Undersampling - Oversampling

Mark,

If we proceed on the basis of the Nyquist frequency, we should oversample at 3.3x our local seeing.

In other words, if our seeing is 2" (FWHM), then we should be shooting at around 0.7" (or better).

Since seeing does vary from location to location, from evening to evening and, in fact, from hour to hour, always pursue an aggressive image scale and you will never regret it.

Anthony.

Ok....this is the last I'll write on this subject until I get round to posting my first picture. But as I said that may be months away. I needed some expert help before buying a CCD - not wanting to buy cheap and loose heart but at the same time not wanting to waste money. Thank you both for your input. I have always used my scope on tripod - but as we speak I have a Mitty wedge en route from the States and hope to put in a pillar in the not too distant future. Either of these outlays will not be wasted, as my scope will still serve me well on this set up for visual....waiting on the camera. When ready....the purchase of the CCD camera will be with the help from your input here. Thanks again.
(For info - I never put this on initial profile - although from Co. Louth area, when not away at work, I live mostly in Spain, where I have many nights of clear skies. My spanish is not good enough for the local clubs/forums - so hope you don't mind me tagging along here.).

Anthony - I vaguely recall reading somewhere (goodness knows where) that empirically, one should aim for an image scale of about half your average seeing. In my case, I used this advice when purchasing my CCD. Average seeing here in Kildare is about 3 arcsec. Therfore, I aimed for an image scale of about 1.5"/pixel. I ended up with an image scale of 1.27"/pixel.

Another thought occcurs to me also. There are many variables in the imaging equation. Pier/tripod stability, mount accuracy (PE), polar alignment, orthogonality (for autoguiding with a guidescope), flexure (guidescope), mirror flop (SCT's especially), focus accuracy, collimation... The list goes on.

If you assume (for those with high end equipment) that these variables are factored out of the equation as much as possible, then sure, over/undersampling becomes a significant variable. However, if these other variables are still significant, surely that's where one needs to direct one's attention at least initially.

I guess according to the Wodaski Rule, if you have a poor performing mount and/or optics, you will get poor images irrespective of the imager you use.

Beacon - I guess you need to consider these other sources of "error" before taking the final plunge. Of course, if you had unlimited cash, you can buy you're way out of it. But I suspect, like most of us, your budget is limited. Nevertheless, do not understimate these error sources. It would be terrible for a purchasr of a pretty high end imager to be (likely) disappointed with the results if putting that expensive chunk of metal and silicon on a so-so mount and OTA combo...

FWIW and my €0.02 ...

Dave

Hi Dave, I decided that I did not know enough to carry on discussing this here - but every piece of info coming in is a help to a final decision. Regardless of my own knowledge (or lack of) - I know that the discussion also helps others - who like me want to move into this. So....regardless re what I said in my last entry.....I'll make another entry here.
Your additional comments are indeed spot on regarding the OTHER factors that come in to the equation. And although I have not sat down and said I must do this and this....I guess I was aware of these things and have also been working on tackling these possible problems. Polar alignment and mount stability are things I was definitely aware of and as stated, am going down the route of changing over to a fixed pier - in time. Also, following up re getting a wedge for the scope, after reading many forum entries (mostly from U.S.) I decided that the idea of a Meade Ultra Wedge for my LX 200 was just not on....so many bad reports. I decided to bight the bullet and go for a Mitty Euro Evolution wedge. This wedge gets great reviews and cost so far has been less than buying the Meade. (I say so far, as it has not come through customs yet....but still think it will be cheaper even though the freight seemed very high.......though the package is 46lbs weight). Anyway,I think the LX200 mount on the pier/wedge with PEC should be OK. This scope has mirror lock. The focusing is another question though and I will have to see how the Meade electronic focuser works on screen images - for the eye it works fine; but I am aware that down the road a very fine focus may be required. Autoguiding may come from the CCD camera itself - if I buy the right one......and this is the point I am at....deciding what to buy.
My initial questions came about after reading this piece...
"The field of view defined by any given optical setup needs to be matched to the size of the pixels of the camera. If you think about it you’ll realize that the wider the field of view (i.e. the more sky you take in with your scope) the smaller the pixels need to be to capture detail on a specific object. A common way of describing this is the term Pixel Scale (in units of arc-seconds/pixel), which is (205 x P)/ FL, where P = pixel size in microns and FL = Focal Length in mm. Astro- photographers note that the right Pixel Scale for good photographs is about 2.5 arc-seconds per pixel" This sent me down the road of wondering whether or not the CCD that caught my eye was indeed matched to the scope. and my initial jump in on any forum to ask you guys for opinions.
Now your last input raises something else....or maybe the same thing put a different way..."average seeing". Perhaps you can explain for me (and others no doubt) what exactly you mean that average seeing in Kildare is 3 arcseconds. How do you arrive at such a figure - (see, told you I did not know very much!). Brendan.

