Go Back   IceInSpace > Equipment > Astrophotography and Imaging Equipment and Discussions

Reply
 
Thread Tools Rate Thread
  #1  
Old 18-05-2016, 11:56 AM
Shiraz's Avatar
Shiraz (Ray)
Registered User

Shiraz is offline
 
Join Date: Apr 2010
Location: ardrossan south australia
Posts: 4,918
short subs - better resolution?

OK, this is what I measured. Left chart shows the first night where I just did a sequence of subs from 5 minutes down to 1 second (with a handful of subs at the shorter settings so that they could be averaged). The second try was more ambitious, with more subs of various lengths taken in sequence, but with the sequence repeated 4 times to ensure that resolution changes were not just due to slowly improving seeing. Focusing was redone regularly, but not synchronised with the sub length sequence (ie, there was no focus bias in favour of any sub length). Same result both nights - shorter subs are sharper as measured by the PI image grading script (also hand checked a few subs with essentially similar results).

Now to try to work out what the mechanism might be!!!

regards Ray

edit: am guiding from well outside of the isokinetic patch, and am beginning to suspect that might be behind this result.
Attached Thumbnails
Click for full-size image (seeing3.jpg)
39.5 KB91 views

Last edited by Shiraz; 26-05-2016 at 08:20 AM.
Reply With Quote
  #2  
Old 18-05-2016, 12:00 PM
gregbradley's Avatar
gregbradley
Registered User

gregbradley is offline
 
Join Date: Feb 2006
Location: Sydney
Posts: 17,877
That's not surprising due to PE of the mount. It affects the subs less in shorter subs. I would put it down to that.

I have noticed sometimes focus subs can be more hi rez than longer exposures due to PE.

Greg.
Reply With Quote
  #3  
Old 18-05-2016, 12:07 PM
Shiraz's Avatar
Shiraz (Ray)
Registered User

Shiraz is offline
 
Join Date: Apr 2010
Location: ardrossan south australia
Posts: 4,918
Quote:
Originally Posted by gregbradley View Post
That's not surprising due to PE of the mount. It affects the subs less in shorter subs. I would put it down to that.

I have noticed sometimes focus subs can be more hi rez than longer exposures due to PE.

Greg.
Thanks Greg. But how does that work exactly - do you think the guide points of our mounts wander about over periods of minutes? (genuine question not sarcasm). there does not seem to be much change in star shape, so the effect applies in both axes - and there is no PE in dec.

Does this mean that, if we can use shorter subs, we can get sharper images?
Reply With Quote
  #4  
Old 18-05-2016, 12:09 PM
multiweb's Avatar
multiweb (Marc)
ze frogginator

multiweb is offline
 
Join Date: Oct 2007
Location: Sydney
Posts: 22,060
Hi Ray, I've noticed that many times as well. It makes sense to me. Taking mount performance out of the equation and assuming you have an hypothetical mount that tracks perfectly the seeing comes and goes in small intervals of time usually tens of minutes. It is very obvious when guiding let's say a series of 5min subs. This was very obvious at the SPSP. The seeing being very good and the temperature stable it was very easy to pick up small variations. So you might end up with very sharp subs and some others not so sharp. That's in a perfect world. Add the mount PE and wind buffeting and other factors on top and you invariably will end up with bigger star profiles the longer you expose. There is also the well depth of your camera. More exposure time, more photons and the star is usually bigger.
Reply With Quote
  #5  
Old 18-05-2016, 12:18 PM
Shiraz's Avatar
Shiraz (Ray)
Registered User

Shiraz is offline
 
Join Date: Apr 2010
Location: ardrossan south australia
Posts: 4,918
Quote:
Originally Posted by multiweb View Post
Hi Ray, I've noticed that many times as well. It makes sense to me. Taking mount performance out of the equation and assuming you have an hypothetical mount that tracks perfectly the seeing comes and goes in small intervals of time usually tens of minutes. It is very obvious when guiding let's say a series of 5min subs. This was very obvious at the SPSP. The seeing being very good and the temperature stable it was very easy to pick up small variations. So you might end up with very sharp subs and some others not so sharp. That's in a perfect world. Add the mount PE and wind buffeting and other factors on top and you invariably will end up with bigger star profiles the longer you expose. There is also the well depth of your camera. More exposure time, more photons and the star is usually bigger.
Thanks for the confirmation Marc. I tried to take varying seeing out of the equation by interlacing the subs - ie if a good bit of seeing came along, they all had a look at it. Similarly if there was wind buffeting, it was as likely to affect a 1 second sub as a 2 minute sub - no sorting was done to remove bad short subs.
I didn't measure star size, just FWHM, which is nearly independent of brightness for unsaturated stars. I manually compared similar stars in a few subs to ensure that star saturation did not affect the result.

