AstroPhoto Insight

Aug / Sep 2007

http://www.astrophotoinsight.com

Page 1

DO YOU HAVE AN IMAGING, PROCESSING, HARDWARE OR SOFTWARE QUESTION... OR NEED TROUBLESHOOTING HELP? Perhaps you're looking for a better explanation than what's provided in the manuals or something you’ve read in the magazine or online? Describe your dilemma or question in an e-mail to the AstroPhoto Insight editors at: [email protected] If you question is used for a column by our experts, we will sent you a “Got Stars?” tshirt. Please be as specific as possible. General questions such as “what’s the best camera?” or “the best way to process and image?” are too vague.

AstroPhoto Insight

Aug / Sep 2007

http://www.astrophotoinsight.com

Page 2

Cover Photo: NGC7380 by Ken Crawford - Taken with a Yankee Robotics Trifid II 6303, IFW Filter wheel, AstroDon -Ha SII OIII LRGB Filters on a RCOS 20" Truss @ F8.3 Paramount ME. Exposure times: Ha =270 minutes Unbinned Red filter 240 minutes Unbinned, SII = 480 minutes Unbinned, OIII = 200 minutes binned 2x2. CCDAutoPilot3 for unattended imaging With MaxDL 4 and processed with CCDStack. You can see a full size/resolution version here: http://www.rdelsol.com/Nebula/NGC7380.html and more of Ken’s work here: http://www.rdelsol.com/

In this issue... 4

Cameras Compared: DSLR vs. CCD

10

Quick Look: Two New Guiders

14

Summer Secrets

16

Canon-Nebulosity Tutorial for Beginners

23

Blair McDonald gives us a quantitative comparison of DSLRs and CCDs.

Craig Stark checks out the CCD-Labs Q-Guide and the Fishcamp Starfish.

Adam Block shares some summer secrets.

Bill Logan’s tutorial helps beginners get started with a Canon DSLR and Stark-Labs’ Nebulosity

Midwest Astro-Imaging Conference Review Joe Ulowetz shares his experience and thoughts on the first annual conference.

28

Readers’ Image Gallery Images from our readers.

AstroPhoto Insight © 2007 Professional Insight, PO Box 464, Woodstock, IL 60098. No portion of this publication may be copied, retransmitted, reposted, duplicated or otherwise used without the express written approval of the author and Professional Insight. Professional Insight makes no warranty or guarantee about the accuracy of the information included in this publication and assumes no liability for the use of the information. All opinions published are those of the authors and are not endorsed by Professional Insight nor do they reflect the opinions of Professional Insight.

AstroPhoto Insight

Aug / Sep 2007

http://www.astrophotoinsight.com

Page 3

A few months ago I acquired a shinny new DSLR after years of CCD photography. Now let me be up front here: I love using a DSLR. Its ease of use and instant gratification, in colour yet, makes for a hard to beat combination. In fact for the last several astrophotographic outings, my CCD camera has been relegated to autoguiding. Now don't get me wrong: CCD photography has revolutionized amateur astronomy. CCD cameras have allowed amateurs to capture images that rival the work done at professional observatories just 20 years ago. Since these cameras have become widely available, amateurs have made significant contributions to science. From supernova search programs to continual monitoring of variable stars: amateurs now have the time and equipment to carry out observing programs the professionals cannot. The issues I have had with CCD cameras led me, as well as many others, to write camera and telescope control software to position, mosaic and capture images in support of our observing programs. Without a viewfinder, astrophotography with a portable telescope becomes a bit of a chore. In this article, I thought I'd put my viewfinder bias aside and attempt to compare CCD and DSLR cameras on a quantitative basis in terms of their noise, sensitivity and productivity. From the DSLR data collected Testing This article compares the results from several Canon DSLR's and two cooled CCD cameras. The image data

collected are from a Canon Rebel XT DSLR and a Meade 416 XT CCD camera. While the Meade camera is getting long in the tooth, it is still a solid performer. It uses a 768 by 512 thermoelectrically cooled Kodak CCD chip. The Canon is an excellent example of a more modern technology. Its CMOS detector provides eight megapixels of very low noise image data collection. Additional data was collected from several Canon DSLR cameras to see what variation exists between models. For test purposes, the Meade camera was kept at ‑20 0C except for producing the noise verses temperature data. Noise vs. Exposure Time Courtesy of some friends at a local university, I was able to get additional noise data on the CCD camera used in the Burke Gaffney Observatory of St. Mary’s University. Additionally, several DSLR users were kind enough to take the time and collect DSLR noise data from other cameras. The CCD cameras tested were temperature controlled models and both were set to -200C. Dark frames were taken at the following exposure times: 30, 60, 120, 300 and 600 seconds. The dark frames were then bias corrected with a master made from the average of 30 individual bias frames. The average noise level was measured using Maxim DL and the data plotted in Figure 1. The results show the noise level climbs with time and that both cameras show some possible second stage noise effects during shorter exposures. This indicates that noise sources other than the detector chip itself (readout noise perhaps) begin to dominate at shorter continued >

AstroPhoto Insight

Aug / Sep 2007

http://www.astrophotoinsight.com

Page 4

Figure 1 - CCD camera noise performance

Figure 3 - 20Da noise vs. time

exposures. The temperature of the SBIG camera was still stabilizing during the shorter exposures accounting for the abrupt upturn in the measured noise floor below 150 seconds. Three DSLR's, all from Canon, were also measured. Data was collected at various ISO settings and is summarized in the series of plots shown below (Figures 2, 3, and 4).

Figure 4 - 10D noise vs. time

Figure 2 - Rebel XT (350D) noise vs. time From the data shown in the plots, most of the cameras have similarly shaped dark signal vs. time plots and the DSLR's have significantly higher noise levels than the cooled CCD cameras. In these plots, the data represent the total dark current collected by the detector

chips, not just the random component, so it’s not a true measurement of noise. Instead, the vertical axis is the average value of the dark signal read from the camera. All of the plots have been bias corrected and show how much dark current signal has been collected for each exposure interval. This is a good indication of the reduction in dynamic range due to dark current. Of the DSLR's tested, the 350D has the lowest dark current while the 10D has the oddest response over time. Data from another Canon 10D shows a pronounced bump in the dark signal curve around the 500 second region at higher ISO settings.

continued > AstroPhoto Insight

Aug / Sep 2007

http://www.astrophotoinsight.com

Page 5

350

300

Average Pixel Value

250

5 Minute 2.5 Minute 1.25 Minute 0.625 Minute 10 Minute

200

150

100

50

0 0

200

400

600

800

1000

1200

1400

1600

1800

ISO

Figure 5 - Canon 10D data with additional points at 400 & 500 seconds The plots clearly show that the dynamic range of the resulting image is reduced by using the higher ISO settings of each DSLR. However, the reduction is very small with a value of 1 dB for the 20Da for ISO settings of 400 and 1600. The following table details the dynamic range reduction, as well as the actual dynamic range, for the DSLR's using ISO settings of 1600 vs. 400 on a 600 second exposure.

ISO 400 1600 Dynamic range reduction

10D 71.7 71.3

Camera 20Da 72.0 71.0

350D 72.1 71.6

-0.4

-1.0

-0.5

Table 1 - DSLR dynamic range in dB With 12 bit ADC's the dynamic range in the absence of noise for these cameras is 72.25 dB while the CCD cameras, with 16 bit ADC's, have a range of 96.33 dB. This additional dynamic range is a clear plus for the CCD cameras allowing longer exposures before saturation. Based on the data in Table 1 and the plots above, some general guidelines can be drawn to act as a guide in ISO selection. First, there is very little reduction in the dynamic range of each of the cameras tested between 400 and 1600. Secondly, when the data is replotted to show dark signal vs. ISO setting for a constant exposure time, a pronounced change in the response is seen with ISO settings below 200. The plots below for the 350D show this effect.

