3 replies to this topic

[james7ca]

This was taken last Friday (March 24, 2023) about one hour after sunset using a Takahashi FC-100DZ and a ZWO ASI183MM Pro camera. The image was created from 256 subs that were each exposed for 250ms and thus it had a total integration time of 64 seconds. The faintest stars that were recorded in the background sky were around magnitude 12 and you can see the brightest (HD18405) at magnitude 7 that is just to the lower left of the moon and only a few minutes from occultation.

 

Since the moon was moving in respect to the stars I had to combine two different registrations (one on the moon and the other on the stars) so that both the moon and stars were reproduced with as much detail as possible. This is probably the sharpest image of earthshine that I've ever gotten and at the original scale it is possible to see structure in the interior of the crater Aristarchus (the below image is about one quarter of the size of the original).

 

Image capture using SharpCap Pro, processing with Autostakkert! (registration on the moon), Registax (wavelet sharpening), SiriL (one star registration), PixInsight (histogram and dynamic range adjustments), and Photoshop 2023 (photo composition, tweaks, and photo frame). There are additional details in the image captions (click on the preview to read them).

 

You can see what the sunlit side of the crescent moon looked like in an earlier post of mine here on CN:

 

  https://www.cloudyni.../#entry12589750

 

Questions, comments and criticisms (QC&C) welcomed and thanks for visiting.

 

[EDIT]

I uploaded a higher contrast version because the smaller size (from the original) that I posted was lacking in detail. I think the new version looks better.

 

Also, I thought of another way to blend the stars into the background and by doing that and with some further tweaks to the contrast and lightness of the earthshine I think I can produce an even better version (for posting maybe tomorrow or the next day).

 

Uploaded new version with improved stars and detail.

[/EDIT]

Attached Thumbnails

• 

Edited by james7ca, 29 March 2023 - 02:01 AM.

[R Botero]

Excellent James!  I need to practice Earthshine shots as I've been told that Coronal streamers for the forthcoming total solar eclipse are around the broadly the same brightness as the Moon's non illuminated areas.  I see you used an colour filter. I need to test without.

 

Roberto

[james7ca]

Here is a treatment using a different sequence where I've tried to maintain detail in both the sunlit and earthshine portions of the moon. In this case I used a lower gain and lower exposure (details in the image captions). The master sequence had 256 subs that were each exposed for 90ms (total integration time just over 23 seconds).

 

This is probably closer to how the moon would have looked visually through a small telescope. The earthshine is pretty weak and noisy (best viewed in a darkened room), but I can definitely see the craters Aristarchus, Plato, Copernicus, and Tycho as well as most of the maria (including Grimaldi, center left near the moon's limb).

 

To produce this image the same master was processed twice to favor either the earthshine or the sunlit portions of the moon. Then, I did a weighted average using PixInsight's PixelMath to combine the two images (I didn't use any masks, the "blend" came from the average of the two images).

 

Image capture using SharpCap Pro, processing with AutoStakkert!, Registax, PixInsight, and Photoshop 2023.

Attached Thumbnails

• 

Edited by james7ca, 29 March 2023 - 06:27 AM.

[james7ca]

Roberto thanks for the notice and similarly to everyone else for the likes.

 

Below are the two masters that were combined to produce the image in post #3 (using PixelMath to produce a weighted average). You can see that I lost a little brightness in the earthshine, but that area of the image is so noisy that I don't think a brighter image would be much better.

 

I also used a fixed ADU to determine the cutoff for the weighted average and that type of programming condition usually results in artifacts at the condition boundary. So, I could probably get better results by introducing a random cutoff that varies slightly between different ADUs (i.e. a smoother blend at the boundary). What I mean is that instead of a cutoff at an ADU of 0.18 it would be better to use a varying cutoff that ranged randomly from 0.16 to 0.20 (or something like that). This cutoff is used to determine whether a particular pixel (brightness) should be weighted more toward the earthshine image or that for the sunlit image.

 

Both of these images were produced from the same sequence (256 x 90ms), they just got difference processing (dynamic range compression, arcsinh and masked stretch for the earthshine version, pretty much a simple histogram stretch for the sunlit version). This does show, however, that it is possible to capture both the sunlit and earthshine portions of the moon using a single exposure setting (although not ideally for both). The ASI183MM Pro is a 12-bit sensor, so maybe one could do a little better with a true 16-bit sensor. But I think you'd still have issues trying to get that 16-bit range to show up properly on a typical computer display or TV.

Attached Thumbnails

• 

Edited by james7ca, 30 March 2023 - 04:45 AM.