Sunday, March 27, 2016

clown nebula, scientific jelly bean pallet

Narrow band imaging is a technique astronomers use to demonstrate the spatial distribution of specific ion emission lines in an astronomical object (usually nebula).  Rather than using red, green and blue filters (which allow relatively broad wavelengths) to construct an image, astronomers use narrow wavelength filters that isolate emissions due to specific ions.  In either case, the digital camera gives a grey scale image (the camera simply counts photons), which is assigned to a specific color in post processing.  For traditional RGB images the choice of color assignments is clear.  But for narrow band images, the imager may record emissions from very different ions which are the same color (as far as the eye can tell).  For example hydrogen Ha, nitrogen NII, and sulfur SII are all red.  If one were to create an image from these filters assigning each to red, the image would convey no information about the distribution of the individual emissions—the data would be lost.  The alternative is to assign several of the emission lines to completely different colors in order to maximize color contrast, making the image more informative re: ion distributions.    The classic “Hubble pallet” assigns SII to red, Ha (also red) to green, and OIII (teal) to blue.  An alternative pallet, often used used by the Hubble team for planetary nebulae, assigns NII to red, OIII (teal) to green, and Helium (royal blue) to blue.  Interestingly the main mission of the Hubble was to image planetary nebulae, but the classic “Hubble pallet” is not used for planetary nebulae.  Here’s a prior post using the classic Hubble pallet.  This technique was initially dubbed “false color” imaging by scientists.  However there was such a visceral public backlash against “false colors” that astronomers (mindful of funding) now use the term “scientifically assigned colors”, a wonderful euphemism.

On coming up with a more meaningful 3 color pallet:
In a previous post on the clown nebula, I discussed the relative contributions of Ha and NII emissions to the image, but used a relatively traditional pallet for the image, assigning NII to red and OIII to teal (the “true colors”).  I subsequently came up with a more meaningful 3 color pallet in order to illustrate the differential Ha and NII emissions (both red): I assigned NII to red, Ha to green, and OIII to blue, then balanced the NII and OIII on the "smile", making it orange.  The net effect was a variation in the color of the NII globules depending on the ratio of NII to Ha with red having more NII, and green more Ha relative to the "smile".  Also note the blue (OIII) section just above the central star, and red NII outer rim of the "nose" upper left.

In more quantitative terms, for the linear images I measured about 200 adu NII in the smile and 700 adu with the 5 nm Ha (Ha+NII), which implies 500 adu Ha, so more Ha signal than NII.  Over all, I find the NII provides more contrast and, arguably, a more interesting image.
Imaging details in previous post.

It also reminds me of jelly beans :lol:

Thursday, March 24, 2016

Double transit and GRS on jupiter

On 3/21-22 Jupiter put on a juggling display
Io, Europa and their shadows crossed Jupiter's face along with the great red spot over the course of a few hours

unfortunately, the seeing was mediocre causing distortions
mid shadow transit, only the double shadows are evident:
ironically, the transiting moons are obscured by the bright surface of Jupiter while their shadows are obvious.

as the moons rotate off to the right edge of the planet,
they become evident casting shadows just before their transits end:

here's the animation of the second half of the show, over ~1.5 hours:

PS i've got confirmation from the amateur community that the GRS has been getting a bit more red since 2012. it's also slowly shrinking according to NASA.

celestron nexstar 8 GPS (8" SCT on a wedge)
27x 3 minute captures with firecapture @ 200 fps (exposure limited at 45% histogram with gain 64)
stacked in autostakkert, Drizzle 3x, then reduced to 1.5x
sharpened in registax 6
Southern California
mediocre seeing
Mid(UT): 3/22/16

Tuesday, March 22, 2016

jupiter's redder red spot, detail on ganymede

jupiter's up and in prime form right now
here's my first of the season:

might be my imagination, 
but the red spot (which had been fading to tan) 
looks more red to me this year

here's an interesting one i think i neglected to post from 2 years ago look closely at Ganymede transiting the face of Jupiter:

The black dot lower left is the shadow of Io which is further left and below Jupiter.
Just below the shadow, Ganymede is crossing the face of Jupiter.  A close look at Ganymede shows the bottom is white while the top is slightly brown, demonstrating detail on one of Jupiter's moons.  Although not much, I never thought i'd get that from a backyard telescope.

