I have made this point in other threads before, although I cannot seem to find them now. As far as I remember, nobody ever seriously challenged my working thesis:

“You cannot improve a lens’s in-air optical performance underwater.”

What I mean by that is fairly simple: any optical system not originally designed specifically for underwater use — with the obvious exceptions of systems like the Nikonos 15mm and Nikonos RS — will suffer once taken underwater. Image quality will always degrade relative to that lens’s native performance in air.

The moment we take a full-frame lens underwater and place it behind a flat port, dome, or a purpose-built underwater correction optic such as OPP, Ivanov, WACP, FCP, or Marelux Aquista, we are introducing a new front glass element into a lens design originally created by the optical engineers at Nikon, Sony, Canon, Sigma, and others.

In other words, we are adding more “filter” to the light path.

We are not performing magic.

We are adding more glass.

We are adding more interfaces.

We are adding more opportunity for compromise.

That does not mean such correction optics are pointless — quite the opposite. A lens may become far more useful, more practical, and more enjoyable underwater when paired with the right optic. But it will not suddenly become sharper, cleaner, or more transmissive than what the best land-based lab tests already show it to be capable of in air.

What it can do is perform vastly better in its actual underwater use case when paired with the correct optic, especially when compared with the same lens behind a flat port or a poorly positioned dome.

So when people speak as though a port or correction optic somehow improves a lens beyond its native land performance, I think that is where the language becomes sloppy. Better underwater than with the wrong setup? Absolutely. Better than its in-air optical ceiling? I do not see how.

So I am putting this out here for discussion and, ideally, confirmation or informed contradiction.

My working assumption has long been that the best way to identify a strong underwater optical system is to start by comparing land-based tests of lenses that already fit key underwater criteria — for example, short minimum focusing distance and other relevant characteristics — and then distill the strongest in-air candidate before bringing it underwater and matching it with the most suitable port or correction optic.

That, to me, seems to be the real algorithm:

start with the best land candidate, then optimize the underwater setup as intelligently as possible.

What about lenses that are not well regarded above water such at the Tokina 10-17 which has quite a following for UW use?

I basically agree with the premise, any additional elements tend to degrade performance to varying degrees, it may degrade a lot or a little to the point it is hard to see, but it entirely reasonable to expect that it won't improve.

As for the statement :

start with the best land candidate, then optimize the underwater setup as intelligently as possible

yes but it still seems that some lenses play better underwater than others, so there is a need to be selective in what you try to take underwater. Just because it is an amazing lens above water doesn't automatically translate into good UW performance. Though it seems that with the latest crop of close focusing mirrorless zoom lenses many of them work quite well UW.

For the specific case of the Tokina 10-17 potentially what we are seeing is that this lens works exceptionally well with small domes and the degradation in performance when taken UW is minimal. But if you compare the UW results with some of the older rectilinear zooms that don't seem to work well UW or perhaps people are using them in too small a dome, the Tokina seems to really shine if you just compare UW shots taken with these older lenses with shots taken with the 10-17. It is also a very flexible wide angle option the ability to zoom in and take reef scenics and CFWA on the same dive can trump ultimate sharpness for many people.