Soft images over time just atmospheric issues?

[sploo]

I've noticed a fairly consistent pattern on the (admittedly relatively few) attempts I've made to capture a number of light frames of a DSO; namely that the first few shots are usually pretty sharp, then over time it looks as though the image has become defocused.

Running the light frames through DeepSkyStacker and looking at the quality scores, I usually get a pattern similar to that shown below. The Y axis is the frame score, and the X axis is the capture number:

Broadly, the earlier captures are better, and later are worse - though that's not 100% true.

Checking the camera lens I see no fogging or condensation on the front element. If I (later in a shooting session) attempt to refocus on my current target, the image is still soft (so I assume the lens losing focus isn't the problem). Occasionally I can swing to another area of the sky and that might be be sharp (or not).

So - my assumption is that it's as simple as clouds/moisture in the atmosphere rolling in; which just happens to coincide some tens of minutes after each of the (few) shooting sessions I've attempted - and therefore there's nothing practical I can do to combat this?

[Adreneline]

Can you tell us a bit more about the equipment you are using, e.g. is the camera a dslr/cmos cooled/ccd cooled? Are you using a telescope or a camera lens.

 

[geoflewis]

Where were you imaging and what duration does the graph cover? If you start imaging close to or after the meridian it could be that the diminishing results are due to the scope steadily pointing lower and lower into the murk as it gets closer to the horizon.

Edited March 31, 2020 by geoflewis

[sploo]

40 minutes ago, Adreneline said:

Can you tell us a bit more about the equipment you are using, e.g. is the camera a dslr/cmos cooled/ccd cooled? Are you using a telescope or a camera lens.

 

Sure. I've seen this pattern with a Canon 7D and a Canon 5D4; both unmodified (though using a CLS filter), and using camera lenses.

37 minutes ago, geoflewis said:

Where were you imaging and what duration does the graph cover? If you start imaging close to or after the meridian it could be that the diminishing results are due to the scope steadily pointing lower and lower into the murk as it gets closer to the horizon.

On the example above, each sub was 30s; so just over an hour. DSS rejected over half the frames; such that only about 58 (~29 minutes) were usable.

This was M42, which means I'm pointing south towards a major city and I do indeed end up moving towards the horizon over a session.

[sploo]

I've just checked the data from another session - this time of M81 and M82; so pointing broadly east, and high in the sky (well away from the horizon).

A similar pattern through the ~350x 30s frames (about 3 hours). Unfortunately DSS rejected all but 64 frames (~32 minutes), so it was a lot of wasted data.

[geoflewis]

5 minutes ago, sploo said:

This was M42, which means I'm pointing south towards a major city and I do indeed end up moving towards the horizon over a session.

M42 is currently transiting around 6pm BST, so assuming that you didn't start imaging until after dark at say 9pm (astro twilight is actually not ending until nearer 9:30pm BST) then it's going to down below 20 deg altitude heading close to 10 deg altitude an hour later, which is going to be a significant contributor to your deteriorating images in that part of the sky. You will also be experiencing atmospheric dispersion with a DSLR (colour) camera at those low altitudes, so there's a lot going against you.

[geoflewis]

4 minutes ago, sploo said:

I've just checked the data from another session - this time of M81 and M82; so pointing broadly east, and high in the sky (well away from the horizon).

Hmm, so that's intriguing, I agree it's not poor seeing at the horizon for that target. Do you acclimatise your camera before imaging or just take it outside and get going? I'm wondering whether you are seeing some focus shift, or if not that, then whether the glass in front of the sensor or the sensor itself is misting over a bit as the evening progresses.

[sploo]

14 minutes ago, geoflewis said:

M42 is currently transiting around 6pm BST, so assuming that you didn't start imaging until after dark at say 9pm (astro twilight is actually not ending until nearer 9:30pm BST) then it's going to down below 20 deg altitude heading close to 10 deg altitude an hour later, which is going to be a significant contributor to your deteriorating images in that part of the sky. You will also be experiencing atmospheric dispersion with a DSLR (colour) camera at those low altitudes, so there's a lot going against you.

That makes, thanks. There's a house "in the way" from my line of sight and it's been interesting noticing how the position of M42 changes over the months. Not something you pay attention to until you actually start "looking".

 

14 minutes ago, geoflewis said:

Hmm, so that's intriguing, I agree it's not poor seeing at the horizon for that target. Do you acclimatise your camera before imaging or just take it outside and get going? I'm wondering whether you are seeing some focus shift, or if not that, then whether the glass in front of the sensor or the sensor itself is misting over a bit as the evening progresses.