More to come, including an animation of last night's double transit.

nexstar 8 GPS (8" SCT on a wedge)
3 minute capture with firecapture @ 50 fps (exposure limited at 95% histogram with gain 64)
stacked in autostakkert, Drizzle 3x, then reduced to 1.5x
sharpened in registax 6
Southern California
3/17/16 fair seeing

Ganymede transit image from 3/30/2014

Sunday, March 6, 2016

clowning around with narrow band filters :)

I'd planned to image a very faint planetary binned x 4, but felt compelled to shoot something binned x 1 when I realized the seeing was good.
NGC 2392 was the only bright planetary i could think of off of the top of my head
so decided to give it a shot unbinned with the NII filter
so decided to give it a shot unbinned with the NII filter which gave great detail in the reds

the source of the NII globules is still a bit of a mystery

Kitchen sink pallet--NII red, Ha green, OIII blue, He magenta:

just for fun i gathered a few exposures with a variety of filters (thanks to bilgebay for the 3 nm Ha data):
minimal processing, just digital development, except the last.
note the lack of detail in the OIII, with even less in the HeII
while NII + OIII is remarkably similar to Ha
(bill w's conjecture  Ha~NII+OIII for PN, ~Ha+OIII for emission nebulae).

Filter digression:
In general, the narrower the bandwidth of the filter, the better the signal to noise ratio (assuming your exposure is long enough to bright out background signal).  I touched on this in prior post covering the rationale behind narrow band filters in light polluted skies.
However, narrower isn't always better.  traditional "Ha" filters typically capture Ha emissions at 656 nm and NII at 658 nm (older papers refer to them as Ha + NII).  The new astrodon 3 nm Ha filters, in theory cut out a portion of the NII signal, which is not desirable when imaging planetary nebulae.

bilgebay recently captured an image with very similar equipment, but used a 3 nm Ha filter (656 nm).
he was kind enough to send me the Ha data for comparison as i was curious to see how much loss of NII signal (658 nm) affected the image.  not quite a perfect comparison as he used >4x the exposure time for his subs, while i probably had better seeing.  you can see the comparison in the mosaic above.  the difference isn't as great as i'd expected, but there is definitely better contrast between the "eskimo's fur" and the circular nebulosity in the Ha.

8" LX200R, SX Trius 694 binned 0.4"/px
astrodon 5nm Ha, 3nm OIII, 3 nm NII, chroma 4 nm He
Ha 5x5 min, OIII 8x5 min, HeII 14x 5 min, NII 30 x 5 min, L 14 x 1 min
eastbluff, CA

no calibration except luminance ;)

Ha 3 nm 7 x 20 min
Celestron C8 Edge HD, Atik 460 EX Mono, astrodon 3 nm Ha
Bilgebay Observatory, Mugla, Marmaris, Turkey
interestingly, sedat used also used an 8" SCT with the Sony ICX694 chip.

Sunday, February 21, 2016

Abell 78 where's the helium?

first, a more aesthetically pleasing version in Ha OIII OIII, then the helium:

Abell 78 is a rare type of planetary nebula who's exhausted central star ran out of hydrogen to burn (fuse) and collapsed, only to reignite--fusing helium rather than hydrogen at it's surface. this is reflected in the unusual shape of the planetary nebula: a smooth outer shell formed initially, followed by a complex inner shell shaped by the much faster helium wind.  note the filaments (not diffraction spikes) streaming from the central star.

faint ring in Ha (hydrogen):

complex shell with inner jets and stream leading to outer shell (bottom) in OIII (oxygen)

very faint in Helium II

Ha OIII He image
with the helium giving not much more than a magenta cast to the outer shell especially upper left

top: Ha OIII He
bottom: HOHe HOO HOO (bright)

a number of sources indicate that the inner ring is mostly made up of helium
i am at a loss to explain this as the helium was so faint relative the the Ha and OIII that i had to bin x 4 to pick anything up (this is what prompted the 4x binning further tested in my last post).
the last link also makes the helium claim and includes the 2d spectrum of the nebula, but by my reading, the He line is much more faint than the Ha and OIII.

where is the all the helium? what am i missing?  are they referring to He I? or the central star itself?
any input would be appreciated.

