Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

Winroddier Vs interferogram comparison?

This topic has been archived. This means that you cannot reply to this topic.
40 replies to this topic

#26 DesertRat

DesertRat

    Fly Me to the Moon

  • *****
  • Posts: 6,266
  • Joined: 18 Jun 2006

Posted 02 December 2013 - 09:47 AM

Some additional references for the curious:

"Wavefront reconstruction using iterative Fourier transforms",
Applied Optics, Vol. 30, No. 11 pg 1325 (April 10, 1991),
Francois Roddier and Claude Roddier

"Testing optical telescopes from defocused stellar images",
Proc. SPIE 2199, Advanced Technology Optical Telescopes V, 1172 (June 1, 1994),
Claude A. Roddier ; J. Elon Graves ; Malcolm J. Northcott and Francois J. Roddier

Glenn

#27 DesertRat

DesertRat

    Fly Me to the Moon

  • *****
  • Posts: 6,266
  • Joined: 18 Jun 2006

Posted 02 December 2013 - 10:35 AM

Wade,
Here's a link to the HST paper:
http://www.hubbleopt...6376C_60960.pdf

Drago,
Roddier is relatively insensitive to seeing. Seeing only needs to be halfway decent. If the image is morphing or shape changing a lot then the seeing is not good enough. Fast seeing is ok with sufficient defocus and summation of many frames or by using longer exposure times with a true astronomical camera.

I have used an artificial star for testing small refractors.

For larger reflectors testing horizontally has difficulties having to do with properly supporting the mirror. Such tests often become a diagnostic of the mirror mount. Lots of information exists on the subject, you could also ask around at the ATM, Optics and DIY forum here on CN. I've seen many IF tests of large mirrors not done correctly. In addition one must account for the added spherical error when the 'star' is not located sufficiently far away. At a good distance ground thermals add up and can cause difficulty, it is best conducted at night or early morning over grass.

Compound scopes like SCT's have a similar problem with induced spherical errors when tested too close to the 'star'. It requires separating the primary from the secondary well past its optimum distance.

One might be able to run WinRoddier and the helper apps on a Linux box using a compatibility layer such as Wine. If you have any success with that let us know.

Glenn

#28 Wade Van Arsdale

Wade Van Arsdale

    Vostok 1

  • -----
  • Posts: 131
  • Joined: 25 Oct 2007

Posted 02 December 2013 - 11:25 AM

Thanks for the link Glenn.

"Morphing and shape-changing a lot" describes all my artificial star attempts to date (6 attempts total so far). Large-scale, slow-rolling warping damaged all my WR artificial star data. As I continued testing, it became clear this was not originating inside the OTA itself, but from the ground and surroundings. I was getting Strehls down in the low .80's on scopes that later tested mid .90's in good seeing using live stars near zenith.

One thing I still want to try at some point is to do an artificial star near dawn when the ground has cooled down to it's lowest temp. Seems like that might help the image stability. Still way too much hassle to transport all my equipment to a remote site though to do what I can do just as well from home on a real star.

Another thing that would be very helpful is to compare artificial star results to a real star from the *same* instrument at close to same time on same night. If the ground thermal and grounds-space issue for proper distance as well as proper point source size can be well-controlled and overhead sky turbulence is fairly low, that would make for a very interesting comparison of methods.

I mostly do WR testing on full-Moon cycles so I'm not using up prime sky time. So far I've seen no ill effects from moonlight on WinRoddier results. This is likely due to the overall brightness and SNR of the test star, even when defocused.

But if you're testing shorter f.l. instruments like f/6-f/8 refractors that don't have the mechanical "slop" considerations that Glenn pointed out (and require far less ground distance than most SCT's), and you don't see the severe ground thermals like I had, the artificial star method should work as long as you can back off far enough from the light source.

A distance that allows for good collimation is generally not far enough distance for WinRoddier analysis of a light point source. There's some math involved to calculate the required distance, but for the sake of time and brevity, I'm leaving that alone for this thread.

Thanks,
Wade

#29 jhayes_tucson

jhayes_tucson

    Fly Me to the Moon

  • *****
  • Posts: 7,388
  • Joined: 26 Aug 2012

Posted 02 December 2013 - 11:36 AM

Thanks for all the help finding the references!