If it is this easy to decrease the effective seeing from 2.6 to 2.1 arcseconds, why haven't I been doing it??
Reply With Quote
  #6  
Old 18-05-2016, 12:21 PM
Atmos's Avatar
Atmos (Colin)
Ultimate Noob

Atmos is offline
 
Join Date: Aug 2011
Location: Melbourne
Posts: 6,980
There are thermal pockets in the atmosphere that change over the course of seconds to minutes. The atmosphere in fluid dynamics works as a convective system. Warmer thermal pockets will rise up causing localised seeing changes, depending on the size of these pockets, where they are in relation to where you're looking and temperature variations (greater thermal differences causes faster moving pockets).

Longer subs increase the chance of imaging through these pockets, ultimately each sub is only going to be as good as the poorest bit of seeing you had over that period.
Reply With Quote
  #7  
Old 18-05-2016, 12:30 PM
Shiraz's Avatar
Shiraz (Ray)
Registered User

Shiraz is offline
 
Join Date: Apr 2010
Location: ardrossan south australia
Posts: 4,918
Quote:
Originally Posted by Atmos View Post
There are thermal pockets in the atmosphere that change over the course of seconds to minutes. The atmosphere in fluid dynamics works as a convective system. Warmer thermal pockets will rise up causing localised seeing changes, depending on the size of these pockets, where they are in relation to where you're looking and temperature variations (greater thermal differences causes faster moving pockets).

Longer subs increase the chance of imaging through these pockets, ultimately each sub is only going to be as good as the poorest bit of seeing you had over that period.
that makes sense.

But, if I take 100 x 1second subs and a 10 second burst of bad seeing come through, I have 10 bad subs- wouldn't they then stuff up the final stack just as much as the same 10 seconds of bad seeing would stuff up a single 100 second sub? - I didn't discard any subs in this exercise.

Last edited by Shiraz; 18-05-2016 at 12:42 PM.
Reply With Quote
  #8  
Old 18-05-2016, 12:34 PM
gregbradley's Avatar
gregbradley
Registered User

gregbradley is offline
 
Join Date: Feb 2006
Location: Sydney
Posts: 17,877
I am not sure if this is a factor but it could also be down to the airy disc.

The airy disc has outer rings and these are not visible in short exposures but they would start to become visible in longer exposures. When you say FWHM is nearly independent of the brightness of the stars I have not found that to be true at all. Usually the lowest FWHM is always a dim not very bright star
and the larger really bright stars have the worst FWHM using CCDstack to measure them. So much so I usually measure a centre star that is small and a bit on the dim side.

This could also relate to the above idea about the outer rings of the airy disc becoming brighter in bright stars in longer exposures. I see a similar phenomenon with small well cameras versus deep well cameras. In short well cameras you have to be alert to overwhelming the wells.

I have noticed the KAF8300 blooms a little bit in 2x2 subs at 10 minutes on a bright star on my Honders at 305mm F3.8. So 5 minutes 2x2 seems to be the go there.

Your finding adds even more value to using larger aperture and faster F ratio optics. They are more demanding in terms of alignments but the gains are there in several ways.


There is a good graphic on CCDware site about autoguding that shows how a star gets pushed around over the course of an exposure and this obviously fattens it. It can fatten in a round manner if the PE is in all directions. So that's why you see longer exposures tending to have larger sizes.


But there are times you want a longer exposure anyway like narrowband or to get the faintest detail and you have to rise about the read noise.

So we are back to the low read noise topic of the ASI1600 again and the appeal of short exposure high QE reasonably widefied astrophotography.

Greg.
Reply With Quote
  #9  
Old 18-05-2016, 12:37 PM
multiweb's Avatar
multiweb (Marc)
ze frogginator

multiweb is offline
 
Join Date: Oct 2007
Location: Sydney
Posts: 22,060
Quote:
Originally Posted by Shiraz View Post
If it is this easy to decrease the effective seeing from 2.6 to 2.1 arcseconds, why haven't I been doing it??
One issue with short subs is the read noise for each sub you download to your PC depending on your camera. The process that empties the CCD buffer, the quality of the USB cables, how fast your PC/notebook CPU is, time lag, background processes, write speed of your HD, quality of the capture software and drivers, etc... So in the end it's always a compromise.