8" LX200R, SX Trius 694 binned x2 to 0.8"/px, binned x4 to 1.6"/px, (final image at .8"/px)
astrodon 5nm Ha, 3nm OIII, chroma 4 nm He
Ha 30x20 min bx2, OIII 48x20 min bx2 (best 26 used for RL deconvolution)
HeII 4x20 min bx2, 1x 40 min bx2, 59x20 min bx4
eastbluff, CA

Sunday, November 29, 2015

the dolphin, sharpless 188: an exercise in faintness

Sharpless-188 (SH2-188, Simeis 22 or the Dolphin Nebula) is one of the largest planetary nebulae known (9 light year diameter).  Extremely faint, it lies in a relatively rich field in Cassiopeia, explaining the many pesky bright stars in the image.  It's an excellent example of a planetary nebula interacting with the interstellar medium.

The tiny blue star to the right and just below the big orange star near the center of the arc is thought to be the central star/white dwarf.  

It is is moving through the interstellar medium at 125 km/s (apparently really fast for such things). This creates a bright bow shock (upper right) and a faint streaming tail (lower left).

Here's an inverted image better showing the faint tail:
An interaction between the slow wind of the dying star, faster wind of the subsequent white dwarf, and rapid apparent wind due to motion through the interstellar medium is thought to account for the closing of the tail (arc connecting trailing lines of wake on either side, closing the ellipse).
More on this here:
The shaping of planetary nebula Sh2-188 through interaction with the interstellar medium
Wareing, C. J.; O'Brien, T. J.; Zijlstra, Albert A.; Kwitter, K. B.; Irwin, J.; Wright, N.; Greimel, R.; Drew, J. E.
Mon. Not. R. Astron. Soc. 366, 387–396 (2006)

Here it is in Ha only:

Here's the OIII which didn't add much to the image:

The faint tail was very difficult to capture, hence the exercise in faintness.  Binning is a way to improving signal to noise ratio by combining groups of 4 pixels in the camera into one big pixel, giving 4 times the signal for the same amount of read noise.  Here are two different images each consisting of 24 x 20 minutes exposures (8 hours).  The first binned x2, the second binned x4.

Ha binned x2, then binned x2 again in software (no read noise advantage) to 4x:

Ha binned x4:
4x binning clearly does a better job distinguishing the faint tail from the background.

Really long winded rant on noise binning etc.  (not for the faint of heart, really)

A digital camera has a chip which basically detects photons.
when a photon hits a pixel, it generates an electron*.
(* most of the time, my astronomical camera has a relatively high quantum efficiency of almost 80%, meaning 80% of the time a photon generates an electron)
At the end of the exposure, the register for each pixel is read and the measured potential is proportional to the number of photons hitting the pixel.

Things get quantum:
With very faint light, you start to see quantum effects in the images--a discrete variation in what should be a smooth area of a nebula.
The issue is actually worse than minor discrete differences from detection to detection.
When things get quantum, the detection of a signal has random variation which can be approximated by the square root of the signal.
Not a big deal for the bright core of a nebula which might give 40,000 photons per minute.
With a noise level of 200 for a one minute exposure, noise will represent only .5% of the detected value.
But with only 100 photons noise will represent 10% of the detected value.
Clearly with only 1 photon per minute, it's going to be very very difficult to separate the signal from the noise.
No problem, all we need to do is get more signal by increasing the exposure time...
most astronomical camera's are designed for very long exposure, with cooling to minimize the effect of thermal noise during long exposures (I run mine at -20C).

Formula, just for fun:
If N is the number of photons detected,
the inherent noise in the measurement will be sqrt(N)
and the signal to noise ratio will be N/sqrt(N) = sqrt(N).
So improving signal to noise is a bit of an up hill battle.
The cool thing about digital cameras is you can take one long exposure, or a series of short exposures and just add them up.  By adding up exposures night after night, you can get a weeks worth of exposure time if you want.