#30 mattflastro

mattflastro

    Vendor - Astrovideo Systems

  • -----
  • Posts: 622
  • Joined: 31 Jul 2009

Posted 04 December 2013 - 05:55 PM

Thanks for all the help finding the references!


Interferograms aren't the only possible reference point to compare with.
The Shack Hartmann needs a lenslet array but the simpler Hartmann mask test doesn't .
SBIG sold years ago (around 1999-2002) a little program called SBIG HARTMANN MASK SOFTWARE . It went for under $100 , sometimes it was given away for free or bundled with CCDOPS .
I still have it running on a windows XP machine.
The SBIG Hartmann software could use a mask with up to 32 holes , placed in front of the telescope aperture, like a focusing mask .
The mask could easily be home made (I used to make them out of cardboard) .
2 mask types could be used, a simple one for checking/adjusting alignment and collimation and the 32 hole one for measuring optical quality.
The Roddier test (which is not voodoo and uses the same principle used in adaptive optics curvature sensing ) is a photometric method , requiring a linear detector transfer function. Saturating stars is not a good idea.
The Hartmann method uses a geometric principle and is a lot less sensitive to image non-linearity such as saturated stars (it relies on centroid positions rather than on brightness distribution) .
There was also a French software named PRISM that had (used it in 2004-2007) very extensive Shack-Hartmann analysis capability built in .
Using several different programs (Roddier vs. Hartmann)and repeat measurements to average and compare results is a lot more accurate.
Due to the different measurement principles (photometric vs. geometric) , it's also a good sanity check. The degree of results correlation indicates also the degree of accuracy .
Other tools have to be used as prerequisites, there's no way to measure a scope with mirror flop , not thermally at equilibrium, uncollimated etc.
The SBIG Hartmann software collimation mask (and other tools like CCD Inspector) must be used first . Scope position must not be changed in between mesurements . Collimation and focusing should be performed before all measurements , no focusing in between . Mirrors shift, focuser tubes sag, CCDs get variable tilt and all results get messed up.
I'd rather pay for 3 different programs than ship a single scope back and forth for an iffy interferogram , which might (or might not) be accurate enough.

#31 William Mc

William Mc

    Mercury-Atlas

  • *****
  • topic starter
  • Posts: 2,783
  • Joined: 24 Nov 2007

Posted 04 December 2013 - 08:18 PM

Interesting stuff about those alternate methods. I hadn't known about them.

As for Winroddier, I hope to soon test all my scopes using Wades new Criteria. (Clouded over for the last week) I know Wade has tested these steps using his scopes that have known good Sterhl figures. I think we need some volunteers that have scopes with known Strehl figures closer to the other end of the spectrum. I don't think this has been tested yet? Would would be interesting to see if the results stay as consistent. Any takers?

#32 Wade Van Arsdale

Wade Van Arsdale

    Vostok 1

  • -----
  • Posts: 131
  • Joined: 25 Oct 2007

Posted 04 December 2013 - 09:26 PM

That's my hope as well, Wmacky. The more data from folks in all kinds of conditions and scenarios, the better.

Actually, I have to say the "Criteria" is all Glenn's, not mine. I'm just the trusty scribe of the project who is the "keeper of the user manuals". :)

Glenn did all the heavy lifting on this, and I contributed a bunch of Roddier data of mine to combine with his so we could put some patterns to things and make sense of it all, sanity-check all the new software etc... Without Glenn, there would be no project. So just want to make sure credit goes to where it's due on this.

Happy scope testing everyone, and Happy Holidays!

Best wishes,
Wade

#33 DesertRat

DesertRat

    Fly Me to the Moon

  • *****
  • Posts: 6,266
  • Joined: 18 Jun 2006

Posted 05 December 2013 - 12:07 PM

The Hartmann method uses a geometric principle and is a lot less sensitive to image non-linearity such as saturated stars (it relies on centroid positions rather than on brightness distribution).

