Astro / ObservingReports / Solar Eclipse Side-2006
Starting off it's advisable to check if focal length and chip size are fitting to the photographic problem.
All photos were taken with a Pentax SDHF 75 (f=500mm, f/6.7) as the lens mounted on a non-modified
Canon 5D digital camera (DSLR) which is equipped with a full-size 35mm chip (24mm x 36mm chip size,
12.8 MegaPixels). See the table given by Espenak for a bunch of
suitable image scales:
Canon 5D camera specs:
- Pixel size: 1 Pixel = 3.5 arcsec = 8.4micron = 0.0084mm
- Pixel number: 4368 x 2912 Pixel = 12.8 MegaPixels
- Chip size: 36 x 24mm
Specs for the complete setup:
- Diameter D=75mm, focal length f=500mm, focal ratio f/6.7
- Field size: 2.7° x 4.1° = 5 x 8 Ds (Ds= solar diameters 0.5°)
- Size of the solar disk: 5mm = 514 pixel
- Pixel size at the sun's distance: 2500 km
- Size of the earth at the sun's distance: 17.3 arcsec = 5 Pixel
- Pixel size at the moon's distance: 6 km
The field size is 2.7° x 4.1°, so the setup can cover the solar corona out to radii of 5-8 Rs (solar radii).
I.e. we should try to take pictures as deep as to show the corona at this distance from the sun. The
size of the solar disc on the chip is rougly 4mm, only. Thus we cannot hope to catch really detailed pictures
of solar prominences, the focal length is too small to allow this. The size of prominences and lunar limb details
can be estimated from the scales given above.
page of exposure times for solar eclipses
is an excellent guide for preparing photographic observations of solar eclipses. Given a specific
radius of the corona that shall be reached by the photo, he provides exposure times as a function of
aperture and ISO speed. As a preparational check I shot some test photos the day before the eclipse,
using my full camera + telescope setup and with an excellent optical foil with neutral density of 5 (provided
by Baader in Germany). All numbers provided by Espenak indeed have proven to be both very accurate
What are the limiting factors for exposing photos of solar eclipses ? At the very short end, shutter speed
is limiting exposure times, although it must be said, that modern DSLRs can reach 1/8000 sec, which is
pretty fast (compared to former times where an analog state-of-the-art Nikon F3 provided 1/2000sec as the
shortest speed). At the long end, we are limited by the rotation of the earth, which moves the image of the sun
across the CCD and spoils image quality at some point. (Of course, you can and should use a motorized mount
to compensate for that, but I could not afford to bring along such heavy equipment, and many people will face the
same problem, I guess). It's easy to calculate the maximal time that you may expose without image shifts
on the detector. This is a function on focal length and of pixel size, of course. Declination effects also do matter,
but we may assume decl=0° for the sun in our case here.
1h = 15 degrees
1 sec = 15 arcsec
1 pixel = 3.5 arcsec = 0.28 sec = 1/4 sec exposure time
So I could not expose longer than 1/4 sec without having the sun smeared out more than 1 pixel by the earth's
rotation. In reality, we can afford a little longer, because of the imperfectness of the air (turbulences, seeing) and
optics, plus the effects of the sampling theorem. It's good to be aware of what the limits actually are in numbers.
If required, you may calculate the limiting exposure time (i.e. defined by shift by 1 pixel) with the following formula:
t = 13680 * dx / f, where you should substitute: t limiting exposure time in sec, dx pixel size in mm, f focal length in mm.
Another problem is the movement of the moon across the disk of the sun. This happens on a longer/slower timescale,
though. The approximate angular velocity can be estimated by dividing the path length (2x solar diameter of 0.5°) by
the time between 1. and 4. contact (155 minutes). This results in an angular velocity of 0.4 arcsec/sec. Given the
scale of my pixels in the focus of my f=500mm lens (3.5arcsec/pixel), the moon moves across 1 pixel in roughly 9sec
of time. This number puts an upper limit on the number of photos that you can stack exactly using the edge of the
lunar disk in front of the sun. My 4 x 3 exposures were made in an interval of 63arcsec, so the total shift amounts
to roughly 7 pixels. So positioning all exposures by using the lunar disk limb will not be an exact procedure in my
case here. One may expect some loss of corona detail, if only the limb is used.