Another thing to consider is that for very faint extended object with a low surface brightness unless you have a perfectly collimated light bucket you'll be stuck with fewer long subs rather than many short ones.
Reply With Quote
  #10  
Old 18-05-2016, 12:41 PM
gregbradley's Avatar
gregbradley
Registered User

gregbradley is offline
 
Join Date: Feb 2006
Location: Sydney
Posts: 17,877
Quote:
Originally Posted by multiweb View Post
One issue with short subs is the read noise for each sub you download to your PC depending on your camera. The process that empties the CCD buffer, the quality of the USB cables, how fast your PC/notebook CPU is, time lag, background processes, write speed of your HD, quality of the capture software and drivers, etc... So in the end it's always a compromise.

Another thing to consider is that for very faint extended object with a low surface brightness unless you have a perfectly collimated light bucket you'll be stuck with fewer long subs rather than many short ones.
How much would you estimate the cable etc you mention affects the noise?

Greg.
Reply With Quote
  #11  
Old 18-05-2016, 12:48 PM
multiweb's Avatar
multiweb (Marc)
ze frogginator

multiweb is offline
 
Join Date: Oct 2007
Location: Sydney
Posts: 22,060
Quote:
Originally Posted by gregbradley View Post
How much would you estimate the cable etc you mention affects the noise?

Greg.
A lot. USB have to be shielded for a start. I use a ferrite blocks on all my USB cables. Also trying to keep them away from power cables. That's for the cabling.

Now my old QHY8 doesn't handle very well interruptions or even interleaves when you download a sub even in slow mode. So I have to be aware of that and pause guiding so the USB cable maximum bandwidth is available when the subs is transferred, then resume when it's done.

Again this all depends on the gear you're using. Some people may never have these issues, some others do.
Reply With Quote
  #12  
Old 18-05-2016, 12:50 PM
rally
Registered User

rally is offline
 
Join Date: Sep 2007
Location: Australia
Posts: 896
Its the basis of Lucky Imaging !
Reply With Quote
  #13  
Old 18-05-2016, 12:51 PM
Shiraz's Avatar
Shiraz (Ray)
Registered User

Shiraz is offline
 
Join Date: Apr 2010
Location: ardrossan south australia
Posts: 4,918
Quote:
Originally Posted by gregbradley View Post
I am not sure if this is a factor but it could also be down to the airy disc.

The airy disc has outer rings and these are not visible in short exposures but they would start to become visible in longer exposures. When you say FWHM is nearly independent of the brightness of the stars I have not found that to be true at all. Usually the lowest FWHM is always a dim not very bright star
and the larger really bright stars have the worst FWHM using CCDstack to measure them. So much so I usually measure a centre star that is small and a bit on the dim side.

This could also relate to the above idea about the outer rings of the airy disc becoming brighter in bright stars in longer exposures. I see a similar phenomenon with small well cameras versus deep well cameras. In short well cameras you have to be alert to overwhelming the wells.

I have noticed the KAF8300 blooms a little bit in 2x2 subs at 10 minutes on a bright star on my Honders at 305mm F3.8. So 5 minutes 2x2 seems to be the go there.

Your finding adds even more value to using larger aperture and faster F ratio optics. They are more demanding in terms of alignments but the gains are there in several ways.


There is a good graphic on CCDware site about autoguding that shows how a star gets pushed around over the course of an exposure and this obviously fattens it. It can fatten in a round manner if the PE is in all directions. So that's why you see longer exposures tending to have larger sizes.


But there are times you want a longer exposure anyway like narrowband or to get the faintest detail and you have to rise about the read noise.

So we are back to the low read noise topic of the ASI1600 again and the appeal of short exposure high QE reasonably widefied astrophotography.

Greg.
Hi Greg. I don't think that we even get remotely close to seeing the Airy disk with our largeish scopes and in typical seeing - the star shapes and sizes are almost solely down to the atmosphere and scattering in our scopes.