Good point. I don't have the camera + lens outside for very long beforehand. I'd always assumed it was misting (so I need to get a dew heater) but every time I've checked the front element of the lens it looks perfectly clear. Perhaps some fogging is happening internally.

[Adreneline]

This may be irrelevant but when I use my 70D - controlled by EoSBackYard - for long exposures I notice that as the imaging session goes on the sensor temperature creeps up - not by very much but nevertheless it gradually increases as the evening goes on. Increasing temperature means increasing noise.

Just a thought.

 

[sploo]

1 hour ago, Adreneline said:

This may be irrelevant but when I use my 70D - controlled by EoSBackYard - for long exposures I notice that as the imaging session goes on the sensor temperature creeps up - not by very much but nevertheless it gradually increases as the evening goes on. Increasing temperature means increasing noise.

Just a thought.

 

I should definitely look into some software for controlling the camera remotely (rather than walking back and forth between the house and garden). Certainly the non-cooled cameras may well suffer, but the issue I'm seeing is almost as if the image is defocussing (stars become larger and dimmer "pools" of light), so I'd assume that isn't a heat/noise issue.

[jimjam11]

Have you looked at some good and bad frames to visually compare? Is the only difference focus? 

[Adreneline]

1 hour ago, sploo said:

the issue I'm seeing is almost as if the image is defocussing

If you use something like BYEoS (or APT) then you can periodically pause the image capture and use the built-in FWHM function to see what is happening to a star in the field of view. It would give you a quantitative value to work with. You can also monitor the sensor temperature as you go along.

Hope you manage to get to the bottom of this - all very frustrating - a word that applies to the whole hobby really!

Good luck!

Adrian

Edited March 31, 2020 by Adreneline
Clarification

[jimjam11]

I think DSS gives the FWHM after it has measured the frames (most likely in pixels)? I had a look at some of my recent frames and FWHM fell during the night which is what I normally see.

This is data from a couple of nights, one with better seeing than the other and towards the end of that set (approx 3am) FWHM were approaching 2"

[sploo]

1 hour ago, jimjam11 said:

Have you looked at some good and bad frames to visually compare? Is the only difference focus? 

It appears to be. I've just picked three frames from an attempt to capture M81 and M82 and cropped a small section from each:

The top image is one of the first captures, the middle is about 47 minutes later, the bottom is 114 minutes into the session.

I've increased the brightness slightly for web viewing. I notice that the bottom image has a much more obvious red hue in the background, and obviously the stars have become blurred "blobs". Maybe it is just atmospheric issues?

[jimjam11]

"Occasionally I can swing to another area of the sky and that might be be sharp (or not)."

Could you have tilt in the imaging train?

Which lens are you using? Some of them can be very difficult to focus accurately, although it is strange you get this pattern every night?

 

 

[sploo]

1 hour ago, jimjam11 said:

Could you have tilt in the imaging train?

Not sure what that is: unless you mean the lens not being parallel to the sensor?

 

1 hour ago, jimjam11 said:

Which lens are you using? Some of them can be very difficult to focus accurately, although it is strange you get this pattern every night?

For these, it's all been a Canon 100-400II, sometimes with a Canon 1.4x or 2x extender.

"Every night" probably covers only a handful of occasions to be fair (I'm a noob), so probably statistically dodgy!

Edited March 31, 2020 by sploo

[Erling G-P]

5 hours ago, jimjam11 said:

I think DSS gives the FWHM after it has measured the frames (most likely in pixels)? I had a look at some of my recent frames and FWHM fell during the night which is what I normally see.

This is data from a couple of nights, one with better seeing than the other and towards the end of that set (approx 3am) FWHM were approaching 2"

How do you get DSS to produce those graphs - or have you copied the FWHM values to another program to make them ?

Very strange problem for the OP.  Having a Nikon 18-200mm zoom lens which won't hold its zoom level if pointed up or down (it 'falls' in or out by itself), my initial tought was to suggest some sort of zoom or focus creep, but if the zoom level stays the same, and attempts to refocus doesn't work, that can't be it.

[jimjam11]

15 hours ago, Erling G-P said:

How do you get DSS to produce those graphs - or have you copied the FWHM values to another program to make them ?

