Blotchy background on my 130 PDS

[Bibabutzemann]

On 25/04/2022 at 19:21, alacant said:

Hi

Ask on the 130 thread?

We keep only stacked flat frames taken with modified pds'.

 

 

I would be also happy to see your stacked flat frame, to see what i can expect from a well tuned 130PDS 😃

[Stefan73]

On 25/04/2022 at 18:11, Bibabutzemann said:

maybe someone could provide me with 20 flat frames they have taken with the 130 pds and a similar size sensor. Then i could rule out if this gradient is only specific to my gear.

Hi,

If it's still useful I'm happy to provide flats taken with a 550D and a completely unmodified 130 PDS.  I think the sensor sizes are the same.  No filters involved or modifications to the DSLR.

[Bibabutzemann]

1 minute ago, Stefan73 said:

Hi,

If it's still useful I'm happy to provide flats taken with a 550D and a completely unmodified 130 PDS.  I think the sensor sizes are the same.  No filters involved or modifications to the DSLR.

yes of course! Would be very thankful!

[alacant]

44 minutes ago, Bibabutzemann said:

stacked flat frame

https://drive.google.com/file/d/17p-lULiGN4fQ_G0mU4rbs9SNs2Dt0P6q/view?usp=sharing

Well tuned? I wish!

 

[Stefan73]

10 hours ago, Bibabutzemann said:

yes of course! Would be very thankful!

There you go:

https://drive.google.com/file/d/1oEGCroWx943lPP1IE0IwpKr1Cgzea4MO/view?usp=sharing

All I did is put a muslin cloth left from when the children were babies over the telescope and a tablet with a light app on that.

[Bibabutzemann]

Thank you @Stefan73 and @alacant !

 i compared all our flats. I did a background extraction of 2nd order in Siril, to see if any complex gradients are left. 

First is Stefans, 2nd is mine and last is alacants.

On all three i could make out that "hook" like structure in the middle of the flat, but it was less pronounced on alacants. 

I tried with flocking and and a baffle ring, but still didnt get rid of those gradients. Maybe im overthinking and i should just focus on doing better flat frames.

 

[alacant]

1 hour ago, Bibabutzemann said:

I did a background extraction of 2nd order

If you apply the master flat to each of the light frames, then register and stack that, then it may make sense to apply background extraction (BE), but only AFTER you have applied de-banding.

JTOL: I don't think applying BE to a flat frame helps in analysing what you think is wrong. 

Anyway, fwiw, using your method after banding reduction, this is what I get. But we're never going to develop/stretch that far anyway. I'll include a normally starched version as well as the histogram max of the original too.

Cheers

Edited April 28, 2022 by alacant

[Bibabutzemann]

17 minutes ago, alacant said:

If you apply the master flat to each of the light frames, then register and stack that, then it may make sense to apply background extraction (BE), but only AFTER you have applied de-banding.

JTOL: I don't think applying BE to a flat frame helps in analysing what you think is wrong. 

Cheers

Debanding is not a tool i use anymore since i dither. 

It only worked on images with much background but did a bad job unbanding nebula photos in my experience.

"I don't think applying BE to a flat frame helps in analysing what you think is wrong. "

Yes, seems like! 

[alacant]

8 minutes ago, Bibabutzemann said:

bad job unbanding nebula photos

Are you sure you applied it to the linear images before registration?

Not sure there have been any changes to the algorithm, but it may be an idea to build the latest version.

Cheers

Edited April 28, 2022 by alacant

[Bibabutzemann]

7 minutes ago, alacant said:

Are you sure you applied it to the linear images before registration?

Not sure there have been any changes to the algorithm, but it may be an idea to build the latest version.

Cheers

Ok maybe i didnt understand correctly. Do you apply unband to every single frame before registration and stacking?

[alacant]

38 minutes ago, Bibabutzemann said:

unband to every single frame

Yes. Just check 'apply to sequence'. 

 

Edited April 28, 2022 by alacant

[ONIKKINEN]

3 hours ago, Bibabutzemann said:

Thank you @Stefan73 and @alacant !

 i compared all our flats. I did a background extraction of 2nd order in Siril, to see if any complex gradients are left. 

First is Stefans, 2nd is mine and last is alacants.

On all three i could make out that "hook" like structure in the middle of the flat, but it was less pronounced on alacants. 

I tried with flocking and and a baffle ring, but still didnt get rid of those gradients. Maybe im overthinking and i should just focus on doing better flat frames.

