The "No EQ" DSO Challenge!

[SteveNickolls]

4 hours ago, SilverAstro said:

Hi Steve very interesting very nice images, amazing in fact

The two of Andromeda - are they full frames or cropped ? Reason I ask is, are the scales of the images as presented comparable ?

Hi,

Thanks for your kind comments. They would both have been subject to some cropping, the one taken with the Canon piggybacked would only have been cropped slightly a few pixels around the edge. The piggybacked image was taken at 200mm FL (on a APS-C camera like the Canon 600D I think this works out at 320mm) while the other was through my refractor (500mm FL). The image through the telescope was cropped more vertically to remove a lot of poorer background. hope that helps.

Cheers,
Steve

[SilverAstro]

1 hour ago, The Admiral said:

My emphasis. Why rotate in software, why not rotate the camera at the beginning of each set so that you get back to roughly the same starting position for each set? Presumably DSS would stack them all as one set?

You may need to do flats for each position if the field uniformity is not axially symmetric, in which case I guess you'd need to stack each set separately.

Ian

Indeed ! that is the thought that I was having that caused me to put in the > Edit : hold !!

that if one did not re-align the camera with the scope (and/or celestial sphere) then it would still be limited to the 'common circle view' that was causing the Nigel crop, so I went away to have a think, ,  and then domestic got in the way So if two of us thunk it it must be right

Any moment now we'll be back to de-rotating mechanisms !

 

Edited September 2, 2016 by SilverAstro

[The Admiral]

19 minutes ago, SteveNickolls said:

My thoughts would be that at that FL the percieved field rotation on the chip was insufficient to cause a streak to appear.

Steve, the trouble is that rotation is rotation, and it won't be affected by the focal length of the lens/telescope. This is not the same as cropping which, although simulating a longer FL, would allow the most affected parts of the image to be removed.

Ian

[SilverAstro]

21 minutes ago, SteveNickolls said:

Hi,

Thanks for your kind comments. They would both have been subject to some cropping, the one taken with the Canon piggybacked would only have been cropped slightly a few pixels around the edge. The piggybacked image was taken at 200mm FL (on a APS-C camera like the Canon 600D I think this works out at 320mm) while the other was through my refractor (500mm FL). The image through the telescope was cropped more vertically to remove a lot of poorer background. hope that helps.

Cheers,
Steve

The reason for the question was this business of the proportion of streak*, it being a degree of rotation, which in the meanwhile of my domestic you are all talking about Cos I was trying to get my head round -> if they are similar scales or maybe even if they are not, the proportion of any streak due (*edit as Ian has just said !)  should be the same and thus not the reason for DSS accepting the one lot compared to the other lot.

and on top of all that they are both taken on the same alt-az mount so it cant be drive jitter sooooo I'm not sure where to go next with  it !

* I like streak, easier to type than "field rot "

Edited September 2, 2016 by SilverAstro

[The Admiral]

3 minutes ago, SilverAstro said:

Indeed ! that is the thought that I was having that caused me to put in the > Edit : hold !!

that if one did not re-align the camera with the scope (and/or celestial sphere) then it would still be limited to the 'common circle view' that was causing the Nigel crop, so I went away to have a think, ,  and then domestic got in the way So if two of us thunk it it must be right

Any moment now we'll be back to de-rotating mechanisms !

 

Ha ha, yes!

And though I may sound convincing, I may not be right . But in this case, having thought about it again, I don't think you'd end up with needing to crop as much by rotating the camera as if you just let it run without a break. Indeed, I suppose if that wasn't the case then de-rotators wouldn't work. Yup, we're back to them!

Ian

[SilverAstro]

1 minute ago, The Admiral said:

Ha ha, yes!

and it is real difficult thinking it / typing it  without being able to wave one's arms about well, I can, but you lot cant see !!!

I think I need a beer , , ,

[jimbo747]

4 hours ago, SilverAstro said:

I cant see it !

You are right, I cant see frame rotation, interesting !

Perhaps you could post a streaky one because I dont think I have seen a picture of this frame rotation wot everyone is on about   (people dont generally post their duds, but they can often be as informative, like what not to do ! as the good ones ) As you say, high in the south is supposed to be the poor direction, very interesting.

Nice dumbell, nice image.

 

Ok, I've attached 3 cropped images, 2 of which from a sub, and 1 from the final processed image to show difference.

1 raw image is what I'd call an 'average' frame, its what at least half of the subs looked like, one is the 'best' frame showing the least amount of streaking. The image has rotated quite a bit in between taking the first frame and the last, presumably because of its position in the sky, maybe that is what helped DSS reduce streaks in the final image. Again, in the final image, only curves, levels and a rough alignment of the histogram were done.

 

Edit - Forgot to add, FL was probably around 180/190mm as I couldn't focus at 200m, had to wind back a tiny bit to get focus.

