Is this a problem with my camera or DSS?

[Steve143]

Hi,

Can anyone help with the images below? They have been created using 120x 60 sec exposures and stacked using flats and biases together (see 1), biases only (2) and no calibrations frames at all (3). They were all taken with a Canon EOS 70D at ISO800 using a 80ED-R fitted with an Optolong-Pro filter (unguided and no dithering). The stacked images were then processed using the same simple workflow in Siril. 

Each image has a cross-hatch pattern, especially in red and green, which is less noticeable in the stack with both flats and biases but this stack then has horizontal bands of light and dark - how can I get rid of this? 

In DSS I used the Intersection mode to stack and selected the detect and clean both remaining hot and cold pixels under the post calibration cosmetic settings. 

Is this caused by my camera or is it a stacking issue - or do I just need more integration time to reduce the noise? This is 2 hours of integration time.

Thanks for any suggestions or advice. 

 

1. Flats and bias used.

2. Bias only.

3. No calibration files used. 

[wimvb]

My guess is that your subs are underexposed. What you see here is mainly the read pattern of the camera. Your scope is probably f/6 - f/7.5, depending on using a reducer or not. The filter reduces the light that enters the camer even more and the result is underexposure. You need an exposure time of a few minutes at ISO 800.

Try to collect more data (2 hrs isn't much on this galaxy). The more light pollution you have, the more data you need to reduce the LP noise.

[Steve143]

15 hours ago, wimvb said:

My guess is that your subs are underexposed. What you see here is mainly the read pattern of the camera. Your scope is probably f/6 - f/7.5, depending on using a reducer or not. The filter reduces the light that enters the camer even more and the result is underexposure. You need an exposure time of a few minutes at ISO 800.

Try to collect more data (2 hrs isn't much on this galaxy). The more light pollution you have, the more data you need to reduce the LP noise.

Yes, I have a reducer/flattener which makes my scope f/5.6. I will collect more data but think I'm going to have to bite the bullet and start guiding to get anything more than 1-2 minute exposures with my setup. 

Thanks.  

[mcrowle]

Can I suggest you try a higher ISO setting?  I used to go no higher than ISO 1600 with my EOS 600D, but during an astrophotography workshop it was suggested I try 6400 - the highest standard setting the camera has. I was shocked to be honest, but tried it anyway - and the results were great, as long as I had a good number of frames to stack of course.

Since then, I reserve ISO 1600 for when I'm piggybacking the camera on the scope or I'm guiding (which I do manually). If I'm not guiding, for whatever reason, and want to keep the exposures shorter, I no longer have qualms about setting ISO 3200 or 6400.

Just my thoughts and experience though - it may not work for everyone.

Regards, Mike.

Edited May 2, 2022 by mcrowle

[Ali Alhawas]

On 29/04/2022 at 19:11, Steve143 said:

Hi,

Can anyone help with the images below? They have been created using 120x 60 sec exposures and stacked using flats and biases together (see 1), biases only (2) and no calibrations frames at all (3). They were all taken with a Canon EOS 70D at ISO800 using a 80ED-R fitted with an Optolong-Pro filter (unguided and no dithering). The stacked images were then processed using the same simple workflow in Siril. 

Each image has a cross-hatch pattern, especially in red and green, which is less noticeable in the stack with both flats and biases but this stack then has horizontal bands of light and dark - how can I get rid of this? 

In DSS I used the Intersection mode to stack and selected the detect and clean both remaining hot and cold pixels under the post calibration cosmetic settings. 

Is this caused by my camera or is it a stacking issue - or do I just need more integration time to reduce the noise? This is 2 hours of integration time.

Thanks for any suggestions or advice. 

 

1. Flats and bias used.

2. Bias only.

3. No calibration files used. 

Hello ,

Is it a cooled camera ?

What was the ambient temperature ?

[Steve143]

16 hours ago, Ali Alhawas said:

Hello ,

Is it a cooled camera ?

What was the ambient temperature ?

No, it's a DSLR. The sensor temperature went from about 15 degrees to 20 during the session. I don't have any way to cool this camera - a cooled camera is on my wish list!! 

Edited May 7, 2022 by Steve143

[Ali Alhawas]

17 hours ago, Steve143 said:

No, it's a DSLR. The sensor temperature went from about 15 degrees to 20 during the session. I don't have any way to cool this camera - a cooled camera is on my wish list!! 

How did you decide the sensor temperature? It is not the same as the ambient temperature ! Its getting hotter ..

If the sensor is not cooled you will see a totally bad images.

Did you dither ? it can help a little but not as affected as cooling.

Post one sub. 60sec here to see what a single image looks like..

What happen is, in the beginning of the session the images are acceptable but sensor is getting hotter after some long exposures and the images getting more heat noise ..

cooling must have unless the ambient  temperature is around Zero or so.

I hope this helps.

 

Geoduck !  

[AstroNebulee]

The horizontal lines would be Canon banding. My 600d suffered with this, there are ways of processing it out I know astronomy tools for photoshop can though depending how bad it is. I'm sure pix insight has a Canon banding module but not sure. 

In the astronomy tools for photoshop there a vertical banding action too though never used that. 

Cheers 

Lee 

[Ouroboros]

Indeed, they look exactly like canon banding to me.  And as you say, AstroNebulee, there’s a canon banding removal process in Pixinsight. Without some kind of banding removal tool I might try to suppress banding with some kind of background subtraction tool or by darkening the background using curves in a photo-processing application. 

I think it significant that using flats made the banding look worse. I think I’d focus some attention on flats. How did you take your flats? Did you set the camera to  AV mode and stay at the same ISO as for the lights? That’s how I do it with my Canon 450D.  Banding can come and go actually. I’m just wondering whether conditions were very different when you obtained the flats. Was the temperature of the camera sensor higher after obtaining the lights? Had you possibly introduced some source of electrical noise when taking flats? A light source perhaps? Was the camera battery running hot by then, or going flat. 

I might be tempted to take some fresh flats and see what happens. OK I know flats are supposed to be taken with the rig set up for lights, but this is just experimental so it doesn’t matter right? 

[Steve143]

On 12/05/2022 at 08:46, Ouroboros said:

there’s a canon banding removal process in Pixinsight.

do you know if there is one in Affinity Photo?

On 12/05/2022 at 08:46, Ouroboros said:

I might be tempted to take some fresh flats and see what happens.

I did just that and the new flats look a lot better. The histogram was just where it should be. I plan to use these new flats to reprocess my lights and hopefully it will improve the end result. I also managed to capture some additional exposure time on this target too which I hope helps with the SNR. So far I've got over 4 hours. 

Thanks again everyone for your comments and suggestions. I really appreciate everyone taking time to respond. 

Steve 

[Ouroboros]

@Steve143  I don’t know about Affinity Photo. I just Googled and didn’t find anything.  My guess is that Canon banding isn’t a big issue in conventional photography, unlike Astro photography in which images are stretched. You can of course use curves in Affinity to darken the background to help hide the banding.

I should say that Canon banding removal in Pixinsight is not necessarily 100% successful, but it helps considerably.  It can work reasonably well when applied to the stacked image, but it is best applied to each calibrated frame prior to stacking. That means setting up a batch process which is not entirely trivial for the computationally challenged like myself.

I’ll be interested to see how your re-processed image looks. Good luck.