Figure 6 - Dark signal vs. ISO setting As can be seen from the plot, with the exception of the 10 minute data, the curves tend to converge and flatten at an ISO setting of 200. This effect was present in all the DSLR's tested and may indicate some second stage noise that is introduced after the variable gain amplifier used to simulate ISO. One possible source of this noise is quantization error in the ADC for the very small signals. This essentially means that for low ISO setting and short exposures, the DSLR's seem to be read noise limited. With the possibility of second stage noise, low ISO settings should be avoided unless the exposure is longer than five minutes, or the subject is bright. Also, because of the small dynamic range reduction for a given exposure time, ISO should be set as high as possible to ensure that the dim areas of the image will be well above the quantization noise levels. In general, I would recommend using the highest ISO setting that does not cause saturation on important parts of the image. Keep in mind that I'm not recommending high ISO settings as a way of shorting the exposure. Signal-toNoise Ratio (SNR) is a function of exposure time. So for best results, the exposure should be as long as possible before saturation sets in. The higher ISO setting simply provides some protection against quantization noise. Additionally higher ISO values will provide more quantization steps for the dimmer portions of the image. This may help to reduce "banding" in stretched images. Sensitivity Here is where the dedicated CCD cameras shine. Figures 7 and 8 are five minute exposures of M97, the Owl Nebula, taken from urban skies. The DSLR image was taken with the 350D at ISO 400. The image was cropped and scaled to match the CCD image and then converted to grayscale. As can be seen, the CCD image

continued > AstroPhoto Insight

Aug / Sep 2007

http://www.astrophotoinsight.com

Page 6

Again, if this were a colour image the exposure time difference is not as great and the DSLR becomes a reasonable option, as the CCD system requires three times the exposure.

Figure 7 - CCD camera image is clearly superior. It has a higher SNR and even the stars are recorded at substantially higher levels. Increasing the ISO on the DSLR from 400 to 800 would make the sensitivities about equal, but the SNR of the 350D image would still be lower than the cooled CCD camera. The line plots in figures 9 and 10 (next page) show the pixel values along a line through the star just above the nebula. They show the SNR differences very, pardon the pun, graphically.

Ease of use For the two types of cameras tested, the photographer (that would be me) was timed doing the various setup functions required to take an astrophoto. This amounted to setting up the camera, focusing, and then framing the object. Here the DSLR has a clear advantage, as shown on the table below. Setup for the DSLR was a “no contest”. Simply remove the lens, twist on the eyepiece adapter, plug in a shutter control cable and pop the camera into the focuser. For the CCD camera, it required connecting the camera head to the control unit, putting the camera in the focuser, connecting the control unit to a laptop, booting the laptop, starting the camera control software, connecting to the camera, turning on the cooler then waiting for it to reach operating temperature. Focusing was done with a diffraction focusing tool for both cameras and no software was used. With the DSLR, rough focus was achieved using the viewfinder and two iterations with the diffraction tool were all that was required. Using the CCD camera, the focuser was continued >

Doubling the DSLR exposure would match the SNR of the CCD camera for the star used to produce the line plots. Keep in mind the DSLR produces a colour image, so while its exposure must double to get the same SNR as the CCD camera, the CCD exposure must triple to produce a colour image! To compare sensitivities when the object is rich in Ha emissions, two images of M42 were taken. Each one is a 90 second exposure. The nebula is just starting to register on the DSLR while it is just starting to saturate the CCD camera as seen in the two images shown below (Figures 11 & 12; next page). Clearly the CCD camera is more sensitive than the DSLR on both planetary and emission nebulae.

AstroPhoto Insight

Aug / Sep 2007

Figure 8 - DSLR image

http://www.astrophotoinsight.com

Page 7

racked out and an image was taken. After inspecting the image, the focuser was racked in a bit and the process repeated until perfect focus was achieved. Now I realize that the rough focus position could have been set using a para-focal eyepiece, but I was comparing initial setup procedures.

Figure 9 - DSLR data

Framing with a DSLR is simple - look in the viewfinder and position the object. If you can't see the object in the viewfinder, take a short exposure and note the position of bright stars in the field, then frame the object based on the same stars seen in the viewfinder. With a CCD camera, framing is a real pain. Take an image and wait for it to download, move the scope and repeat until the object is positioned properly. Conclusions It was interesting carrying out the testing; as each type of camera definitely has its pros and cons. The cooled CCD camera is the hands down winner if sensitivity is your prime concern. Without a doubt CCD's are more sensitive and have a lower noise floor than the DSLR's tested. The images shown in this article clearly show the advantage of a cooled camera over an un-

Figure 10 - CCD data

continued >

Figure 11 - DSLR M42 image

AstroPhoto Insight

Aug / Sep 2007

http://www.astrophotoinsight.com

Page 8

Figure 12 - CCD M42 image cooled DSLR. In addition, the CCD system has 16 times the dynamic range of the DSLR due to its 16 bit ADC. If exposure time is your primary concern, and if you are taking tri-colour or LRGB images then the DSLR becomes a viable option. It will require about two thirds the exposure time of the CCD (for the same SNR) as all Function

CCD

DSLR

Setup

15 minutes (it had to cool) 10 minutes

1 minute

5 minutes

.5 minute

Focusing on bright star Framing

5 minutes

Table 2 - Setup times colours are captured at once. Keep in mind that this is for colour imaging only. If comparing the DSLR to a monochrome CCD image, then the CCD's greater sensitivity and lower noise floor delivers much shorter exposure times. When it comes to ease of use then the DSLR is the clear choice. For instant gratification the DSLR is hard to

AstroPhoto Insight

Aug / Sep 2007

beat, and the combination of an LCD display, a viewfinder and no computer or complicated hookup is very nice indeed.  Acknowledgements The author would like to thank several individuals for their help in compiling data for this article. Mr. John Walton for data on the 10D and for his ear while attempting to analyze the data. Mr. Burton Briggs for his efforts in collecting the data for the Canon 20Da. Mr. Eric Jacob for collecting the Canon 10D data presented in the text. Mr. Daniel Majaess of St. Mary’s University, Department of Astronomy and Physics, for his collection of SBIG CCD camera data.

Blair McDonalds heads an R&D group at an east coast Canadian company that designs and manufactures rugged satellite radios for search and rescue. As a long time member of the Royal Astronomical Society of Canada (RASC) he has always been active in astrophotography. With the advent of digital imaging systems, he has been able to combine the techniques of digital signal processing from his day job with his passion for astrophotography.

http://www.astrophotoinsight.com

Page 9

Disclaimer: Craig Stark operates Stark Labs and is the author of the PHD Guiding program (freeware) provided with both cameras described here. He is also the author of Nebulosity, one of the capture and processing programs provided with the Fishcamp Starfish. He served as a beta-tester for both cameras but has no financial interest in either Fishcamp Engineering or CCD-Labs. As the author of PHD Guiding, I am often asked by users for my opinion on various guide camera solutions. My advice is typically to favor cameras that are a) capable of exposures of several seconds, b) monochrome, and c) have a reasonable sized chip. Unfortunately, webcams, the cameras may want to use for guiding don't fit the bill. Most are limited to about 30 ms exposures and almost all use small color chips. The short exposures and presence of color filters over each pixel conspire to limit the choice of guide stars to bright stars (e.g., mag 4-5). Given the small swath of sky covered by a small chip, makes finding a suitable guide star even more challenging. When standing out in either a freezing cold or hot and mosquito-infested field, the last thing I want to be doing is hunting around for a suitable guide star. Here, we take a Quick Look at two relatively new entries into the market, the CCD Labs QGuide (http://www.ccdlabs.com/Qseries/qguide.htm) and the

AstroPhoto Insight

Aug / Sep 2007

Fishcamp Starfish (http://www.fishcamp.com/ starfish.html). Both cameras satisfy these requirements quite well and have performed admirably for me in the field. The two cameras are designed with two very different philosophies and aimed at two different markets, however as we'll see below. Commonalities Both cameras use the Micron MT9M001 black and white CMOS sensor and stream raw, uncompressed data from the camera over a USB2 interface. It is a 1/2" format (6.6 x 5.3 mm) sensor with 1.3 megapixels (1280 continued >

Front view of the Starfish (left) and Q-Guide.

