Telescope Reviews: SNR: Image Sampling

Well - the heavy qualifier makes this more of a dismissive classification than a form of praise. I am specifically targeting the advantages of video guiding, novel centroiding, and low latency - not for the sake of video - but in order to achieve optimal guiding without AO.

As for your discussion of the merits of longer exposures, etc., please remember that I am very results driven - and I am puzzled that in autoguiding - advice is given based on arguments and theory rather than results. If there is a place where you post high resolution work with mid-range equipment, demonstrating your experience in this realm - please point it out. I have looked for such a site - like I have with MG - but I haven't found it. Without actual experience guiding in the low fwhm range, you are just speculating based on theory. If you have confidence in your approach but your results are limited by seeing, as many claim of their own methods, you are still speculating and don't have direct evidence based on end results. If you prefer not to make the effort for small fwhm - that is perfectly fine - but it does make you less of an informed voice on how to achieve good results in that realm.

You make a separate argument based on limited isoplanatic angle that often comes up with AO work - but it is irrelevant here because it misses yet another point. I am not chasing the seeing - I am not chasing periodic error - I am chasing the gearbox and bearings. I try to remove the effect of seeing so that I have a good centroid that lets me chase the mount. For mid-range mounts, a key source of noise is the 1-10s wiggles from the gearbox and bearings - and guiding on an accurate centroid from a single star will remove wiggles from a wide field of stars and tighten them up. Subtleties like this are part of my different 'take' on autoguiding - and why I am so insistent on actual results to back claims and interpretations. And they amplify the need for low latency, short exposure guide images with an accurate centroid.

With a high end mount, you can be very casual and get tight stars. With a mid-range mount, you can use AO at fairly low rates of 1Hz or so and get very good improvement across a wide field. This is part of what made me focus on this high speed low latency realm - but without AO. And I think my results - which I regularly post - show that I am onto something.

As for sampling - you have suggested that "mere mortals" cannot achieve small fwhm easily - but again my goal is exactly that - for motivated amateurs with humble equipment and free software to achieve results comparable to much higher end equipment and costly software. I think they should feel empowered to compete with higher end equipment if they so choose, and I would not underestimate their abilities or potential.

A key thing to clarify is whether to sample based on seeing or to sample based on REALIZED FWHM. If your guiding is good, then you can use seeing as a reference because it limits your fwhm - but if your guiding is bad - you cannot. Many nice looking images have fwhm's in the 4-6" range, independent of seeing, and it would make no sense to sample finely with such large stars. But I and others aren't really interested in imaging if the results are above 3" because the images are so much more detailed at smaller fwhm - which is why people spend so much to achieve such results.

I don't encourage beginners to start with a long focus SCT and a dslr - I would encourage them to start with a short refractor guided by a guidescope. Later, when if they want to move to an sct, I would encourage them to use OAG rather than a guidescope because they would likely get better guiding - which is why high end work tends to be done that way. The problem is - many beginner imagers already start with an SCT - so it is a choice between buying a guidescope and encountering flexure - or starting out with OAG and its steeper learning curve, but more consistent results.

No matter what - they need to focus well - which is not trivial.

If they are getting non-round stars at long focal length - that isn't an issue of over-sampling - that is an issue of poor guiding or flexure. Under sampling would only serve to mask the problem - not solve it. With OAG and good focus, long focal length is much easier to handle than with a guidescope. I'm not sure if you have done much OAG guiding, but if you haven't, your pessimism of easy results may be more linked to guidescopes than long focal length.

I have high praise for PHD - although I have never seen it in use. I gather it is focused on an enjoyable experience for the user, particularly a beginner, and people not only achieve results quickly, but enjoy the process. In contrast, I would call the MG interface somewhat user belligerent, though I tried to improve it in the last release. I also like that PhD is available on both windows and mac.

When I first came out with MG I was looking forward to scientific and engineering discussions on autoguiding that would eschew anecdote and myth and focus on results. In such an environment, anyone could have a crazy idea that departed from the herd, and the only way to separate myth from progress is by the results they get.

This vision was pie in the sky and instead I see the same anecdotes repeated and no sense of validating assumptions. I don't think it's too late, and I'm happy to discuss things more, probably offline, if you would like. Ultimately I'm interested in people getting good images using good technique - no matter what software they use - and I think much of what I have learned would have application to guiding with PhD.

Frank