Sunday, April 20, 2014

Image Scale of 400mm Lenses with Kenko 1.4x

This post shows some tests of the "reach" of a Canon 60D using a combination of a 400mm lens and a 1.4x extender compared to that of a SX-50 HS super-zoom camera. Note that the focal lengths printed on the lenses need to be interpreted correctly and cannot simply be compared. First, the values printed on DSLR lenses are approximate;  no company is going to market a 396.45mm lens. Second, on non-DSLR lenses, the printed focal length typically refers to the size lens which, when fitted to a full framed DSLR (35mm sensor), would give the same field of view.

Since the focal length, image aspect ratio, sensor size, and number of mega-pixels all differ between the two systems, I needed a meaningful basis of comparison. Since my primary application is birding, where images are almost always cropped, the field of view is not relevant. What is important is the number of pixels "on the bird," or resolution of the image. In astrophotography, this is called the "image scale" and measures how much of the target is covered by one pixel. Thus, the question I want to answer is:
What lens would I need to use with my DSLR to give the same reach as the SX-50 HS?
To test the effective reach empirically, I mounted the cameras on a tripod and placed a target about 25 feet away. I took an image with each configuration and imported them into Photoshop as separate layers. This aligned the images based on matching pixel size. Since the SX-50 has fewer pixels on its 4000 x 3000 sensor, its image frame, shown as a green rectangle, is smaller than that of the 5184 x 3456 sensor of the Canon 60D, shown as a purple rectangle. To measure relative resolution between any two images, I used the Photoshop measuring rule tool to determine the distance in pixels between two well defined points in the target image. For a given camera, this pixel distance varies linearly with the focal length.

Here are the three test images:

Canon 60D, EF 400mm f/5.6L with Kenko 1.4x MC4 extender 
Canon 60D, EF 100-400mm f/4.5-5.6L with Kenko 1.4x MC4 extender
Canon SX-50 @ 1200mm.
The first result is that the focal length of the two DSLR lenses differ by about 5%; neither is likely exactly 400mm. However, if we assume for comparison that the EF 400mm f/5.6L prime lens is actually 400mm, then the EF 100-400mm f/4.5-5.6L zoom lens fully extended would be 380mm. With the extender attached, the prime lens would be 560mm and the zoom lens 532mm.

The second result is that the image of the SX-50 HS has about 10% less resolution (is more zoomed out) than the Canon 60D - EF 400mm f/5.6L - Kenko 1.4x combination. So again, if the prime lens is 400mm, the image produced by the SX-50 HS has a resolution equivalent to the Canon 60D with a 504mm lens attached or a 360mm lens withna 1.4x extender. Thus, using this camera, either 400mm lens with an extender gives more reach than the SX-50 HS.

Of course, results will differ for other DSLR cameras, including those with APC-C sensors having more or less than the 18 mega-pixels of the Canon 60D. For example, our Canon 450D has an APS-C sensor with 4272 x 2848 pixels. Each pixel is therefore 21% larger. Using this camera, the SX-50 HS would have a resolution equivalent to using a 612mm lens or a 437mm lens with a 1.4x extender. So, in contrast, using this camera, either 400mm lens with an extender gives less reach than the SX-50 HS.

So, in answer to my question, the reach of SX-50 HS is equivalent to either:
  • 20.2 mega-pixel APS-C with 480mm lens ( 343mm lens with 1.4x )
  • 18.0 mega-pixel APS-C with 504mm lens ( 360mm lens with 1.4x )
  • 12.0 mega-pixel APS-C with 612mm lens ( 437mm lens with 1.4x )
Note on Auto-Focus - On both the Canon 60Da and the Canon 450D, auto-focus worked reasonably quickly on either 400mm f/5.6 lens when combined with the Kenko 1.4x MC4, despite the fact that the resulting f/8 aperture is not officially supported by the phase-based auto-focus circuitry. Though it works, I found the auto-focus to be picky about the target and lighting conditions. In a poor contrast target, it would hunt and fail to lock-on focus more often than the naken lens. Contrast-based focusing, i.e. LiveView Focus, should not be affected ... but is limited to non-moving tripod targets. Would be fine on the Llano eagle nest, for example.

No comments:

Post a Comment

Relevant comments and questions are welcome but submissions with spam-links will not be published.