Jump to content


Photo

Double Star Imaging Eternathread

  • Please log in to reply
45 replies to this topic

#26 Cotts

Cotts

    Just Wondering

  • *****
  • Posts: 4945
  • Joined: 10 Oct 2005
  • Loc: Toronto, Ontario

Posted 15 July 2014 - 10:29 PM

Perseus, my $0.02...

You are using a 6-inch I see. With about 1000mm of focal length (you may need a barlow or powermate...) you can get nice double star images of pairs down to around 2 arc sec. Below is an example from my 5 1/2 inch refractor, focal length 980mm of Gamma Delphini which is 8.9" separation.

I know very little about your camera but you should, at least, try it on some brighter, wider pairs (i.e. > than 6th mag and > 6" or so...) before spending some hard-earned cash. If your camera takes movies you can do lucky imaging with it. A planetary camera will be optimized for high definition work, I think, which makes it a possibility for double star imaging..

Dave

Attached Files



#27 gregj888

gregj888

    Viking 1

  • -----
  • Posts: 609
  • Joined: 26 Mar 2006
  • Loc: Oregon

Posted 15 July 2014 - 11:14 PM

Perseus, Cotts,

At the risk of opening a can of worms, what Cotts says is a very good suggestion.

If you want to get sub-seeing measurements, then things change, and that's my area of interest. All the cameras used for planetary work can do double stars. Limiting mag will depend on the camera, seeing and the f/#. Diameter of the objective controls minimum separation you can measure.

If you are looking for a $500-ish camera for sub-seeing double star work, my earlier post is my personnel best suggestion.

For stars wider that the seeing, virtually anything will work including a DSLR.

What is thought to be STF1536, no filter f/50 0.8m, Flea3 in really bad seeing. Stacked, best 25% of 1000, no filter. Seeing also effects delta mag.

Greg

Attached Files



#28 Perseus_m45

Perseus_m45

    Messenger

  • *****
  • Posts: 465
  • Joined: 16 Jul 2011
  • Loc: pittsburgh area

Posted 16 July 2014 - 08:15 AM

I have tried the Orion StarShoot Solar System Color Imaging Camera IV on stars and its a no go for it.But as you see with the moon picture it does take a nice clear image. It can also pic up Jupiter. But Mars blinks in and out its right at the limit for the cam I guess .
Has anyone tried the Mallin cam products?
mike h

#29 Ed Wiley

Ed Wiley

    Apollo

  • -----
  • Posts: 1045
  • Joined: 18 May 2005
  • Loc: Texas, USA

Posted 16 July 2014 - 02:20 PM

Mallincams have, I think, large pixels: that is what makes them work for live video feed. I might be wrong. You want your pair with pixels in between, not falling on the same pixel. Besides, you can pick up a quite reasonable used starter DMK21 for cheap.

For close pairs (1"=5', 8"-11" aperture) I use a off the self Image source DMK21 with the 618 chip. The Skyris is made by image source and has the advantage of USB3. Don't know about the chip in that camera. might be better or the same as the DMK21. You want small pixel size and as sensitive a chip as possible. You are taking video at settings like 10-100 milliseconds and 60 frames/second.

For wide pairs (wide being defined by your imaging system)any DSLR or CCD should work. You plate solve the image and have the software determine angle and separation. So long as you do not saturate and have decent separation the measure should be reasonable. You have to experiment.

Color is not as good as Mono, but will work. Try your present camera on some pairs that are wide (plenty of those!)and see how it goes(closer in magnitude helps). In computers its "hello world." In doubles its Mizar.

Ed

#30 ssmith

ssmith

    Vostok 1

  • *****
  • Posts: 134
  • Joined: 28 Jun 2012
  • Loc: Colorado

Posted 16 July 2014 - 04:15 PM

Here's a shot of 32 Eridani.

