High Plains Drifter

A test of super-ultra long exposure with Nikon D3.  After work on Sunday night, July 6th, I decided to drive south to my favorite nearby photography location — Big Basin Prairie Preserve.  I wanted to give the Nikon D3 + 14-24mm f/2.8 lens combination a try in some really low light conditions — the night sky.  I really was curious to try a super-ultra long exposure (greater than 30 minutes) to see how the sensor noise was leaving the shutter open for so long.  Digital imaging sensors will accumulate "hot" pixels the longer the shutter is left open.  As technology has improved in digital imaging sensors, the duration one can leave the shutter open without introduction of noise has greatly improved over the years.  I remember my very first digital camera was a point and shoot Nikon Coolpix 950.  You couldn’t take 8 second exposures without introducing all sorts of horrible "hot" pixels dotting your image.  When I then got my first digital SLR (Nikon D70) one of the first things I noticed was the absence of hot pixels at 8 to 30 seconds!  They were absolutely clean of noise at low ISO.  That being said, image sensor noise ("hot" pixels) started showing up at super long exposures of 5 to 10 minutes.  Also, this thing called "amp glow" also started showing up in the form of a pink blotchy area at the corner of a frame due to the sensor heating up.  Internal camera digital noise reduction can eliminate much of this by a simple technique called frame subtraction (whereby taking another black-frame exposure of the same time length and subtracting that from the original image).  The only kind of photography really needing this kind of long exposure is star trail photography, which require 30 minutes to 3 hours worth of exposure time, depending on how much trailing you are interested in for your shot.  

Along comes the D3.  I was really curious how long I could keep the shutter open on this camera before the hot pixels and/or "amp glow" became too much of a problem.  On the old D70 I had, the amp glow started really becoming a problem at about 8 to 10 minutes exposure time.  Below are two versions of the same image from the D3 I shot last night with a shutter speed of 3,650 seconds (1hr, 0min, 50sec).  The first is the processed version adjusting the exposure compensation, levels/curves, noise reduction in Adobe Lightroom:

In the 2nd image, which is unprocessed, you can barely pick out the slight purple-ish hue at the very bottom of the frame.  You really have to look closely, but it’s there.   This is a ONE HOUR exposure, and you have to squint to find the amp glow.  I think Nikon finally got it right with the D3 and hot pixel/amp glow digital sensor issues for exposures longer than 10 minutes.  Some of the star trails, though, have some holes in them, as if it wasn’t exposing properly for a bit during the one hour.  I can’t entirely explain what was going on there.

But here is what really excited me…

Milky Way looking south.  Nikon D3, 14-24mm f/2.8 lens @ 14mm.  Exposure 99s @ f/2.8, ISO 1250 

I couldn’t believe the detail I was able to capture here.  Of course, the artificial lighting on the horizon was a bit of a pain, but given this shot, I am excited to get out on another clear, moonless night absent of any horizon lights to shoot the Milky Way again.  A full album of images from last night (just 7 images), is available. 

Mike U 

Incredible High ISO Performance!

I did a poor man’s experiment to see how well very high ISO values of the D3 compared to my D200.  I was blown away.  The end result was that I was getting 640 ISO (from the D200) performance at around 3200 ISO!!  This is in the territory of 2 to upwards of 3 stops performance improvement in noise handling!  

In the example above, I photographed one of my large prints to use as a gauge for noise.  The top half of the image is shot with the D3 at an ISO setting of 2000.  In the bottom half, I shot with the D200 at an ISO setting of 640.  The noise on the D3 is quite a bit finer, which is easier to clean up.   More thorough reviews/tests of the D3 High ISO performance can be found below:

http://ishootshows.com/2008/01/07/impressions-on-the-nikon-d3-high-iso-performance/

http://www.kenrockwell.com/nikon/d3/vs-5d-180mm.htm 

All about speed! 

Here is an example of two images of the same scene — The first is with the Nikon D3 and the 14-24mm f/2.8 lens at full wide angle.  The second image is with the Nikon D200 and the 12-24mm f/4 DX lens, also at full wide angle.  The test here is to demonstrate a similar shooting situation in a low lit storm environment and you couldn’t set up a tripod.  These are both handheld, both shot at the minimum possible exposure for handheld without getting destructive motion blur (usually the ‘rule of thumb’ for minimum hand-held shutter speed is the reciprocal of the focal length of the shot… i.e. for 14mm lens, about 1/15s exposure).  In the two shots, I tried to create the same exact exposure image with "reciprocal of the focal length" shutter speeds:

The two images below are 15% of the full size, embedded JPEG file from the RAW (using DCRaw to retrieve the JPEG)

Nikon D3 with Nikkor 14-24mm f/2.8 @ 14mm.  1/15s, f/2.8, ISO 400. 

Nikon D200 with Nikkor 12-24mm f/4 DX lens @ 12mm (18mm virtual focal length).  1/20s, f/4, ISO 1250

This is a 100% crop showing the more detail of the first image with the D3 + 14-24mm f/2.8.

This is a 100% crop showing the more detail of the second image with the D200 + 12-24mm f/4. 

Obviously, the first image is much cleaner thanks to the lower ISO.  This is due to the lens, not really the camera.  The next test in part III will be an ISO side-by-side comparison using the same lens on the D3 and D200.  This will give me an idea how far I can push the ISO out in the field to my liking.

Below are some examples comparing Nikon D3 + Nikkor 14-24mm lens with Nikon D200 + Nikkor 12-24mm DX lens.  The Nikkor 12-24mm DX lens was built solely for Nikon DX SLR cameras (the smaller sensors).  This is a smaller lens, and it will only project to DX size sensors.  Now, a DX lens will still work on an FX digital camera, like the D3, but it severely limits the capability of using an FX camera.    

All the images below are 15% of the full size, embedded JPEG file from the RAW (using DCRaw to retrieve the JPEG) 

Nikon D3 with Nikkor 14-24mm f/2.8 lens @ 14mm.  1/640s, f/8, ISO 250, focal length 14mm

Nikon D200 with Nikkor 12-24mm f/4 DX lens @ 12mm.  1/640s, f/8, ISO 250, virtual focal length 12×1.5=18mm 

Nikon D3 with Nikkor 14-24mm f/2.8 lens @ 14mm, but this time I set the D3 to capture in DX sensor size, effectively turning the D3 into a ~6 megapixel camera.
1/640s, f/8, ISO 250, virtual focal length 14×1.5 = 21mm

Now, here is what a DX lens (the 12-24mm f/4 in this example) would look like when used on an FX camera (Nikon D3 in this example) in FX mode!

To spin this another way — You would need a ~9.2-16mm on a DX camera to replicate the angles of view that this 14-24mm lens gives on FX and film cameras.