http://www.astrophotoinsight.com

Page 10

x 1024 array of 5.2? pixels). The sensor has a relatively high quantum efficiency (QE: Percent of photons captured and converted into electrons) of 56% with a broad spectral response (sensitivity falls to half of the peak at ~370 nm and ~780 nm). This contrasts quite favorably with another CMOS-based guider on the market, the Yankee Robotics Trifid Nugget ($495). It's Kodak KAC-9619 sensor is smaller (1/3" format of 4.9 x 3.7 mm), has fewer pixels (648 x 488 array of 7.5?), and a lower QE (27%). The Micron sensor is approximately The port connections on that back of the Starfish (left) and Q-Guide. the same size as the Kodak KAF-0402ME (765 x 510 array of 9? pixels) used in the SBIG ST-402ME ($1395), but acts as a Value Added Reseller of the QHY CCD verit's 56% QE is clearly lower than the KAF-0402ME's QE sion of this camera, the QHY5. They provide USA distriof 85%. bution, support, and have pushed the development of the camera (addressing driver issues, providing support Both cameras have onboard ST-4 style autoguider for PHD Guiding, etc.). They also provide the camera output ports. These RJ-11 (telephone-style) ports allow fully ready to go, complete with nosepiece, cables, inyou to send the guide commands to your mount's guider structions and a CD with drivers and Windows-based port (if present). As software running on your computer software . detects the star is beginning to move from the target location, signals can be sent out along this port to your The QGuide is very nicely put together and is a very mount's motors to bring the star back in line. This, compact (2.5" circle, 1.25" deep) and lightweight unit therefore, provides the same manner of guiding your (4.2 oz). This latter aspect is not to be under-rated as mount provided by the popular ShoeString adapters and flex in the guide scope mounting and focuser is a real is an alternative to sending guide commands to your cause for star trails and something often overlooked by mount via ASCOM and the serial (or USB) port on your computer. Both can attach to your telescope via either a imagers as they attempt to get autoguiding working well. Putting a heavy camera on the end of an inexpensive provided 1.25" nosepiece or standard T-threads. refractor's focuser is a recipe for differential flex between the image on the guide chip and the image on the main Both cameras also require a computer and software camera's chip. Guiding software may keep the star to actually collect and analyze the images and send the locked on the same position in the guide chip, but this guide commands to your mount (the no-longer-available flex makes the image still move on the main camera's SBIG ST-V was the only guide setup that did not require chip. Fixing this flex alone in my personal rig let me this). Both cameras ship with PHD Guiding and have move from 2 minute exposures to over 15 minute expodrivers available for guiding in other software (Maxim sures. DL, CCD Soft and additionally for the QGuide, AstroArt). Q-Guide The CCD Labs Q-Guide is designed to do one thing, do it well and do it cost-effectively. It hits on all the recommendations noted above and costs $270. CCD Labs

The camera needs no power other than that provided by the USB port and has connections on the back for a USB cable and a guide command output. There is also a single red LED that flashes during camera activity. Over a USB2 connection, full frames are downloaded in continued >

AstroPhoto Insight

Aug / Sep 2007

http://www.astrophotoinsight.com

Page 11

about 70 ms. USB 1.1 is supported, but the ability to speed up image transfers by downloading only a regions of interest (ROI) is limited to 500 x 500 pixels or larger. On USB2 connections, this is a non-issue, but it does keep the camera from being highly responsive on older USB 1.1 connections. Some may view this as potentially suitable for lunar or planetary imaging as well as guiding. CCD Labs is clear to note that this is designed as a guide camera only, however, and that it is not optimized for solarsystem work. While the camera is specified to give exposures as fast as 1/10,000 s, this is done via a "rolling shutter". With this shutter scheme, each line may be exposed for a very short duration, but it still takes a fixed amount of time - 1/15 s in the case of the QGuide to expose the entire chip. This limits the camera's ability to deal successfully with atmospheric turbulence and freeze a clear image during brief moments of steady seeing. Likewise, while the camera is capable of 1 minute exposures, its noise characteristics and the fact that data come off in 8-bit format do not make for a good DSO camera. It is designed to fill a specific need - the need to autoguide - and to do that well. Starfish The Fishcamp Starfish is also designed primarily as a guide camera, but where the QGuide cuts corners to provide a basic, highly-functional guider, the Starfish is designed to be the ultimate guide solution. Fishcamp has pulled out all the stops to get the most out of the Micron CMOS sensor and provide the cleanest possible image and the highest level of functionality. Where the QGuide consists internally largely of the sensor and a USB interface, the Starfish adds a tem-

AstroPhoto Insight

Aug / Sep 2007

perature-regulated TEC (capable of -15º C from ambient, regulated to 0.2º C) and an on board CPU (processor) and memory. Where the QGuide has ST-4 output and basic "activity" LED, the Starfish has an ST4 output, an RS-232 serial output, and a full array of status LEDs. These features and more put it in a different league than the QGuide and provide a justification for the higher price tag ($995). The Starfish is provided with software for both Windows and OS X. In Windows, PHD Guiding is provided for autoguiding and Nebulosity is provided for longexposure image capture and processing (both from Stark Labs). In OS X, users have a choice of applications. Both PHD Guiding and Nebulosity are provided along with Fishcamp's own StarLink application. StarLink provides full guiding and image capture functions as well. As noted above, drivers are provided for Maxim and CCD Soft. Physically, the Starfish exudes the feeling of a very well-built device. It is larger than the diminutive QGuide, measuring 2.75" in diameter and 3.125" long and weighing in at 10.9 oz. The rear of the camera starts showing off several of the added features. You get not only a basic "camera activity" LED, but separate LEDs indicating when each of the data lines on the onboard guide port are active (a very handy feature when troubleshooting). In addition to the ST-4 output and USB ports, the back also has an RS-232 serial port (to connect to your telescope or filter wheel) and a power plug. The power plug takes 12V input and is used to operate the TEC (the camera functions without power just fine). Here you'll note an example of Fishcamp's attention to detail providing a coax socket that has a screw-on collar to lock the cable in place. Inside the camera, the onboard CPU handles image processing, guider timing TEC cooler regulation, and camera diagnostics. Images come off the sensor in the full 10-bit format the Micron sensor is able to provide and are passed into an internal memory buffer. The CPU then takes care of bias correction (to remove fixed pattern noise) and additional processing tailored to the specific sensor and the current imaging parameters to reduce the noise. These all act to reduce the read noise of the Starfish to levels well below the QGuide and makes it very tolerant of issues with multiple devices sharing the same USB bus. The Starfish can also act as a multi-purpose camera. While it still uses the rolling shutter (part of the Micron chip), it has several internal tricks up its sleeve to get better lunar and planetary images out of it. The pixel clock is continually adjusted based on the imaging de-

http://www.astrophotoinsight.com

Page 12

DSOs"? Some will point to the 10-bit limitation on the data and this is certainly a shortcoming. However, uncooled 16-bit cameras waste many of the bits on thermal (and other) noise. In addition, since there is no such thing as a fraction of an electron, image sensors incapable of holding at least 65k electrons do not need 16 bits to digitize the image. Most amateur CCDs don't have fullwell capacities nearly this high. Finally, as one stacks images, the error associated with limited bit depth - quantization error - is reduced and overall bit depth is increased (so long as one stacks with greater bit depth). Thus, while 10 bits is a bit lower than one would hope for in a DSO camera, it's not as bad as one might initially think. Having seen a number of very clean DSO shots from the Starfish, it's clear that it's up to the task. Conclusions

mands up to the maximum rate of 48 MHz (twice the QGuide's fixed rate), dropping the time to read the data off the chip in half to 1/30 s. Full ROI control further drops this duration when cropping the frame during capture. In addition, the onboard image buffer and CPU let the camera's exposure duration be independent of the current load on the computer you're using. All timing is handled inside the camera, which makes for more consistent and accurate exposures. When combined with the reduced noise and 10-bit range, the Starfish takes on the appearance of a lunar and planetary camera as well. While few would mistake it for a high-end DSO camera, the Starfish can actually do a very respectable job here as well. The TEC's cooling is regulated down to 0.2º C and the camera is capable of 5-minute exposures. When combined with the very low noise (the manufacturer has quoted me ~7 e- of read noise) and respectable QE, one starts to wonder "why not shoot

AstroPhoto Insight

Aug / Sep 2007

I have used both cameras for several months now, with each getting a number of nights of guide-duty. Both have functioned as advertised and guided my mount well. The relatively large monochrome sensor used has made finding and locking onto guide stars a painless process using my 66 mm guide scope. Despite both using the same sensor and being both being designed as guiders, comparing them for the purpose of choosing between the two is an apples-andoranges situation. The Starfish is cleaner and more versatile. Even when guiding, the effect of cooling and the onboard noise reduction show through to provide a cleaner image of the star field and to guide on slightly fainter stars. The Starfish also works perfectly on my Mac. While I've not tried DSO or planetary imaging with it, I've seen a number of very nice images in both categories. It's also nearly four times as expensive as the QGuide. If you're looking for a basic guider that is nicely up to the task of guiding your mount, the QGuide is hard to beat. It's compact, inexpensive, sensitive, has onboard guider outputs, and just plain works. It's an order of magnitude better as a guide camera than even long-exposure modified webcams or cameras like a Meade LPI or Celestron NexImage. The nice thing is, we've now got two more excellent choices when shopping around for cameras. 

