What are facts?

[ollypenrice]

This arises from a discussion on the imaging board. Yves and I had our first taste of long focal length, high payload imaging this week and met with some good news and some bad news.

The bad news was the guide trace. It looked wild. The good news was the perfect roundness of the stars from a 2.4 metre focal length on a night of bad seeing and high wind. In thirty minute subs, that is.

I suddenly lost all interest in the guide trace...

For me the facts are the round stars. Sure, a truly random guide error could deliver round stars while still smearing the resolution but even the very small stars are round. I think the guiding is terriffic.

Part of the problem is my lack of familiarity with PHD but, although I don't like it because I don't seem to be able to disable Dec guiding (or am I wrong?) it is delivering round stars. It stays!

Also from the earlier thread there's the issue of 'Rubbish in, rubbish out.' I think this is essentially correct. I'll use rubbish in selected parts of an image but they are small parts. (The corners of my TEC images need Ps bodging if there are stars there because I don't have the glorously expensive flattener and I'll use soft data to for noise reduction in the faintest parts of an image in chasing marginal nebulosity. I first heard the 'rubbish in, rubbish out' expression from Tom O'Donoghue. Say no more.

So what floats your boat, the statistics or the pictures? Both are facts.

Olly

[swag72]

Interesting one this Olly.

I am a bit of a slave to the guide gaph, although I am trying my hardest not to be. I generally get a pretty flat trace, but if I don't then I recalibrate and see how it is next time round. If it's jumping all over the place, then I'm not a happy bunny.

Do I go ahead and image anyway and see how the guide trace has in fact affected the data? Sadly I don't, but I know that I should!!

I have posted a similar type of comment in a couple of threads where people have posted a rather hilly looking graph, coupled with nice round stars. I can't help thinking that I have become a slave to the damn guiding graphs that are there to make life a little easier! There is loads of time spent on tinkering etc to get the flattest graph possible, yet there are examples out there that suggest we could be better spending that time on catching data.

I think I will make that my 2012 resolution - Look at the image first, take it no matter what Maxim may be telling me and then make an informed decision from there!!

Great post Olly.

[NickK]

I think I've experienced something close to this where the stars are rounder than the trace would show. I'm using an OAG at considerably lower fl and weight. However the parameters we normally use are small focal length and not the 2+ m focal length you're using.

I think the only way to do that is provide a pixel level view of a single star that is round an look at the distribution of pixel values. You may perceive it to be round when it's not.

Pick a set of stars, crop, take the FITS and use PI to show the 2D plot of values (use a 1:1 rather than the default 6:1 pixel mapping). For good data that will appear as a sharp point but you may find that the data shows a different story.

However - that's the difference between a graph (science) and perceived roundness (art).

[ollypenrice]

Sarah, this is the first time in my imaging life that I'm NOT a slave to the guide trace!! (I quite like it, it's very liberating - maybe like burning your bra - something I have never done...)

Nick, yes, art and science. If it looks round then, to an aesthetic imager like me, it is round. However, some resolution might be slightly compromized eleswhere. Think of Greek architecture: why does the Parthenon look so incredibly symmetrical and rectilinear? Becasue it is asymmetrical and curved. But is that art or science? The illusion is based on a precise understanding of geometry and mathematics. AAARRRGGGG!!!!!

Back to the point, I had no idea that many people have, it seems, gone through this strange long FL/OAG/Bad Guide Trace experience and come to the 'just get on with it' conclusion.

Olly

[RobH]

Having a little experience of long FL imaging, a guide graph in PHD that looks somewhat Himalayan but with round stars isn't unusual, especially when using an OAG. Remember, what you're seeing is the maximum correction, not the actual movement. I have my RA aggressiveness dialed right down to something like 20% when using long FL's.

Also, remember that at long FL's, a lot of the star movement will be down to aatmospheric conditions, so be sure you're not trying to chase the seeing.

I also have my polar axis fractionally misaligned so that all of my DEC corrections are either N or S, depending on which side of the meridian the scope is.

This eliminates any backlash issues (you don't get these with the Mesu of course!) and stops the see sawing of the DEC axis as it tries to take up backlash and then overcompensates.

To turn DEC guiding off in PHD, go to the 'brain' icon...this allows control of all the guiding parameters including DEC.

If you have a recent version of PHD, you can bring up the graph from the 'tools' menu and adjust parameters on the fly from there.

Cheers

Rob

[RobH]

BTW....I'm firmly in the 'if it looks round, then it is round' camp

[AndyUK]

The guide graph isn't an indication of how round the stars are going to be? Oh no.... !!!! I've sometimes spent 30 mins fiddling around with getting the PHD settings right to produce a flat guide graph!

Not that I know much about this at all, but I personally couldn't care about how the graph looks if the result is small round stars (art before science), but my own paranoia needs something to convince me that guiding is as it should be as it's in capture mode - I know it's stupid, but after calibration I find myself staring at those two lines willing them to stay flat for possibly a couple of minutes, and if a line does move, I'm horrified if it moves the same way again or starts to "drift". I'm not sure how I'd feel if the guiding graph was flipping around all over the screen, even if the image came in with the tiniest/sharpest stars...

The strange thing is that when I was using a Synguider, I simply trusted it as there were very limited settings to play with (once calibrated, just halve the calculated aggression) and I happily left it to get on with it (no stress). Now that I have more settings and a graph to look at, it seems as though I MUST tinker with them to get the lines anchored to that central line...

[swag72]

Right ...... See you folks later, I'm off to burn my bra!!!!!

[Deneb]

@ Andy, on a good stable night....

1000mm Focal length :evil6:

[vdb]

Indeed when you click on the brain there are a lot of settings, turning of dec is one.