I used FWHM, because it is independent of star brightness and size - if it isn't, the detector must be operating in a non-linear region, or the star profile detector algorithm is approximate. In PI, the FWHM generally stays within about 10% on my raw images (that's an impression, haven't measured it specifically), over a wide range of star brightnesses. That was why I used it as a measure of sharpness. I am surprised that you have found FWHM to be unreliable, but I guess that must depend on the software?

yes, this arose from the 1600 discussion - I wanted to see just what would occur with short subs and was surprised at the result - still am.

Last edited by Shiraz; 18-05-2016 at 01:11 PM.
Reply With Quote
  #14  
Old 18-05-2016, 12:55 PM
codemonkey's Avatar
codemonkey (Lee)
Lee "Wormsy" Borsboom

codemonkey is offline
 
Join Date: Jul 2013
Location: Kilcoy, QLD
Posts: 2,058
Quote:
Originally Posted by Shiraz View Post
But, if I take 100 x 1second subs and a 10 second burst of bad seeing come through, I have 10 bad subs- wouldn't they then stuff up the final stack just as much as the same 10 seconds of bad seeing would stuff up a single 100 second sub? - I didn't discard any subs in this exercise.
Interesting thread, Ray, thanks for posting.

I'm inferring from the above post that you integrated the images. I assume that the total exposure for each integration was normalised? Sorry if that was already posted and I missed it.

In addition, if the images were integrated, there's potentially a lot of things happening here, for example, Lanczos 3 interpolation is the default in PI for registration and this has a slight sharpening effect... if you integrate more subs, does that impact the overall FWHM?

What rejection algorithm did you choose? I'd be particularly interested in comparing results with no rejection at all; as pure a mean as possible would be of most interest I think.

A note on FWHM: this is calculated very differently in different applications. Some applications, like PixInsight do it "properly"; calculating it via fitting to statistical distribution models.

Other applications, like PHD, approximate it by calculating the intensity slope between adjacent pixels.

Even applications using statistics as the basis for their calculations will vary a lot depending on the distribution model (gaussian vs moffat for example). Some might even use multiple and use a fitting algorithm to find the best fit.

Ray is correct in that a true FWHM (on an unsaturated star), should be independent of brightness value, assuming linear sensor response, but I'd be skeptical about the assumption that the reported FWHM values calculate it correctly anyway.
Reply With Quote
  #15  
Old 18-05-2016, 12:56 PM
Atmos's Avatar
Atmos (Colin)
Ultimate Noob

Atmos is offline
 
Join Date: Aug 2011
Location: Melbourne
Posts: 6,980
Quote:
Originally Posted by Shiraz View Post
that makes sense.

But, if I take 100 x 1second subs and a 10 second burst of bad seeing come through, I have 10 bad subs- wouldn't they then stuff up the final stack just as much as the same 10 seconds of bad seeing would stuff up a single 100 second sub? - I didn't discard any subs in this exercise.
Those 10 subs shouldn't make too much difference, they may a tiny bit but certainly not to the same extent.

When stacking a lot of short exposures together, they must be registered for small wobbled between each sub. I also wonder how well star registration works in fraction of a pixel changes in short subs. Maybe a large stack of short subs can be almost as bad as a large stack due to an inability to correctly register such small changes in stellar location. Someone else will know more about how small of a movement the PI Registration can work at.
Reply With Quote
  #16  
Old 18-05-2016, 01:01 PM
Shiraz's Avatar
Shiraz (Ray)
Registered User

Shiraz is offline
 
Join Date: Apr 2010
Location: ardrossan south australia
Posts: 4,918
Quote:
Originally Posted by codemonkey View Post
Interesting thread, Ray, thanks for posting.

I'm inferring from the above post that you integrated the images. I assume that the total exposure for each integration was normalised? Sorry if that was already posted and I missed it.

In addition, if the images were integrated, there's potentially a lot of things happening here, for example, Lanczos 3 interpolation is the default in PI for registration and this has a slight sharpening effect... if you integrate more subs, does that impact the overall FWHM?

What rejection algorithm did you choose? I'd be particularly interested in comparing results with no rejection at all; as pure a mean as possible would be of most interest I think.

A note on FWHM: this is calculated very differently in different applications. Some applications, like PixInsight do it "properly"; calculating it via fitting to statistical distribution models.

Other applications, like PHD, approximate it by calculating the intensity slope between adjacent pixels.