Hi Brendan - very succint response!

True - I find that I tend to get a lot of info from threads that I didn't start or even contributed to. So this kind of discussion helps others and is well worth thrashing out.

So, in response to your points...

1. Pier - sure, a solid pier is probably the most stable but I wouldn't rule out a good solid portable pier/tripod. Personally, I have an AstroPhysics portable pier - totally rock solid but then it *is* carrying a lot of weight. So you *do* have time on your side whilst you figure out the other variables.

2. I'm not an expert on fork-mount OTA's but I have heard, like you, that good wedges are hard to come by. A 46 lb wedge sounds pretty solid!

3. *Do* spend the time on getting the persiodic error sorted out. I don't know what, if any software, meade supply to tackle that. If there isn't anything, take a look at PemPro. I don't know how much it is but its top notch.

4. Collimation will be important - Bob's Knobs are well worth the small price.

5. Focusing - any electronic focuser should be way better than by hand. Use FocusMax (free) - another excellent software package. I think the Meade's focus lock on their SCT's is supposed to be good. So mirror flop shouldn't give you any problems.

6. Autoguiding is a subject in its own right. I don't need to do it so I've little to offer you - other than an imager with an internal guide chip or an off-axis guider seemingly are preferable to a seperate piggy-backed guidescope.

7. Image scale - I really can't remeber where I read the reference. It could have been a photometry book I have or from another source. I'm sure it was a reputable source though. Bear in mind that I was clear about my objectives - science. I'll leave the "pretty pictures" to others. Though, pretty pictures is a gross understatement when talking about Anto... So, my reference was "half you average seeing". Where did I get that? By taking images and inspecting them to give me average FWHM's. Over a period of a few months in the autumn/winter I kept on eye on the numbers and found that on average, 3 arcsecs FWHM seeing was the order of the day. On the odd occasion it would dip to around two. On literally two or three occasions it dropped below two. So my image scale of 1.27 arcsec/pixel relates to an average seeing of about 2.5. Which is nicely in the middle. You could find your average seeing either by taking a few images over a few weeks, asking astronmy folks in the locality or even checking with a local observatory. You may have good seeing there in Spain - I really don't know.

8. One final point, being in Spain, do make sure you can get sufficient cooling on the imager. I personally (well, the CCD) can go 70 deg c below ambient but that's probably overkill for Ireland. Also consider humidity - we're plagued with high humidity here which can lead to dewing of optics and CCD coverslips (which can least to frost and then recharching of desicants). Just another gotcha to watch for when imaging...

I hope this gives you food for thought. It certainly took me a long long time (2 years) to make the decisions on what to buy. What I have I'm happy with but there is always room for improvement - budget permitting of course...

Dave

p.s. I'd be more than happy to come over and hlep you set up the rig

dmcdona wrote:

Anthony - I vaguely recall reading somewhere (goodness knows where) that empirically, one should aim for an image scale of about half your average seeing. In my case, I used this advice when purchasing my CCD. Average seeing here in Kildare is about 3 arcsec. Therfore, I aimed for an image scale of about 1.5"/pixel. I ended up with an image scale of 1.27"/pixel.

Hi Dave,

Research has been done on fidelity signals and one can get very good approximation once we reach 3.3x and above. In fact, I will sit down and find you a reference to this effect (really for everyone's benefit).

With 2x oversampling, there is no doubt in my mind you are getting good data. However, I am the type that likes to push things to an extreme so that I can squeeze anything and everything there is to squeeze and, hence, for my stance on aggressive oversampling.

My seeing is usually between 1.6" and 2.5" FWHM and I am delighted at my results taken with an image scale of 0.58"/pixel. I am eagerly waiting for notification for the AP305/f12.5 Mak-Cas so that I can push that down to below 0.40"/pixel.

The good thing about oversampling is that we can always downsample if it seems we have pushed things further than what was possible. Of course, the converse is not possible and, as such, we have a win-win situation by oversampling.

Let me look for the reference on the minimum requirement of 3.3x oversampling for a signal to be adequately replicated.

Anthony.