also seem to be a nice alternative.... Michael Here are the full jpegs of the stack:     |
