- This topic has 10 replies, 4 voices, and was last updated 1 hour, 23 minutes ago by
.
-
Posts
-
I’d asked for advice on the BAA Comet Sections email discussion list and this question stems from that.
——-
I think the best version of the Comet from data I already have is now attached and on my member page; someone might be able to get more from the data and I’ve had a kind offer for someone to look at this which I will follow up. I’ve put the lights, darks and awful flats in a folder on Dropbox. https://www.dropbox.com/scl/fo/gpuillmqfefit55yjzjah/AE7sWTZkTxTwIBN1ExKHAHc?rlkey=ulsgudix2wq5dni93lxvloot4&dl=0
From reading the comments sent, and watching some more online tutorials etc, I think there are a number of areas I need to improve on, and I suspect it is a case of making small improvements in a number of areas, rather than one big improvement [others may disagree, and suggest I just stick to anaesthesia rather than trying to image comets!].
I think my areas to improve on include the bellow which I am sure is an inconclusive list:
1. I think I need my raw images to be a bit brighter, maybe I was too stingy on the ISO (800) for a Canon 6D. I would be interested to see how bright others’ individual subs are.
2. I need better flats. The image attached has been processed without any flats as my flats seem to introduce banding, which is more annoying than the vignetting or dust bunnies which remain otherwise. I suspect I need a quality flat panel rather than the thing I am using which must be flickering at an inconvenient frequency. I’ve included the flats in the Dropbox folder, if you could comment on their brightness (whether too bright or too dark that would be handy); I know there is banding which needs addressing.
3. I think my focus could be tighter. I do have live view and a mask, but without a bright star it is hard to get this spot on. With the DSLR it may be better for me to use the laptop and Canon Utilities and inspect the diffraction pattern more zoomed in. I suspect it would be better/easier to just use a dedicated CMOS camera and so it all on the laptop, though my FoV will be smaller.
4. I do need to learn more about processing. I’ve got GIMP which is free and seems very good, and watching more tutorials on curves and levels has helped, but there are a myriad of other tools there which will probably help, but are at present a mystery to me.
5. Everyone has lovely dark backgrounds to their images which contrasts nicely with the comet in the foreground; by striving to get the background darker, I am sure I am loosing some data from the comet. Addressing the other issues listed here may well help with this.
6. I am unsure how long individual subs should be, and the advantages and disadvantages of longer or shorter subs so would value thoughts on this.
7. I am also unsure how long the total run should be. Looking at the last 30 or so images on the BAA gallery of comets the mean integration time is 6.8 minutes and the median time 3 minutes. Having seen time lapse of comets the tail is dynamic and presumably exposing for too long will just cause a blur of the tail.
8. When using DSS and trying to get both comet and stars in focus and not streaking, unless the total integration time is less than about 2 minutes I am finding the core of the comet creates a streak and does not stay as a compact entity.
9. I suspect a DSLR is just not as “good” as a dedicated astrophotography CMOS camera.As I say, work in progress…
Any replies to the forum post will be grateful received.
Thanks
James Dawson
Attachments:
I’m no expert but would start with the flat field issue – that seems a decent easy-win as the background brightness variation from vignetting is hiding the extent of the comet.
How many counts were in your RAW flat field frames?
I agree with Grant. You want to get your cal frames right before you go any further. This is really important for something faint and fuzzy like a comet. I tend to take 50 each of flats, flat darks and darks at the same temperature/gain/ISO. You need to make sure that the flats are taken at the same focus and aperture settings. I expose the flats to have a median of around half the saturation level. All of the cal frames are then averaged before you use them. The calibrated subs are then:
calibrated sub = (light – dark) / (flat – flatdark)
I tend to use sky flats but I do have a cheap artist’s tracing panel that I bought from Amazon which works well if I can’t get a sky flat.
One other thing. For Bayer colour sensors I tend to debayer the (flat-flatdark) to extract the green channel before then using it to divide the (light – dark). This is because I tend to use sky flats which are generally blue and which would therefore change the white balance. There are lots of ways to do this which ultimately work the same way but the key thing is to start with high quality cal frames by averaging lots of them.
In terms of what ISO to use with your camera have a look online at astro website for your particular camera. Assuming that you are using raw format there is an ISO which is generally optimises read noise and dynamic range. An example for the Mark 1 Sony A7s is here: http://www.astrophoto.fr/sony_a7s_measures.html.
Thanks both. My flat panel is an illuminated artists sketch pad from Amazon; I wonder if I should have just made the pad brighter so that I could have reduced the exposure length which may have avoided the flickering. I’ll experiment.
The DSLR doesn’t use counts so I used the histogram and aim for a range between 25-50% and take 5 or so at each setting to later review on the PC and use – I only use one exposure batch, not flats from different exposure settings.
Thanks again.
James
Hi James,
I ran your darks, flats and lights through DSS, selected comet and stars tracking (star freeze effect) and got the attached image. It’s a pleasing view but it didn’t track the comet’s coma as well as I hoped. I’ll try it again.
You had 6 darks and 3 flats. As Nick advised, you need double figures of those to create good master darks and flats. DSS does a lot of the work for you but if you have only a few darks and flats, and if the masters are not good quality, they can produce a final image which can be worse than simply stacking all the lights.
Cheers,
Alex.
Attachments:
Alex, thanks for this. The image you have created is much more akin to what I was hoping to get. You’ve managed to bring out the tail in lots more detail and there is structure visible within it, without blowing out the nucleus. I can still still the impact poor flats are having on the data, but I am really impressed. I’ve made an animated gif to show the image you could get out of my data vs the image I got out. Given it is the same data, I think there is something you are doing in the stacking process and in the post-processing step which I am not doing. Maybe sometime I can watch what you do and learn from that. I am so grateful for your time.
Attachments:
Hi James,
You’re welcome. On the baa-comet forum Nick Haigh outlined the workflow in DSS for comet stacking. This is also what I use.
To minimise saturating the coma yet bring out details in the tail you use the Luminance tab. Use your mouse to draw a box around its head and part of the tail, then gradually tweak the pairs of sliders for Darkness, Midtone and Highlight, each time clicking Apply to see the change – or Reset.
This is finicky work – it’s almost a Dark Art – subtle changes in the values can make a big difference to the image.
I’ve taken some screen shots of my steps in DSS, so I’ll retry comet stacking (stationary stars+comet option) and send you them (offline) in a PDF, either later today or tomorrow.
Cheers,
Alex.
Hi James,
Not a detailed workflow, rather the key steps to enable DSS to stack comet images. See Nick Haigh’s advice here:
https://www.simplelists.com/baa-comet/msg/25949574/
Alex.
9. I suspect a DSLR is just not as “good” as a dedicated astrophotography CMOS camera.
James,
A DSLR can be used to get very good widefield images of comets, whole disc images of the Sun and Moon, and details of sunspots and lunar features. Your DSLR has a full-frame 14-bit sensor which can record a greater tonal range than a 10-bit sensor such as used in the Seestar S50, for example.
I have some dedicated planetary imaging cameras yet it’s still fun to attach a DSLR to a ‘scope and get some nice ‘large frame’ lunar vistas.
Alex.
Attachments:
James – As Alex says uncooled cameras like DSLRs or mirrorless cameras can do a very good job for generally less money than a cooled camera. Modern CMOS sensors are very low noise even at room temperature. Here is an image of the comet taken with a Sony A7s:
https://britastro.org/observations/observation.php?id=20241020_080600_25df8a3cd03b0737
I think a lot comes down to good processing rather than cooled or uncooled sensors.
- You must be logged in to reply to this topic.