Canon introduced the
RF-S 10-18mm f4.5-6.3 IS STM ultrawide zoom for APSC RF mount cameras in
2023. This is a nice compact stabilised lens but is constrained by the small and
variable aperture which limit its appeal indoors and in low light.
The constant maximum aperture Sigma 10-18mm f2.8 DC DN/C STM
became available for Sony E mount around the same time and now in 2024 is also
available with Canon RF mount.
Some people say an ultrawide lens is good for landscape
photos but that is not my experience at all unless my subject is actually an
eye catching foreground with the background being relegated to minor status in
the frame.
I find ultrawides most useful for architectural interiors,
with or without a madding crowd.
For handheld interiors a very useful lens is a compact ultrawide
with a wide and preferably constant maximum aperture (small f number) and either an image stabiliser in the lens or
in body image stabiliser (IBIS) in the camera body.
When mounted on a Canon EOS R7 the Sigma 10-18mm f2.8 DC
DN/C meets these requirements in fine style. This light, compact, moderately
priced body/lens combination is capable of delivering high quality professional
results equal to or better than anything we could produce a few years ago with
much larger, heavier more expensive gear which required a tripod for best
results.
Aspheric elements, special types of glass and post capture
corrections have enabled lens makers to produce smaller, lighter and optically
better lenses than was previously possible.
Description
The Sigma 10-18mm f2.8 is remarkably light and compact. Diameter
is 72mm, length 62mm, mass 260 grams.
It is designated Contemporary which in Sigma world is
the entry level with models designated Art
and Sports being more expensive. Entry
it may be but the quality is nevertheless very high. The full frame equivalent focal length range
is 16-29mm. Unlike the Canon RF-S 10-18mm it does not have to be unlocked for
use.
Build quality, fit and finish are very nice. The zoom action
is very smooth and rotates the same way as other Canon lenses.. As with several
modern zooms the lens is physically shortest at the longest focal length. There
is a focus/function ring but no switches on the lens barrel. The mount is metal
with a rubber O ring seal around the edge.
A standard 67mm screw in filter can be fitted. I have a Hoya
HD.2 UV filter mounted on my copy of the lens permanently and all tests were
done with the filter in place.
A petal type lens hood is supplied. This mounts with a
straight push-in-click action and is removed with a tiny little twist. There
are markings to show how to orient the hood for outwards and reverse mounting. Align
the bar for hood out, the arrow for hood reversed. It’s a system I have not
seen before and it appears to work well with no problems thus far.
The constant f2.8 aperture of the Sigma lens is 1.3 stops
wider than the Canon RF-S lens at the short end of the zoom and 2.3 stops wider
at the long end.
The Sigma lens has no image stabiliser.
Function
Single shot autofocus is very fast and consistently accurate
even in low light and with heavy backlighting, even with the sun shining
directly into the lens. Human/animal/face/eye detect all work just fine with no
problems encountered by me.
The lens supports the maximum frame rate of which the R7 is
capable.
Optics
There are 13 elements in 10 groups. Four elements are
aspheric including the front element, three FLD, one SLD.
The lens is sharp in a large central circle right from f2.8
at all focal lengths.
At 10mm the long sides of the frame and corners reveal mild
softness and contrast loss on the test chart although this is less obvious in
real world photos. The softness improves at f4 and is almost gone by f5.6.
The peripheral softness at f2.8 becomes less evident as we
zoom out to 18mm focal length.
Overall sharpness is excellent on this lens which can be
used at f2.8 with confidence.
With Raw capture and conversion in Adobe Camera Raw, the
distortion present in uncorrected files is almost perfectly corrected, making
the lens very suitable for architectural subjects.
Corner shading is also largely corrected.
Some mild blue/purple fringing is sometimes seen at high
contrast boundaries such as foliage against a hot sky at the wide end of the
zoom and wide apertures. This is easily corrected in Adobe Camera Raw. In most cases however color fringing is
insignificant.
I can induce veiling flare by placing the sun at or just
inside or outside the frame however focussing and image integrity are retained.
It is possible to induce a few shaped and colored flares but I have to
deliberately place the sun in the worst
possible position relative to the frame to achieve this.
I also note some bleed-over flare of very bright highlights
into adjacent dark shadows in areas where there is very high subject brightness
range locally.
With sensible camera work flares are not a serious problem
with this lens. This is quite an achievement for Sigma as ultrawide zooms have
in the past tended to be quite flare prone.
The out of focus rendition (bokeh) such as it is with this
type of lens looks smooth with no nasty artefacts that I have seen.
I did notice a bit of coma towards the corners in photos
with a lot of tiny Christmas lights in the frame. This could be relevant to
astro work which I have not yet tried with this lens.
I saw no apparent focus breathing when racking focus back
and forth.
I test the front-to-back and side-to-side distribution of
sharpness of my lenses by photographing cut grass at 45 degrees and applying
the find edges filter in Photoshop. Some wide and ultrawide zooms
exhibit very substantial shifts in the shape and distribution of the zone of
sharpness as the lens focal length and aperture are altered. Fortunately the
Sigma 10-18mm is well behaved in this regard and will not present the user with
any nasty surprises.
When using Raw capture and Adobe Camera Raw the field of
view in the output photo at 10 mm focal length is 4% linear larger than we see
in the viewfinder. This suggests that the actual focal length of the lens might
be around 8-9mm at the wide end. The JPG output size is the same as that seen
in the viewfinder. I see this behaviour
quite often with current model wide angle zooms.
Summary
The Sigma 10-18mm f2.8 DC DC/C is a very welcome addition to
the catalogue of lenses compatible with Canon RF mount APSC cameras. It is
particularly suitable for the R7 which has IBIS and a high pixel density
sensor.
Some enthusiastic reviewers have described the lens as
“perfect” which it is not quite, but it comes very close with no substantial
faults or failings in any aspect of optics or operation.
Alternatives
* The Canon RF-S 10-18mm f4.5-6.3 IS STM is a compact, well
behaved alternative with a stabiliser at around half the price of the Sigma.
The Canon lens is smaller when not in use but has to be unlocked for use so it
ends up almost as long as the Sigma.
The Sigma is sharper than the Canon at all focal lengths and
comparable apertures, but only on the long sides of the frame and the corners.
On my tests they are equally sharp in the center and on the short sides of the
frame.
* As I write this in November 2024 Tamron has just announced that its 11-20 mm f2.8 Di III-A RXD (do they actually
pay someone to come up with these confusing designations ?) for RF mount is now
in production.
This lens has been available in Sony E and Fujifilm X mounts
for some time and has been favourably reviewed many times. It is not stabilised.
It offers a slightly longer focal length range and is both
longer and heavier than the Sigma. I have no personal experience with this
lens.
Suggestions
I suggest the Canon when price is an issue, when usage will
be mainly in reasonably well lit locations and when mounted on a camera body
without IBIS (R10, R50). The Canon is actually a very good and useful lens,
better in practice than its modest specifications might lead us to expect. It can be used in very low light if there is
not much movement in the frame. On the R7 I have gotten sharp pictures hand
held at 0.4 seconds with this lens in near darkness.
The Sigma is best used on the R7 which has IBIS. The wide
aperture and overall slightly better image quality make this the preferred
option when a wide aperture and/or best optical quality is required. This is
the preferred option for the user who regularly photographs interior
architecture and similar subjects.
At this point I am not sure what the Tamron has to offer
which might make it preferable to the Sigma.
Conclusion
Three years ago I thought that Canon might be on a pathway
to eliminate APSC models from their catalogue. Their support for the now
defunct EF-M mount was tokenistic and their initial offerings for the RF-S
catalogue were underwhelming, to put it mildly. To make it worse Canon refused
for a long time to allow third party lens makers access to the RF mount.
But the entry of Sigma with two zooms and four primes and
now Tamron with one zoom initially has given a new lease of life to the Canon crop
sensor system.
There are rumors of an EOS R7.2 coming sometime in 2025 with
a stacked sensor. I think such a model
if well executed as to performance and ergonomics could prove very attractive
to many photographers.
I hope these rumors are correct because I have sold off all
my full frame RF mount cameras and all my compact and bridge cameras and fully
committed to the R7 and crop sensor lenses.
No comments:
Post a Comment