M51 reprocess

[silentrunning]

With this revision of my last M51 post I have given up trying to drag out detail in the noise and forsaken that detail for hopefully a less noisy result. Still no noise reduction just a more conservative stretch on the L layer. I would be interested in any comments comparing this with the original shown below this one.

 

 

and the original...

 

 

 

[Clive]

The original gets my vote.

[+Graham]

Thats much easier on the eye John.

 

[silentrunning]

Thanks Clive and Graham. 

 

Trying to process the noise out drives me nuts. The answer is as usual more data. I most spend more time on an individual target rather than swinging to the next one at the first opportunity - been saying that to myself for years!

[red dwalf]

new one looks amazing and much better a tiny bit of noise reduction in Photoshop, maybe despeckle and a smart sharpen and it would be better on the noise although there is little of it left. what processing software did you use ?

[silentrunning]

3 hours ago, red dwalf said:

new one looks amazing and much better a tiny bit of noise reduction in Photoshop, maybe despeckle and a smart sharpen and it would be better on the noise although there is little of it left. what processing software did you use ?

 

Thanks Rob.

 

Main s/w for processing is PI and I've tried all the NR tools it has to offer. I have also tried combinations of PI and PS tools and numerous free and bought plugins and actions for PS and the odd standalone program. I've worked in non-linear and linear modes and applied all sorts of combinations of NR between the RGB and L layers.

 

I honestly think I'm just asking to much given the limited data acquired. The original shows deeper levels of the "fans" of stars thrown out by the collision but you can see how deep the extreme extensions of these fans are in the noise. To keep them in and reduce the noise they are essentially made up of is a big ask. 

 

Personally I also prefer the latest one. As Graham sums up nicely it is "easier on the eyes" and I find I can do so much more with the image in terms of more pleasing contrast, tone and vibrancy once I have given up trying to drag every last detail out of the noise. I have also tried just about every masking technique to reduce the borders between NR'd areas of low contrast and brighter ones which don't require it. These borders can become the main distraction after some NR routines have done a superb job on the selected areas.

 

I have tried despeckle and smart sharpen in PS but usually stick with PI as the routines are actually built for the type of noise we astrobods encounter - not to say PS doesn't do a superb job on some images. Having said that I do usually finish up in PS using things like Vibrancy adjustments as they are single step very effective tools.

 

The same applies to sharpening. I tried all sorts of packages. I mainly stick to deconvolution in PI and also a PI script called AdvSharpening which again is attractive because it does in a single step what would take some time with the individual PI processes.

 

The new image has no NR applied. I just did a stretch to the L layer that did not reveal the very faint stuff before combining with the RGBHa data. Not exactly clipped but less of an actual stretch.

 

 

 

 

 

 

Edited May 10, 2020 by silentrunning

[+Derbyshire Dave]

Really great stuff John, congratulations on POTM

 

I missed this when you posted it, really glad to have seen it now