Sony FE 20-70mm f/4G: Assessing Distortion at 20mm, 24mm, 35mm, 50mm, 65mm
Related: distortion, distortion correction, Dr S, optics, Sony, Sony FE, Sony FE 20-70mm f/4 G, Sony Lenses
This page looks at distortion uncorrected and corrected of the Sony FE 20-70mm f/4 G at 20/24/35/50/65mm.
Comparison lenses include the Sony FE 12-24mm f/2.8 GM, Voigtlander FE 35mm f/2 APO, Voigtlander FE 50mm f/2 APO, Voigtlander FE 65mm f/2 APO. While those latter three are reference-grade lenses, that’s the point—the difference.
Sony FE 20-70mm f/4 G Distortion @ 20mm, 24mm, 35mm, 50mm, 65mm
Includes corrected and uncorrected images, and a comparison/reference lens at each focal length.
Dr S writes:
Your comment tells it all:
“But if it’s for casual snapshots or vlogging or any task where high resolution is not a goal, then turn on distortion correction and be happy.’
Cam/Lens manufacturers have been building in distortion corrections for years. Yes, that screws up high resolution. Sony programmed a huge amount in the 20-70/4 to sell a bunch of lenses......enough quality to satisfy the masses but not nearly enough for many of us. Obviously there must be a large "Sony" market for a wider range zoom. Just glancing at some of the forums one can see a ton of people with their new lenses. Some are concerned about the excessive WA distortion, while others accept a lower quality image but the question, as you alluded in your blog entry, is what are the images used for?
I still shoot primes and will continue to do so for all my shots (well, I did get the GF 20-35 for the hobby part of me) For some tele shots for events I use a 70-200. Sharpness there is not primary.
Don't know where the industry will ultimately end up but it seems to me the manufacturers are in a higher trajectory to build in a lot more software correction for imagery beyond distortion. Going even further I'm waiting for the day soon when one will be able to buy a Sony MILC with a built-in telephone. Maybe one exists but the massive amount of computational photography built into mobile phones has not been lost on the eyes of the major cam manufacturers. Interesting time.
DIGLLOYD: I’m OK with software correction for things like lateral chromatic aberration, though it is not ideal; the effects are not equally good as a well corrected lens and can never be. Especially since longitudinal chromatic aberration has a substantial perceptual effect.
Distortion and the negative effects of distortion correction are a radically different case because of degraded sharpness and micro contrast—together with grossly misleading MTF charts that account neither for diffraction nor distortion correction, a defiance of how such designs MUST be used. It’s flat-out dishonest IMO. Simple math is all that’s required to understand the losses (no pictures required!). Though I’ve also shown directly (many times) how much sharpness is lost with/without correction.
These points have clearly been evaded by the Amazon-owned folks of the dpreview crew, but whether it is incompetence along with conflating unrelated types of optical design compromises, or intentional misrepresentation using inappropriate examples for “air cover” for advertising sponsors in reviews to come (the article hints as much), I don’t know. But I do know that rationalizations are more profitable than criticisms.
BTW, I cannot (yet) show the sharpness losses with the Sony FE 20-70/4 because ACR as yet lacks a lens profile for correcting it. Indeed, the 20-70mm as it stands is unusable with RAW because there is no easy way to correct the extreme distortion (manual correction can help but it is only an approximation).
