hangs on wipe

Bugs, glitches and unexpected behaviour.
User avatar
admin
Site Admin
Posts: 3367
Joined: Thu Dec 02, 2010 10:51 pm
Location: Melbourne
Contact:

Re: hangs on wipe

Post by admin »

Lou wrote:Hi Ivo,
I upgraded my old Desktop PC today and I run a quick test - Laptop vs. Desktop PC.
11min15sec vs. 4min 50sec for the final noise reduction step.

My Desktop PC is operating with 16GB Ram and a AMD Ryzen 5 2600, 6x 3.40GHz processor. No new graphic card.
Very happy with the improvement and no more problems with the "hangs on wipe" :D

CS
Lou
Fantastic choice! The Ryzen series CPUs are beasts! :twisted:
Ivo Jager
StarTools creator and astronomy enthusiast
happy-kat
Posts: 372
Joined: Sun Feb 01, 2015 11:31 am

Re: hangs on wipe

Post by happy-kat »

I too am having trouble with hanging on wipe using latest bleeding edge version.
Windows 10 64bit
Wipe does work for me but as soon as I add a mask, either before selecting the wipe module or when in the wipe module on picking Do for the wipe the software hangs.
Slightly frustrating as I am trying to protest my globular cluster from being wiped.
User avatar
admin
Site Admin
Posts: 3367
Joined: Thu Dec 02, 2010 10:51 pm
Location: Melbourne
Contact:

Re: hangs on wipe

Post by admin »

happy-kat wrote:I too am having trouble with hanging on wipe using latest bleeding edge version.
Windows 10 64bit
Wipe does work for me but as soon as I add a mask, either before selecting the wipe module or when in the wipe module on picking Do for the wipe the software hangs.
Slightly frustrating as I am trying to protest my globular cluster from being wiped.
Hmmm... Does the same cause (and solution) apply to you?
Ivo Jager
StarTools creator and astronomy enthusiast
happy-kat
Posts: 372
Joined: Sun Feb 01, 2015 11:31 am

Re: hangs on wipe

Post by happy-kat »

Patience was the same of the game.
With a mask set walk away do something else then go back and it will be done.
But
The mask was ignored and the entire image was wiped.
I could repeat this, the mask is always ignored on wipe.
The mask is still there just ignored the whole image is effectively wiped.
The same happens in version 1.3.5.289 as it does in 1.4.332

Wipe is also putting the image background way too close to the black point and effectively clipping the image when using wipe colour and brightness option.
I am going to try process again to see if the clipping is coming about from the Contrast module I used directly after wipe.
User avatar
admin
Site Admin
Posts: 3367
Joined: Thu Dec 02, 2010 10:51 pm
Location: Melbourne
Contact:

Re: hangs on wipe

Post by admin »

happy-kat wrote:Patience was the same of the game.
With a mask set walk away do something else then go back and it will be done.
But
The mask was ignored and the entire image was wiped.
I could repeat this, the mask is always ignored on wipe.
The mask is still there just ignored the whole image is effectively wiped.
The same happens in version 1.3.5.289 as it does in 1.4.332

Wipe is also putting the image background way too close to the black point and effectively clipping the image when using wipe colour and brightness option.
I am going to try process again to see if the clipping is coming about from the Contrast module I used directly after wipe.
Hmmmm.... I've been trying to replicate this, but to no avail. I'm wondering if there might be a misunderstanding as to what Wipe is doing with the Mask?
Masking out in Wipe does not mean the area will be unaffected in the end-result. It just means that the area is not used as input for modeling gradients.
The fewer pixels you give Wipe to work with, the harder it has to work to create a model for the complete image; the pixels you mask out still need values calculated for them to subtract. They will still be corrected; all pixels in the image will be corrected.

You should use a mask for Wipe in two cases;
  • You have dark anomalies (e.g. darker-than-real-background pixels) in your image that could confuse Wipe.
  • You have a large area of contiguous nebulosity (without background poking through) in your image which Wipe may start mistaking for a gradient at higher Aggressiveness settings; Wipe may start removing the nebulosity a little in a larger-scale context.
To see whether you even need a mask, see what Wipe produces with- and without a mask. You may find there is virtually no difference at the default Aggressiveness settings.

