Since we are talking about low light performance...

... I thought I'd resurrect my study of The Great Nebula in Orion (M-42) that I did a couple of years ago, At the time, there sere some members who were (heroically) doing deep sky photography with Foveon-sensor cameras.

M-42 is the brightest nebula visible from the northern hemisphere; most other deep sky objects (DSOs) are much, much dimmer. So being able to render M-42 in a pleasing manner should be a very low bar to entry into the world of astrophotography.

I imaged M-42 using the Sigma 150-600 mm lens (I have this lens in both Sigma and Nikon mounts), and took 30 second exposures at ISO 100 (Sigma; 800 for Nikon) and f/6.3, on a tracking mount. For 30 sec exposures one doesn't need to guide, and typically one will take many, many such exposures and average them to increase the S/N of the nebula.

The images below are each a single 30 sec exposure, and they might have been the stars in the movie 'the good, the bad, and the ugly'

Shown here in reverse order:

The left-most is the Sigma SD1M, and the obvious color blotching will make signal averaging difficult (to say the least). The middle is a run-of-the-mill 24 mp Bayer-sensor camera (Nikon D7200). Nikon reds are not particularly strong, but the random nature of the noise means that signal averaging will be possible.

The third image is a Nikon camera (D5600) modified for better red sensitively. You will appreciate why one might be willing to spend a few hundred dollars for this mod, since it  means that a good DSO image might take only a couple of hours of light collection rather than a couple dozen hours.

The mod improves the sensitivity to red so much that using just this one 30 sec exposure, a lot of detail can be brought out:

LHS-unfiltered. RHS filtered and additional detail brought out

(BTW, those streaks are geostationary satellites. Since the camera is moving to track stars, the geostationary satellites appear to be moving.)

David

p.s. So my definition of "low light" is "not visible to the unaided eye". You can see just the central part of M-42 with the naked eye - it's that part in the above image that is over-exposed.

A fascinating study, David.

I'm not sure if I understood you completely... did you actually try stacking the SD1M images?

I was under the impression that the SD1M makes a dark frame exposure for shutter speeds 30 secs and over but I don't know how you would access those frames. So I imagine that you would have to take your own dark frames anyway.

I've been planning on using my GFX 50S II for astro work, but I haven't done so yet. The longest lens I have is the 100-200 f/5.6 which is probably a bit slow.

It's too bad you couldn't do any of this kind of photography anymore, now that all those Starlink satellites have ruined the night sky.

Scottelly wrote:

It's too bad you couldn't do any of this kind of photography anymore, now that all those Starlink satellites have ruined the night sky.

They've made it harder, no doubt. But I believe that the stacking software has an option to delete a trailing light (like a plane or satellite) if it only appears in one image. I haven't tried that yet.

David

SigmaChrome wrote:

A fascinating study, David.

I'm not sure if I understood you completely... did you actually try stacking the SD1M images?

I was under the impression that the SD1M makes a dark frame exposure for shutter speeds 30 secs and over but I don't know how you would access those frames. So I imagine that you would have to take your own dark frames anyway.

I've been planning on using my GFX 50S II for astro work, but I haven't done so yet. The longest lens I have is the 100-200 f/5.6 which is probably a bit slow.

I didn't show the stacking results because the further away you go from the raw image, the more opportunity to introduce some bias in the workflow.

But, you asked, so I redid the processing for stacks of 14 images (all that I took with the SD1M), making all the steps as similar as possible,

LHS single SD1M exposure,, Middle 14 SD1M exposures, RHS 14 D5600 exposures:

I left the background above black so the noise (and faint, real features) are visible.

David

DavidWright2010 wrote:

SigmaChrome wrote:

A fascinating study, David.

I'm not sure if I understood you completely... did you actually try stacking the SD1M images?

I was under the impression that the SD1M makes a dark frame exposure for shutter speeds 30 secs and over but I don't know how you would access those frames. So I imagine that you would have to take your own dark frames anyway.

I've been planning on using my GFX 50S II for astro work, but I haven't done so yet. The longest lens I have is the 100-200 f/5.6 which is probably a bit slow.

I didn't show the stacking results because the further away you go from the raw image, the more opportunity to introduce some bias in the workflow.

But, you asked, so I redid the processing for stacks of 14 images (all that I took with the SD1M), making all the steps as similar as possible,

LHS single SD1M exposure,, Middle 14 SD1M exposures, RHS 14 D5600 exposures:

I left the background above black so the noise (and faint, real features) are visible.

That looks pretty impressive to me, David. The star colours look fairly good too.

Edit: It was very late last night when I responded. I just re-read your post. For some reason I thought the RHS was the final SD1M result, but now I realise it's not.

1 1/2 years ago I captured comet neowise using just a DP2s:

https://www.dpreview.com/forums/post/64170315

not nearly as good as your image but still it is possible to do low light even with a simple camera like the DP2s