I have just received a bunch of email bounce notices which seen to come from German speaking accounts.
I can assure you that I had nothing to do with these emails and you should treat them as fraudulent.
Ich habe gerade eine Reihe von E-Mail-Bounce-Hinweise, die aus dem deutschsprachigen Raum Konten zu kommen gesehen.
Ich kann Ihnen versichern, dass ich nichts mit dieser E-Mails zu tun hatte, und Sie sollten sie als betrügerisch zu behandeln.
I was very lucky during the transit because an early morning haar burned off just before the transit and we had mostly clear blue skies above Inverness, Scotland for the rest of the day. This let me capture most of the transit until the Sun sunk below my roofline. Here is the animation I produced from the results (some jerkiness is caused by occasional clouds).
Equipment and Software Used
Technique
I had intended to use my PST at prime focus to capture the transit but I changed my mind just after the start and zoomed in with my barlow. Mercury made a perfect autoguide target for DSSR and I only had to guide manually when a telephone cable went in front of the Sun. DSSR automatically recorded 669 consecutive videos, each 30 seconds long which were batch stacked and processed in AviStack.
I used DVS to stabilise and animate the AviStack images and add the clock, text and scale Earth. The scale Earth threw me for a while because Mercury is only 38% the size of the Earth and it looked far too big when I added my standard scale Earth. It took me a while to realise that Mercury and the Sun are at different distances from Inverness so I added 2 scale Earths.
I love pushing the limits of low cost astro equipment so I am delighted that my £70 Pipo X7 PC was able to act as my observatory workhorse.
I have always admired the fantastic Milky Way mosaics that you see on the web and here is how to try it for yourself.
Click for HD version. Original mosaic was 11,440 x 2,800 pixels.
Field Work
I shot this in Scotland in October using only a fixed tripod and a Canon 1100D with the stock 18-55mm lens. At this time of year, the Milky Way passes close to the zenith which makes this method easy to use. You could adapt it for other dates by tilting your tripod head so that it follows the Milky Way as you tilt the camera.
First, get to a dark site with clear views of the Milky Way from horizon to horizon. Set your tripod level and take some test shots to test focus and exposure. I set the lens to 18mm, wide open aperture, ISO 6400 and exposure to manual at 30sec.
Now aim at where the Milky Way crosses one horizon and take your first shot. Tilt the camera up about 15 degrees and take another shot. Repeat until the camera is pointing straight up which should give you a series of 7 shots or thereabouts. (The angle is not too important – you just need a good overlap between shots. Now turn your camera round to the other horizon and repeat for a second series.
Shooting all the images took me about 10 minutes.
Create the Mosaic
I used the amazing and free Microsoft Image Composite Editor (ICE) to stitch the mosaic. Start by loading the first series of images (I used the raw .cr2 files). Then click Stitch and select the Transverse Mercator projection when it has finished stitching. Click Crop and then No crop and then click Export and select the .PNG image file option. Click Export to disk and select where to save your first half of the Milky Way. Repeat for the second series of images which will give you 2 mosaic halves like these.
You can now use ICE to join these 2 halves to make a complete mosaic. Load them in ICE and Clicking Stitch will give you a result like this.
We need to rotate this a little to bring the Milky Way vertical so insert a value of 10 for Roll to give this.
You can now click the Crop button and adjust the crop window to suit like this.
You can now export this image and edit it in your favorite image editor.
You can also load both sets of original images in and let ICE stitch them all together. Below is how this looks using the Fisheye projection and colors tweaked in Gimp.
Not bad for 10 minutes of field work?
I was asked to help out some users align their image sequences of the recent eclipse of the Moon and this is the result using DVS.
[vimeo 141665784]The following is my idea of the best way to proceed.
1. Use DSSR to Guide (optional)
DSSR can guide on full disk Moons and this has the advantage of minimising the frame to frame drift which speeds up DVS alignment. To do this, zoom out the video screen until the Moon is small enough that you can select the whole disk as a guide target. Reducing the size of the guide target reduces DSSR’s cpu hit and makes guiding easier.
2. Set Similarity to 10
Similarity is used to eliminate false matches in cases like sunspot groups where multiple possible matches occur. For eclipses, there will only be one Moon per frame so the Similarity filter is redundant. Setting it at 10 turns it off.
3. Set Size Parameter to Exactly Match Moon Diameter
You need to make the anchor box the same size as the lunar disk. First set the Drift parameter to 1 to minimise the time taken to display the anchor box after you click. Then click on the centre of the disk and adjust the Size parameter until the anchor box is the same size as the disk.
You can then adjust the Drift setting to suit your image set wobbliness or turn on the search full screen option.
4. Optimise Anchor Swap Setting
This is the KEY parameter for eclipses. It determines how often the current alignment target is swapped out for the current best match. The quickest way in the long term is to make a second set of images at a reduced scale. DVS processing time goes up roughly as the 4th power of image size. Using half size images will align 16 times faster and quarter size images will align 256 times faster.