By day, Craig Stark, Ph.D. is a professor of Cognitive Neuroscience at Johns Hopkins University. By night, he is an amateur astrophotographer and developer of software for astrophotography

http://www.astrophotoinsight.com

Page 13

The portion of our galaxy that appears high in the sky during the early evening hours for the northern hemisphere is filled with remarkable nebulae. Some of these clouds of gas and dust show us the beginnings of the stellar evolutionary cycle. Natal clouds of gas and dust glow expectantly with the creation and unveiling of newborn stars. Other smaller gas structures hint at the deaths of stars with both “planetary nebulae” and supernovae remnants.

these details can be detected with telescopes greater than 2000mm in focal length. However, do not despair if you happen to have a smaller/faster system. Since this object is bright, amateur imagers can do something that not even the Hubble Space Telescope can do. The image on page 15 (Figure 2) shows the Egg Nebula in two different wavelengths. Neither of these images are RGB (wide band/true color) images. Thus, this is information in the first image that is not easily discerned in the HST images. The color of the Egg nebula shown above can be detected with even relatively small telescopes. Note that half of the nebula is bluish and the other half is distinctly yellow. These colors instantly give us an interesting piece of information. We know that this object is intrinsically the same color throughout. It is basically a bluish object because we are looking at scattered light like a reflection nebula. So why is half of it yellow? It is because it is tilted from out vantage point. The bluish half is in the foreground and the yellowish portion is the rest of the nebula that is slightly behind the thicker clouds of dust (behind the bar). Thus we can imagine its orientation in space just based on its color from our amateur images; and this bit of information is difficult to perceive even with the astonishing resolution that HST achieved!

Here two selected stars illustrate this stellar evolution and offer the amateur imager some exciting goals to strive for. The first rather strange (and perhaps esoteric) object is the Egg Nebula (Figure 1). Its name is derived from its discovery and examination in Infrared radiation. Astronomer Mike Merril (now at NOAO) thought it looked egg-like in those low-resolution images. A higher resolution (amateur) image can certainly show many interesting structures. This object is an example of a proto-planetary nebula. This star has already begun to shed its outer atmosphere. However, the winds are not yet strong enough to clear away the gas and dust it has begun to release. Thus one easy goal is to detect the brownish bar of dust that hides the central star. Further details include the two “searchlight” spokes that emanate from the brightest portions. Higher resolution images (less than 2 arcseconds) can also show the radial concentric rings that very few other celestial objects have. All of Figure 1 - RGB image of the Egg Nebula AstroPhoto Insight

Aug / Sep 2007

The other summertime secontinued > http://www.astrophotoinsight.com

Page 14

cret is NGC 6781 (Figure 3). Though well-known to visual observers and imagers alike, the details this object offers the practiced CCD enthusiast are striking and rewarding. First and foremost is the wreath of red that envelopes the entire planetary nebula. Both low and high-resolution images show this red portion to be made of a smooth bubble and thinner asymmetric outflows. More difficult to detect is the faint teal of the central bubble. In low resolution images is it often lost in the overwhelming redness of the nebula. This becomes more easy to process and display with over-sampled and high S/N data. The stars in the field represent an interesting processing challenge. The sheer numbers of them create a masking effect (confusion) that de -emphasizes the nebula. One trick that is often employed on fields such as these is to use an erosion filter that replaces bright edges with the darker adjacent color. For stars this appears to “shrink” and deemphasize them slightly. The other stellar challenge is that the stars are quite bright compared to the nebula. Often a delicate balance of brightness is necessary to show the nebula with a pleasing amount of contrast WITHOUT making most of the stars look white. Most critical is to retain color of the bluish central star. It is the super-blue of this star (ultraviolet light) that is makes the entire nebula glow in the first place. Finally good images will also show slightly dark concentrations of the bubble that are in the foreground and attenuate the interior light. These darker clouds require reasonable seeing and luminance data that may exceed an hour in total duration. 

Adam Block developed and crafted the nightly observing programs (NOP) and Advanced Observing Program at Kitt Peak National Observatory. Adam’s images and tutorial DVDs are available on his website: http://www.caelumobservatory.com AstroPhoto Insight

Aug / Sep 2007

Figure 2 - Hubble image of the Egg Nebula

Figure 3 - NGC 6781 Note: The details on the two images shown in this article can be found at: http://www.caelumobservatory.com/obs/n6781.html and http://www.caelumobservatory.com/obs/egg.html http://www.astrophotoinsight.com

Page 15

I am often challenged learning new software. Eventually I learn by asking a lot of questions on various astroimaging online groups. In an attempt to help others, I have written this little tutorial on using Nebulosity, available from Stark Labs (http://www.stark-labs.com), with a Canon Digital SLR. Hopefully I have written it in an easy step-by-step guide that you'll find useful. Although Nebulosity can be use with many cameras, this guide is specifically designed for those who are using a Canon DSLR camera with Nebulosity. Before beginning, it’s recommended that you disable any screen savers and adjust the power settings for your laptop so that it won’t go into hibernation due to a lack of activity. Also, you should disable any Wi-Fi or Internet connection while using Nebulosity. Depending on your computer’s performance, it may be best if Nebulosity and PHD autoguiding are the only programs running . I will assume that you have your camera attached to your scope and are powering it with an A/C power supply (ACK-700). This is helpful as long exposure photography tends to drain an NB-2LH battery in no time. You should also have the Canon EOS USB cable connected to the computer and camera, and a Shoestring Astronomy DSUSB cable (or similar) to control the shutter connected to the remote shutter control of the camera and the computer. There’s not much to Camera control as Nebulosity does all the work. However, there are two things that you need to do before starting. Set the Canon EOS to “M” manual, and the auto-off control to “off”; otherwise the camera will turn itself off after 2 minutes Don’t worry about ISO setting or image size. Nebulosity will control those items.

AstroPhoto Insight

Aug / Sep 2007

Editor’s Note: Bill Logan passed away in June 2007 of a heart attack. He submitted his Canon-Nebulosity Tutorial for publication prior to his passing. “Bill Logan was a long time acquaintance and a good online friend. During my long tenure in Flagstaff we often discussed astronomy and astrophotography, especially the ‘old school, film-based’ kind. I was surprised when Bill moved over to the digital imaging world, but he always wanted to see what was on the other side of the mountain. His latest images revealed his dedication to perfecting DSLR photon collection. He was clearly reaching new levels of expertise. I was shocked to learn of his passing, as he had never mentioned a heart condition. Bill represents for me the infectious good spirit that runs throughout the amateur astronomical community across the world. All of Bill's many friends will miss him. Here's to Bill, a comet who has passed us by.” ~ Max Oelschlaeger

Important USB information: Some laptops only have one or two USB ports. This will present a problem for using all the USB controlled items such as the mouse, remote shutter control, autoguiding, image capture, and camera control. Newer laptops only have three or four ports which may not be enough. I recommend you purchase a 7-port “powered” USB hub. A typical laptop’s USB port only has 5VDC @ 500 mA; which on its own would not be enough to power all the accessories. On your laptop, I recommend that you create a separate folder for your target object. That will make it easier to organize your images for image collection and proccontinued > http://www.astrophotoinsight.com

Page 16

essing and reduce the risk mistakes during processing. Capturing Images Step #1: Turn on the Canon EOS camera. The Canon Utilities window will automatically open. Allow it to open and then click “Quit”. We found that if you try to run Nebulosity with Canon Utilities open, you will get an error message. Now double click the Nebulosity icon. The Nebulosity window will open with a large black screen. Click the “Files” tab drop-down menu at the upper left corner of the window. Near the bottom of the list, click “Preferences.” Click “Acquisition Mode” and click the down arrow on the right and select, “RAW acquisition.” Next, click “DSLR Long Exposure Adapter.” Click the down arrow on the right and click “Shoestring DSUSB” (or whatever is appropriate for the longexposure adapter you’re using if you’re not using the one from ShoeString). If you don’t select this option, the camera will only give you a maximum of 30-second exposures. Leave all the other settings alone unless you

know what you’re doing. You only need to set these once as Nebulosity will remember the settings. Warning! If you do not make the necessary selections in the “preferences” section to RAW Acquisition, the Canon EOS will take the photos in JPG color format. Processing in RAW is not only more accurate but uses less memory than processing in full-color. Step #2: On the right side of the screen under the “Camera” heading, click the down arrow for the dropdown menu. Select “Canon DIGIC II DSLR.” It will take a second or two for the computer and camera to establish communications. If a cable connection is bad or you simply forgot to connect the camera, you’ll get an error message. Point your telescope to a relatively dim star near the target. Don’t chose Sirius, Venus or any other ridicucontinued >