SW100ED F9 & Olympus E-PL1 @ prime focus
ISO 200 1/2 sec exposure

Processing: re-sampled from 314 pix/in to 628 pix/in
Levels -> color curves -> sharpening

Attached Files



#31 Perseus_m45

Perseus_m45

    Messenger

  • *****
  • Posts: 465
  • Joined: 16 Jul 2011
  • Loc: pittsburgh area

Posted 16 July 2014 - 06:57 PM

Steve buddy you got your act together, nice work .
mike h

#32 Ed Wiley

Ed Wiley

    Apollo

  • -----
  • Posts: 1045
  • Joined: 18 May 2005
  • Loc: Texas, USA

Posted 16 July 2014 - 08:04 PM

Steve:

Nice work, lovely color difference!

Ed

#33 Rich (RLTYS)

Rich (RLTYS)

    Fly Me to the Moon

  • *****
  • Moderators
  • Posts: 5300
  • Joined: 18 Dec 2004
  • Loc: New York (Long Island)

Posted 17 July 2014 - 05:38 AM

That is nice!

Rich (RLTYS)

#34 drollere

drollere

    Surveyor 1

  • *****
  • Posts: 1588
  • Joined: 02 Feb 2010
  • Loc: sebastopol, california

Posted 17 July 2014 - 03:38 PM

cotts, this is a great thread idea ... i've enjoyed the posts.

two comments as a casual reader: it would be useful to include the specifics of system magnitude and photographic settings. some images have been made with video stacking, but without mention of the video camera used!

perhaps you could recommend the basic info that posters here should include with their images.

it would also be interesting to read about the full process. not for every post, but some imagery requires trial and error, or adjustments from first settings, and this is valuable information about where the important decisions have to be made in order to get a good image.

keep 'em coming!

#35 R Botero

R Botero

    Apollo

  • -----
  • Posts: 1291
  • Joined: 02 Jan 2009
  • Loc: Kent, England

Posted 18 July 2014 - 04:09 AM

Florent Losse has provided me with a copy of REDUC and I measured my images of Eta CrB with it. I have not calibrated my setup in REDUC so the values I'm obtaining are not representative but the program managed to reduce and identify the two components in my image after using the procedure outlined in its tutorial for close pairs. The algorithm used is a surface algorithm developed by Guy Morlet and Pierre Bacchus for the French Astronomical Society that creates a mathematical model of the brightness distribution of the components. The model resulting from my reduction is the below. I will report back once I calibrate my system and measure this pair again.
Roberto

Attached Files



#36 Ed Wiley

Ed Wiley

    Apollo

  • -----
  • Posts: 1045
  • Joined: 18 May 2005
  • Loc: Texas, USA

Posted 18 July 2014 - 09:41 AM

Hi Roberto:

You are on your way! For those who may not be familiar with the surface algorithm -- using standard methods one will conclude that a close pair is closer than it actually is due to overlap of their point spread functions. "Surface" compensates for this. Bob Buchheim discusses this in the Argyle book along with many other topics, including polar alignment and calibration.

One note: The closer you get the less accurate the angle measure. I am finishing off some 2013 measures. My C11 Rayleigh limit is ca 0.5". Using interferometric techniques on a perfect night with a equal pair I should be able to resolve to the limit. Two of my pairs are ca. 0.75" and resolved well. However one pair's angle was + 1.7" over and the other pair about 1.7" under the calculated positions. So F30 was not the right F-ratio to image these pairs for a good measure, I should have gone of F50. This is a fairly well-known phenomenon, see Napier-Munn and Jenkinson in the latest JDSO.

Good measures!

Ed

#37 Ed Wiley

Ed Wiley

    Apollo

  • -----
  • Posts: 1045
  • Joined: 18 May 2005
  • Loc: Texas, USA

Posted 18 July 2014 - 09:57 AM

A note on filters

Greg mentioned filters. Yes, professionals use filters, very narrow band filters. But we must remember that they define a small telescope as something in the 1 meter range and they use $13K cooled Andor (or similar) high-end cameras. My experience with "tiny" telescopes and amateur webcams is that non-photometric red and IR filters may be of great use for doubles that are low in the sky. (Florent has some very informative comments.) But the price is picking only brighter pairs ("brighter" being defined empirically by experience). So my recommendation is that if you wish to image and measure doubles with something in the 0.25-meter or less class scope, start with no filter and see how it goes.

