question about differential flexure and image stacking

[DarkAntimatter]

I was wondering if differential flexure is important if one usually takes relatively short subs.  To be more specific, if a setup exhibits differential flexure problems over an hour or so, but for a 5 minute sub the effects are not visible in the images (tight, circular stars), can the flexure problems be ignored for the case of stacking lots of the short images?  

In googling this I've only managed to confuse myself as the answer seems to be yes and no.  Thanks for any help you can provide.

[ollypenrice]

Firstly, be aware that round stars do not indicate good tracking. This common assumption is incorrect because random but equivalent errors in both axes will produce round stars. This is not idle theory, I have seen it happen when calibrating a new mount. You can have round stars and poor tracking. There is no doubt about this.

All sources of guiding error will be diminished in accordance with reduced sub length. How do we define the consequences of different sources of guiding error? One way is to consider one rotation of the worm wheel. On many mounts this is around eight minutes. If you can do eight minutes you should be able to do sixteen or thirty-two.

The amplitude of diff flexure may be very variable. It may arise from a tight cable, for instance, in certain positions in the sky. It will get worse as you approach those positions. (So over time it's effects won't be linear.)  Or, in a reflector guided without OAG, it might be at its worst at angles which introduce the most mirror flop. That might be around 45 degrees elevation but who knows?

Olly

[vlaiv]

It can be done provided that you don't encounter what Olly said - different rate of differential flexure.

I've done it with 1 minute subs - not enough to cause any trailing over the course of multiple hours. It can even be beneficial - you don't need to dither in that case, as there is "natural" dithering between frames. But do check your guiding and mount - measure FWHM to see if your guiding is good in the first place.

[DarkAntimatter]

Thanks, olly and vlaiv,  this makes sense and seems logical.  

I think what confused me most is one site where (I can't remember which one now) it was claimed that stacking would align the stars but not fuzzy things like nebula or galaxies, so even short subs would not help for those images.  But this seems like it would mean the alignment software moved each individual star and not the whole frame to match the previous frames, which doesn't make sense to me.  

[ollypenrice]

10 minutes ago, DarkAntimatter said:

Thanks, olly and vlaiv,  this makes sense and seems logical.  

I think what confused me most is one site where (I can't remember which one now) it was claimed that stacking would align the stars but not fuzzy things like nebula or galaxies, so even short subs would not help for those images.  But this seems like it would mean the alignment software moved each individual star and not the whole frame to match the previous frames, which doesn't make sense to me.  

Whoever said this was deeply confused. Ignore them!

Olly

[Mognet]

17 hours ago, DarkAntimatter said:

I think what confused me most is one site where (I can't remember which one now) it was claimed that stacking would align the stars but not fuzzy things like nebula or galaxies, so even short subs would not help for those images.  But this seems like it would mean the alignment software moved each individual star and not the whole frame to match the previous frames, which doesn't make sense to me.  

It's more a case that stacking software aligns on stars rather than just aligns the stars. As stars, nebulae and galaxies are all positioned relative to each other, even if the field of view moves from image to image, the alignment will improve the signal to noise ratio making it easier to bring out the faint fuzzies captured in the images. The stacking programs I've played with certainly seem to work this way, and it's the way I had intended to write one too. The only exception may be AutoStakkert, which is great for lunar and planetary images. That one seems to do something different for alignment.