AstroPhoto Insight

Aug / Sep 2007

http://www.astrophotoinsight.com

Page 17

lously bright object. To properly focus the camera you’ll need a star about 7th magnitude or dimmer. This will take some time and practice, but you will be amply rewarded with exceptionally sharp images if you take your time with focusing. Craig Stark has made focusing very easy in Nebulosity as the pixel algorithm does all the work for you. Step #3: Focusing the camera. On the right side of the screen, under exposure>duration, select "1 or 2 seconds” or any short duration you feel is appropriate. (If the duration in parenthesis says, “(ms)”, go back to file>preferences> uncheck the box for “use msec not seconds.”) Leave Gain and Offset alone and leave the number of exposures to 1. Nebulosity will continue to exposure your focus star at the time you have selected until you press “Abort.” To get a rough focus, look through the Canon viewfinder and focus visually as best you can. Now press “Frame and Focus.” Don’t be alarmed at the image if it doesn’t look good. The goal here is to get close to focus and to frame the target. (Caution: If you see “dancing” stars, you are looking at hot pixels. Remove the lens cap!! I know because I’ve done it!). Nebulosity shows an overlay with crosshairs and something like a bull’seye. When you start, the crosshairs will be offcentered to the right. Frame and Focus uses the “small” image size from your DSLR, but even these are probably bigger than the screen. There are two ways to get it centered. You can either use the “-“ for zooming out located under the histogram or you can use the sliders on the right side and bottom of the images until the cross-

AstroPhoto Insight

Aug / Sep 2007

hairs are in the middle. Adjust your focus and move the mount to get things looking sharp and centered. Once you have this, click “Abort”. Now click “Fine Focus”. The screen will do nothing. Next, select a dim star and click on it. The screen will suddenly go blank and your star will appear in the upper left hand corner along with three letters with numbers. M, m and S. (Note: If the window is too small, use the zoom-out feature under the histogram) “M” is for maximum pixel intensity; “m” is the average of three maximum pixels; and “S” is for sharpness. You can use any or all of these numbers to get a fine focus as they all increase when focus is best. I like using “M” - the top number - because it has a larger range and is more sensitive. Turn the focus knob of the imaging telescope until you have acquired the maximum number. The “M” maximum that you can get while focusing is 65535. That’s why you don’t want to use Sirius or other bright star for focusing because it will always be at 65,535 regardless of your focus. Remember that we’re using a relatively dim star. It may never reach 65535. The object is to reach the maximum possible number. You will also have a running graph of each letter that shows you when you’ve achieved the best focus. Remember that we said focusing requires a lot of time? Here’s why. Each time you touch the focuser, the telescope moves causing the numbers to be very erratic hence the reason why some folks like to use motofocusers. Allow the telescope to settle down and give you an accurate reading. These numbers will also be erratic when the seeing conditions are bad. Once you have reached the maximum numbers, tighten focuser set screws and then click “Abort.” Step #4: Previewing your image live on the computer screen. Slew the telescope to your target object if you’re not already there. Click “Frame and Focus.” I like to zoom all the way out so that I can see the entire field of view. With your telescope hand controls move the telescope so that the target object is composed in the center or wherever you want it. Remember that on some telescope/camera combinations, the edges will have coma (caused by having a flat CMOS surface that doesn’t coincide with a curved optics) so it’s important to center your target as best as possible. Once you’re satisfied with the composition, click “Abort.” Step #5: How long should my exposure be? To start, change the exposure duration to 30 seconds and click “Preview.” After each image, look at your histogram. As the tutorial on the Stark Labs Wiki says, “Is all of the data to the far left? If so, your entire image is faint and you should increase your exposure.” Let’s say you are trying to image the very faint Rosette Nebula from a dark site. At 30-seconds, the histogram will show almost all of the data on the left. It wasn’t until I increased the exposure to 180 seconds (3 minutes) continued > http://www.astrophotoinsight.com

Page 18

sibly damaging the internal components. Many dedicated astronomy CCD cameras however, address this problem with Thermal Electric Coolers called TECs. Offset: This is automatically set by Canon EOS series DSLR cameras and is grayed-out. Some other cameras let you adjust this. Step #6: Taking a series of images. Well it looks like we’re ready to image. First thing you need to do is tell the computer what the name of the object is and where will the RAW images will be stored. Go just below the Abort button and give your images a name. If you are taking photo of the Trifid, get rid of “Series 1” and type in “TrifidRaw.” (Note: It is a good practice to change the name of your images at every step in the process just in case you want to go back and reprocess) Just below that, click the “Directory” button and select the folder that you named previously, “Trifid.” If you forgot to make a folder, click the “Make New Folder” button and under “My Computer” name it, “Trifid” that I finally got a decent histogram. Ideally, you want a maximum signal on the left and gradually decreasing to the right without having the right side of the histogram cut-off as it hits the right side. Exposure time is something you must experiment with. Remember: the longer the exposure and/or the higher the ISO number, the more noise you will have in the chip (giving you more “hot pixels”). Gain - ISO setting. ISO is like the old ASA settings of film. The higher the ISO number, the higher the sensitivity which produces more grain (noise). By default, the gain will be set to level 4, ISO 1600. Personally, I like level 3 - ISO 800 for providing faint details while keeping the noise down. However, it’s a personal preference. Experiment with different settings to your liking. The lower the level – the less sensitive the CMOS and more exposure time will be required. The more exposure time; the more heat the chip picks up, creating more noise. Somehow you have to find that happy medium of high sensitivity and low pixel noise. If you can keep the chip cool you can use higher ISO settings. Just a side note here; some people apply dry ice to the camera in an attempt to keep it cool, but this is not recommended for humid conditions, as there is the possibility that condensation will form on the INSIDE of the camera – pos-

Let’s say that you have determined the exposure time to be 60 seconds; and have decided to use gain at level 3 - ISO 800. Set your number of exposures. We personally like 30 exposures for ease of processing, but there’s nothing wrong with trying 10, 20 30 60 or as many as you like. Depending on your situation, it may be better to take 60 exposures at 30 seconds than to take 30 exposures at 60 seconds. Both are 1800 seconds (30 minutes) of data but 60 exposures at 30 seconds will have less thermal noise and fewer problems with your mount mis-tracking. 30 at 60 seconds will have less read noise. There are plenty of trade-offs here, but the important thing is to get as much data as possible. Set your number of exposures. If you want to keep your CMOS chip cool, allow it take a break and select 60 second between exposures. This will also allow you to take a short break if you’re manually guiding. For those who autoguide, this is the time to start your autoguiding program such as PHD. Select a star and insure that the telescope is going stay on the selected star during the time of capture. If you are ready and the autoguider is working well, click “Capture Series.” Step #7: Taking darks. Some people don’t take any darks, flat or bias frames and have excellent images, but if your camera has been imaging all night, it is possible continued >

AstroPhoto Insight

Aug / Sep 2007

http://www.astrophotoinsight.com

Page 19

that the chip may have gotten some hot pixels. You may want to take some dark images to get rid of the hot pixels during processing. Place a lens cap on the imaging telescope. Change the number of exposures to 10 or 20. Change the name of the images to “TrifidDarks” and now click “Capture Series.” The directory will stay the same until you change it. It’s important to take your darks at the same exposure time and temperature of your light frames. If you took 180-second exposures, then take 10 or 20 dark frames at 180 seconds. Step #8: Now that you have all your data, it’s time to process them. Suggestion: Do not attempt to process your images on the same night. If you’re like me, you’ll be fatigued. Taking digital images requires a lot of mental exercise. To get excellent images from your data, it’ll take hours to process images taken with a large format digital camera such as the EOS series. Each of your raw images will be a staggering 15.5 MB. Files that big require a lot of processing time and you should have at least 1 GB RAM. Processing Images Stacking Dark Frame. Start “Nebulosity” and when the window opens, look at the menus on top. The first thing you want to do is combine your darks. You will use this process to combine your flats and bias frames also, however to keep things simple we will do just the dark frames. From the “Processing” menu, select “Align and Combine Images.” A new window will pop up and you will be given a choice of different alignment options. When stacking Dark, Bias or Flat frames you will press the “None (fixed)” button. Click “OK” and a window will come up. At the top of the window find the folder with your images. Select all your Dark frames by clicking the first frame and then press “Shift + left click” on the last Dark frame. All of your darks will now be highlighted in blue. Click “OPEN.” The screen will now show you all of the dark frames as they are opened. If you see a box with a crosshair, close the entire program, reopen and make sure you press the “None (fixed)” button and try again. Once all the dark frames have been opened, a new window will appear. At the very top, make sure your Save-In is the correct folder. Click the filename box and save your stacked darks as “DarkMaster” or whatever name you wish to give it. Now click “Save.” The screen will blink and if you’ll look at the very top, it will show the name of the newly saved file, “DarkMaster.fit” Congratulations! You have just stacked all of your dark frames.

