Specific noise from DSLR

[UCzerw]

Hi! 

I am starting an adventure with a Canon 600D with an astromod. And I noticed a terrible noise. Below a stack 30 frames (from Siril, zoom and auto stretch) of 180 seconds, iso 800 with calibration frames.

I checked earlier photos taken with Nikon without astromod. There has never been anything like this.
Where does it come from? How to fight it?

[AstroNebulee]

Which mount are you using and scope? 

[david_taurus83]

What temperature was it when you took this? Did you flip out the screen?

[alacant]

46 minutes ago, UCzerw said:

with calibration frames.

Hi

Not really enough information, so some general stuff...

Best without dark frames on a 600d.

Dither between each frame.

Don't forget to use bias on both light and flat frames.

Stack with a clipping algorithm.

Take more frames.

HTH.

[Seelive]

As mentioned above, the sensor temperature can make a big difference to the noise and keeping liveview on will increase the internal camera temperature. You can download EXIFLOG from http://www.covingtoninnovations.com/dslr/exiflog.html that will show you the temperature of the DIGIC chip when the image was taken (not the actual sensor temperature, but probably won't be far off).  Ideally your darks and lights should show the same temperature. If their is a large difference (>5°C?), try stacking without the darks.

With my 200D I take at least 80 x 180 sec images to get the noise within (my) acceptable limits but that may be due to my Bortle 5-6 sky.

[UCzerw]

1 hour ago, AstroNebulee said:

Which mount are you using and scope? 

HEQ5, SW Evostar 72 with reducer/flattener 0,85

1 hour ago, david_taurus83 said:

What temperature was it when you took this? Did you flip out the screen?

I opened the screen while taking phots. APT shows lights at 23-24 degrees Celsius.

 

1 hour ago, alacant said:

Hi

Not really enough information, so some general stuff...

Best without dark frames on a 600d.

Dither between each frame.

Don't forget to use bias on both light and flat frames.

Stack with a clipping algorithm.

Take more frames.

HTH.

I use exactly this algorithm. With bias, darks, and flats. Today I want to try dithering. I may succeed.

1 hour ago, Seelive said:

As mentioned above, the sensor temperature can make a big difference to the noise and keeping liveview on will increase the internal camera temperature. You can download EXIFLOG from http://www.covingtoninnovations.com/dslr/exiflog.html that will show you the temperature of the DIGIC chip when the image was taken (not the actual sensor temperature, but probably won't be far off).  Ideally your darks and lights should show the same temperature. If their is a large difference (>5°C?), try stacking without the darks.

With my 200D I take at least 80 x 180 sec images to get the noise within (my) acceptable limits but that may be due to my Bortle 5-6 sky.

Darks EXIF

Light EXIF:

Biases EXIF:

 

[barbulo]

1 hour ago, alacant said:

Best without dark frames on a 600d.

I've read that a few times, but could you explain why dark-flats instead of darks in a 600d?

Thanks in advance.

Edited June 5, 2021 by barbulo

[alacant]

9 minutes ago, UCzerw said:

I use exactly this algorithm

Mmm. That's the procedure for the whole calibration and stacking process.

Anyway, when you have dithered, lose the dark frames and try sigma clip: low 5, high 2. Try with and without normalisation.

Cheers.

[alacant]

1 minute ago, barbulo said:

dark-flats instead of darks

Pragmatic approach with modern DSLRs...

Always use a bias as a bias and a flat as a flat. No darks of any type, but don't forget the dither. 

We found dark flat frames made no difference to the stack. If anything, they created banding. Of course, YMMV so by all means try.

Cheers.

 

[AstroNebulee]

Like alacant said don't bother with darks, I have a Canon 600D and 72ed ds pro and don't bother with them, I just take my light frames as usual then 20 bias and flats, jobs a goodun, no dithering for me 

[Seelive]

17 minutes ago, alacant said:

... try sigma clip: low 5, high 2...

Cheers.

Low, high? I've always thought that the first number refers to the desired standard deviation and the the second number to the number of iterations to try to achieve it (I'd guess that 1 is normally sufficient in most cases) but I suppose the definition could vary between software.

