DXOMARK Score for Lenses

Writing this post to learn more. I don’t know enough about the subject and don’t have the resources to make my own real-life comparisons. The following post prompted me to start this thread:

@jrp is this when you look at the lenses as a whole or would you say the individual measurements do not match (e.g. difference in sharpness is greater than in the test results)?

At anyone who would care to help educate me (and perhaps others):

What happens with sample variations for example? Are they taken into consideration when the scores are published? Would the even matter when the scale of measurement is considered?

How is it even measured? I can understand better when camera and lens are combined. But how do you set the value of a lens on it’s own, when it is restricted to certain mounts i.e. sensors?

Can you even really compare lenses outside of the systems the were intended for? I mean I could have a lens that would qualify as excellent on my APS-C (I wish :smiley: ), which would not be useful on full-frame.

I would also be happy if someone could share some links to say “Read this and this…”

Thanks in advance for the answers.

Have you looked as their information about their measurements?
Perhaps that can shed some light on things.

Can I assume that you already knew that DxO Labs, the maker of PhotoLab is a completely different company from DxOMark? They went their separate ways in late 2017. Questions about DxOMark testing procedures posted here are unlikely to yield any definitive answers.


That is what I thought too. I also recall something like that being said in a recent discussion between a YouTuber and DxO. It was in French though.

Based on @jrp’s comment I got the impression that the measurements used for PhotoLab’s Optic Modules can be very off. Citing their ownership of the above mentioned lenses as proof that the Optics Modules aren’t reliable.

General remark: Measurements don’t count unless they are repeated a few times to eliminate measurement errors. Then saying “this lens model is good” (I’m simplifying) is not worth much unless several samples have been tested in order to characterise sample variations.

Both DxOMark and DxO Labs (relating to their camera, lens and module processes) have never commented on those qualities as far as I know. The one site that comments its measurements of lenses is lensrentals.com. They rent lenses and want to make sure that the quality of their gear is up to spec. Check out and search their blog pages, eg this one: Lens Rentals | Blog.

1 Like

Thanks yes, I had read that. But not reinforced from anywhere else. No mater how much I like using PhotoLab (and other DxO tools), I’m always a bit cautious with what manufactures say about their own products.

Often texts are written by people with the technical knowledge and then given a marketing polish before going online. I don’t know DxO well enough yet to decide if I can simply trust what they say.

Thanks. I have visited seen that site before (I think it was from a link you posted :smiley: ).

In this case it was really the comment by @jrp that apparently got me confused. Specifically the part preceding the lenses:

It made me realise, I am happy with the results, but I’m also doing some things in good faith and asked myself if I could face issues because of this in future.

My questions are about the DXOMARK score and the statement be @jrp about the score (Lenses) and how it’s supposed to relate to the Optics Modules in PhotoLab (Lens + Body).

So is it correct to say that the lens score on it’s own, has nothing to do with the Optics Modules?

In which case then still just my initial question at @jrp

The 35mm APO Summicron is a sharper, better corrected lens than the (much lighter and more compact and, of course, cheaper) 35mm f2.8 Sony/Zeiss lens. They are not in the same class. (They were designed years apart, the Summicron for 100Mpx bodies, the former at a time when 24Mpx was the top of the hill.)

Anyway, the lens corrections in Photolab are a benefit (although I’m not so sure that they are worth waiting for, in comparison to the (geometry / CA) corrections described in the RAW file by the manufacturer).