frame. Click “Dark” and select your “DarkMaster” that you just made, then click “Done.” A new window will appear and ask you to select your light frames. As before, click the first light frame “TrifidRaw_001.fits” and at the last frame, press “Shift + left click” and all the light frames will be highlighted. Click “OPEN.” It will now dark-subtract all the frames and label them as “PPROC.” This may take a couple of minutes. Color Conversion and Squaring. A message will appear at the bottom notifying you that the images have been processed. The next step is to convert them to color and square the pixels. Under the “Processing” menu, select “Batch Demosaic + Square”. Now select all of the “PPROC” frames then click “OPEN.” As each image is being processed, Nebulosity will label them as “RECON.” Again, this will take some time so relax or find something else to do. Sometimes watching the process is about as exciting as watching an apple turn brown. However, increasing the RAM in your computer can speed up this process considerably. Now it’s going to get exciting. It’s time to stack! This continued >

Dark Subtracting. Now you are going to apply the “DarkMaster” to the light frames. Under the “Processing” menu, select “Preprocess B&W/RAW. At this time, don’t be concerned with the Bias, Flat or Autoscale Dark. We’re just going to do the MasterDark

AstroPhoto Insight

Aug / Sep 2007

http://www.astrophotoinsight.com

Page 20

is the best part and Nebulosity makes it so easy. From the “Processing” menu, select “Align and Combine Images.” The “Alignment/Stacking” window will pop up, but this time click “Save Stack” for the Alignment Method, select the “Translation” button. Lets pause here for a moment to explain the choice of options. “Translation” means shifting (up, down, left, and right) only. This is appropriate if all your images were taken, one right after another, through, or piggybacked on, an equatorially mounted telescope. “Translation + Rotation” means shifting and rotating. This is appropriate if your telescope was tracking in alt-azimuth mode or if you think the camera or lens may have twisted slightly between exposures. “Translation + Rotation + Scaling” means all of that the above, plus stretching (changing size). This is rarely needed, but it might be appropriate if the image size of your telescope changes when you touch up the focus. Make sure “Adaptive Scale Stack to 16 Bits” is checked. Click OK and a window will pop up. Select all of the “RECON” frames and then click “OPEN.” After the computer has retrieved all of the “RECON” images, the cursor will change to a crosshair. Use the zoom-in button so that you can see a close-up of the stars in the center of the frame. Your goal is to identify the same star on each good frame and delete bad frames from star movement especially if you have manually guided. If you have zoomed in enough, you can actually see the pixels of a dim star. Move the cursor over that star and click it. An hour-glass will appear indication that the image is being processed. When finished, you should see red circle around the same star making it easy to identify (if you selected Translation + Rotation, you will select the first star and when finished with all frames, will ask you to pick the second star located well away from the first star. When the second star is selected, there will be a red circle with an “X” on the first star. The second star will have a green circle). If you have good tracking, you can press Ctrl-Click to keep the guess through all the frames. If you see a frame that has star movement, an airplane or a satellite, you can press “Shift + left click” to skip the frame. This is a good time to take a break, as the computer will now be doing some serious number crunching. When the stacking is finished, a window will appear and as before, name your object such as “TrifidStacked.” Be sure that it is being saved in the same “Trifid” folder. Press “Save.” At this point, you have a stack of all your images. It’s time to clean it up a bit. Odds are there is a dark gray or

black border around the image. This happens during stacking as it moves the frames to line up with each other. Zoom out so you can see the whole image, select the clean area and pull down Crop from the “Image” menu. Also, odds are the background sky is an odd color. Under the “Image” menu, select “Adjust Color Background (offset)”. Don’t play with the sliders. Just click “DONE.” Again, this will require some number crunching so take another break. Now it’s time to bring out the detail. Under the “Image” menu, select “Digital Development (DDP).” Our goal is an image that will be easy to process with PhotoShop or similar programs. At this point, we still have two problems: The midtones are too dark relative to the rest of the image; that is, the image needs gamma correction. Depending on whether they were stretched earlier, the pixel values may not use the full range (0 to 65, 535) of a 16-bit TIFF image. A quick way to address both of these problems is called digital development. This is a combination of gamma correction and unsharp masking (sharpening). continued >

AstroPhoto Insight

Aug / Sep 2007

http://www.astrophotoinsight.com

Page 21

Digital development was invented by Kunihiko Okano. It is a rather quick algorithm to do what would otherwise be a multi-step process. In Nebulosity, the digital development algorithm usually works well with DSLR images. Just go to Digital Development (DDP), and let the algorithm do its work. We have had good results by simply pressing “Done.” Although the defaults are usually reasonable, you can further improve the image by adjusting the sliders; so feel free to experiment with them. Now for a time-saving tip. Before selecting “Levels/ Power Stretch in the next step, with the mouse, left click and hold down to select a small portion of your image. That way, the Levels/Power Stretch processing will only affect the selected portion for faster previews. At this point you probably don’t quite have a finished image. You can do some more stretching using the “Levels/Power Stretch” function. Under the “Image” menu, select “Levels/Power Stretch”. Some imagers suggest sliding the bottom slider left until the background is a very dark gray. The middle slider should be moved about ¼ or 1/3 left. Finally, move the top slider to the apex of the histogram. Keep watching both the image and the histogram as you stretch, trying to bring out all the detail in your image without clipping off the left or right sides of the histogram. This is where you can bring out a lot of the detail and is the last step before post processing in a photo editing program such as “Microsoft Picture It!” or Photo-

Shop. When in Nebulosity, you’re working with the image in 32 bits per color per pixel (96 bits per pixel total). When you go to something else, you may be working at 8 bits per color per pixel. This is why you’ll be better off doing most of your stretching in Nebulosity first. Save your image under the files drop-down menu. Select “Save JPG file as displayed” if you want something that looks like it does on the screen now and “Save 16/bit color TIFF file” if you want something with the most detail possible and if your post-processing software supports it (e.g., PhotoShop CS). At this point, you can go on to post process in other software. You may also want to try some of the tools in Nebulosity. It has tools to bin the image to make it smaller, sharpen the image using a special “Star Tighten” routine, etc. Here, I’ve covered basic pre-processing to get you going, but there are other things in Nebulosity to explore. Before stacking you can automatically grade files based on their sharpness or manually preview a bunch of files to select just the best ones. You can also stack using Drizzle to increase the resolution or using Standard Deviation stacking to decrease the noise (both of these are very demanding on your computer). But, with this tutorial, I hope to help get you get you past the initial challenges and up and going. I’d like to express a special thanks to Kip Pauls, John Armstrong, Michael Covington, Ed Hall, Michael Garvin, Craig Stark, and many others for their support. 

Binning The practice of grouping CCD pixels to create a larger virtual pixel from smaller real pixels. For example, if you bin 2x2, you get a four-pixel array that is treated as if it were a single pixel for output. When taking short exposures, such as for rough focusing, binning increases the sensitivity of the camera, and it also allows for much faster downloads. This allows you to visualize what is in your field of view in 5-10 second exposures and quickly assess your framing and field of view. Binning is also useful when you are using a camera/telescope pair that would otherwise have too much resolution. Seeing limits the resolution you can actually obtain. If the camera has higher resolution than this, it is wasted. Binning reduces the resolution, increases the sensitivity of the camera, and allows you to make productive use of your time. Dark Frame A dark frame is an exposure taken with the shutter closed. A dark is normally taken with the same exposure duration and cooling temperature as the light frame to which it will be applied. The purpose of a dark frame is to record the system noise of the camera. It is subtracted from a light frame to remove that system noise. For best results, take multiple dark frames and median combine them. This will reduce the amount of random noise added to the image by the dark frames. Although a dark frame is used to remove system noise, it can add it a small amount of random noise at the same time. From the CCD Glossary originally compiled by Ron Wodaski and currently host at http://www.skyinsight.com/r/?a=ccdglossary AstroPhoto Insight