[UCzerw]

2 hours ago, AstroNebulee said:

Like alacant said don't bother with darks, I have a Canon 600D and 72ed ds pro and don't bother with them, I just take my light frames as usual then 20 bias and flats, jobs a goodun, no dithering for me 

I made a stack with no darks (and no dithering) below. The effect is still weak. But I found the histogram strange/ragged after auto stretch.

I checked on older material from my Nikon, without mod. The histogram was fine. And noise. 

And my earlier attempt with this Canon. Same stacking methods as M101.

Edited June 5, 2021 by UCzerw

[edarter]

You could also try different sta king software, I had awful trouble with banding from my 600d (though horizontal, not vertical like yours) when stacking with DSS. At about the same time I made the jump to APP and the banding immediately disappeared.

I know you aren't using DSS but the stacking software is something else to consider as a cause. 

[UCzerw]

2 hours ago, edarter said:

You could also try different sta king software, I had awful trouble with banding from my 600d (though horizontal, not vertical like yours) when stacking with DSS. At about the same time I made the jump to APP and the banding immediately disappeared.

I know you aren't using DSS but the stacking software is something else to consider as a cause. 

I tried this material in DSS as well. Maybe it was slightly better.

Does dithering require any special stacking options?

[alacant]

30 minutes ago, UCzerw said:

Does dithering require

any special stacking options

Hi

Yes. Please see here. 

Having tried DSS and Siril for stacking, IMO Siril has the best stacking. There's a nice comparison between apps here.

Cheers

 

Edited June 6, 2021 by alacant

[UCzerw]

23 minutes ago, alacant said:

Hi

Yes. Please see here. 

Having tried DSS and Siril for stacking, IMO Siril has the best stacking. There's a nice comparison between apps here.

Cheers

 

Thank you so much. 

But to be sure, I will ask... Should I use bias frames for lights before stacking in pre-processing? At Siril, I only use biases frames for flat frames in the pre-processing. As they show in tutorial. 

Will I find a better tutorial somewhere? Because from what I can see they are using Winsorized Sigma C. I don't know how much different that is.

[alacant]

You're not going to be using dark frames (which have the bias included) so you should subtract the bias from the light frames too before stacking.

In Siril, this is easy. Choose your light sequence and pre process with both the master bias and master flat selected. Register, throw out the bad frames (which may be introducing noise) then stack.

1 hour ago, UCzerw said:

Winsorized Sigma

Yes. That's fine. Especially if you've dithered.

 

[UCzerw]

OK thanks. Previously, I only did preprocessing with biases for flat frames and pre processing for lights with stacked darks and stacked pre-pocessed flats.

I will try as you say in Siril. Yesterday's attempt to dither every 3 frames below. From DSS (no dark frames, only biases and flats).

The background looks good I think. I have to extend the time between dither and shot, because the stars after dithering were not perfect. And I'm still struggling with mount and backfocus. 

[UCzerw]

Still the same problem, even with dithering.... Some "smudges" of noise.

[bottletopburly]

1 hour ago, UCzerw said:

Still the same problem, even with dithering.... Some "smudges" of noise.

How many pixels are you dithering , try a dither of at least 15 pixels you have walking noise so possibly not dithering enough .

Edited June 11, 2021 by bottletopburly

[UCzerw]

1 hour ago, bottletopburly said:

How many pixels are you dithering , try a dither of at least 15 pixels you have walking noise so possibly not dithering enough .

I started with 10 pixels. Maybe more frames with shorter time will be a better option?

[bottletopburly]

What programs are you using to dither out of interest 

[UCzerw]

31 minutes ago, bottletopburly said:

What programs are you using to dither out of interest 

PHD2 and APTool (older version)

[bottletopburly]

What’s your guide scope Fl , imaging scope fl, imaging camera pixel size and guide scope pixel size ?. 

[UCzerw]

23 minutes ago, bottletopburly said:

What’s your guide scope Fl , imaging scope fl, imaging camera pixel size and guide scope pixel size ?. 

Gude scope 60/240mm, guide cam 3,75 upix

Imaging scope 72/420 + 0,85 reducer;  camera 4,3 upix

Scale 1,3x probably