Ed

#38 Ed Wiley

Ed Wiley

    Apollo

  • -----
  • Posts: 1045
  • Joined: 18 May 2005
  • Loc: Texas, USA

Posted 18 July 2014 - 11:16 AM

Bruce was interested in the process. So... here is goes. This is for science stuff: astro-art of doubles is a completely different skill set and one I am not good at. Also note that this is the way I make observations and not meant to suggest it is the only way or even best or most efficient way.

1. Two important steps. First, complete collimation. I use MetaGuide. Second, polar alignment. If you are not exactly polar aligned it will throw your camera angle off as you move from one region of the sky to the other. Tip: try to re-calibrate when you move the scope to a new region.

2. Stuck at the end of the Crayford focuser is a flip mirror with eyepieces fitted with rings to be parfocal to the video camera and the video camera in line with a 2x-5x barlow. (F10? -- minimum is 3X; F20? you can use 2x) Focus of the video camera via a mask on a close bright star at regular intervals if possible.

3. Eyepiece/flip mirror find the double. I do go-to: this is science not pleasure: get to the target by the faster means possible. Experience shows where the double is relative to the camera (don't think its always at the center shown by the cross hairs!).

4. Flip the mirror, you should see the double integrating at, say, 250ms. If not increase the integration until you acquire. Prepare to be frustrated -- acquiring doubles at F30-F50 on a 1/4" chip is a acquired skill.

5. Begin decreasing the integration time. Do so until you can barely see both the primary and the secondary. As integration time decreases the stars will begin "boiling" or "dancing." We are sampling within the point spread function and seeing "super speckles." If there is a magnitude difference, concentrate on the secondary. This can be anywhere from 1ms to 1000ms depending on the magnitude. Wide pairs do not require fancy techniques, the closer the pair, the faster the integration required. Experimentation is necessary. The main worry is saturation of the primary. When beginning take video stream of different integration times and keep a record of saturation versus integration time. Remember -- this is science not art: as you decrease integration time the pictures start to look really ugly, some of mine are just scatters of pixel hits and others look like boiling masses.

6. Take 1000-4000 frames for close pairs, as few as 100 for wide pairs. If you can take them as fits (Firecapture), so much the better, If not, then you will have to post-process to bmp. (I am assuming a PC, ask Dave about Macs).

7. For an actual session the first pair imaged is a calibration pair. You want something wide, not close and in the region you are working. Rectilinear pairs or wider Grade 1 or grade 2 orbital pairs are good, many of the WDS calibration pairs are more suited for larger scopes. I try for 10-20" separation and 100-200 frames.

8 Finally, program stars. I usually work one region at a time and if I make large move I take another calibration pair. In fact, I take calibration pairs at regular intervals.

9. At the end of the run, take another calibration pair to insure no camera movement.

10. Post-processing and measuring is well covered in the REDUC tutorials by Florent Losse if you use that program.

Now, much depends on those calibration stars as they are used to determine camera angle and plate scale. Best practice is to take several. Check to see how well each agrees with the calculated theta and rho for that night and take the average. Odd balls do pop up, so do not go blindly into the night. The best way for the beginner to begin is to measure only brighter and well separated (>5") rectilinear pairs. Why? Because you can compare your results with the predicted results of professionals. If your measures fall within the errors of an O-C, then you gain confidence that your measures have scientific worth. You can calculate the predicted theta and rho simply, it ain't rocket science.