Aug / Sep 2007

http://www.astrophotoinsight.com

Page 22

Joe Ulowetz The first Midwest Astrotelescope mounts and equipment. Imaging Conference occurred at He emphasized that successful St. Joseph College in Rensselong exposure imaging depends laer, Indiana, July 13 and 14, on the mount, and that there are 2007. For a first-time event, it many relatively inexpensive opwas well attended and included tions here for the budding astroan impressive lineup of speakphotographer. He also stressed ers and topics. The conference “weight management” of your was held in conjunction with the equipment since the fine guidance EPOCH 2007 Star Party, loneeded for imaging is very sensicated about 14 miles from the tive, and shaving a few pounds St Joseph’s College in conference site. Both the imhere and there can sometimes Renssalaer, Indiana aging conference and star vastly improve results. On this party were made possible by the sponsorship of “20/20 point I concur with Craig, as I have had poor results Telescopes”. when imaging with my home-made Newtonian OTA on a Losmandy mount, but when I switched to a commercial Although future conferences may be held elsewhere, SCT tube half its weight, my guiding became rock solid! my initial impression of this year’s venue was very favorable. Upon arrival at the convention; attendees received Craig also had many practical tips such as offa CD-ROM of the presentations plus some additional balancing your telescope (make it slightly heavier on the reference material, including all past issues of Astroeast side of the pier so the drive gear doesn’t “float”), Photo Insight magazine. The presentations took place and the kind of lubricating grease to use (lithium). I also in the college’s auditorium. There was easy access to took another suggestion of his to heart: learn your equipelectrical outlets which allowed many of us to run our ment. If you’re like me, you probably already feel that laptops and follow along with the presentations’ material on the CD-Rom. Another nice feature of the venue was that the auditorium lights could be dimmed for the presentations (and the geek in me was impressed because I had never seen florescent lights on a dimmer before and didn’t know that was possible). When talking about a subject such as astrophotography, each individual’s skill levels vary so it’s hard to target everyone at once. But speaking for myself, the presentations here were all very helpful to my current skill level, and I learned much that I was able to apply to my imaging techniques immediately. Friday The conference started off with Craig Stark speaking on “Affordable DSO Astrophotography.” Craig develops imaging software products, including “Nebulosity”, “PHD Guiding”, and has extensive hands-on experience with AstroPhoto Insight

Aug / Sep 2007

Conference attendees enjoy a generous workspace at St. Joe’s modern facility

http://www.astrophotoinsight.com

Page 23

you know how to use your equipment, but there is always some new and potentially important aspect that can be discovered by experimentation. I had this advice in mind when I decided to use a clear evening recently after the conference ended to experiment with my off-axis guider instead of imaging. I’d been using this equipment for a year now, but in the technique I’d developed I had forgotten how much the adjustment on the tilt of the pick-off mirror can expand the usable region for finding a guide star. It’s one of those smack-yourself-in-the-head moments, but I’d been concentrating on what worked for me and completely forgot to include this adjustment in my search for guide stars. By following Craig’s advice, I tripled the area from which I can find guide stars. That’s one of the nuggets that made the conference so stimulating and valuable for me. The second session of the morning was Jim Burnell’s “Seeing the Light: The Nature of Astronomical Color and the Use of Filters in CCD Imaging.” Jim, co-author of the book “Handbook of Astronomical Image Processing” and it’s accompanying software AIP4Win, started off by describing how the human eye perceives colors and how color imaging with CCD chips is different. He gave a very “illuminating” example, reminding us that we have all seen photographs showing hydrogen alpha regions as pure red. He then demonstrated with a small lamp and a hydrogen filled tube. When he energized the lamp and the hydrogen began to glow (yielding pure hydrogen alpha emission); surprisingly, it was not red at all! It looked to me more like an electric pink. You can’t argue with a demonstration like that! Jim went on to describe the different types of color filters: colored glass that works by absorbing other wavelengths (inexpensive but poor efficiency), and diachronic filters that work by reflecting other wavelengths (steep cutoff, very effective, but very expensive). A helpful tip is that filters should be par-focal so that refocusing isn’t needed when switching between filters. Another very important point that CCD users should keep in mind is that CCD’s are sensitive far into the infrared, but some color filters do not block IR so you may need to add a separate IR blocking filter. He also pointed out that some colors, such as yellow, don’t exist in space. There are no strong emission lines in that part of the spectrum, so there will never be a yellow Ring Nebula. The 3 strongest emission lines encountered in astronomy are Hydrogen Alpha (656.3 nm), O III (495.9 and 500.7 nm), and S II (672.6 nm). Not only is Hα not pure red, but with color photos from the Hubble Space Telescope, they

Jim Burnell discusses “Seeing the Light”. assign Hα as green and S II as red. A different standard, the Canada-France-Hawaii (CFH) pallet, does the opposite with Hα as red and S II as green. At least everyone uses O III the same way, mapping it to blue. He also explained the use of “G2V” stars in calibrating our images, and that the double star 16 Cygni consists of two stars of this color, and is conveniently located overhead on summer evenings. Jim demonstrated how AIP4WIN software automates calculating the color coefficients from a G2V image, so that we can adjust the color balance on our other images. Since the balance doesn’t shift much over time, he said a yearly check is probably sufficient here. When capturing an LRGB image, he suggested esposing the RBG images using 2x2 binning. This will allow shorter exposures for the color frames, allowing more time to take additional images during the imaging session. And since more images stacked mean less noise, this is a good thing! I changed my technique to do this as soon as I returned home from the conference. After a complimentary lunch, Alan Chen spoke on “Imaging with a One-Shot Color Camera”. He explained that these cameras operate by using a color filter matrix over the detector. The design most often used is the “Bayer matrix”; and since the human eye is most sensitive to green light, the Bayer matrix has twice as many green pixels as red or blue. There is some contention in the astroimaging world about whether one-shot color cameras can obtain as high a resolution as a monochrome (unfiltered) camera. The answer is not simple. Although the Bayer matrix exposes each pixel to just one color, the resulting image includes data for all 3 colors for each pixel continued >

AstroPhoto Insight

Aug / Sep 2007

http://www.astrophotoinsight.com

Page 24

useful. In these conditions he also uses a lower shutter speed, and a green filter. He noted that brief periods of this level of seeing can be found sometimes right after sunset or before dawn. He noted that conditions of 9+ Pickering require a tropical isle (don’t laugh, I know of at least 2 people who post great planetary images online and live in the tropics). Under these conditions it is possible to use an actual webcam because all the colors are sharp and good seeing tends to come in waves.

Alan Chen shares his insights on One-Shot Color Cameras. location. The process, called debayering, uses the data from neighboring pixels to estimate the colors that a pixel wasn’t sensitive to. There are a half-dozen different algorithms for how to do debayering, all resulting in different blends and making the issue of actual resolution very complex. And this is all just for the raw image data before any processing is done. Alan said he often takes three to five evenings to process one image in Photoshop, and that he often restarts three to five times on an image. Next on the agenda was Alan Friedman with “The Art of High-Resolution Planetary Imaging.” This was a topic especially close to my heart since I can usually be found outside on any clear night when planets are visble, trying to squeeze every last bit of planetary detail from my equipment. Alan recommends using the longest effective focal length possible; with his 10” f/14.6 scope he usually images in the range of 7-11 meters using Barlow extension. He uses a DMK camera from The Imaging Source which produces 60 frames per second. He also uses filters for color information since he feels that a Bayer matrix limits resolution.

After capturing images, he stacks them and uses multiple alignment points, sometimes as many as 30 different points in an image. He also creates a luminance frame while processing Jupiter or Saturn by mixing equal amounts of red and green to make the luminance layer, and then combines this with the RGB color data in a separate layer in Photoshop using Luminance blending. In my personal attempts to improve planetary imaging results, this talk gave me lots to think about, especially with Mars returning this Fall. I’ve been using a webcam for several years and getting good images, but to do even better I will need to use a monochrome camera. To finish out the first day of the conference, Warren Keller spoke about “The Art of Image Processing.” He reminded us not to ignore composition and framing when we image, and that something as simple as rotacontinued >

He often starts his observing sessions with a double star to help judge seeing conditions, which for him are usually 5-6 on the Pickering scale from his home in Buffalo, NY. In such conditions, only 15-20% of the video frames are usable. He explained that “in planetary imaging, a color filter acts to settle the air.” For average seeing conditions, he recommends using a red or IR filter to see the most detail. For nights with seeing at the level of 7-8 Pickering, which he said is the best we can expect here in midnorthern latitudes, he finds 25-35% of the video frames

AstroPhoto Insight

Aug / Sep 2007

Alan Friedman (left) and Alan Chen

http://www.astrophotoinsight.com

Page 25

tion of the target can result in a more artful image. He also warned against over sharpening the image, especially for the stars. He prefers a softer look for the stars and recommends only sharpening the luminance layer. Warren actually separates the stars from a nebula in Photoshop to process them separately;. A tool he suggested for Photoshop is “Neat Image” which will automatically smooth out background noise. He also uses Gradient Xtractor as do many other astroimagers, although he pointed out that this tool needs sky background in the image to work, so a full-frame image of M42 would not benefit from it. That wrapped up the first day of the conference. On the lawn outside, someone set up a Coronado solar scope and several of us spent time admiring solar prominences. Later that evening people car pooled over to the EPOCH star party site for dark sky observing. A few clouds moved through at times that evening, but there was plenty of sky available to explore. Saturday The second day began with Adam Block speaking on “A Question of Contrast”. For someone like me, still struggling to figure out how to process my images, this was a great session. Adam demonstrated techniques using Photoshop, which has become the de-facto standard for image processing among most astronomers. I had Photoshop running on my laptop and was able to follow along with Adam’s techniques, applying them to an image that I had taken a few days before. One key technique I learned here was that limiting the range of

