Carlo_M
Offline
• •
Upload & Sell: Off
|
 p.1 #10 · Lenses that "require" software correction. | |
I'll add to what @Mike_5D wrote. First, some general points so we have a basis for understanding:
Just about all RF lenses, if you're after near-perfect and near-invisible vignetting, require *some lens correction*. That's every zoom lens (by virtue of having to cover multiple focal lengths), and yes even the highly respected RF 85 1.2L (for vignetting, not so much for distortion). I am no camera/lens historian, so I'm sure there are some "optically perfect lenses" that exist, but we're in the Canon forum, and in the current RF line there basically doesn't exist a lens that doesn't improve with *some* level of profile correction.
Second, there are in my view, two (okay technically three, when you see how I split the first type) general types of lenses requiring profile corrections, and I admit this is my conjecture as I'm not a lens designer nor do I play one on TV.
Type 1 is the kind of lens that was "well we did the best we could to create a lens to meet a price point, but there are still some imperfections which we will correct via a lens profile."
What I think of as Type 1A is for expensive lenses like the RF 85 1.2L, they built that thing to the best of their ability, it requires very little correction, and is priced accordingly. RF L zooms also fall under this type but require more correction because of the complex design nature with creating a zoom vs. a prime lens.
What I think of as Type 1B is for affordable lenses, they are built with significantly lower quality than L glass, and are so optically imperfect that while profile correction may improve their geometry and vignette, it comes at a noticeably visible cost in day-to-day shooting results (corners appear noticeably soft even when stopped down, noise from having to eliminate vignette through the profile correction starts to show through even in properly exposed--or slightly underexposed--images that are brought up in post).
And then there's a (relatively) new Type 2. Lenses that are *designed* with profile correction in mind. The new VCMs fall under this. I do not have the new non-L (but nearly L quality) zooms that Canon has just released within the last year and are garnering positive reviews, but my guess is that those lenses may also fall under this category. These are designed so that even after profile correction for vignette and geometric distortion, the results are equal to or better than older established, respected lenses. I own the 20mm and 35mm VCMs, and the 50 1.2L and until a month ago I owned the 24-70L IS USM. I posted videos in other threads in this forum, but when shooting both against test charts and at museums shooting art work, I found the 35mm VCM to be slightly sharper and with better microcontrast in the middle, and noticeably superior sharpness in the corners than the 24-70L IS USM. And comparing the 35VCM to the 50 1.2 at identical settings (shutter speed, aperture and ISO) while keeping the artwork size as consistent as I could in-frame...the 35 VCM again showed slight sharpness/microcontrast improvement in the center, an advantage which increased towards the corners (though the margin of improvement was smaller than the 35VCM vs. 24-70L @ 35mm).
For a Type 1B lens, I'd likely not be interested in these lenses, as the corrected results still show mediocre (or worse) corner sharpness/noise after correction.
For Type 2 lenses, I'm all in. I love my 20 and 35 VCMs to the point where I easily parted with the 24-70 and have zero remorse.
|
Of course this may entail cropping out black corners (depending on the crop, you can still retain either the extra horizontal or vertical coverage) or using content-aware fill (adding more 'make believe' to the mix).