My much shorter ride on the Christmas Horsey

[symmetal]

To complement Olly's posting, here's my 84 minute ride on the horse, comprising 168 exposures of 30 sec. using the ASI2600MC on a RASA 11v2. As many objects are very bright I opted for shorter exposures. Even these well exceeded the 5x read noise swamping factor on the darkest areas, and I have SQM 21.4 skies on average. No where near the depth of Olly's 4.5 hour effort though. 🙂

Alnitak caused some rather noticeable diffraction effects including a curved arc, which is likely related to the cable routing shape, and I had a go at removing these with the 'Patch' tool in Photoshop. This is similar to the cloning tool but blends the original patch edge colours across the area you're cloning from. Also there were two odd blue tinted patches which I removed fairly well with masked layer blending. The original image is at the end for comparison.

I think the default Star Reduction amount in the latext SXT is too high as even with a higher than normal star layer stretch they are still tiny. There are no blue star halos on mine compared to Olly's though. Whether this is due to only 1/6 the individual frame exposure coupled with an Astronomik L-3 filter, I'll have to test. 

Here's the original image with the diffraction effects around Alnitak, and the circular blue patch near the centre and the blue vertical patch at top centre. I assume they're flares from Alnilam just out of frame top left but why it creates a circular flare seems odd. I took new flats the following day so it's not dust, and the fast scopes don't seem to show dust bunnies anyway.

Alan

[tomato]

Great detail in the nebulae and Alnitak is really well controlled.

[symmetal]

In my PI SXT processing of the above image it removed the stars but left the extended diffraction effects from Alnitak etc. on the background layer as I expected and so had to carefully remove them in PS. The little galaxies as seen on the annotate image above were left on the background layer, again as I expected.

I thought I'd reprocess and bin the data by 2 straight away in PI rather than downsample the final result by 2 in PS as my above image is. This should give better S/N in the image.

However when I ran SXT this time it removed all the diffraction effects too and put them on the star layer along with the surrounding blue flaring leaving a really clean background layer with no residual star bits. This should make removing the diffraction effects much easier. However BXT also put all the little galaxies on the star layer too which is not so good, as it'll make the galaxies too dim when I add the stars back in at a lesser stretch. This means I'll have to manually move the little galaxies back to the background layer before stretching. Here's a PI screenshot showing the layers.

This is the first time this has happened in my SXT processing. Has anyone one else had anything similar. I've always binned the 6200 mono data by 2 in PI and it's always behaved as I expected, leaving diffraction effects and galaxies on the background layer.

Alan

[ollypenrice]

That's a very nice rendition.

For the record, our cables are routed thus;

Regarding arc-like artifacts, I lasso them fairly closely in Ps CC and then use content aware fill. This seems to work astonishingly well.

Do you ever get a kind of grid-pattern artifact from Star Xt? It only appears on certain images but is a royal pain when it does. (Edit: I always use Large Tile Overlap.)

Olly

 

 

Edited December 23, 2023 by ollypenrice

[tomato]

I have only ever managed to get SXT confused once but when it did, it lost it big time.😊 Do you like the ghoulish figure in the centre, the Conspiracy Theorists would have a field day with that.

[symmetal]

5 hours ago, ollypenrice said:

Regarding arc-like artifacts, I lasso them fairly closely in Ps CC and then use content aware fill. This seems to work astonishingly well.

Do you ever get a kind of grid-pattern artifact from Star Xt? It only appears on certain images but is a royal pain when it does. (Edit: I always use Large Tile Overlap.)

I'll try content aware fill and see how it compares. For filling in narrow areas it sounds a good method. I haven't had the grid artifacts from SXT. It's always behaved itself until now, when it's moved the galaxies too. 

I have a similar shape cable router though clipped to the scope front edge. I carefully designed it so that the area in front of the corrector was exactly 180 degrees. The RASA 11 has a larger central obstruction so there's room to run the cables down the side of the camera.

5 hours ago, tomato said:

I have only ever managed to get SXT confused once but when it did, it lost it big time.😊 Do you like the ghoulish figure in the centre, the Conspiracy Theorists would have a field day with that.

That's impressive, and indicates what the AI was 'thinking about' at the time. Maybe when it was being trained with analysing thousands of star images a rogue image got inserted for fun. 🙂

Alan

[symmetal]

I did a test with a Coma Cluster image and indeed SXT does remove galaxies below a certain size even though they are still elliptical shaped. Top is original, bottom left is SXT background and bottom right in binned 2x2 SXT background. The smaller core must look more star like and the elliptical overall shape is assumed to be a distorted star. This explains why the annotated image may indicate small PGC galaxies which aren't visible on the image. 

Alan