Does this help?
Ivo Jager
StarTools creator and astronomy enthusiast
happy-kat
Posts: 372
Joined: Sun Feb 01, 2015 11:31 am

Re: hangs on wipe

Post by happy-kat »

Yes thank you.
Mask has different effects in different modules, I had wrongly assumed the unmasked area would be ignored both in the calculation and the application of the wipe.
User avatar
admin
Site Admin
Posts: 3367
Joined: Thu Dec 02, 2010 10:51 pm
Location: Melbourne
Contact:

Re: hangs on wipe

Post by admin »

happy-kat wrote:Yes thank you.
Mask has different effects in different modules, I had wrongly assumed the unmasked area would be ignored both in the calculation and the application of the wipe.
Great! If you have any trouble with Wipe, let me know (preferably separate thread).
Clear skies!
Ivo Jager
StarTools creator and astronomy enthusiast
ajh499
Posts: 13
Joined: Sun Dec 09, 2018 9:23 am

Re: hangs on wipe

Post by ajh499 »

I'm having the same problem with wipe when a mask is present, it's unbelievably slow, to the point that it's useless with a mask.

I had originally thought that StarTools was just crashing, but it's not, it just takes a really long time.

I tested it today with a 5500 x 3600 TIFF file. Without a mask it takes about 10 seconds to do a wipe, with a mask it took over 20 minutes! That surely cannot be right.
It's definitely due to the size of the mask, the fewer pixels included, the longer it takes.
ecuador
Posts: 98
Joined: Sat Jan 24, 2015 10:23 pm

Re: hangs on wipe

Post by ecuador »

ajh499 wrote:I'm having the same problem with wipe when a mask is present, it's unbelievably slow, to the point that it's useless with a mask.

I had originally thought that StarTools was just crashing, but it's not, it just takes a really long time.

I tested it today with a 5500 x 3600 TIFF file. Without a mask it takes about 10 seconds to do a wipe, with a mask it took over 20 minutes! That surely cannot be right.
It's definitely due to the size of the mask, the fewer pixels included, the longer it takes.

Huh, you appear to be correct. I tried masking out most of the image and the processing time skyrockets. I am sure that the reason this has gone un-noticed though is that this is not a common way to use the wipe module - you'd normally want to mask out some sort of dark artefacts like dust bunnies etc so that they don't mess with the gradient model. Why would you want to only include a small part of the image, are you sure Wipe is the correct module for what you are trying to do?
Also, processing at 5500x3600 should only be done if you are sure you don't have oversampling (which is quite rare for a 20MP image - or impossible if we are talking about a color bayer matrix). If you bin, all your processing will be faster (even using unusual masks for wipe) and with less noise.
ajh499
Posts: 13
Joined: Sun Dec 09, 2018 9:23 am

Re: hangs on wipe

Post by ajh499 »

Huh, you appear to be correct. I tried masking out most of the image and the processing time skyrockets. I am sure that the reason this has gone un-noticed though is that this is not a common way to use the wipe module - you'd normally want to mask out some sort of dark artefacts like dust bunnies etc so that they don't mess with the gradient model. Why would you want to only include a small part of the image, are you sure Wipe is the correct module for what you are trying to do?
Also, processing at 5500x3600 should only be done if you are sure you don't have oversampling (which is quite rare for a 20MP image - or impossible if we are talking about a color bayer matrix). If you bin, all your processing will be faster (even using unusual masks for wipe) and with less noise.
Yeah, I wouldn't actually use that image without binning it normally. This was a worst case example to demonstrate the enormous increase in time, with and without a mask.

Maybe there should be a warning that processing will take a long time if a large area has been masked out.

And, of course, a progress bar that actually does something rather than sitting there for 20 minutes blank would be helpful too :-)
Post Reply