Warren Keller (left) with speaker and co-host Jeff Terry brightness in an image – for example, setting the black level – should be done close to the end of processing, not in the beginning like I have been doing. He emphasized that every adjustment we make to an image reduces the amount of data contained in the image. He demonstrated what happens when the dynamic range gets too small: the comb effect, also referred to as “posterization”. In terms of workflow steps, once images are captured in a software like MaxIm DL, Adam does the dark and flat calibration, hot pixel removal, and stacking/ combining images still in MaxIm. He ends up with a luminance image and a separate color (RGB) image. He then takes these two images into Photoshop for the remaining processing. He said that most of the effort should take place on the luminance image, and that little work should be done on the color image. The two images are combined at the end using the layer blending mode of “luminance.” The amount of material Adam covered was necessarily vast because of the numerous tips and techniques necessary to bring out the best in an image, but he just focused on the subject of contrast in this presentation. In this short report, I can only give a flavor for Adam’s material. The best way to go into this subject in more depth is with the tutorials Adam sells on DVD. He has an entire volume devoted to Photoshop, two other volumes on image processing topics, and has just introduced a new volume covering use of the CCDStack program.

Adam Block (right) with conference co-host, Al Degutis

AstroPhoto Insight

Aug / Sep 2007

The final afternoon of the conference started with Greg Piepol speaking on “Solar Imaging Techniques”. Greg shared his enthusiasm for the sun with us, describing different equipment and aspects of solar observing.

http://www.astrophotoinsight.com

Page 26

For example, Coronado sells a solar telescope to view the sun in the Calcium K emission line; Greg pointed out that this line is technically outside the range of human vision and older observers sometimes see nothing through this telescope, although photography would still work. He also pointed out that solar observing can be tough on the equipment, and that solar filter can have a limited life so check the warranty. The final speaker was conference co-host Jeff Terry speaking on “iCCD and Remote Imaging.” Jeff is the author of the Macintosh software iCCD and initially developed it for his own use to automate the steps he was doing repeatedly. He makes the software freely available to the Mac community. Jeff proudly pointed out that the Sloan Apache Point 3.5 meter observatory is entirely run by Macs and is just one example of Macs used in astronomy. One of Jeff’s suggestions was that deconvolution should only be done on a luminance layer; it should not be run against the color data. He also pointed out that deconvolution adds noise for each iteration run, so don’t use more iterations than needed. Plus, if you only have RGB frames without separate luminance data, you can create a synthetic luminance image from the color data and then sharpen that. The other part of Jeff’s talk, remote imaging, was

Even after the conference concluded, the speakers continued the discussions in the hallway. (Left to right: Adam Block, Jeff Terry, Warren Keller, Alan Friedman and Craig Stark) dark skies as well as to southern sky targets. Jeff makes the valid point that before image processing can be done, you first need a raw image to process, and a facility like Global-Rent-A-Scope allows anyone - regardless of personal equipment or location - to capture high quality raw image data. There is a cost involved in using this equipment, but it is much less than it would cost to travel to a site like that for many of us. This is certainly a good option to keep in mind for anyone who wants to get into astrophotography but doesn’t currently have the equipment or location to do it. Overall, this conference covered everything from capturing the initial image data, through processing it, with adding some artful touches, and lots of topics in between. In addition to the presentations, it was a great opportunity to meet with – and learn from – other astroimagers. I left the conference filled with energy and ideas that I couldn’t wait to try out myself. I certainly hope the conference continues as a yearly event because, at least for me, it was very educational and yet still only just scratched the surface of this complex and dynamic topic.

From left to right: Warren Keller, Craig Stark, Greg Piepol, Alan Friedman, Jeff Terry, Adam Block and Al Degutis about his experience renting time on remote telescopes for taking astrophotos. “Global-Rent-A-Scope” is a company with remote-controllable telescopes located in New Mexico, Israel and Australia. This gives users access to

AstroPhoto Insight

Aug / Sep 2007

Joe Ulowetz, hooked on planets since the Mars opposition in 1971, has been actively observing ever since. He works as a computer programmer to support his astronomy habit and family (although not necessarily in that order). He started actively pursuing planetary astrophotography, especially Mars, 3 years ago from his back yard in Northbrook, Illinois. For more information on the 2008 conference go to: http://www.skyinsight.com/maic/

http://www.astrophotoinsight.com

Page 27

These images can also be seen at http://www.skyinsight.com/gallery See page 31 for information on how to submit an image for the gallery.

M82 with a touch of Ha - Franck Bugnet Taken with a SBIG ST-2000XM through a RCOS 12.5" at f/9 on a Paramount ME mount with Astrodon filters from a fairly light polluted area, in France, 30 km south of Lyon. Taken in Feb during 3 consecutive excellent nights (seeing). About 30 hours total exposure, including 9h of Ha. Processing guidance by Jay GaBani.

AstroPhoto Insight

Aug / Sep 2007

http://www.astrophotoinsight.com

Page 28

Barnard's Star Rob Johnson Barnard's star in Ophiuchus taken at intervals over the past 16 years. The latest images were taken through a 14" Newtonian reflector with an Artemis 285 CCD camera with a 10sec exposure. Earlier images had been taken through the same telescope but with Hypered Tech Pan film (1991) then a Starlight Xpress SX Parallel port CCD for the intervening years.

Antares region John Drummond T a k en w it h a S B I G STL11000M CCD and a Sigma 70-200mm f2.8 APO lens. I took 10 minute sublights - 15 x Luminance, 8 x RGB. The camera was 20*C and 1x1 binning was used. It was mounted on a Meade Starfinder 16" Newtonian. Acquired and stacked in Maxim; processed in Photoshop.

AstroPhoto Insight

Aug / Sep 2007

http://www.astrophotoinsight.com

Page 29

M8 - Michael Fulbright Taken with a modified Canon Rebel Xt, Astronomik CLS filters on a GSO 200mm f/4 Newtonian with Baader MPCC and an Atlas EQ-G. Guiding with a Toucam webcam in tandem mounted Orion ST80 refractor, acquired with Maxim DL. Exposure: 22 x 1 min at ISO 800. Processed with DeepSkyStacker and postprocessing with Photoshop CS3 and PixInsight Standard Beta Sunset - Nedeljko Markovic Taken with a Nikon FE2 and Fujicolor film; Russian MTO1000 Macsutov-Cassegrain telephoto lens

Bubble Nebula (NGC 7635) Robert Vanderbei The image was taken with a Starlight Express SXV-H9 camera on a 10" RitcheyChretien from RCOS. The image combines 240 minutes of H-alpha and 120 minutes of O-III. Raw images were guided 20-minute exposures. L=Ha+OIII, R=Ha, G=OIII, B=OIII.

AstroPhoto Insight

Aug / Sep 2007

http://www.astrophotoinsight.com

Page 30

Shuttle STS-117 Launch - Doug Neal Taken from the NASA Causeway on June 8, 2007 - 7:38 PM EDT with a Canon XT and Sigma 70 -300mm lens from approx. 6-miles across the lake from Pad 39 A.. Exposure: 1/500 sec., F10 and ISO 200 - Lens set @ 300mm.

We invite you to submit your images to the Readers' Image Gallery. The number of images included in each issue will depend on the amount of space available. We cannot guarantee that your images will be published, but if you don't submit any we can guarantee that they won't. If your image is used for the magazine cover you will be awarded $25 USD. Submit your images to [email protected] and be sure to include your complete contact information (full name, mailing address, phone number, and e-mail address), as well as the details of your imaging setup (kind of telescope, camera, film, or filters used, etc.), exposure times and any special techniques involved (e.g. tricolor, mosaic, multiple exposures, stacking techniques, etc.).

AstroPhoto Insight

Aug / Sep 2007

http://www.astrophotoinsight.com

Page 31

M81 - Franck Bugnet Taken with a SBIG ST-2000XM through a RCOS 12.5" using an AP reducer (working at 0.76x) on a Paramount ME mount with Astrodon filters from a fairly light polluted area, in France, 30 km south of Lyon. Exposures: 34 x 10 min for the luminance and 17 x 10 min per color; binning 2x2.

AstroPhoto Insight

Aug / Sep 2007

http://www.astrophotoinsight.com

Page 32