In order to reach good signal to noise ratios in the final images, I planned to take several exposures with
identical shutter speed and stack those on the computer. In addition, you need to take exposures of different
exposure times to bracket the whole (or at least a bigger part of the) dynamical range of the corona. During
testing it turned out however, that the total time needed quickly reaches the total eclipse time available.
But I wanted to have time to look as well, of course ! So I compromised on a total time of 2min to spend for
photography, resulting in the plan to shoot 4 different shutter speeds, each with 3 identical exposures.
I also planned to take photographs only in the first half of totality, then take off the camera and observe visually.
Doing photography first was useful, because I could use the non-critical time before totality for focusing and
arranging everything. The foil filter (on top of the front lens) does not shift the focus of the telescope, so the
pre-totality focus could be reused in totality, too.
The series of 3 identical exposures was shot with a Canon timer device and using mirror locking (6 timer impulses with
1 sec of delay in between). This procedure reduced the danger of making errors. Locking the mirror avoids residual
movements of the telescope at the time of exposure.
One should keep in mind that DSLR may run into buffer full problems when shooting large and fast series of RAW
exposures. Waiting for the write buffer to be freed while precious totality time is wasted - could be a potential nightmare.
So I tested everything in beforehand. I did not encounter any problems, simply because you need some time to change
the shutter speed and so you do not shoot continously. Storing in Jpeg does ease this problem also. I had absolutely no
problems with my Canon 5D, even when using RAW (10 MB per image), although I had no especially fast CF card installed.
I should also remark that the clock of my camera was mis-adjusted. I had a shift in time of 215sec and still had a time
setting on CEST, instead of UT or local time. So image times are converted to local time by adding 1 hour and
subtracting 215 sec (local time in Turkey = CEST + 1h = UT + 3h). I measured this shift by calibrating with the
exposure of the 2nd contact (imgae #2528 take at 11:58:32 CEST).
Single Shot Images
In order to capture a large dynamic range , I took exposures as follows (EV differences
provide the difference in aperture steps):
- prominences (1/1000 sec)
- inner corona (1/60 sec, +4 EV)
- outer corona (1/8 sec, +3 EV)
- dark side of the moon (1/2 sec, +2 EV)
All exposures taken together are covering a dynamic range of roughly 9 aperture steps. When adding the
internal dynamics of one picture to this, we end up with a total dynamic range of roughly 15 aperture steps.
Below you find some non-processed single images taken during totality. They were binned by a factor of 5 and have
medium (80%) Jpeg quality only. All images were taken at a setting of 200 ISO. No further processing (except
camera-internal processing of course) was applied.
Click on any image to enlarge it !
Totality image taken at shutter speed 1/1000 sec (left, EV=0) and 1/60 sec (right, EV=4).
Totality image taken at shutter speed 1/8 sec (left, EV=7) and 1/2 sec (right, EV=9).
Below you find two composite images composed from all shutter speeds (1/1000, 1/60, 1/8, 1/2 sec), rebinned
to 1/3 of the raw resolution.
Each single shot was processed with Pellet's method. This works as follows: first of all, images are filtered/smoothed in tangential
direction only, erasing most corona structures, which tend to have radial direction.
(Photoshop filter Blur/Spin, which corresponds to rotating the image around the center of the solar disk). These unsharp masks
get subtracted from the original image, resulting in images with radial structures only. These images are multiplied into the
original (non-filtered) image in order to enhance radial structure. Finally, all images of different exposure times are overlaid
and added into one single composite. All processing steps were performed in Adobe Photoshop CS2.