This speed increase allows you to quickly run a series of alignments with different swap settings and see which best aligns the moon. I tried settings of 1, 5, 10, 20 and then with the parameter unchecked. Setting 5 gave the best results.
5. Produce Intermediate Image Set
The great thing about step 4 is that you now have a good rough alignment of your animation which you can apply to your original image set. Load the original set and then click the Data button on the alignment module. Select the best data file from step 4 and enter 2 for half sized images, 4 for quarter sized images, etc.
Now export your frames as an image sequence.
6. Final Alignment
Load the image set from step 5. You can use a smaller drift value because the Moon should be aligned to within a few pixels. This greatly speeds up processing time and you can then do any manual alignment tweaks, crop, add overlays and export as normal.
This may seem tedious but I found that a set of 328 images from Jim Fakatselis of Peppermill Skies Observatory took 2h30m to align from scratch. Producing the 5 runs in step 4 took a total of 25 minutes and the final alignment took only 36 minutes because it was already roughly aligned. The method above was 2.5 times faster and the results were optimised because of step 4.
7. Tricks and Tips
Finally, here is another of Jim’s eclipse sequences captured on a second scope. Note how the Moon moves about 2/3rds thru due to the overexposed limb.
[vimeo 141995085]
I have been working with Sylvain Weiller to implement his drift method of video capture into DSSR5 beta. This method moves your scope around a target feature while recording video. Stacking apps can integrate this video to produce an image of the whole scanned area. This effectively quadruples the area of your imaging sensor and also removes the dreaded Newton’s rings.
DSSR5 allows you to scan along single axes (right ascension or declination) or dual axes with the option of backlash take up. The dual axes mode moves around a rectangle 1>2>3>4>1 with a sawtooth edge profile (to eliminate stacking artefacts) like so:
Dual Axes Scan Path
You can see a 4x speeded up selection of scans in the video below:
[vimeo 132827265]We are still working out the best processing workflow to handle these scans so I have uploaded a full disk and detail scan (zipped AVI in Y800 codec) so others can experiment. Please let me know thru my user group or other groups I posted this link to on how you get on and what worked for you. The files are here:
Sylvain managed to process these files in Registax 5 before he went on holiday and his results are below. Notice the extra imaging area compared with the normal 1/4″ ccd field of view and the total absence of the Newton’s rings which were very obvious in the detail video.
Click for full size image.
Update July 13th
I had a go myself using AviStack and here are my uncropped results. Very close to a good result but I need to close up the missing pixels.
Click for full size image
Andrew Cool of SkippySky noticed an interesting result from the AviStack processing results. The Frame Shift graph actually captures the scanning path with sawteeth very well. It also reveals that my mount needs to be more accurately aligned because the red and blue lines should end on zero.
SmarTrak is DSSR’s helper that steps in to guide your scope when the guide taget is lost due to cloud or other obstructions. I seem to have broken it in DSSR5 and I have added some logging features to help me fix it. These can also give you an insight into how DSSR is controlling your scope.
When DSSR loses the guide target it logs the guiding history like so:
08:22:17.838 Target lost – SmartTracking
08:22:17.839 SmartTrak totalTms:4638563 ppmsH:-0.01001258 intH:-49937.2 pH:-500.0 ppmsV:-0.03671094 intV:-13619.9 pV:-500.0
08:22:17.839 ,0001,6834,0,0
08:22:17.839 ,0002,7170,51,305
08:22:17.839 ,0003,7668,103,508
08:22:17.839 ,0004,8266,155,711
08:22:17.839 ,0005,8662,-310,-1322
08:22:17.839 ,0006,65300,-310,-1220….. etc for all stored guide history points.
Each red line contains the log time, point index, time in milliseconds since the last guide pulse and the pulse times in milliseconds for the RA and declination axes. You can copy and paste these points into a spreadsheet and chart the guide pulses in each axis against time. Below is the RA chart for one of my sessions.
Click for full image
This shows the plus or minus RA guide pulses have a recurring period of about 10 minutes. I don’t use periodic error correction on my HEQ5 Pro but this chart shows that I probably should. When I play back my session videos I can clearly see the target moving back and forward in RA around a mean position during SmartTraking.
SmartTrak applies the mean pulse (-50ms red dotted line) but you can see that the polynomial trend line runs from -20ms at 78 minutes ago to -70ms when SmartTrak kicked in at right on the chart. This suggests that I need to cut my number of history points down to cover say just 2 cycles or 20 minutes. This will give a better mean pulse time for SmartTrak to use.
The fact that the mean line is not at 0ms suggests that my mount still has an alignment error. You can see a similar alignment error in the declination chart below.