1) least amount of streaking

2) 'Average' sub

3) final image

 

Edited September 2, 2016 by jimbo747
pic order

[SilverAstro]

5 minutes ago, jimbo747 said:

The image has rotated quite a bit in between taking the first frame and the last, presumably because of its position in the sky,

Thanks for those, I'll have a think about them, (the middle one is well rotated by 90deg but I guess that it a software/posting glitch !? )

cos I was imagining a streak due to the 'frame rotation' due to the nature of an alt-az mount that it would appear to be like a bull's-eye display, the object in the middle with a set of arc stars round it getting bigger the further out. ( Like a star trails picture round Polaris in fact ! )

[jimbo747]

5 minutes ago, SilverAstro said:

Thanks for those, I'll have a think about them, (the middle one is well rotated by 90deg but I guess that it a software/posting glitch !? )

cos I was imagining a streak due to the 'frame rotation' due to the nature of an alt-az mount that it would appear to be like a bull's-eye display, the object in the middle with a set of arc stars round it getting bigger the further out. ( Like a star trails picture round Polaris in fact ! )

I think if the object was bang on north, and you took shots for several hours you probably would get that, aim at polaris or something.

 

I noticed after I uploaded that the image has rotated - must be DSS that does this? The camera shots are in portrait due to how I connect camera to mount so all subs are in portrait, stacked image before any PS is in landscape.

Edited September 2, 2016 by jimbo747
edut

[The Admiral]

I agree with SilverAstro, they don't look like rotation, more like just the normal effect I see on subs as a result of the mount moving.

But the fact remains that field rotation is not visible, when I'd have thought it would have shown up. I don't really know how kappa-sigma clipping works, but will DSS just ignore those stars that are badly trailed?

Ian

[Nigel G]

2 hours ago, The Admiral said:

 

 

4 hours ago, SilverAstro said:

I cant see it !

You are right, I cant see frame rotation, interesting !

Perhaps you could post a streaky one because I dont think I have seen a picture of this frame rotation wot everyone is on about   (people dont generally post their duds, but they can often be as informative, like what not to do ! as the good ones ) As you say, high in the south is supposed to be the poor direction, very interesting.

 

Quote

Here's one with slight frame rotation,  it's an image of  Polaris. Taken 6 months ago 4 hours of 30 second subs

? ? 

Nige.

 

 

[SilverAstro]

11 minutes ago, Nigel G said:

Here's one with slight frame rotation

Hehee ! Nice one , yep just a little ! 

Shame about the clouds

[The Admiral]

22 minutes ago, Nigel G said:

 

 

 

Presumably these are just stacked and not corrected for field rotation.

Ian

PS. Sorry, I see I've not entered into the spirit of your post!

Edited September 2, 2016 by The Admiral

[SteveNickolls]

I can't articulate this very well but the idea I have is if the light from a small star in a set period of time could be detected on two adjacent pixels, that would be the start of a trail. However if the light in twice that period of time remained in the same pixel (because the field of view was so much wider) then you wouldn't see any apparent motion. It is possible then that longer periods of time (when using shorter focal lengths) could also show no apparent movement.

My head hurts. I'm off to have a curry and some beer.

Cheers,
Steve

 

[The Admiral]

2 minutes ago, SteveNickolls said:

I can't articulate this very well but the idea I have is if the light from a small star in a set period of time could be detected on two adjacent pixels, that would be the start of a trail. However if the light in twice that period of time remained in the same pixel (because the field of view was so much wider) then you wouldn't see any apparent motion. It is possible then that longer periods of time (when using shorter focal lengths) could also show no apparent movement.

My head hurts. I'm off to have a curry and some beer.

Cheers,
Steve

 

You are right Steve, so long as you are referring the the same star. But, a star on the edge of the frame will appear nearer the centre when used with a shorter focal length. Another star, again on the edge of the frame, will still move by the same amount as your original star with the longer focal length. Another way of thinking about it is imagine a line passing through the centre of the frame. If the camera, or image, rotates by, say, 1 degree, that line will appear to rotate by the same amount, irrespective of the fl of the lens attached.

Curry and a beer? The answer to most problems!

Ian

[Herzy]

8 hours ago, The Admiral said:

Unlike a finely tuned EQ mount, there are two main problems with the Alt-Az mount: field rotation and tracking performance. In order to minimize the manifestation of field rotation on one's images one needs to use short exposures, only a few tens of seconds, so a total exposure of 4 hours or more will produce hundreds of subs. Star streaking will be evident in a percentage of these subs, the more as the exposure time is increased (even if that time is acceptable from a field rotation point of view), and DSS will reject them, so the integrated exposure will be less than one might hope for. I think it's also true that the overall tracking performance of Alt-Az mounts is somewhat below that of the correctly adjusted EQ mount, so if one did embark on a long exposure session it would have to be done in a number of shorter bites with mount realignment in between.

I'm not saying that all this couldn't be done, but the reason I'm using Alt-Az imaging is that I have to set up the gear each time I use it so it needs to be simple and quick to set up (and I can't see the Pole star from my observing position). As Steve alluded to, this is because here in the UK the weather seems so unpredictable that often the decision to image is a last minute one. Personally, I'm not striving for salon-quality images, but just to reveal the glory of what I can only see as a grey smudge when viewed by eye. That is not to say of course, that I don't want the best output that I can achieve within the constraints imposed by this style of imaging. But one has to be realistic about the law of diminishing returns.