Light pollution:
Here's why light pollution is such a huge problem.
Increasing exposure time increases signal, but also it increases the signal due to light pollution.
Suppose we have 1 photon per minute from a nebula and increase the exposure time to 20 minutes for 20 photons.
Light polluted skies can easily give you 2500 photons which gives an inherent noise level of 50,
So the noise due to light pollution exceeds the nebula signal itself.

The skinny on narrow band filters:
But wait, all is not lost.  For the case of an emission nebula, the signal typically consists of light at a single wavelength (or a few wavelengths).  By using a narrow band filter which blocks all wavelengths but those that the nebula emits, the sky signal (which spans the entire spectrum) can easily be reduced by a factor of 100 (usually much more because light pollution is not uniform, for example there is a yellow peak due to sodium lamps).

Read noise:
Every time a digital camera takes a picture, there is a small amount of noise associated with reading the registers, typically on the order of a few electrons for a high-end astronomical camera.
Given the fact that there is noise generated by each exposure, a single long exposure, theoretically has less noise than a series of short exposures. The key to optimal exposure length is to make sure that the noise due to sky signal (light pollution in my case) far exceeds the read noise.  If this is the case, the noise associated with a series of short exposures approaches that of a single long exposure.  so sky noise can be your friend ;)

To bin or not to bin:
One issue that arises with narrow band filters is that the sky noise can be so completely suppressed that read noise becomes dominant, requiring extremely long exposures in order to efficiently stack a series of exposures.  A problem I encountered with my new camera using ultra narrow band filters at long focal length was that the background signal due to light pollution was essentially reduced to 0 even with 20 minute exposures.  Enter binning.  Binning allows you to group 4 adjacent pixels together as one super pixel giving you 4 times the signal with the read noise of a single pixel.  So you increase the ratio of signal to read noise at the expense of image resolution.

Proof is in the dolphin:
Theory's all well and good, but the question is: does binning make a difference?
For this extremely faint target I discovered binning x 4 does.

Lucky imaging in the near future:
Around the corner are cameras with zero read noise and near 100% quantum efficiency.  These cameras allow stacking of extremely short exposures, potentially keeping only the lucky few captured during the best seeing.

8" LX200R, SX Trius 694 binned x2 to 0.8"/px, binned x4 to 1.6"/px
(final image at 1.6"/px)
astrodon 5nm Ha, 3nm OIII
Ha 24x20 min bx2, 30x20 min Bx4, OIII 29x20 min bx4

Tuesday, November 10, 2015

NGC 6210 in NII-OIII: Can superturtles fly?

Here's NGC 6210, the turtle nebula in hercules:
This fairly sharp RGB-OIII image from 2007 shows at least two pairs of jets or ansae (wings).  they appear to be curving slightly, perhaps due to rotation.  Wondering if the condensations in the longer pair of jets were red FLIERS, I decided to try a deeper image in nitrogen and oxygen (NII and OIII):

This NII-OIII image suggests that the upper condensation is a red flier while the lower is not, as the lower condensation is absent in the narrow band image, but present in RGB-OIII image and the luminance (below) --probably a superimposed star.

Here's a blink of luminance (broad band including all visible wavelengths), followed by a green continuum filter (no narrow band emissions), then a stretched NII image.  Which suggests that the upper condensation is an NII red flier, while the lower a broad band star.  Not sure why i'm picking up the central glow with the continuum filter.  reflection nebula? IR leak?
NGC 6210 Luminance-Continuum-NII

There is certainly a lower condensation in the OIII, almost looks like a smeared attempt at a red flier, also note the faint outer shell to the right:
NGC 6210 OIII stretched

The NII-OIII core may represent a letter in the krypton alphabet befitting our herculean superhero*:
NGC 6210 NII-OIII linear

*Terry Pratchett fans claim to see 4 elephants (link) on the turtle's back, particularly in NII

Lastly here's a collage showing various filter images:
top: NII, OIII, NII-OIII color; linear stretch
mid: NII, OIII, NII-OIII color; non-linear stretch
bottom: continuum, luminance, NII-continuum
Answering the initial question:
with 4 wings, but only one flier, the superturtle can fly, but slowly.
QED ;)

8" LX200R, SX Trius 694 0.4"/px
astrodon 3nm NII, 3nm OIII
NII 33x20 min, OIII 35x5,6x20 min, L 155x1 min, 545x50 79x2min