Adreneline

Thank you Adam.

A frustrating end to a frustrating year. This is 30 x 2min of Ha. I had hoped to get some OIII and maybe even some SII but as with so many imaging sessions this year it was not to be. ASI1600MM-Pro + Canon 200mm unguided on an CEM25-EC all controlled from an ASIair. The star shapes are not great and next year I plan to start guiding the iOptron and dithering. A Happy New Year to one an all on SGL and thanks for looking at this image and for all the help and feedback on my images throughout 2019. Adrian

I went for a Baader DIamond Steeltrack and it transformed my ED80DS-Pro. Prior ro fitting the Steeltrack focussing was very hit and miss and completely non-repeatable. To make things worse at high dec setting the drawer tube would just slide out of the OTA complete with camera, focus wheel, etc., resulting in hours of duff out of focus subs. The Baader was easy to fit and cured all the problems. HTH

No Christmas prizes for guessing which galaxy. Took a whole bunch of 30s reds, greens and blues and threw in a few Ha's as well using my Canon 200mm and ASI1600. All processed with copious glasses of wine and mince pies. Does make you think though - is there a speck of dust in there somewhere where they are celebrating their equivalent of Christmas. Merry Christmas to one and all and thanks for looking. Adrian

This is what it says in the attached Readme file (cut 'n pasted): STRIDE parameter comes from the fact that input into the neural network is NOT the whole image, but a small square tile of it. So image is transformed in small parts, size of which is STRIDE x STRIDE. So if you set it to 64, the image will be transformed in tiles 64 by 64 pixels. What is important, is that there can be inconsistencies between neighbouring tiles (this is usially happens near large stars), which will result in vertical and horisontal edge artefacts. The larger the STRIDE, the more of these artefacts you will likely to see, the smaller STRIDE, the smoother the result. BUT, with too small STRIDE, the running time might mecome an issue: running time is quadratic function of STRIDE, which means that if you decrease STRIDE by a factor of two, running time will increase by a factor of four. I find that STRIDE 64 works well. I use values like 128, 64, 32, but you can try any EVEN number. HTH Adrian

Magical. Adrian

I am sure the data is very good. Processing is such a challenge and there is no "one rules fits all" either. With data in short supply we have to make the most of what we've got. I do like subtle colours though so your image works for me. Adrian

Hi James, I prefer the second version. Hope you don't mind but I've had a crack at it as well. I did a little bit of noise reduction to try to reduce the coloured star rims and then bit more noise reduction to reduce noise overall in the image. I then had a fiddle with selective colour. I also reduced the star size across the image by a very small amount and tweaked the saturation a very small amount. (All in Photoshop). As you say it is purely subjective. Hope you don't mind Adrian

I really like starless images when it helps you to appreciate the structure of the nebula. Thanks for sharing. Adrian

Thanks Ragnar. Space might be spectacular but I personally don't believe it is garish. Thank you. A family portrait - just what we all need at Christmas Thanks for the kind comments. Adrian

Thank you Gina. I always use a gain of 139 and an offset of 56. The LP levels here are not conducive to pushing the gain any further. I've tried 200 gain and not really noticed any difference in the end result. Sticking to 139 and 50 makes life simpler with calibration frames. The Canon 200mm is really very good and I feel better than the Samyang 135mm in terms of star shapes in the extremities of the image. Definitely worth having a go with your f4 I would say. Adrian

That sounds worth a look at - thanks very much. Adrian

Hi Alan, Thanks Alan - I'll go in search and see what I can find. My basic RGB subs don't seem to contain that much colour - probably because they are so short (10s and 30s). I tried using ArcsinhStretch in PI and that introduces lots of colour but the stars all look a bit untidy and ill-defined. By the time I've restored the star shapes most of the colour seems to have evaporated away again so that when I come to layer them into the NB image it is very hard to see any real difference. I've also used basic HistogramStretch on the RGB image but again the colour is not really there and by the time it starts to show the stars are all getting a bit bloated. It's all a bit of a challenge. I think I favour little stars with a lack of colour to bloated stars with colour. Adrian

Thanks for the kind comments @assouptro, @alan potts and @geordie85. I took this image over two weeks ago and have been messing around with the processing ever since! Despite having everything thrown at it (filters and software) I still seem to have successfully lost all the star colours More cloudy nights forecast so perhaps I'll have another bash. Thanks again. Adrian

This is R, G, B, Ha, OIII and SII all taken with a ZWO-ASI1600MM-Pro fitted to a Canon 200mm f2.8 lens. The R, G and B were all 10s and 30s combined in Photoshop. The NB was all 30s, 60s and 120s, again layered in Photoshop. This is essentially RGB combined with HOS, pre-processed in APP, post-processed in PI and combined in PS. Thanks for looking. Adrian

What's to say? How about awesome! Adrian

Like what I haven't got! The Ha was pretty distinct but the OIII was really hard to get. Oh well - worth a go though. Thank you.