One cannot hope to realistically show the extreme luminosity and dynamics of the solar corona in any such image (especially
not so because of the limited colour resolution of web-based 8-bit images). This is obvious from the considerations of corona
dynamics above. Total dynamics exceeds 15 aperture steps (so >15bit of depth in luminance would be required). So by necessity,
the image is shown in a pretty much compressed mode here.
It is interesting to note the big prominences (at least 4) on the left image. The minimum corona shows strong structure, with
pronounced axis symmetry (the maximum corona shows central symmetry instead). The sun's rotation axis is oriented from
lower left to upper right. The left image very nicely shows the visual impression we had when looking at the totality with
binoculars or in the telescope at low power.
The picture on the right had a little bit more enhancement with respect to radial structure, in order to optimize corona detail
in the outskirts. Coronal spikes can still be detected out to at least 3 solar radii.
Please click on the images for a full-frame view and switch your browser to full frame mode (using F11 in MS InternetExplorer).
Two image composits (each made of 4 single exposures) taken during totality in Side.
Both are processed with Pellet's method and reproduced at 1/3 of original resolution.
The left image fits very well the visual impression we had of the eclipse. The right picture puts more emphasis
on the outskirts of the corona showing the prominent spikes reaching out to large radii
(Please click on images for full frame views!)
We arrived on sunday just 3 days before the eclipse and were spending our days in a kind of
'alerted' mood. We did some sports, enjoyed the excellent food, went to Manavgat and Side
and finally also made tests of our equipment and procedures that we had figured out. (However,
I am going to be real silent about the 'Raki' business of Andy's here !). The hotel got more and more crowded
as time went by and
the day before the eclipse it seemed completely booked out. I guess that 80-90% of all people had
especially come to see the solar eclipse. Although, some really lucky guys dropped in accidentially
and all of a sudden wondered what all the camera and telescope equipment in the gardens was to mean.
We had many groups from Germany, some 50 Indians, 30 US people, guys from Austria and the Netherlands...
Weather conditions were improving more and more in the days before. The satellite images
indeed looked very good on the very day of the eclipse. On the morning of March 29th, the sky showed no
clouds at all and we were all very happy to be that lucky (there was an initial estimate of 40:60 for
good : bad weather). We had figured out a good site just 100m off the hotel buzz on the upper part of the beach,
just 30m away from the sea with free view to southeast. This was the direction where the shadow was predicted
to come flying towards us. The beach runs rather accurately in east-west direction here. So we guessed that
we had a pretty optimal site for observing the eclipse at elevation of 55° in this direction.
The observing staff at this site consisted of my wife Sabine and my colleagues Andy, Michael + Christa, Peter H. aus N.
:-) and Alex (Christina + Reinhard had gone off to find a different site). A group of guys from the Netherlands installed
their equipment just next to us and we had some nice talks on what was going to happen. All other hotel guests set up
themselves inside the hotel area or on the beach right in front of it (about 200m away). We set up the equipment around
10:30 local time, and I re-tested my camera and timer setups for a final time. Focusing turned out to be more
tricky than I had thought before, because I had no point sources at the daytime sky, of course. Finally I made up
my mind to use a ship far off the coast and/or the (point-like) light reflexes on the sea far out. I re-checked this with
the details on the sun's surface. We were lucky to have some clear sunspots (even a small group) available this day.
Exactly as calculated, at local time 12:38:20 (local time = UT+3h) the eclipse starts with the first contact. The moon
starts to shift its dark disk in front of the solar disk from the lower right. We can hear lots of people clap and cheer
somewhat further up the beach. The world-wide eclipse community seems to love parties anyway ! I am now shooting
the first exposures and plan to keep on doing so regularily every 15 minutes (through the pentax equipped with the
ND=5 filter foil). Actually it's good having a watch with alarm function. Still 1h 20m to go till totality. Some cirrus has
appeared now at the sky, but it seems pretty small and localized and only in the west. Atmospheric turbulence
seems to be OK, as we can see from the reasonably sharp visual images at our telescopes. Michael's Coronado
image shows Hα details on the sun's photosphere pretty well, too. From using the Coronado we already can tell now,
that we are going to see some prominences in totality, too. Although they seem pretty small. One group of sun spots
also shows clear flares in its surroundings. Our photos already show that the lunar limb is not exactly round and smooth.