Very strange problem for the OP.  Having a Nikon 18-200mm zoom lens which won't hold its zoom level if pointed up or down (it 'falls' in or out by itself), my initial tought was to suggest some sort of zoom or focus creep, but if the zoom level stays the same, and attempts to refocus doesn't work, that can't be it.

I used Pixinsight for that, but looking in DSS it does have a FWHM column so I assume it is measured as part of the analysis (albeit in pixels rather than arcsec).

 

[jimjam11]

So:

1. This happened every/most nights you attempted to image?

2. It happens in all parts of the sky?

3. You are confident you are accurately focussed? If so, how are you judging this?

4. Presumably you are guiding? If it was high clouds the guide star mass would vary like this:

 

[sploo]

6 hours ago, jimjam11 said:

So:

1. This happened every/most nights you attempted to image?

2. It happens in all parts of the sky?

3. You are confident you are accurately focussed? If so, how are you judging this?

4. Presumably you are guiding? If it was high clouds the guide star mass would vary like this:

 

1. Yes, but that's only a handful of times, and at one location - so perhaps not that significant (in the sense of being able to say it always happens with that equipment)

2. Maybe not. I'm pretty certain there have been a couple of times where I've moved the camera to point at a different area of sky and got a better result

3. Generally I'm using the liveview autofocus on a bright star (though the 5D4 is much more reliable than the 7D), switching the lens to manual, then swinging to the target

4. As in - a separate autoguiding camera? No - just using the Skywatcher Star Adventurer for the moment

I think a comment made earlier in this thread (about acclimatising the camera before shooting) is pertinent. It makes sense that focus might change as the lens and camera cools (assuming I've set up focus when the camera had just come outside from a warm room).

However, when I've experienced the problem of the soft images - and been present to check the camera - I'm pretty certain I've never got the image sharp again by attempting to refocus. The front element of the lens has never shown any fogging in those scenarios, so unless there is something fogged up inside the camera then maybe it is just atmospheric issues. I.e. I've just been unlucky in that every time I have gone out to image I've had some cloud roll over a few tens of minutes after starting.

[Notty]

Was there much of a temp drop between start and end of imaging? I’d always understood correct focus does change with temperature so as it drops you’d need to refocus

[sploo]

37 minutes ago, Notty said:

Was there much of a temp drop between start and end of imaging? I’d always understood correct focus does change with temperature so as it drops you’d need to refocus

Good question. It's not something I recorded (or was aware of) so couldn't say for sure.

It's interesting that on both of the quality graphs I posted earlier there's a strong fall off in the image score over the first ~50 frames (25 minutes) and then it tends to flatten out. I assume that a camera brought outside from a warm room will experience a higher rate of temperature shift initially, then settle down. I think the important test next time the skies are good here will be to get the camera outside for a good 30 minutes before I start to shoot.

[geoflewis]

15 minutes ago, sploo said:

Good question. It's not something I recorded (or was aware of) so couldn't say for sure.

It's interesting that on both of the quality graphs I posted earlier there's a strong fall off in the image score over the first ~50 frames (25 minutes) and then it tends to flatten out. I assume that a camera brought outside from a warm room will experience a higher rate of temperature shift initially, then settle down. I think the important test next time the skies are good here will be to get the camera outside for a good 30 minutes before I start to shoot.

Yes, all gear (telescopes, cameras, etc.) used for astrophotography (actually also for visual) needs time to acclimatise to ambient outside temperatures.

[vlaiv]

+1 temperature shift.

Let the gear cool to ambient before you start and then refocus roughly at every 1C of ambient change (if you have slow scope - you can go for 2C, but anything F/5 or faster - odds are you'll need to refocus on ~1C).

If you have motor focuser - make your sequence check for focus every hour and on filter change of course (even if you have par focal filters).

[sploo]

43 minutes ago, vlaiv said:

+1 temperature shift.

Let the gear cool to ambient before you start and then refocus roughly at every 1C of ambient change (if you have slow scope - you can go for 2C, but anything F/5 or faster - odds are you'll need to refocus on ~1C).

If you have motor focuser - make your sequence check for focus every hour and on filter change of course (even if you have par focal filters).

A motor focuser is one of the things on the cards. I've just started putting together a prototype for a tracking mount for the 300P scope (stepper motor driven - from your advice on a post of mine some weeks back).

It did occur to me that it should be possible to control focus with another stepper; certainly at high magnifications it's so easy to lose a target simply by bumping the focuser (and very hard to focus when the target is moving around).