Dark frame Problem

[julian489289]

I have a Canon 650D DSLR camera which I use for AP.

When I take RAW dark frames I get a white area to the right hand side of the frame.

I always cover the view finder and the front of the scope and take darks directly after the imaging session whilst it is totally dark.  The LCD screen is closed and off and I power the camera through a dummy battery with an external powerpack bought from Simon FLETCHER.

I always image at ISO 800 and the white area starts to become visible at around 90 secs exposure and gets gradually more the longer I expose.

I enclose a still of  dark frame at ISO 800, 120 secs

Has anyone got any ideas as to what this may be please and maybe how i can resolve this?? 

 

[daz]

Probably amp glow....

I'm not a DSLR imager, but would think flats would solve it...

[Ouroboros]

I wonder if it's amp glow caused by the amplifier chip which is attached to the side of the sensor?  One for more experienced imagers. 

I used to see something like this on images from my canon 450D. It was in the light files though. Funny thing is, I don't see it anymore. Might be because I tend now to shoot at a lower gain of ISO400. 

I seem to recall various suggestions to help this. One was to allow a short period of cool down between images. The other was not to hold the shutter open (live view) whilst gathering images as this acts as a source of thermal energy. 

Butvithers more experienced than myself will know more. 

 

[julian489289]

Thanks guys

I normally allow 60 sec between frames for cool down.  Would this be enough?? 

I always ensure the live view is off which is fairly easy with the 650D and articulated screen.

[ollypenrice]

You're not alone. I think you'll find this thread instructive. 

If it's there in your darks is it not there in your lights? In which case the darks should do their stuff and remove it. Have you actually tried applying the darks with the glow?

Out of interest, in the linear, unstretched, dark what is the typical ADU value of the brightest bit of the glow compared with the background average?

Olly

[Ouroboros]

1 hour ago, julian489289 said:

Thanks guys

I normally allow 60 sec between frames for cool down.  Would this be enough?? 

I always ensure the live view is off which is fairly easy with the 650D and articulated screen.

I didn't find it made much difference. Or I wasn't willing to wait long enough between shots.   As I say, it was an intermittent problem even within an imaging session. Some lights and darks would contain this feature. Others wouldn't. No rhyme or reason to it as far as I could tell. It can be processed out of the final stacked image or just cropped off anyway.  I used to frame the images to keep away from the problem edge. 

Incidentally, my amp glow was on the bottom long edge of the image, not the short side like yours. Are the amplifiers on different sides of the sensor on the 450D and 650D?

[ollypenrice]

4 hours ago, Ouroboros said:

 

Incidentally, my amp glow was on the bottom long edge of the image, not the short side like yours. Are the amplifiers on different sides of the sensor on the 450D and 650D?

In the thread to which I linked it does look as if the 700D has the amp glow in the same position as the OP's.

Olly

[julian489289]

Olly, I did trial pixinsight and when I did the global stretch I could see the same glow on the light frames.  It was also there in the final stacked image.  I have just been cropping it out but that loses me 1/3 of the frame

I do not have any software that will give me the ADU values.

[Davey-T]

17 minutes ago, julian489289 said:

Olly, I did trial pixinsight and when I did the global stretch I could see the same glow on the light frames.  It was also there in the final stacked image.  I have just been cropping it out but that loses me 1/3 of the frame

I do not have any software that will give me the ADU values.

I would expect PI to have some way to measure ADU values ?

Dave

[julian489289]

Unfortunately the trial ended and I could not warrant the purchase of the full license

[Davey-T]

22 minutes ago, julian489289 said:

Unfortunately the trial ended and I could not warrant the purchase of the full license

Me too, but may still end up buying it 

The camera temperature is stored in the image file but you may need some software capable of accessing it.

Dave

[Oddsocks]

30 minutes ago, julian489289 said:

Unfortunately the trial ended and I could not warrant the purchase of the full license

What software are you currently using for post processing?

ADU or Pixel values are displayed in most image processing software packages.