Hartman is sensing wavefront slopes. It is very sensitive to the placement of the sub aperture sampling points as well as to small uncertainties in defocus. The number of points places a limit on any measure of wavefront smoothness.

It has some utility for larger scopes, where you have plenty of light. For a small scope a mask can be useful for focussing, but not in wavefront sensing and optical analysis. I have not seen much in the way of optical evaluations of smaller scopes using Hartman.

Roddier does have its limitations as well. But saturated stars are not one of them. In Roddier enough defocus is used to get thousands of sampling points and yet remain sensitive enough to detect wavefront curvature. At that level of defocus its quite difficult to overexpose, as the intensity is something like 1/10000 of that seen in focus. Defocus accuracy only requires the extra focal images to be within a few pixels or so of each other in size.

Its is important that the image be a true linear capture. That is why we emphasize not to do anything in the way of image processing except cropping. For the very best analysis, careful flat fielding techniques can be employed.

Roddier is an active project. A number of people are working to make it even more useful for the amateur community.

Glenn

#34 drago

drago

    Messenger

  • -----
  • Posts: 431
  • Joined: 11 Jan 2008

Posted 05 December 2013 - 12:32 PM

Drago,
Roddier is relatively insensitive to seeing. Seeing only needs to be halfway decent. If the image is morphing or shape changing a lot then the seeing is not good enough. Fast seeing is ok with sufficient defocus and summation of many frames or by using longer exposure times with a true astronomical camera.


yes, i partly agree on telescope position.
on other hand, take a shots with DSLR ir way more easy, imho - it not require to have computer on the field, as so as DSLR is widely accessible and not high priced item. astronomy, and even webcam is different story - it have to a purchase, and not always it be a cheap, there also must made a attachment this camera to focuser, and the computer with wires on the field...

#35 DesertRat

DesertRat

    Fly Me to the Moon

  • *****
  • Posts: 6,266
  • Joined: 18 Jun 2006

Posted 05 December 2013 - 01:57 PM

Drago,
As a planetary imager I don't think twice about carrying all the trappings to the backyard. But I can certainly appreciate the freedom a DSLR can provide from that chore.

Capturing with a DSLR can be performed for Roddier. The use of linear RAW format is required. Be careful how the RAW image is converted. The default RAW conversion to TIF of most programs use a S curve or other gamma correction that destroys linearity. The result will be an invalid analysis. During capture be careful there is no clipping of the data.

Since the DSLR pixels are so small software binning may help to enhance signal to noise and in general keep the image to a workable size. Several iterations are required in Roddier for better analysis, and the processing time can become fairly long for large images. Use care in calculating the effective pixel size.

If sufficient interest using a DSLR for Roddier exists we could document the needed techniques to avoid making any gamma corrections to the data.

If you have done this DSLR work please share with us what techniques you found as best practice.

Glenn

#36 mattflastro

mattflastro

    Vendor - Astrovideo Systems

  • -----
  • Posts: 622
  • Joined: 31 Jul 2009

Posted 05 December 2013 - 06:22 PM

The Hartmann method uses a geometric principle and is a lot less sensitive to image non-linearity such as saturated stars (it relies on centroid positions rather than on brightness distribution).

Hartman is sensing wavefront slopes. It is very sensitive to the placement of the sub aperture sampling points as well as to small uncertainties in defocus. The number of points places a limit on any measure of wavefront smoothness.

It has some utility for larger scopes, where you have plenty of light. For a small scope a mask can be useful for focussing, but not in wavefront sensing and optical analysis. I have not seen much in the way of optical evaluations of smaller scopes using Hartman.

Roddier does have its limitations as well. But saturated stars are not one of them. In Roddier enough defocus is used to get thousands of sampling points and yet remain sensitive enough to detect wavefront curvature. At that level of defocus its quite difficult to overexpose, as the intensity is something like 1/10000 of that seen in focus. Defocus accuracy only requires the extra focal images to be within a few pixels or so of each other in size.

Its is important that the image be a true linear capture. That is why we emphasize not to do anything in the way of image processing except cropping. For the very best analysis, careful flat fielding techniques can be employed.

Roddier is an active project. A number of people are working to make it even more useful for the amateur community.