Is this alread evidence for the mountains on the limb ? Several photos taken with some time in beetween show the same
structure, so yes, it seems we are seeing mountains on the moon here (and this is not atmospheric turbulence or
something artificial) ! This can be checked with this
profile of the lunar limb .
Some 40 minutes later, we are halfway gone, 50% of the sun is covered by the moon. We already notice a change in
the character of the daylight. It seems dimmer (like at late afternoon), but it's not correspondingly reddish, rather
a little bit more blue and with less colour saturation. If you calculate, you see that we've already lost 0.75mag (a factor
of 2) in light now (if you neglect scattering from the sky + only calculate ratios of the areas covered). We can also feel a
drop in temperature, already. The radiation field simply has faded (colleagues have measured a total drop of 7 Kelvin
in the shadow, during the eclipse). Still, our eyes compensate for those effects by adaption and you have to consciously
examine things to notice them.
Another 20 minutes later, effects have become more outstanding now. Light has dropped by another factor of 2 and
another 0.75mag are gone. Changes seems to accelerate notably now. We are also getting more nervous,
feeling that totality is coming towards us. I'm re-checking the critical parts of my equipment. The time taken to cut
the amout of light down by a factor of 2 is now getting shorter and shorter. This acceleration by itself already creates some
kind of low-level panic. This gets real real massive when totality is only 1-2 minutes away. You feel like falling into a hole,
a true collapse. People around us are now clapping and cheering louder and louder. We're definitely bound to drop down now...
Looking to WSW we see it's getting threateningly dark. We can not see a clearly defined shadow approaching, but anyway
my pulse frequency goes up considerably.
The sky gets dark like during twilight after sunset and some stars start peaking out. Venus can be seen in the southwest.
Just 30sec before totality I take off all filters from the binoculars and the telescope. Bailey beads are as big as the sun at first,
but it's dimming down rapidly during the final 20sec. I'm now shooting the first 3 exposures through the Pentax with manual control
(exposure times 1/250sec ... 1/800sec at ISO 200). Although you can still see a piece of the solar photosphere, the corona now
is celarly visible...! At 13:54:55: BANG - we have 2nd contact, start of totality ! The sky very suddenly gets dark like 1h after sunset.
The only thing remaining is the corona shining bright and the dark side of the moon in its center. Wow, it's crazy to watch
this happening - it's still like a shock ! You feel like some huge guy far above is pointing his finger towards you...
People on the beach run crazy like on new year's day (only those rockets and stuff are missing). I have no time to
watch since I first need to run through my photographic schedule, shooting 3 exposures with 2sec spacings in between,
controlled by the remote exposure controller. Then increase exposure time, check position in the camera viewer and
off goes the next short series of exposures. This happens 4 times. In a hurry I now take the camera off the telescope
and get the eyepiece in place for visual observing. 'Now, DO get rid of the camera', I think to myself - but instead I wonder and wonder
if I should shoot another exposure with the 50mm standard lens... but finally I remember that I have enabled mirror locking
and I would need to zap through the camera menues to switch off the mirror locking mode.... I finally manage to get rid of
the camera, and store it in my flight case.
The first glimpse on the corona through the binoculars is absolutely fantastic. The corona isn't round like 1999 but elongated
and only axis-symmetric. It is coverered all over with radial hair-like stripes and details, and on top of this 4 or 5 long spikes
are reaching out diagonally to large radii. In the pole regions the hair is pretty short. Below this bright detail there is also a non-structured
round and diffuse distribution of light. At the upper edge of the sun I can now clearly see the prominence that we saw in the coronado
narrow-band images before, but it's now much more saturated, luminous, bigger and with much more detail. The colour is a
screaming mix of pink and orange. There are even more prominences to be seen nearby. I am just looking at the details of
all this in the telescope when someone shouts '...watch out, only 30sec left to go!' It's incredible how the time flies by...