For purely reading EXIF data or basic image processing there are several free programs available, here is link to a revue page:

http://carlcheo.com/best-exif-viewers

FastStone and ExifPro provide basic EXIF data readout:

http://www.faststone.org/FSViewerDetail.htm

http://www.exifpro.com

Photoshop, even in the "lite" home, versions reads and displays ADU as pixel values in the tool bar or in the information window.

[julian489289]

Thanks but I thought ADUs were the average pixel values in the image not temperatures??

I use PS CS6 but cannot find anywhere that if I hover over the image I can get a reading of the ADU like in Maxim DL??  (Don't have that either!)

[Davey-T]

54 minutes ago, julian489289 said:

Thanks but I thought ADUs were the average pixel values in the image not temperatures??

I use PS CS6 but cannot find anywhere that if I hover over the image I can get a reading of the ADU like in Maxim DL??  (Don't have that either!)

That was in answer to your previous post about knowing how long to leave between subs.

You can use the info window in P'Shop but I think it may only be 8 bit so 0 to 255 so somewhere around 170.

Dave

[Oddsocks]

On 10/05/2017 at 14:14, julian489289 said:

I use PS CS6 but cannot find anywhere that if I hover over the image I can get a reading of the ADU like in Maxim DL??

This should explain the method to read ADU with Photoshop:

 

 

[julian489289]

Sorry mate, yes it was forgive me I am a little slow in all of this voodoo dark art

Thanks for that explanation it's appreciated.

 

[julian489289]

On 09/05/2017 at 14:18, ollypenrice said:

Out of interest, in the linear, unstretched, dark what is the typical ADU value of the brightest bit of the glow compared with the background average?

Olly

I can now try to reply to Olly's query later

[Oddsocks]

On 10/05/2017 at 15:15, julian489289 said:

I can now try to reply to Olly's query later

When opening your RAW images in PS make sure to always select 16 bit working space in the RAW import screen that pops up when you first open the file, this will give you much greater control over the post processing tools and work steps.

You still need to manually check and set the info screen to 16 bit too.

Image below explains.

 

 

[julian489289]

Thank you that is of great help

[julian489289]

Hi Olly

Thanks to Oddsocks i can respond as follows;

Linear unstretched dark frame typical  background average ADU 240-250

Brightest bit of the amp glow is ADU 2292 

[Oddsocks]

On 09/05/2017 at 14:18, ollypenrice said:

Out of interest, in the linear, unstretched, dark what is the typical ADU value of the brightest bit of the glow compared with the background average?

Just read your last post Julian. 

If you want to alert the writer of a previous post to your reply then highlight a piece of relevant text in that persons post, like Ollies above, release the mouse key and a tiny little pop-up will appear saying "Quote this". click on the pop-up and that quote will appear in your reply box, then when you complete your new post that person will see a flag at the top of the forum home screen and will know you have quoted them. Otherwise your replies may get missed by the person you were trying to reach, doesn't always work of course but better than nothing...

Edit...I just noticed you used the quote function in one of the earlier posts so I guess you just missed it this time.....

1 hour ago, julian489289 said:

Linear unstretched dark frame typical  background average ADU 240-250

Brightest bit of the amp glow is ADU 2292

 Ollie has more experience with imaging than me and gets to see more DSLR images too, but given the similar sorts of dark frames produced by the D700 I would say those ADU values were not unusual for the D650 camera.

There is nothing you can do to the hardware to resolve amp glow issues but as Ollie mentioned earlier dark frames of the same exposure time should calibrate out and remove most of the brightening in the lights saving you from cropping excessively.

As you are no longer using the PixInsight trial then you can calibrate for free in DSS, it will take Canon raw frames, lights, darks, flats and bias, direct from the camera, and output in 16bit TIF format that can go directly into P.S.

This is the link to DSS if you have not tried it before:

http://deepskystacker.free.fr/english/index.html

If you are trying to keep things simple at the moment, say using the camera for milky way-landscape shots etc and don't want to bother stacking and calibrating then check in the Canon setup menu for long exposure compensation settings, I am a Nikon user so not totally familiar with Canons but I seem to remember the canons come equipped with an internal automatic dark frame function, you set an exposure of say three minutes with long exposure compensation set to on, the shutter opens for three minutes to take the shot, then immediately afterwards the camera takes another three minute shot automatically but with the shutter closed, during this time the camera seems to be doing nothing but after the three minutes is up the camera software will automatically subtract the dark from the light and output a dark calibrated image that should have no amp glow.

