NGC 1333 lum.....done?

[Rodd]

25 minutes ago, vlaiv said:

Mind posting fits stack of lum (integral version of 411, but you can post also this reduced version) - I fancy having a go at stretch and noise control myself?

Here you go.  Remember, crop and DBE already performed, and in the case of  Lum 305, a touch of mure denoise as well (a very small amount).

Lum-305.fit

Lum-411.fit

[wimvb]

9 minutes ago, Rodd said:

I remember.  Good idea (imogi with eyebrows moving up and own rapidly like some one considering a devious plot...couldn't find that one)

[vlaiv]

I think there is a lot to be improved by using appropriate noise control mechanisms.

This is Lum with 411 subs in it:

[carastro]

I have been imaging this target on and off for over 3 years and even putting bits of data here and there together from various dark sites it is STILL noisy.  

I think it is just one of those targets that needs really dark skies and loads of data.  

It's still a WIP and I have lost count of the amount of Lum data I have for it.  Probably half the amount your have as I have limited opportunities to get to dark locations, but mine is with a CCD camera.

https://www.astrobin.com/full/366073/0/

Going to have a bash with your linked files if I can open them.

Carole 

 

 

[Rodd]

2 hours ago, ollypenrice said:

Do you actually need to stretch it quite so hard?

How about this Olly--lesser stretch.  This is the 305 sub stack.  2 versions.  The second was with a slight curves adjustment (shallow S)

[carastro]

OK I managed to open your files.

Is this any better?

I treated one of the stacks to a stretch and then a noise reduction in dust and scratches, more than I would normally to an image, and then a gentle stretch to the other file, and then pasted the gentler stretched one on top of the noise reduced one, and then blended them at 50% for the top layer in Photoshop.

What do you think? 

Carole

Lum-411& 305 combi.tif

[Rodd]

31 minutes ago, vlaiv said:

This is Lum with 411 subs in i

I think the Balrog (NGC-1333  shows a bit too much NC.  if you can notice NC--its too much.  Its almost like the smell of mercaptan in natural gas.  But I see you point.  My recent efforts are closer to this

Rodd

[Rodd]

3 minutes ago, carastro said:

What do you think? 

I think you are on to something.  Pretty nifty.

Rodd

[vlaiv]

13 minutes ago, Rodd said:

I think the Balrog (NGC-1333  shows a bit too much NC.  if you can notice NC--its too much.  Its almost like the smell of mercaptan in natural gas.  But I see you point.  My recent efforts are closer to this

Rodd

Which part do you think is showing too much NC?

A, B or C?

[Rodd]

4 minutes ago, vlaiv said:

A, B or C?

The region defined by all 3.  the central figure (I think it looks like a Balrog) is a bit soft.  If you draw a line from A bisecting a line drawn from C to B, the line from A passes through a region that has allot of details that could be sharper.  Maybe a bit of sharpening will fix it.  Or protecting this region when you apply NC.  Remember high signal areas need much less NC.

Rodd

[Rodd]

12 minutes ago, vlaiv said:

A, B or C?

 

Here is a crop from another attempt--not the last one I posted, which may be better, but this is just an example.  The outer regions are noisy--we know that--so ignore them.  Just the central feature.  I want to protect the structures in there and reveal some sharpish edges. 

[Rodd]

18 minutes ago, vlaiv said:

Which part do you think is showing too much NC?

This one is the previous one cropped--better, less noise as I did not tweak with Hist.

[vlaiv]

Don't know, it looks that all features captured are there, here is simple levels/curves stretch of that region without any noise reduction / control what so ever:

As you see, A area has sufficient SNR so no noise reduction was necessary - I masked that off when doing noise reduction.

You might be talking about local contrast enhancement - nebulosity has some variations and you may want to bring out those (I personally don't like it over emphasized). As far as I can tell - there was no loss in features or sharpness with noise reduction / control - all stuff present in this simple stretch is still there. Also - no artifacts associated with noise reduction are present as far as I can tell - there is sill some noise left in there, but it looks "natural" not blocky or jagged edged.

Maybe we just perceive differently noise and its effect on the image, and also there maybe slight mismatch in monitor calibration (just occurred to me that there will be some variation in both viewing conditions and monitor calibration).

[Rodd]

7 minutes ago, vlaiv said:

Don't know,

This is not the same crop you posted before.  It has less NC and the features are sharper.

[vlaiv]

1 minute ago, Rodd said:

This is not the same crop you posted before.  It has less NC and the features are sharper.

No it's not - it is simple stretch without any noise reduction, I posted it as a comparison to previous one - to my eyes all features are there and only issue with sharpness is star halos present due to a bit stronger stretch (which is possible with noise control).

[vlaiv]

Maybe this will help figuring out if any features / details are missing from NC version:

[ollypenrice]

1 hour ago, Rodd said:

How about this Olly--lesser stretch.  This is the 305 sub stack.  2 versions.  The second was with a slight curves adjustment (shallow S)

Yes! This works for me. I don't see that the softer stretch is missing anything at all. The problem with stretching the entire brightness range and then bringing in the black point each time is that you take the darkest parts above the noise floor. That's why, when my darkest regions have hit the noise floor, I only stretch above them. This is particularly applicable to images which contain 'sooty' regions which are significantly darker than 'empty' background sky. Such background sky is actually quite bright compared with genuinely dark nebulae, some of which are to be seen in this image.