Don't forget also that much of the UK is blighted by light pollution, so that is yet another issue we have to contend with.

Ian

I see. Even though I work with an EQ mount, I still use 30s subs for the most part because my mount isn't the most reliable. I usually end up with like 500 images, which causes trouble for my computer when I try to stack. 

I see what your talking about though. I dealt with the same problems not very long ago. 

Just keep in mind the potential of your setup if you were to go for "hardcore imaging" of several hours. Although it'll be difficult, the results are rewarding.

---------------

Btw, Steve, those images of Andromeda are pretty awesome. I've heard it's bright, but is it an easy target? I've never imaged it before and I'm thinking of giving it a go.

Edited September 2, 2016 by Herzy

[SteveNickolls]

13 hours ago, Herzy said:

Btw, Steve, those images of Andromeda are pretty awesome. I've heard it's bright, but is it an easy target? I've never imaged it before and I'm thinking of giving it a go.

Hi,

Thanks for the kind comments on M31. Yes it is an easily accessible target, it can be seen in the finder scope so composing your image is much easier than say with M33 which required taking snapshots to check its position in Live View. Right now if you are using an Alt-Az mount and have to consider the implications of field rotation M31 is well placed in the sky both in terms of azimuth and altitude. If you had an equitorial mount I guess you'd wait until it was at its highest position. With my SkyWatcher Synscan Alt-Az mount I have been able to take individual exposures of M31 lasting 60 seconds and the mount overall has produced a high percentage of frames acceptable to DSS (76%). Definitely go for it!

Cheers,
Steve

Edited September 3, 2016 by SteveNickolls

[SteveNickolls]

2 hours ago, The Admiral said:

You are right Steve, so long as you are referring the the same star. But, a star on the edge of the frame will appear nearer the centre when used with a shorter focal length. Another star, again on the edge of the frame, will still move by the same amount as your original star with the longer focal length. Another way of thinking about it is imagine a line passing through the centre of the frame. If the camera, or image, rotates by, say, 1 degree, that line will appear to rotate by the same amount, irrespective of the fl of the lens attached.

Curry and a beer? The answer to most problems!

Ian

So if I crop out the star at the edge...I get to image for longer?

It's the beer talking.

Cheers,
Steve

[happy-kat]

Would it work if you image say for an hour and then another hour but in DSS use the tabs for the separate batches

[SilverAstro]

DSS can not compensate for those parts of the sky that have rotated out of the camera frame during the 2nd and 3rd hours

Ah ! what we need are circular sensors !! doh

[The Admiral]

Hmm, that's an interesting question Steve, needing a bit of thought! My gut feeling is that the answer is yes, but with a caveat. If the pictures are presented at the same size, then the cropped one would need to be enlarged more, and this would reverse the gains you've made by cropping. I guess the exception would be that if you've cropped to the point where the rotation is less than a single pixel, then it wouldn't make any difference how much it is enlarged. Then again, I imagine that would be fairly extreme cropping.

I reserve the right to change my mind at any time 

Ian

Edited September 2, 2016 by The Admiral

[SilverAstro]

5 minutes ago, The Admiral said:

If the pictures are presented at the same size, then the cropped one would need to be enlarged more, and this would reverse the gains you've made by cropping.

I reserve the right to change my mind at any time

Yes, I agree with you (for now !)  I think it is all in the matter of scale. But we still need to sort out why Steve's DSS was happier with his whateverit was compared to his thingumy on Andromeda !

I reserve the right to change your mind for you   and Steve is free to crop at the edge, as long as he doesnt crop the middle, for a crop there would create a black hole and then who knows what might happen.

 

[Stub Mandrel]

2 hours ago, happy-kat said:

Would it work if you image say for an hour and then another hour but in DSS use the tabs for the separate batches

If you use the same flats, darks and bias frames and all the lights have the same exposure then you don't need to use groups. It will align lights that are significantly mis-aligned even from different nights. Groups just make sure that day 1's images are used with day 1's flats and so on.

[SteveNickolls]

Sorry to have taken so long to post these images. They are what I promised to do-restack the M31 and M33 images I posted before in DSS and reprocess using StarTools to see what visible differences might be seen when omitting dark frames but using kappa-sigma in DSS. I would class the M31 image as being easy to process in StarTools indicating the data quality was better (for a brighter object).

First M31 with darks and 'median' setting in DSS-

And now without dark frames but using sigma-kappa in DSS-

And now M33, first with dark frames and 'median' setting in DSS-

And next without dark frames and using kappa-sigma in DSS-

Changes in cropping were down to me.

Comments welcome.

Cheers,
Steve

[happy-kat]

I thought what about no dark frames but with bias and flats.

The second image looks better without the darks but it is not so clear cut with the first. Thanks for experimenting.