 

This blotchy look is because of running the tool, not somethig that exists in the flats themselves.

Your flats are fine, they just dont match your lights because of mechanical problems, and using someome elses flats is guaranteed to be worse. 

[Bibabutzemann]

6 minutes ago, ONIKKINEN said:

This blotchy look is because of running the tool, not somethig that exists in the flats themselves.

Your flats are fine, they just dont match your lights because of mechanical problems, and using someome elses flats is guaranteed to be worse. 

Im not sure about that. When i do a 2nd degree background extraction, the software cant create such complex patterns. Just like a 2nd degree function can only make one curve.

Also these patterns are dependant on orientation and focus postition.  So i would say they are correlated to the light train.

I was just unsure if this is specific to my setup or something normal that i shouldnt think too much about.

For now i will focus on doing better flatframes.

 

[pipnina]

I think I notice a similar colour pattern in various scopes. To varying degrees.

I use a mono camera (IMX571) with a TS-PHOTON 8" f4. So a very different setup to yours. However I also get some blotching issues and struggle to get working flat frames.

I think for us cheap newtonian users, if modifications to the scope can't solve our flats, we need to adjust HOW we take our flats. I recall an anecdote from another frustrated newt user, although I'm sadly unable to find the post right now. They used an illuminated panel and sky flats like I have. But as has been said here before, the panel being right up close to the telescope aperture can cause light to bounce around in the tube and reach the sensor. They only solved their issues by buying a large projector screen and illuminating it with... I want to say a projector, from memory? This was effective because the screen was considerably larger than the telescope aperture (so easy to point at without the image containing another object) while only creating light at angles similar to the FOV of the telescope, as the screen was placed across the garden, so a fair distance away.

I am looking at something possibly cheaper but hopefully just as effective. Using a little photographic upper-body backdrop, and my current flat panel (an A3 artist drawing pad) to illuminate it evenly.

I bought the black velour material from FLO, and put a big patch opposite the focuser behind the secondary, and covered the un-blackened edge of the secondary with it as well. Weirdly, this didn't yield a difference in my flats that I could tell, but it can't have hurt either!

This is a recent image of mine where I tried to image LBN 599, a very faint nebula. It exists in the image as some dusty clouds, but (especially in the saturation-boosted version) you can tell that my background has a lot of non-nebula blotching!

 

So here is what the flat-calibrated stack looks like without any Pixinsight DBE

Over-correction, and I think in the top left a major over-correction where the filter gasket for one of my filters goes in front of the sensor slightly, more visible in the processed shots above. You can see how the colour changes through the image from green in the middle to pinky on the left side and yellowish on the right.

Here is the stack without calibration:

I've obviously increased the saturation a lot. But the background blotching exists in calibrated frames too!

I think it's possible that in my case the coma corrector catches light from the sky / flat panel and the light bounces around in there and off of the lens elements, eventually winding up in my sensor. Narrowband images are less prone to this but evidence for my theory is still there however...

Flat calibrated autostretch of a H-alpha dataset. Notice the gradient runs to the top right corner.

Flat calibrated autostretch of an OIII dataset... I'm sensing a pattern here! If the coma corrector was reflecting light into the sensor, it would make sense that it would hit one side of the corrector wall, then bounce to the opposite side of the sensor more than the close side.

But if we stretch the OIII data when it ISN'T calibrated, we see no background blotching from the broadband RGB images, and we don't see the strong gradient to the top left. Suggesting that the flats are bad and the light from outside the frame is interfering with the images as well. With narrowband being a sticking plaster over the issue as it rejects so much non-target light.

I think the only solution is, as people above have suggested, very good baffling of the scope. I think a rigid baffled dew guard, larger blackened edge secondary and short coma corrector might be an improvement here (the TS-GPU sticks out into the tube, no protection hiding inside the focuser barrel!).

I've considered spending lots of money on a high end reflector like the ONTC carbon fiber newts by TS, or an Orion AG. But I think newtonians just suffer from their right angled focuser, it makes perfect baffling near impossible. Maybe a slight change, where the focuser is at more than 90 degree to the primary, shielding the coma corrector from light entering at the aperture but requiring a more eccentricly shaped secondary.

 

Bit of a ramble, but for sure I share your pain with these issues with newts, and despite having 6+ grand worth of equipment, I still get shoddy results due to my telescope choice!

Hopefully we both find the solution to our scope's dodginess. Be it baffles, distance flat correction, etc.