Cheers, Ed

#39 drollere

drollere

    Surveyor 1

  • *****
  • Posts: 1588
  • Joined: 02 Feb 2010
  • Loc: sebastopol, california

Posted 18 July 2014 - 02:47 PM

ed, thanks for the collegial information. i would be pleased to learn that there's a book in the works. (the recent SAS proceedings also have some great imaging papers.)

both the "wildlife" imaging that cotts does, which i have to say i envy with an admiration, and the science work that you do, which i have to say i envy with a great respect, are both cutting edge in my view.

there's an imaging conference in vienna next year ... just saying ...

#40 Perseus_m45

Perseus_m45

    Messenger

  • *****
  • Posts: 465
  • Joined: 16 Jul 2011
  • Loc: pittsburgh area

Posted 18 July 2014 - 08:48 PM

guys do you prefer reduc over registax6 ?
mike h

#41 Ed Wiley

Ed Wiley

    Apollo

  • -----
  • Posts: 1045
  • Joined: 18 May 2005
  • Loc: Texas, USA

Posted 18 July 2014 - 11:14 PM

Hi Mike:

REDUC is dedicated to doubles and powerful. It is designed for measuring. I have never used Registax6 for doubles but it certainly is good for other things it is designed to do well, like stacking and processing planetary and lunar images.

Since REDUC is free on request, it is what I recommend.

Ed

#42 Ed Wiley

Ed Wiley

    Apollo

  • -----
  • Posts: 1045
  • Joined: 18 May 2005
  • Loc: Texas, USA

Posted 19 July 2014 - 10:06 AM

Hi Mike:

I tried to post this one last night, If it shows up sorry for the double post.

REDUC is dedicated to double star data reduction while Registrax is a more general imaging processing program. I have no idea if you can do things like measure angle and separation in Registax, such operations are simple to do in REDUC. If you are measuring wide pairs and can plate solve there are other alternatives: just about any program that plate solves can measure doubles (e.g. MPO Canopus, CCDSoft/Sky, Sky, Maxim, etc, etc). But REDUC has operations such as autocorrelation to "reduc" the close ones.

Ed

#43 Perseus_m45

Perseus_m45

    Messenger

  • *****
  • Posts: 465
  • Joined: 16 Jul 2011
  • Loc: pittsburgh area

Posted 19 July 2014 - 01:38 PM

Ed, Now let me ask you from when you gather your data from the scope, and you do your Reduc processing to come out with a complete measurement of a double ..How much time do you have invested per star?
mike h

#44 azure1961p

azure1961p

    Voyager 1

  • *****
  • Posts: 10462
  • Joined: 17 Jan 2009
  • Loc: USA

Posted 19 July 2014 - 06:10 PM

Probably about the same time it takes to process both stars together.

Pete

#45 Ed Wiley

Ed Wiley

    Apollo

  • -----
  • Posts: 1045
  • Joined: 18 May 2005
  • Loc: Texas, USA

Posted 20 July 2014 - 12:08 AM

Ed, Now let me ask you from when you gather your data from the scope, and you do your Reduc processing to come out with a complete measurement of a double ..How much time do you have invested per star?
mike h


That depends on the pair, and I have never done a time analysis of the work flow. Off hand--It can take several minutes if you are processing 4000 images, 1000 a pop (N=4) because they have to be aligned. And that assumed you have fits or bitmap files ready to load. On the other hand, if you have 100 images of a well separated pair its seconds. If you have good ccd imges (say four), its lightening fact as it will measure each and give you the angle, separation and error in 1-2 seconds. Of course, you can log the data to a file (three files, in fact).

If you are taking video, then it takes some time to do the conversion to bmp. That is a PITA and suck up a lot of time.

Ed

#46 Ed Wiley

Ed Wiley

    Apollo

  • -----
  • Posts: 1045
  • Joined: 18 May 2005
  • Loc: Texas, USA

Posted 20 July 2014 - 12:11 AM

BTW: Do not take flats with REDUC. Darks are OK, but flats see to not work. Also, the delta magnitude estimate may work for ccds, but it certainly does not work for webcams, Forget about it.

Ed






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics