This script now has its own page, Sky Limited Exposure

I’m evaluating MaxIm DL and PixInsight at the same time because I only want to purchase one image processing package. The first significant obstacle I’ve come across is that PixInsight doesn’t seem to scale dark frames very well. I’ve run the exact same calibration process on the same image and I get completely different results with the two packages. I’m testing using a 600 second raw image, a 300 second master dark frame, and a master bias frame.

This is a picture of the half moon from early in January 2009. This was taken with a Canon 10D through a Takahashi FS102 refractor at prime focus.

There have been a lot of reports of the 5D’s trouble with bright point lights. I recently updated to the new firmware (version 1.0.7) and took some test shots with the 5D Mark II and a Takahashi FS102 refractor. I also wanted to test out the sensitivity at various ISO values. All these pictures were taken from my driveway on a day with fairly bad seeing and transparency. I got out a ladder and threw a blanket over one of the street lights but it didn’t help much with the sky glow.

I recently had some of my old astrophoto slides scanned with a Noritsu slide scanner so I can put them into permanent storage and finally move into 100% digital photography. I’ve learned a lot about image processing and the tools are much better now than they were when I originally took the photos. There are three major changes that affect the quality: The slide scanner is much better (although the lab gave me JPG files instead of 16bit TIFF files, what the hell?

It’s been at least three years since I’ve done any astrophotograpy but last night I lugged out the equipment and shot a few frames. I only made it out as far as the driveway and the sky isn’t very dark in northern San Diego but it’s amazing what kind of results you can get with a Canon 10D. Traditionally you would use either slide film or a cooled, black and white CCD sensor.

I ended up at the Red Bull Air Race last weekend. The tickets were $25, but we just stood at the convention center for free and watched from high-up. These guys are crazy, it was like MotoGP plane racing.

My Gallery2 embed plugin for MovableType kind of sucked before. It dove right into the SQL database to grab information which makes it brittle if the schema changes. I’ve updated it to use the GalleryEmbed API to grab image blocks. It works a lot better. I also threw in some support for floating thumbnail boxes to make the pages look a little more interesting. I’ve started uploading some macrophotographs into a new album.

While playing around with the Aperture database I decided to try to make something useful. The result is an Automator workflow that looks for duplicate master image names. Here’s the result. Disclaimer: it’s not pretty and the output goes so TextEdit instead of some fancy Cocoa application or Aperture album. Please give it a try and let me know if anything goes wrong. Find Aperture Duplicates v0.1

Most of the time we want more resolution in our images so they can be printed on bigger and bigger formats. However, for web publishing it’s important to have quality downsized version of our images. A question about how Aperture does this was raised on the Aperture discussion group about image downsizing. I used a few test images from Oshyan Greene’s website to find out. Here are the results. It’s hard to tell exactly what algorithm Apeture is using, but it’s probably bilinear filtering since it’s very easy to do using the GPU.

I’ve tried a couple Aperture folder hierarchies and this is the one that seems to work the best for me. Initially I imported my entire image collection into one library with subfolders for the date. I quickly realized that I was going to hit the 10,000 image-per-project limit. I’ve now decided on a different organization that has worked really well. The basic idea is major category folders at the top, year folders under that and all projects below that.

I’m working on decoding the Aperture database structure. Just playing around in the database has been very interesting. Update: this was written with Aperture 1.0 in mind. This is probably completely inaccurate now These tables are incomplete. However, they do contain the columns that I used in my test queries. Update: I just found a cool tool called SQLite Database Browser that makes it much easier to look around the db than using the command line sqlite3 tool.

Apple’s new Aperture program is awesome. It’s easily the best image cataloging program available, and it’s in version 1. However, the performance leaves quite a bit to be desired. Here are the things I’ve discovered. I used a combination of the fs_usage tool and the Thread Viewer tool to profile Aperture. Both excellent tools are included in the Developer package. I wanted to find out why it always takes 10 seconds for the search filter dialog to open up, causing the application to become unresponsive the entire time.

Just added some new surfing photos. Shooting against the sunset makes the color a bit odd, but not bad overall.

After getting sick at the XBox developer conference (and dosing up on plenty of dextromethorphan) I decided to move my website content over to MovableType. I had been using a home-brew system but it was nowhere near as feature rich as the MT system. However, I’m really happy with my old photo gallery system and wanted to be able to show photos from the gallery in my MT entries. It only took me about three hours to adapt my image embedding tags into a MT text formatting plugin.

This is a tutorial on replacing the bearings on the Losmandy G11 telescope mount. This may or may not improve the tracking ability. It’s kind of fun to do anyway. The Losmandy G11 mount is a great piece of hardware. However, the included bearings seem to fall short when compared to the rest of the mount. Unfortunately these bearings can play a large role in the tracking accuracy of the mount.

Takahashi FS102 Refractor This Apochromatic telescope provides very sharp images with no color fringed stars. When reduced to F/5.9 you get about 3 degrees of coverage in the sky. I prefer to take pictures of large, wide field objects and this is just about the right focal length for those objects. I also use the extremely well made Takahashi camera angle adjuster to make framing easy. I use a collection of camera bodies and lenses along with the telescope tube for pictures.

Because most CCD cameras made for astronomy are monochromatic it is neccessary to place a series of filters over the camera and take seperate images that each cover a different spectral region. These images can then be combined to create a “true-color” image. The problem with placing filters over the camera is that it reduces the amount of light that strikes the chip. This means you need to take longer exposures.

Update: I now use Noise Ninja for all my noise reduction. It works really well. Using Photoshop Adapted from a paper by mwcook@cris.com The idea behind this method is that the low-light areas of the photograph need to be smoothed in order to remove grain artifacts while the bright stars and nebulosity should remain unchanged. To achieve this in Photoshop you need to construct a mask for the Gausian Blur filter.