Even applications using statistics as the basis for their calculations will vary a lot depending on the distribution model (gaussian vs moffat for example). Some might even use multiple and use a fitting algorithm to find the best fit.

Ray is correct in that a true FWHM (on an unsaturated star), should be independent of brightness value, assuming linear sensor response, but I'd be skeptical about the assumption that the reported FWHM values calculate it correctly anyway.
Hi Lee. No I wanted to get around all of the complications of stacking by taking just the raw subs and averaging the FWHM for the handful of interleaved subs used at each sub length (from 4 subs total at 300sec up to 200 subs at 1 sec (from memory - don't have the data on this PC) . Nothing has been stacked.
Reply With Quote
  #17  
Old 18-05-2016, 01:05 PM
Shiraz's Avatar
Shiraz (Ray)
Registered User

Shiraz is offline
 
Join Date: Apr 2010
Location: ardrossan south australia
Posts: 4,918
Quote:
Originally Posted by Atmos View Post
Those 10 subs shouldn't make too much difference, they may a tiny bit but certainly not to the same extent.

When stacking a lot of short exposures together, they must be registered for small wobbled between each sub. I also wonder how well star registration works in fraction of a pixel changes in short subs. Maybe a large stack of short subs can be almost as bad as a large stack due to an inability to correctly register such small changes in stellar location. Someone else will know more about how small of a movement the PI Registration can work at.
Thanks Colin. maybe this is suggesting that short subs will need care in stacking. My thinking was going down the same route - there may be local star-to-star distortion within the short subs that doesn't show up in individual sub FWHM, but will affect the final stack, unless a distortion correcting transform is used. Clearly needs more work, but is sounding plausible.

edit: thinking further, the star-to star distortion (if it exists) could have an origin in the small size of the isokinetic patch - stars will wobble differently in different parts of the image - that wobble will be averaged for longer subs, but will be partially frozen by shorter subs. Interesting that there seems to be a fairly smooth change from 5 minutes down - no sign of any plateau.

Last edited by Shiraz; 18-05-2016 at 01:54 PM.
Reply With Quote
  #18  
Old 18-05-2016, 01:09 PM
gregbradley's Avatar
gregbradley
Registered User

gregbradley is offline
 
Join Date: Feb 2006
Location: Sydney
Posts: 17,877
What's an interleaved sub? I haven't heard that term before. Interleave means to insert something between 2 other things repetitively.

Greg.
Reply With Quote
  #19  
Old 18-05-2016, 01:16 PM
multiweb's Avatar
multiweb (Marc)
ze frogginator

multiweb is offline
 
Join Date: Oct 2007
Location: Sydney
Posts: 22,060
Quote:
Originally Posted by gregbradley View Post
What's an interleaved sub? I haven't heard that term before. Interleave means to insert something between 2 other things repetitively.

Greg.
Like interlaced for a TV picture for instance. Some process download half of the sub (every second scan line) then go through the rest of the remaining lines. Some others just stream the pixels in a continuous stream from left to right in one long uninterrupted data string. This depends on camera drivers implementation, hardware type and capture software used. Some implementations are good and can buffer the stream, some others (like my old QHY8) don't buffer that well and you end up with missing lines or horizontal artefacts in your downloaded sub.

Sorry - cross post. Ray's talking about something else. Still valid explanation.
Reply With Quote
  #20  
Old 18-05-2016, 01:17 PM
Shiraz's Avatar
Shiraz (Ray)
Registered User

Shiraz is offline
 
Join Date: Apr 2010
Location: ardrossan south australia
Posts: 4,918
Quote:
Originally Posted by gregbradley View Post
What's an interleaved sub? I haven't heard that term before. Interleave means to insert something between 2 other things repetitively.

Greg.
I tried to set it up so that the subs of various lengths would be fairly well mixed up - eg, a the 5 minute subs were not all taken together, but were spaced out with other length subs taken in between. The intention was to minimise the effect of slow changes in seeing - if the seeing changed, it would affect the averages for all sub lengths by roughly the same amount.

sort of like shuffling a deck of cards so that all the kings are not together.
Reply With Quote
Reply

Bookmarks

Thread Tools
Rate This Thread
Rate This Thread:

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT +10. The time is now 05:04 AM.

Powered by vBulletin Version 3.8.7 | Copyright ©2000 - 2024, Jelsoft Enterprises Ltd.
Advertisement
Bintel
Advertisement
Testar
Advertisement