Full size or less than that, as an ideal? When I'm doing small targets at modest aperture (TEC140 at 0.9"PP) it has to be full size or the target is too small on the screen. This means lots and lots of data. Widefields, and most especially mosaics, don't have to hold up to full size but I'm motivated to try. However, using a higher resolution and presenting it at less than full size does not give remotely the same result as obtaining the same image scale at lower resolution and full size. The stars in the downsized higher res image will be smaller and tighter and the whole image will have more finesse and subtlety.

Olly

[Rodd]

9 minutes ago, vlaiv said:

to my eyes all features are there and only issue with sharpness is star halos present due to a bit stronger stretch (which is possible with noise control).

Yes the features are there--but they are soft(er).  I had to pic myself off the floor (I have epilepsy)...just kidding.  You flickering demonstration not withstanding (a very good one I must say), the crop you originally posted looked soft to my eye.  It makes no difference if that is untrue and just a perception.  When images are being looked at, they do not flicker between versions.  We enjoy an image for what it is.   

But in this case this debate is moot, because an image processed from this data will have a deficiency somewhere.  the data is just not good enough, or plentiful enough, to render a truly exceptional image stack).  What's better, too much NC but less noise and s0fter features, or less NC, more noise and sharper features.  that is a matter of taste.  I really am averse to several things in AP:

1) too much noise control.  I am very sensitive to it.  I may use too much---but I know it and don't like it.  If its necessary, then I will not be satisfied with my image, no matter how good other aspects turn out.

2) Too much star control.  I like natural looking stars even though I may not be that good at rendering them.  It just means I am unsatisfied with my work (no surprise there).

3) Too much sharpening....something I do often.  I am trying to change.  

4) Unnatural saturation....I make very good cartoons.  Again, working on it.

5) Green in NB images.  For once I am satisfied.

Rodd

[Rodd]

5 minutes ago, ollypenrice said:

However, using a higher resolution and presenting it at less than full size does not give remotely the same result as obtaining the same image scale at lower resolution and full size. The stars in the downsized higher res image will be smaller and tighter and the whole image will have more finesse and subtlety.

Struggling to understand this.  Are you talking specifically about crops of images from widefield compared with crops from long FL made to look the same size?  For instance I think my image (any of these versions) look OK at normal viewing.  Its kind of like print--many defects are hidden at that level of zoom.  But at full res viewing on the forum--they break down.  I guess down sizing so full res is no closer in than normal viewing fixs this.  The only problem is a truly good image stands up to full res easily.

Rodd

[wimvb]

And here's my feeble attempt. There were black pixels in the original, so I used cosmetic correction on the image.

Noise reduction before stretching (an oldie but a goodie: AtrousWaveletTransform)

Downsampled before posting

(click for a larger version)

[ollypenrice]

2 hours ago, Rodd said:

Struggling to understand this.  Are you talking specifically about crops of images from widefield compared with crops from long FL made to look the same size?  For instance I think my image (any of these versions) look OK at normal viewing.  Its kind of like print--many defects are hidden at that level of zoom.  But at full res viewing on the forum--they break down.  I guess down sizing so full res is no closer in than normal viewing fixs this.  The only problem is a truly good image stands up to full res easily.

Rodd

Sorry, I wasn't very clear.

Take two images which present Field of View X at a certain size on your screen and have comparable noise levels. One was taken at higher resolution but is too noisy to stand up to being viewed at full size, so the imager presents it resized downwards.  The other is a lower resolution image shown at full size. Remember that, to the person looking at them, the noise level now seems the same. My contention is that the two images will look very different, with smaller stars and more finesse from the downsized image.

To do this comparison in reality you'd either need one image to be a crop, or one of the cameras to have a larger chip, or one image to be a mosaic. As an example, here's a downwardly sampled mosaic of the NAN with 'x' level of perceived noise. I could shoot the same field, single frame, in a camera lens and get to 'x' level of perceived noise in that, but I think the images will look very different. Most importantly I think the camera lens will produce fewer but much larger stars.

My conclusion is that it isn't the end of the world if you can't present the image at full size.

Olly

 

[Rodd]

1 hour ago, wimvb said:

And here's my feeble attempt

Not feeble at all.  What was the purpose of downsampling?

Rodd

[Rodd]

17 minutes ago, ollypenrice said:

My conclusion is that it isn't the end of the world if you can't present the image at full size.

Now I see.  I bet this image could stand up to a much bigger size than you have it currently!

Rodd

[wimvb]

5 minutes ago, Rodd said:

What was the purpose of downsampling?

Just to get the file size down, no other reason.

[ollypenrice]

10 minutes ago, Rodd said:

Now I see.  I bet this image could stand up to a much bigger size than you have it currently!

Rodd

It can, yes, but it isn't in its comfort zone at full size. Mind you, there are processing decisions to be made which depend on the intended presentation size. If you're not going for a full size presentation you can do more extreme star reduction, for instance.

Olly