www.kengchang.com/Sensor.JPG
This is 100% zoom of ashot with F4.5, ISO 100, exposure time 21 minutes and 20 seconds.
It looks alright if I didn't zoom in but looks horrible if I look at it actual size.
www.kengchang.com/Sensor.JPG
This is 100% zoom of ashot with F4.5, ISO 100, exposure time 21 minutes and 20 seconds.
It looks alright if I didn't zoom in but looks horrible if I look at it actual size.
The dots? Thats sensor noise. To get a long exposure like that without nosie you will want to use something like a ti-86 calculator (google) to automate the shutter release to take multiple short exposures, to combine them later in photoshop (same star trail length, less noise).
Before you get right to it, I'd read up on issues that are had with automated exposures such as star trails that appear to 'pulse' due to the program not triggering the shutter release quick enough compared to the rotation of the earth.
Those are called hot pixels. There are three ways to remove them:
* Enable long exposure noise reduction in-camera. This is not really "noise reduction" in the sense we think about it most of the time. It's actually darkframe subtraction. This is not a good solution for star trails because it requires doubling the exposure time.
* Do your own dark frame subtraction. The only raw converter software I know of that supports this currently is the kind built for astrophotographers, and they tend to be difficult to use. (This is what I do.)
* Use a raw converter that has a hot pixel killer built-in. Adobe's raw converter does this automatically, and others, like Bibble, give you the control over how the hot pixels are killed.
The last solution is probably the best for you.
Originally Posted by nickds7
I agree that 20 minutes is a bit on the long side and the OP would be better off with an automated shutter release of some kind (I use a $30 chinese one off ebay).I However, that wont really solve the hot pixel problem because many cameras have hot pixels even at very short exposure lengths (e.g. 3 seconds). So ideally the OP would investigate something to solve the hot pixels and something to do interval shutter release.
I use the first of Daniel's three solutions (in-camera noise reduction) and find that it works surprisingly well. The downside is that if you do a 20min exposure, the camera must do a 20 min dark exposure when you are done (ie, it takes another 40 minutes to take the picture).
Taking several shorter pictures and stacking does alleviate this somewhat, but you must use a timer. This is because if you take a 60 second picture and then a 61 second picture in succession, the camera does two darks. If you take two 60 second pictures, it does only one and applies it to both. With a timer, you could have taken seven 3-minute exposures, then waited only 3 min more. (And then, of course, stacked the pictures).
Many people prefer stacking because you tend to get less read noise (using say, ISO1600 for 16 exposures as opposed to ISO100 for one), and because you get more range. More importantly for me, if I mess up and get a bad frame, I waste eg 3 minutes of exposure rather than 20 (though I suppose if you're capturing star trials, missing a frame does pose a problem).
Ok, I just wasn't expecting this at ISO 100. I was using timer to do time lapse.
Here's the result, please watch it in 1080P
http://www.youtube.com/watch?v=VYoPsmb1OpU
It is a common feature of long exposures. It's why people shell out $10K+ for frozen CCD's. (Lowering the temp helps a lot).
Here's the result, please watch it in 1080P
I love it! Were those night shots all 20 min exposures? Must have been pretty wide angle.
Yep - most of my microscope cameras have Peliter-cooled CCDs. I ran across this DIY solution for Canon cameras...
Okay, that is cool. I've never seen anything like that. If I had even a little conviction or self confidence, I would try something like it
Posting Permissions
Reply With Quote