Star Test of Canon 70-200mm/f4.0 L USM Lens

This post summarizes my testing of the Canon EF 70-200mm/f4 L USM lens for use in astrophotography. This lens has been consistently well reviewed on-line for its sharpness and low chromatic aberration.  At around $700, it is also very affordable for an "L" series lens.  I would very much welcome comments by other users of this lens as to whether the imperfections described are expected.

Here are the MTF curves which Canon published.  As I am using an APS-C sensor, the region of the curve beyond 13mm does not matter for me.

Imaging Tests

For this set of tests, I used a Canon 60Da camera mounted on an equatorial mount and auto-guided to remove any star-trailing considerations. I gathered images at both ends of the focal-length range, 70mm and 200mm.  For each, I tested the lens with the aperture wide open at f/4.0, reduced by a full stop at f/5.6, and reduced slightly at f/4.5.  Each of these images consisted of a 120 second exposures at ISO-800 centered on the bright star Procyon.

At each focal length, I focused the lens wide open using the electronic lens control in Astro Photography Tool, minimizing the FWHM metric.  I did not refocus for the stopped-down tests.  I captured and saved the images in Canon RAW format, debayered in MaximDL, and saved as uncompressed 16 bit TIFF files.   In Photoshop, I adjusted the black-point and applied a gamma correction, brightening the f/4.5 and f/5.6 images further to compensate for the differences in EV.  I cropped various small regions ( about 8% of full image width) along each of the diagonals and saved these as high-quality JPEG files.  A sub-set are presented below.

Full Frames

The following two images show the full APS-C frames at 200mm and 70mm.  This gives an impression of the overall field of view

Canon 70-200/f4 L USM at 200mm f/4.0 (full frame)

Canon 70-200/f4 L USM at 70mm f/4.0 (full frame)

200mm at f/4.0

The following three images show different regions of the 120sec ISO-800 frame taken at 200mm f/4.0.

Canon 70-200/f4 L USM at 200mm f/4.0 (center of frame)

Canon 70-200/f4 L USM at 200mm f/4.0 (half way from center to bottom-left corner)

Canon 70-200/f4 L USM at 200mm f/4.0 (bottom-left corner)

200mm at f/4.5

The following three images show different regions of the 120sec ISO-800 frame taken at 200mm f/4.5.  I believe there may have been a glitch in this test as the focus is appreciably worse than in either the f/4.0 for f/5.6 test.

Canon 70-200/f4 L USM at 200mm f/4.5 (center of frame)

Canon 70-200/f4 L USM at 200mm f/4.5 (half way from center to bottom-left corner)

Canon 70-200/f4 L USM at 200mm f/4.5 (bottom-left corner)

200mm at f/5.6

The following three images show different regions of the 120sec ISO-800 frame taken at 200mm f/5.6.

Canon 70-200/f4 L USM at 200mm f/5.6 (center of frame)

Canon 70-200/f4 L USM at 200mm f/5.6 (half way from center to bottom-left corner)

Canon 70-200/f4 L USM at 200mm f/5.6 (bottom-left corner)

200mm Reflection Artifact

In each of the 200mm tests, I noticed an imaging artifact near Procyon.  The image below shows the cropped central region.  The diagonal black lines cross at the center of the frame.  It is clear that the artifact is symmetrical with Procyon and might be some form of reflection artifact.

In performing the tests, I had not removed the Pro-Master 67mm Digital Protection Filter which is attached to the front of the lens.  I will need to repeat the test with this filter removed in order to see if it is the source of the artifact.

Possible reflection artifact, opposite from bright star Procyon, at 200mm

70mm at f/4.0

The following three images show different regions of the 120sec ISO-800 frame taken at 70mm f/4.0.

Canon 70-200/f4 L USM at 70mm f/4.0 (center of frame)

Canon 70-200/f4 L USM at 70mm f/4.0 (half way from center to bottom-left corner)

Canon 70-200/f4 L USM at 70mm f/4.0 (bottom-left corner)

Canon 70-200/f4 L USM at 70mm f/4.0 (bottom-right corner)

70mm at f/4.5

The following three images show different regions of the 120sec ISO-800 frame taken at 70mm f/4.5.

Canon 70-200/f4 L USM at 70mm f/4.5 (center of frame)

Canon 70-200/f4 L USM at 70mm f/4.5 (half way from center to bottom-left corner)

Canon 70-200/f4 L USM at 70mm f/4.5 (bottom-left corner)

Canon 70-200/f4 L USM at 70mm f/4.5 (bottom-right corner)

70mm at f/5.6

The following three images show different regions of the 120sec ISO-800 frame taken at 70mm f/5.6.

Canon 70-200/f4 L USM at 70mm f/5.6 (center of frame)

Canon 70-200/f4 L USM at 70mm f/5.6 (half way from center to bottom-left corner)

Canon 70-200/f4 L USM at 70mm f/5.6 (bottom-left corner)

Canon 70-200/f4 L USM at 70mm f/5.6 (bottom-right corner)

Conclusions

At 200mm, this lens performed well, even wide-open.  The stars are round to the corners of the frame and have consistent focus.  There is a hint of chromatic aberration in brighter stars which was more evident away from center.  Stopping down slightly to f/4.5 improves the CA but does not appreciably change the sharpness or shape of the stars.

Imaging wide open at f/4.0 has the advantage of eliminating the diffraction pattern on bright stars as the iris opening is round.  The iris blades form an octagon as the lens is stopped down leading to a well formed pattern around bright stars.  I don't always appreciate these "maltese crosses" in my images so I am glad that imaging wide open is an option with this lens.

At 70mm, the lens performance is less impressive.  This was not unexpected as edge-effects are typically worse at wider angles. Wide open at f/4.0, the chromatic aberration is much more noticeable at 70mm than at 200mm - see specifically the star above Procyon in the center image.  At 70mm, stopping down to f/4.5 has a much more pronounced effect in removing the blue halos.

Star distortions are also much more pronounced at 70mm.  The stars in the bottom left are smeared radially into ellipses, similar to what I would expect of a refractor with uncorrected field curvature. The stars in the bottom right are slightly distorted into crosses.  Stopping down the lens improves the "crosses" but not the "ellipses."

Update 2013-04-11 - Imaging Test

This image of the belt and sword region of Orion shows a more representative test, although the blog has compressed the image.  This image consists of 16 stacked exposures, each 60sec at ISO-800. These were taken with the lens at 200mm, f/4.5. Post processing in Photoshop to apply gamma stretching, gradient removal, and noise smoothing.