I can't be sure of the exact name for this function so you would need to read through the Canon manual to find it (if it is there).

best of luck...

[Oddsocks]

9 hours ago, julian489289 said:

Linear unstretched dark frame typical  background average ADU 240-250

Brightest bit of the amp glow is ADU 2292 

While bored watching the observatory do its stuff during the night I looked at the latest Canon EOS D650 manual on-line. The name for the automatic dark frame subtraction is called "Long Exposure Noise Reduction", details are on page 125 of the manual:

http://gdlp01.c-wss.com/gds/3/0300008673/02/EOS_650D_Instruction_Manual_EN.pdf

This works well for ISO speeds up to 1600, above 1600 the resulting image has more noise than usual but for astro imaging stacking will reduce this.

The advantage of using this is that the dark frame is taken automatically after each light frame so the temperature of the sensor for both images will be a better match. As the imaging session continues and the camera temperature changes then each new image will always have a matching dark frame automatically applied so it should produce a better calibrated image than taking a series of lights first followed by separate darks later into the session or at some other time. In post processing, using DSS for example, these automatically dark calibrated lights can still be calibrated with flats and stacked, you just don't bother loading separate dark frames for the lights but still load flats and a bias master renamed "master dark for flat" which is loaded into DSS as if it were a dark frame. When post processing you have less data to deal with so PC hardware is not so heavily taxed when handling these large raw DSLR files.

The disadvantage of the Long Exposure Noise Reduction is that each exposure takes just over double the exposure time set so you end up with fewer images taken during any observing session rather than just taking as many lights as possible and taking darks at some other time.

Might be worth looking at trying the Long Exposure Noise Reduction function and see if this suits your post processing workflow better.

HTH.

[ollypenrice]

15 minutes ago, Oddsocks said:

While bored watching the observatory do its stuff during the night I looked at the latest Canon EOS D650 manual on-line. The name for the automatic dark frame subtraction is called "Long Exposure Noise Reduction", details are on page 125 of the manual:

http://gdlp01.c-wss.com/gds/3/0300008673/02/EOS_650D_Instruction_Manual_EN.pdf

This works well for ISO speeds up to 1600, above 1600 the resulting image has more noise than usual but for astro imaging stacking will reduce this.

The advantage of using this is that the dark frame is taken automatically after each light frame so the temperature of the sensor for both images will be a better match. As the imaging session continues and the camera temperature changes then each new image will always have a matching dark frame automatically applied so it should produce a better calibrated image than taking a series of lights first followed by separate darks later into the session or at some other time. In post processing, using DSS for example, these automatically dark calibrated lights can still be calibrated with flats and stacked, you just don't bother loading separate dark frames for the lights but still load flats and a bias master renamed "master dark for flat" which is loaded into DSS as if it were a dark frame. When post processing you have less data to deal with so PC hardware is not so heavily taxed when handling these large raw DSLR files.

The disadvantage of the Long Exposure Noise Reduction is that each exposure takes just over double the exposure time set so you end up with fewer images taken during any observing session rather than just taking as many lights as possible and taking darks at some other time.

Might be worth looking at trying the Long Exposure Noise Reduction function and see if this suits your post processing workflow better.

HTH.

It would be interesting to see whether twice as much data without the auto NR was better or worse than half as much with it. I don't know the answer, but if we were nattering about this in the pub after a couple of pints I'd probably put a tenner on twice as much data!

Olly

[StuartJPP]

On 11/05/2017 at 08:01, ollypenrice said:

It would be interesting to see whether twice as much data without the auto NR was better or worse than half as much with it. I don't know the answer, but if we were nattering about this in the pub after a couple of pints I'd probably put a tenner on twice as much data!

Olly

Gathering twice as much data would be my guess too. I haven't used or worried about darks since I started to dither between subs.

What you may find is that for a small number of subs the image may appear cleaner with ICNR but I bet all those lovely bits of faint fluffy data is being thrown away. Though I can't say for sure as I am no expert, I just do what I believe is right for me.