I don't know whether APP is infallible or not - so far it's not failed me. This was assembled in APP: All the Ha and OIII data for all the sessions was fed into APP in one go and the linear version of the above popped out. No attempt has been made to hide boundaries cos' I couldn't find any (as it is uncropped it is pretty obvious where they ought to be if you want to go looking). I also have the SII data but decided against using it. Adrian

I agree entirely. Part of me feels that if the object of the exercise is to image the nebula then it is the nebula I want to see. If we have the means to remove the stars (which I am not trying to image) then why not? - especially if it helps to reveal detail and structure that was not as apparent with the stars present. I've done my own starless Horsehead and Orion image but have been hesitant to post. Well done to @AbsolutelyN for sharing this excellent and revealing image. Maybe there should be a separate Starless Thread for those who like starless images - no stars allowed - a star-free zone. Adrian

Thanks Carole. The Oiii subs are very feint and I was reluctant to over stretch them. I could probably have gone to 180s subs but with the light pollution I’m not sure I would have gained anything. I’ll try and grab some more Oiii next time it’s clear and moonless - hopefully in my lifetime! The CEM25-EC is great for unguided use. I do a careful PA and a two-star alignment with the ASIair and then plate solve the target. Plate solving with ASIair is amazing compared with SGPro. The ASIair typically takes only 4 or 5 seconds to plate solve and sync with Sky Safari on the iPad and then it’s a case of centre the target and off you go! Adrian

My first attempt at this target. Not the easiest because surrounding trees and buildings really limit the time you can get on target. Furthermore IC 2177 never really clears the light pollution zone from a neighbouring town. All that aside this is 17 x 120s of Ha and 10 x 120s of OIII, calibrated and stacked in APP, processed in PI and Levels applied in PS with a small amount of noise reduction. Taken with a Canon 200mm lens with an ASI1600MM-Pro at gain 139/offset 50, a ZWO-EFW and ZWO-EAF; the whole thing controlled from an ASIair, unguided. Comments and criticism welcome. Thanks for looking. Adrian

Nice result, especially considering the moon being in attendance! Hope you don't mind but if you back off the black point a little quite a lot more of the surrounding nebulosity starts to appear. I've attached a .tif of your image with the black point set at 23:23:23 or there about followed by a little bit of noise reduction and a tiny tweak of the blue histogram to try to line them all up. Adrian m42.tif

Those are seriously impressive results especially recognising where you took them from. I cannot imagine imaging on the side of the street with all that light pollution. Well done!

Hi Helen, Not sure I can answer your question but yesterday I used my Atik414ex-osc with SGPro for the first time in some months and also had problems with cooling and image acquisition. In fact the cooler just went off on one and took the camera down to -26.37 degrees! - and then refused to re-warm. The only thing I had changed since last using the camera was to update the Atik Core software to the latest version - which I did because I wanted to try the latest version of Dusk. I similarly found the 414 was not correctly identified in SharpCap and I could not capture images properly either. In the end, because I did not want to waste the first clear night inages, I gave up trying to use the 414ex (and Dusk) and spent the evening using my ASI1600. My experience may not be of any help but there again .... Adrian

This is an attempt to combine RGB and NB data. The image consists of 30 x 120s of Ha and OIII plus 10 x 120s of R, G and B taken with a ZWO-ASI1600 + Canon 200mm. Image acquisition was all done using an ASIair controlled from an iPad. Pre-processing was completed in APP. PI was used for background extraction. Non-linear Ha and OIII were combined in PixelMath R:G:B - Ha : 0.3*Ha+0.7*OIII : OIII). Linear RGB were combined in ChannelCombination and then stretched using ArcsinhStretch. The two images were then combined in Photoshop and tweaked with Levels and a little noise reduction. I am sure it must be possible to get more star colour. I find it really hard to combine RGB and NB so any critical/helpful comment or advice would be most welcome. Adrian