Click for full image
This shows that all the guide pulses are negative and there seems to be a much shorter periodic error. The mean pulse is -175ms but the polynomial trend is at -150ms when SmarTrak kicks in at right. Again this suggests that fewer history points are needed to give a more accurate mean value.
The above charts have been very useful to help me see what DSSR is doing during autoguiding and I hope to have SmartTrak fully operational soon.
You can try this out for yourself by changing your screenshot interval to 5 seconds, turning on logging, checking the SmartTrak box and then guiding for an hour say. Then raise the minimum guide similarity to 999 to force SmartTrak to kick in and leave it running until the target leaves the field of view. Make a session video as per Appendix E of the manual and scrub it back and forward in VirtualDub to see how SmartTrak is faring.
You can then use a spreadsheet to construct the charts above. I used the free LibreOffice Calc spreadsheet for mine.
The graphic below shows how to calculate the size of the scale Earth to use as an overlay in DVS animations.
Click on image for full size.
I have just been testing DSSR5’s rewritten scan feature. This effectively triples the area covered by your imaging chip and also completely eliminates Newton’s rings. Here is how it works on a PST – DMK21 – PowerMate 2.5x setup.
Click for full size scanned result.
It also allows you to take full disk images of the Sun with a PST and DMK21 at prime like below.
Click for full size scanned result.
DSSR5 is currently only available to members of my user group but will be released publicly soon.
The scanner uses the same file format as DSSR captures so that DVS can automatically add the date and time like the 1123×673 pixel scan below.
DSSR5 has a new scanning feature which enables cameras with small chips to produce images of the entire sun. I tried it with my DMK21 (1/4″ CCD) at prime focus on my PST which looks like this during capture.
Raw video frame
I then used the new scan function to automatically record a 5 minute video of the Sun while DSSR steered the scope to cover the entire solar surface. Registax then did its magic to produce this full disk image below.
Full solar disk produced by Registax 20150417 @ 09:30UT
This is a lot easier than mosaics and it looks like the method smooths out the PST hot spots as well. (It’s also a lot cheaper than a 1/2″ chip camera:-)
A very big thanks to Sylvain Weiller for suggesting this feature and helping me bring it to fruition.
Below are some recent scans – all taken with a 1/4″ DMK at prime focus on a PST.
Scan 20150418 @12:33UT
My old XP desktop in my garden observatory finally started chugging so I checked out the price of an upgrade to a newer version of Windows. Doing so, I found this little beauty which is a quad core PC with Win8.1 for a few dollars more than Win8.1 on its own.
Setup was a doddle and the stock SSD had 21GB free after updating Windows. I then installed ASCOM, DSSR, DVS, DFM, DSR, AstroKam, TIS driver, TeamViewer, Cartes du Ciel, AviStack, StarTools, Krita, Sony Vegas, Microsoft ICE, LibreOffice, VirtualDub, 7Zip, DeepSkyStacker, AdBlockPlus, Canon DPP and Utilities, AS2Cull, EQMOD, EQDIRECT and SPC900 drivers. This left the SSD with 14GB free.
It starts from cold in 20 seconds and is very nippy because all programmes are on the SSD. Power consumption is 8W with no camera and 11W with my DMK21 attached. I can afford to leave it running 24/7 which will cost me around $12 a year. There is no fan so it is totally silent which makes it great as a media player. The aluminium case doesn’t even get warm to the touch.
It is roughly 40% as powerful as my 8-core AMD 3.1GHz, 8GB, 64bit W7 desktop PC, based on these comparison processing times:
DeepSkyStacker image stack: Pipo = 34Min, AMD = 15min.
AviStack batch video processing run: Pipo = 36min, AMD = 13min.
However, the Pipo wins hands down on a bang per buck and bang per watt basis.
It runs my HEQ5 Pro and DMK21 under DSSR with no problems which is why I bought it. It can autoguide all day long while constantly capturing video clips with no dropped frames. One thing to note is that you need to be wary of how much power your devices suck from the USB ports. I have to run my DMK21 camera from a front port and a portable HDD capture disk from a rear port. A powered USB hub should solve this problem.
I added an un-powered 4 port USB hub which lets me add a mouse, keyboard and DSLR together with my EQMOD cable, DMK21 camera and 500GB capture disk.
Here is one of the its first capture and processing tests. Intermittent clouds makes for a jerky video with stacking artifacts but I think it performed to spec.
[vimeo 123507654]
I can see the Pipo being a lot of use in astronomy. You could strap it to your scope tripod with a battery pack (or solar cell) and control it via TeamViewer. Or, have a bunch of them running in a cupboard as a video and imaging processing farm.
Would I buy it again? I already have, and #2 is my new media PC. Netflix, YouTube, iPlayer, etc all play 1080p video without a hitch. The great thing is I can also do video or image processing on it overnight.
ps I have heard rumors than Win10 will be offered as a free upgrade to Win7 and Win8. This would make the Pipo even more of a bargain.