Focussing.....help!

[Marky1973]

Hi All

I posted some pictures a little while ago of Pleiades asking about focus. The stars looked bloated and, upon blowing the image up, they looked like little donuts, suggesting the focussing wasn't quite right.

I popped out last night for a go at Orion and the Jpeg of the very lightly tweaked image is attached. I like the image and think I can probably tease a bit more out of it but I noticed that the stars are bloated again and they appear as donuts when blown up.

I set my kit up and used the focussing aid on APT to very slowly adjust the dual speed focusser. I didn't have the Bhatinov mask out last night, but will try that with APT next time. I just hadn't been out for ages so wanted a go!

Using APT, my understanding is that the figures shown for FWHM and HFD should be as close to zero as possible. I selected a bright star and overlayed the "target" on APT and started focussing. I got it as low as 0.95, although the figures seem to fly about all over the shop so I am never very clear what the figures are. But I got it so the first figure was around "1" so felt I was getting close.

Fired it all up and took my 40x30sec subs, and the stars come out like this.

Am I doing something wrong, or do I need to get the Bhantinov mask out as well? Am I reading the figures in APT wrong? I'll get the manual out tonight for another read through, but I was struggling, last night to see how I could get closer with the focussing. I wondered if there was a tracking issue with this image, but the rest of it looks okay and I was only using 30sec subs, so presumed this would not be a major issue. I did polar align.... I guess I need to try drift alignment next.

Or, ultimately, with a bit more time, I need to take some test shots and then adjust the focus accordingly?

Many thanks

Mark

[Marky1973]

Oh and the image is cropped from the original so already slightly enlarged.

[Marky1973]

I was just playing around with my images from last night and found some gash ones in the file....and then I remembered. My first run was interrupted by the neighbour's cat pulling on the USB cable, so I guess this might have interfered with my focussing - although I had locked the focusser, I guess a quick yank from the cat could have shifted it enough to just nudge it out of focus!

Fingers crossed I will have better luck next time - although still happy to take any tips on getting the focus spot on!

I was worried about people wandering past and taking an interest in my kit, when I should have been worried about the local cats!

[Alien 13]

From my own experience with camera lenses and non APO refractors I find getting focus is difficult because there isnt an exact point of perfect focus so I use liveview to ensure that the dimest stars in the FOV are as tight as possible (use of ISO shutter speed to ensure the  target star is only just visible), this for me gives better results than the software tools/bat mask.

Alan

[Atreta]

I don't image using a telescope, but i guess it's practically the same way.

With my kit lens i used go to a bright star and do a zoom in liveview in the camera or zoom with apt and move the focus slider until i see a pinpoint star, then take short exposure  test shot and see how it looks like. BTW, my measures with fwhm fluctuate from 1 to almost 2.

[Marky1973]

Thanks chaps, might be a case of just getting it as close as possible and putting up with the donut stars until I can upgrade the scope then! (At least I have a good excuse now). Will try some of those tweaks next time I am out! Will just keep of trying different things until I find what works for me....and get a water pistol for the cats!

Cheers!

[Alien 13]

I am assuming that you are using the 70mm ED listed in your sig for imaging if so you should get decent results if the focus is correct. My WO does have some star bloat but using the focus method i described seems to minimise its effects too.

I would say that your original image is not too far off regards focus .

Alan

[Marky1973]

Cheers Alan - I think I just need more practice - as long as it isn't some major miss-step that I am making. I'll keep at it!

[glowingturnip]

focus is out i'm afraid, those dimmer stars are definitely doughnutty (mmm, doughnuts)

your fwhm reading in APT would have been jumping around because of the seeing - one moment in focus, the next not.  iirc, there's a setting in the fwhm focus aid to increase the number of frames APT stacks on the fly to calc the fwhm, which would then average out the seeing.  You could try increasing that which will gve you a more stable reading.  Or try with the bahtinov mask + APT bahtinov focus aid

[Marky1973]

Thanks Stuart - working my way through the APT manual....will have a look at those settings....I'm still blaming the cat, but anything else that helps is more than welcome! :-)

[SkyExplorer]

I blame our cat as well 

Before APT, I used a Bhantinov mask and live view to get the focus just right.  I just started to use APT, but still use the mask and zoom in a bit more with live view in APT.

One day I'll get around to learning the Bhantinov mask focus aid in APT.

[Carina Lass]

I've been looking at the live focus in APT with a bahtinov mask.

But it seems that my ability to get the thing to balance focus wise is enough to give me pretty tight stars.

This is a crop of m79 which is a pretty small target.

12 mins at iso800

[swag72]

A Bahtinov mask is brilliant to help with focus. As you are already using a PC, the why not download Bahtinov Grabber http://www.njnoordhoek.com/?p=660 - This takes every bit of guess work out of focusing and I bet you'll be surprised how far out you really are when you think it's right! 

When I was using manual focus I always used this combination - Got me bob on focus, without fail,  every time.

I'd also suggest that it's worth checking during the run whether your focus has changed - As the temperatures change you will often find that focus does too. Some scopes are not prone to this, but others are more so. It's worth seeing if your scope and focus is affected - A quick refocus in the middle of the run may make all of the difference.

I think that focus can make or break an image so it's worth spending time to get it right. 10 minutes to focus was often routine, but worth every minute. I guess what I'm saying is don't scrimp on focus!

[frugal]

And when you have got it into focus, lock it in place with the little knob on the underside of the focusser. Otherwise the heavy DSLR camera will slowly pull the focus out, not that it has ever happened to me...

[Marky1973]

Thanks guys!

Sara - I'll give that programme a look when I get the chance! Anything to help!

And I know I need to check the focus in the run....I need to get less lazy in 2015! ;-)

Frugal - Yeah, I have been locking - although you wouldn't believe how long it took me to figure out what the similar knob was for on my first scope when I got it last year....at least I was prepared with the refractor! :-) The mechanism on the 70ED seems pretty solid, but I'm not taking any chances - however, the added weight of a cat on the end (even with the locking nut tight) probably didn't help!

I'll have a play with all these techniques next time there is something to focus on and then post the results if they look half-decent!

[jefrs]

Doughnuts or donuts are characterisic of an out of focus reflector lens. The DoF in front of the (any) lens is shorter than the DoF behind it. One can use the visual appearance of doughnuts to aid focussing by finding the mid-point between their formation. The stars are effectively at infinity but focussing can go past infinity (diverge), this applies to extra-long camera lenses too.

Of possible interest the DoF of a 800/F8 reflex lens at 3 metres is approx 1/2-in :- not very much. Consequntly telescopes can be a pain to keep focussed.

Whilst I can and do use the camera's focus confirmation on the moon (it will even AF on with the right lens), the moon is not at infinity, near enough but not quite good enough for deep sky.

The camera won't confirm focus on stars although I did get lucky with the Orion Nebula when it did a pinpoint AF on it for me with an extra-long staibilised (OIS) lens: that indicated to me that the lens was not on its end-stop.

tpyos are feature :-)

[Marky1973]

Hi All - dropping in for an update and a few more questions if you'll all humour me a bit longer.....

Had another go with Orion a few nights ago. Pushed the exposures up to 60 seconds and spent the best part of 40 minutes playing with focus and taking test shots before I settled into an imaging routine. I also checked it half-way through to make sure it was still (as far as I could tell) in focus. I used APT, live view, focus aids and a Bhatinov mask - I've not had a chance to play with Bhatinov Grabber yet, but that is next on the list.

I have attached four shots from the start, middle and end of the run (30x60s subs). These images are, perhaps still a little out of focus but,when zooming in on the stars, they do not exhibit the same donutting as the ones above or the ones in the last picture which is a crop of the stacked images - DSS and then into GIMP.

The reason I am a little confused is the lack of donuts (to the same degree) in the original pictures as opposed to the processed one. So the question is, am I perhaps doing something wrong in DSS/GIMP when stack and then processing the image, are the donuts exacerbated by processing if not careful? Would one or two images with trailing/blacklash effects in them badly affect this if they were not discounted by DSS? Or am I just clutching at straws and just need more practice with focussing - I am happy if that is the case - well, not "happy" as such, but at least I have something to aim at (excuse the pun).

If anyone could have a look and offer any advice, I would be grateful. I appreciate this may be getting annoying as I ask the same question over and over... 

So, four images from the beginning, middle and end.....

And the final stacked/processed and cropped image

I want to get this right as the next step is to take some shorter subs for the core and try composite images....the core is a little blown!

Thanks again....save me from myself!

[Carina Lass]

Over processing, particularly raising the curves or doing the s too deep too far at once can produce a dark centre and clipped star edges.

Contrast pushed too far can do it too.

[Marky1973]

Thanks Carina. I have been stacking in DSS and then doing a very basic bit of processing within DSS by shaping the curve and aligning the RGB peaks together on the upwards turn of that curve, as per a tutorial I found. Perhaps that is something I shouldn't be doing? I have then been processing in GIMP, although I have started playing around with the evaluation version of StarTools, which looks great, but I have a long way to go with that.

Perhaps I should be stacking and then taking the unchanged FITs file out of DSS and putting it in StarTools to see how that works? I have found that it seems harder to get anything out of GIMP if I don't do a bit of processing in DSS beforehand, but I may just need to be taking more time in GIMP/StarTools to tease out the detail?

And, if you'll indulge me one moment longer, with processing, when you mention doing the s curve too much at once, am I right in thinking that processing is cumulative process. So that changing a setting by a degree of 10 in one go is less subtle than doing it in steps of 1? If that makes sense?

Sorry to be a pain! I've promised myself an HEQ5 for my birthday in August if I can figure enough of this stuff out, so I am keen to learn! I'll keep reading and watching tutorials!

[kalasinman]

Yes +1 for not doing ANY processing in DSS.

[frugal]

One of the problems with GIMP is that it only works with 8 bit images. Your 1100D will be producing 14bit images, so you are already throwing away loads of information. Try the trial version of PixInsight, or Photoshop and see if you can get more detail out. PI is £213, which sems quite a bit of money, but it it less than 10% of my scope plus mount plus camera. Adobe now do a photographer version of their creative cloud subscription that has lightroom, bridge and photoshop for £8/month which is an alternative.

[Marky1973]

One of the problems with GIMP is that it only works with 8 bit images. Your 1100D will be producing 14bit images, so you are already throwing away loads of information. Try the trial version of PixInsight, or Photoshop and see if you can get more detail out. PI is £213, which sems quite a bit of money, but it it less than 10% of my scope plus mount plus camera. Adobe now do a photographer version of their creative cloud subscription that has lightroom, bridge and photoshop for £8/month which is an alternative.

It's about 50% of my scope and mount! ;-)

I have the beta version of GIMP which I think works with 16 bit images, but I may be wrong. I am slowly playing around with StarTools which does seem to give some better results, but I don't really know what I am doing half the time! Will try the Pixinsight trial as well sometime when I have a chance to watch some tutorials. 

Sounds like I need to avoid DSS and stick to unprocessed images in StarTools. I believe I will be better off with a FITS file and telling StarTools that the image is Linear/Bayered and Non-White Stretched, if I leave it alone in DSS - my 1100d is unmodded.#

I'll have a play tonight!

Thanks everyone!

[lensman57]

It's about 50% of my scope and mount! ;-)

I have the beta version of GIMP which I think works with 16 bit images, but I may be wrong. I am slowly playing around with StarTools which does seem to give some better results, but I don't really know what I am doing half the time! Will try the Pixinsight trial as well sometime when I have a chance to watch some tutorials. 

Sounds like I need to avoid DSS and stick to unprocessed images in StarTools. I believe I will be better off with a FITS file and telling StarTools that the image is Linear/Bayered and Non-White Stretched, if I leave it alone in DSS - my 1100d is unmodded.#

I'll have a play tonight!

Thanks everyone!

You can and should save your stacked files in Fits format ( DSS can do this ) without any adjustments applied. In StarTools you then choose, Linear, DSLR and Debayered and go from here on.

A.G

[Marky1973]

Cheers! Just to check - because I'm a numpty - if the DSLR is unmodded I go with bayered? Right? Or am I getting confused....it happens! 

Will have another play.

[jefrs]

Cheers! Just to check - because I'm a numpty - if the DSLR is unmodded I go with bayered? Right? Or am I getting confused....it happens! 

Will have another play.

Yes.

Pretty much all stock condition cameras have Bayer.

A few models have unusual configuration but amount to the same.

Quite a few camera models are getting rid of their anti-aliasing filter (for Moire) to increase sensitivity.