I just read Min motion should be a lot higher at long focal length, we experimented with 0.15, maybe it should be even 0.4 ... in our case that still is good enough as we use pixel binning on the main camera and that pixel is 2x the size ...

But as mentioned again lets look at a set of lights taken under good conditions.

Yves.

[AndyUK]

@ Nadeem - Not bad... I had this with the finderguider once when using the MN190 :

(This doesn't happen as often as I would like, but I believe it's easier getting flatter graphs like this with finderguiders than it is with "proper" guidescopes though?)

[Psychobilly]

Whats the vertical scale on the graphs form PHDGuide... if it's pixels then wouldn't a shorter FL guide scope produce a flatter graph as each pixel covers a larger angular distance...?

I think ultimately I would still go by the "quality" of the stars in image rather than the guiding graph...

Peter...

[RobH]

Whats the vertical scale on the graphs form PHDGuide... if it's pixels then wouldn't a shorter FL guide scope produce a flatter graph as each pixel covers a larger angular distance...?

I think ultimately I would still go by the "quality" of the stars in image rather than the guiding graph...

Peter...

You are correct Peter.....guiding graphs from a short FL scope are nothing to go by at all. In fact, if you guide at too short a FL, you can have a lovely graph but stars that aren't round.

Rob

[AndyUK]

Cheers Rob/Peter - That's what I thought, but thanks for confirming. Unfortunately I couldn't get on top of the flexure I was getting with a "proper" piggybacked guidescope (). Perhaps that's why I'm now so paranoid about graph tweaking/watching...

[morimarty]

I can remember my first outing with a c11 and a oag using phd. I was going for 300 second subs @fl 2800. when i looked at the graph i was horrified, like rob says( himalayan) it was a bit of a culture shock having regularly flat lined an ed 80. but when the first image was downloaded to my astonishment i had nice round stars. I do think at these focal lengths the seeing as a lot to answer for. so i too am definitely in the quality of stars camp.

ROB, thats interesting that you slightly offset your PA. this could be good news for portable setups?

ATB martyn

[vdb]

Offset of PA van be of use but is dangerous you need to calculate your error,

here a nice site with calculator webpages:

On Polar Axis Alignment

Click on the appropriate equation really nice!

Yves.

[RobH]

The offset only needs to be tiny, and I've never run into any field rotation issues. I just drift align at as long a FL as possible, get the PA absolutely right and then tweak the RA axis just a fraction.

Rob.

[Catanonia]

I totally ignore the PHD guiding trail. As long as the star stays in focus in PHD and it is guiding, the mount and software are doing thier job fine. Unless you get a massive blip on PHD, don't worry about it.

Assuming you are polar aligned to fairly well.

[Deneb]

Point taken chaps, sometimes the graph looks really good, but most of the time the stars look good & the graph is all over the place...

[Darth Takahashi]

Oh dear oh dear,

I see that I'm going to be the only one on this side of the discussion!!!

FACTS are FACTS but in this case its not 1 or 0 (True or False).

I'm not interested in opinion or what appears to be round that isn't round etc... "Point of fact; lots of stars are not spherical - round"

As stated in the previous post I'd like to see the statistics!!! the guide log; the mean and standard deviation etc... For those that don't already know AVERAGES lie!!!

Everything in our images come right back down to statistics whether you like it or not; the size and roundness of the stars = diffraction, seeing = the "PSF" Point Spread Function, and even the noise S/N ratio = Poission function!

*** I missed FWHM function for focusing ***

Nothing you do in this hobby can be done without statistics whether you like it or not, sorry for that.

And unfortunately here is another shocker; pretty images tell you anything about the performance of a telescope or mount. In fact, extended images; anything other than a true point source effectively smear the truth and deep down Olly already know that fact and has alluded to it in his post.

Rubbish in Rubbish out is a computer adage! don't expect a computer + software to magically make things better, right? It can be applied to imaging since we all want to start with good data but sometimes have to dress up bad data!!!

Please let’s get this over and done with; post the guiding data with all the provisos that you have mentioned and then we can draw our own conclusions for better or for worse, right???

The only thing I asked for is some independent information from the manufacturers claim and you are in the best position to do that, that all?

[JohnC64]

It might be my simplistic understanding of PHD but I've always understood that as long as the square in the middle stays green, then PHD is guiding within parameters. So if the graph is "hilly", phd just needed to do more work to keep the mount/scope pointing at the same point but it managed it. If the square goes orange then PHD couldn't move the mount quick enough or far enough so you will get star trails. I have had guiding working in the past and its never been a perfect graph but i do get reasonable tracking.

I would go with the end results and if a sub looks good then it is good enough to be stacked and used. :-)

Cheers John

[Darth Takahashi]

Whats the vertical scale on the graphs form PHDGuide... if it's pixels then wouldn't a shorter FL guide scope produce a flatter graph as each pixel covers a larger angular distance...?

I think ultimately I would still go by the "quality" of the stars in image rather than the guiding graph...

Peter...

Exactly right! which is why the guiding log is important.

So when you 0.51 arcseconds / pixel if make a huge difference than 5 arcseconds / pixel.

Without measuring you cant control, if you cant control you cant improve!

[Darth Takahashi]

Here another fact;

Don't achieve perfect focus and guess what your stars will be ROUND!

Its a pointless discussion to keep measuring the guiding performance of a mount by one metric, the roundness of the stars it produced, think about it!!!

I think the real issue is that too many people are scared of probabilities and statistic.

[vdb]

So you want a graph, here is one ...

Telescope Reviews: MESU- mount #2

Yves.

[Demonperformer]

Sorry to ask a total novice's question (and possibly go a bit off-opic), and the graph you link to looks quite good, never reaching 2", but how would that compare with something like the EQ6? Does anyone have a similar graph that those of us who have no idea can compare ?

Thanks.