It's been too short, simply much too short ! I'm angry about having spent so much time 'in vain' on photographs rather than
looking and experiencing...
My chronometer reminds me: watch out; the sun's gonna pop out any moment now. People on the beach are clapping and cheering
once again and BANG, here comes the first beam scratching the lower lunar limb! Again we can see the corona shine,
despite Bailey beads shining bright just next to it. The stars still can be seen.. Venus .. some others ... I simply forgot to
look for mercury, Mars, Formalhaut. But now, I urgently need to get my filters back to protect my lenses and our eyes.
It's getting brighter and brighter very very fast, and the beach people are still very excited, clapping, shouting... Venus remains
to be easily seen for a considerable time - we almost have normal daytime light already. I again take off the filter from the Pentax
trying to check out Venus, zooming in. We have a half venus today ! Christa's really astonished, 'hey, looks like moon ?!'
Sure Christa, - Venus has got phases ! :-) I negotiate with Sabine to let me have her little Powershot A80 DigiCam to
make some automated shots into the zoom eyepiece on the Pentax (just holding the cam behind the eyepiece and watching
the TFT sceen). In fact, we can seen venus on these shots quite nicely. Great. Even 5 minutes after 3rd contact we easily find
venus with the naked eye (you need to remember the position on the sky though). In the telescope we easily see it
even 15 minutes later...
With time, our excitement settles down slowly but steadily. All body functions back to normal - but real slowly ! All the stress (that started
with second contact, and even before already) fades away, finally. It's incredible that everythings seem to be over already.
And, coming back with some remorse to my camera, I am finally very glad of course: I've spent much time taking some
pictures, but they turn out to be excellent when browsing through the files on the camera's built-in 2.5inch TFT. Especially the
pictures taken of Bailey beads seem really wonderful on first sight. Yeah ! I quickly forget my bad wording towards my photo
equipment (I felt really bad about time running out with ending totality).
Everybody's relieved now, and people are checking their photos and videos. I'm staying busy with shooting some
photos of the partial phase now with the Canon + 50mm lens through the Pentax with eyepiece projection - just for fun.
Now I also can afford a little walk with taking pictures of all the people around. Everybody is still upset and checking and
discussing their images. 'Yeah, that one's really great !' Andy crying out loud, because he has shot down his WebCam
programs just at the wrong moment - aaarrrggh ! But still, he has got something, and he will get some nice pictures after
stacking the video data... Peter H. has taken a video including sound, and you can hear all the excited people on it...
A recording that reminds you really intensely of the atmosphere and the beach party during totality !
We are spending the partial phase really relaxed. Taking some photos. Some detail look on the lunar limb, where you
can see the mountains ! Finally, at 15:13:32 the moon detaches from the sun. A big final cheer and thank you from
everybody around. We are all very happy, about our lucky weather, about the impressive observations and nice photos that
we have taken. It also warms up now again. And, really thinking about it, we are finally not unhappy to have the sun back again. :-)
In the afternoon everybody meets on the terrace for a coffee - a nice thing to compensate for the exhaustion during totality.
We enjoy lots of cookies, as most people (unusually) did not have lunch today. We get known to some new friends and
colleagues from all around the world, exchanging images and email addresses.
It's been really really wonderful ! - the next eclipse, where and when did you say it'll happen ?
...and some pictures !
Please click on the picture for a larger view !
Partial phase 1, shortly after 1st contact
Partial phase 2
Partial phase 3
Partial phase 4
Bailey Beads 1
Bailey Beads 2
Bailey Beads 3
Bailey Beads - details
Totality starts, 2nd contact
Composit image of the corona
Totality on the beach with Venus in the southwest
Venus after 3rd contact, photograph taken with a simple DigiCam
through an eyepiece on the Pentax
© 2007-2012 by P. Surma