Glenn

Glenn, who is writing the actual code for the Roddier algorithms?

#37 DesertRat

DesertRat

    Fly Me to the Moon

  • *****
  • Posts: 6,266
  • Joined: 18 Jun 2006

Posted 06 December 2013 - 04:22 PM

Hi Matt,
The development team is outlined at the project page along with contact information:

http://www.astrosurf...ier/projet.html

For an understanding of the algorithm I recommend the text:
Optical Shop Testing 3rd edition by Daniel Malacara

Beginning in section 11.5, wavefront retrieval and curvature sensing is described.

The flow chart of the iterative Roddier process (figure 11.30 pg 431 of Malacara) describes the algorithm quite well.

That along with the references including the historical work begun by Jacques Beckers of the European Southern Observatory provide good background information:

"Interpretation of Out-of-Focus Star Images in Terms of Wave-Front Curvature"; Beckers J. M; J. Opt. Soc. Am. A., 11, 425-427 (1994)

Beckers did a lot of the early work evaluating the Multiple Mirror Telescope in Arizona.

Glenn

#38 mattflastro

mattflastro

    Vendor - Astrovideo Systems

  • -----
  • Posts: 622
  • Joined: 31 Jul 2009

Posted 07 December 2013 - 12:44 AM

Hi Matt,
The development team is outlined at the project page along with contact information:

http://www.astrosurf...ier/projet.html

For an understanding of the algorithm I recommend the text:
Optical Shop Testing 3rd edition by Daniel Malacara

Beginning in section 11.5, wavefront retrieval and curvature sensing is described.

The flow chart of the iterative Roddier process (figure 11.30 pg 431 of Malacara) describes the algorithm quite well.

That along with the references including the historical work begun by Jacques Beckers of the European Southern Observatory provide good background information:

"Interpretation of Out-of-Focus Star Images in Terms of Wave-Front Curvature"; Beckers J. M; J. Opt. Soc. Am. A., 11, 425-427 (1994)

Beckers did a lot of the early work evaluating the Multiple Mirror Telescope in Arizona.

Glenn

Glenn, thank you.
Matt

#39 Wade Van Arsdale

Wade Van Arsdale

    Vostok 1

  • -----
  • Posts: 131
  • Joined: 25 Oct 2007

Posted 11 December 2013 - 11:17 AM

A new revision (1.4) of the latest WinRoddier Ver. 3 User Manual has been posted on the Yahoo Roddier Group and at my alternate download site:
http://www.compubuil...User_Manual.pdf

Glenn and I have expanded a little bit on the image capture instructions and some of the WR theory behind the setup instructions to make it more clear and help the end-user optimize their setups for best results in WR3.

Note: I just noticed the Yahoo Files Section of the Roddier Group is not functioning normally. So use the above link to get the latest manual revision for now. The latest revision number is Ver. 1.4 of the manual, dated 12/11/2013.

Thanks,
Wade Van Arsdale

#40 DesertRat

DesertRat

    Fly Me to the Moon

  • *****
  • Posts: 6,266
  • Joined: 18 Jun 2006

Posted 11 December 2013 - 12:48 PM

Hi Wade,

Good work!

Yeah I noticed the discrepancy. You might have to delete the old v1.3 and reload 1.4. Don't know for sure if that will work but its worth a try.

Yahoo mail is not working right either at the moment - so thats why I replied/posted here!

Moving the group to google is an option. But hopefully yahoo will get things sorted out eventually, and the devel team still refers to the yahoo group.

Glenn

#41 Wade Van Arsdale

Wade Van Arsdale

    Vostok 1

  • -----
  • Posts: 131
  • Joined: 25 Oct 2007

Posted 12 December 2013 - 03:19 PM

Glenn, the file name-change workaround did the trick. I slightly changed the user manual name and got the latest revision to upload successfully. So folks who prefer to get the downloads directly from the Yahoo Roddier Group Files section can get this latest revision from there now.

(My own download link is still available for anyone getting the user manual from there...no plans to take that down anytime in the forseeable future):
http://www.compubuil...User_Manual.pdf


Thanks,
Wade


CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Recent Topics






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics