With a load of architectural style images I took last night, I'm more than happy to say that the lens is very good from a rectilinear POV. With 50MP images from my 5Ds, I've found I needed relatively little work to make some 22" x 33" prints.
CA at a few pixels is easily fixable and vignetting drops a lot by f/4 (to less than the ef11-24 at f/4). Shooting at f/5.6 gives very acceptable fine detail and residual corner sharpness is very amenable to (masked) sharpening with tools like Piccure+
I'll be interested to see how they market the lens (something I don't have to worry about!)
For people unconcerned with it being fully manual, I can see it being a great addition to the camera bag - then again I've the EF11-24 F4L, which has AF (but at quite a hefty price and bulk).
arduluth wrote:
When this lens was announced, all of the ad copy I read indicated to me that "Zero-D" meant "rectilinear" rather than "no distortion." I remember a blurb from Laowa that stated something like "unlike most lenses this wide, this new lens won't distort with a fish-eye effect." I haven't been able to find it yet, so I may be mistaken.
Here's the original thread-- which I started-- where you can go back and read i) one poster explaining that the marketing material is specifically referring to zero/negligible distortion (later updated to 0.2% distortion), and ii) me arguing that such marketing language is going to turn out only to be marketing language (negligible distortion will come at the cost of other IQ characteristics like coma, CA, resolution in the corners).
And here's a quote from someone in that thread: "Actually if you look at the Chinese site the zero distortion is not about rectilinearity, but about distortion. They are claiming .2% barrel distortion, which is indeed close to zero...."
I don't have a lens test lab, but careful checking of straight edges around the frame suggests that they have indeed managed to keep distortion down to very usable levels.
Of course I'm testing a pre-release lens so I appreciate that what ships may change a bit.
From my own POV as architectural photographer, rectilinear implies straight lines, and barrel/pincushion distortion means curved lines, so the argument about what Zero-D 'means' is a bit of an academic one...
At the FOV of this lens, the inherent perspective distortions of a rectilinear projection are very obvious and an integral part of the challenges of composition. I regularly create stitched images with lenses from 200mm to the EF8-15 fisheye, where post processing changes of projection geometry give a wide range of choices in how a scene is mapped to a flat sheet of paper. If for example, I need a rectified image (for measurement) I'll create one from multiple shots combined with data from a total station on the same tripod.
With my more scientific hat on I'll be interested to see some actual numbers even if, as I often realise, they mean precious little when it comes to actually taking photos that clients like ;-)
The Close-to-Zero distortion claim is not just marketing language only. From Keith's excellent picture of architectures, you can have a feel on how straight lines are retained. But the distortion we are referring only limits to Optical distortion (Barrel/Pincushion/Moustache) but not perspective distortion. Hope it clarified.
venuslaowa wrote:
The Close-to-Zero distortion claim is not just marketing language only. From Keith's excellent picture of architectures, you can have a feel on how straight lines are retained. But the distortion we are referring only limits to Optical distortion (Barrel/Pincushion/Moustache) but not perspective distortion. Hope it clarified.
Thank you for clarifying! When the lens was announced, I never saw any straight reference to distortion on a spec sheet in the English materials, which made it bard to know if it was marketing speak for "rectilinear" or not.
Keith -- Thanks for the excellent examples and your evaluations.
Do we have any idea of a release date and list price? This looks like a lens I'd very much like to own for architecture and maybe even some landscape photography.
When you mention fully manual, does that mean like the Samyang, no lens exif supplied to camera?
Yes, -fully- manual ... the camera thinks there is no lens there.
On my 5Ds that rather messes with metering, but I'm OK with guesstimating exposure in most situations - that and a check with a histogram to be sure.
If I had a lens I'd likely get an AF confirm chip from eBay to attach to the EF mount. This allows you to set some info for the camera, that makes stuff a bit easier. I was shooting tethered yesterday and the exposure to get live view to work OK was miles away from what worked for photos. The upside was that I got to see how wide 12mm is when focussed six inches in front of the lens...
At 12mm you are also fine setting focus to infinity for outdoor shots beyond a few metres at f/5.6 or f/8.
In looking at the Laowa I note the degree of coma wide open, which limits its use for astro shots wide open, but the minimal geometrical distortion, which makes it just fine for my own work.
I'd contrast the design decisions of Laowa with the Samyang 14mm, which bends straight lines like a coke bottle, but has cleaner coma and other off axis distortions.
So there we go: 1.1% distortion. Very nice for such a wide angle lens. Zero Distortion? Yeah, marketing... I'd call it low-distortion. Actually, I wonder how it stacks up against the Voigtländer 10mm and coming 12mm lenses (thought they are MUCH slower) regarding the same.
I note they do point out the need for the test chart to be very close in the PZ tests
From my own outdoor shots with buildings, I'm minded to think it will be somewhat less than 1.1% when measured at a longer focus distance. Enough that I've not one test shot where I felt it would have benefitted from any correction in processing.
Slight up or down tilt of the lens leading to tilted verticals for buildings, now that is a very different matter and emphasises the difficulty of using such a wide lens, even on a good tripod ;-)
