Sharp, nice bokeh -- great lens!

medon78 wrote:

This is the kit-lens for my RP. So far I like it quite a lot - it is sharp already at f/1.8, only getting better when stopped down.

Bokeh is nice, has a little bit of character, but so far I did not in a single shot see something that had to be described as "busy" or wishing me to re-take the shot with different aperture or focusing distance.

Focus is not super-snappy as with USM-lenses, but that's OK with me. It is still plenty fast.

Some testers say, this lens would exhibit extreme vignetting. Well, looking at the example below (far from scientific, true!) I can see a 1.5EV difference in the corners between f/1.8 and f/5.6 in the RAW file. That does not bother me too much, since in-camera correction reduces this even further.

You are comparing different exposures and different ISOs... Optical Limits measured 3.8 EV vignetting, which is pretty extreme. I checked their numbers of other lenses (I do not own the RF 35) in RAW, and their numbers are quite close to mine. DPP measures extreme vignetting as well.

Price: in the kit with an RP or so the lens is reasonably priced. The stand-alone list price of ~500-550 EUR as of late 2019 seems a little high, though.

medon78 wrote:

J A C S wrote:

medon78 wrote:

...

Some testers say, this lens would exhibit extreme vignetting. Well, looking at the example below (far from scientific, true!) I can see a 1.5EV difference in the corners between f/1.8 and f/5.6 in the RAW file. That does not bother me too much, since in-camera correction reduces this even further.

You are comparing different exposures and different ISOs... Optical Limits measured 3.8 EV vignetting, which is pretty extreme. I checked their numbers of other lenses (I do not own the RF 35) in RAW, and their numbers are quite close to mine. DPP measures extreme vignetting as well.

Yes you are right, the ISO were different in the real world snapshot.

I re-checked with fixed ISO against my white ceiling, with 4 variations -- lens corrections turned ON/OFF, focus set to infinity vs. focus at 0.5m.

Seems the both variables "focus" and "lens corrections" do not affect vignetting, maybe the focusing distance changes corner exposure by ~0.2 EV. Not too significant.

Result: Yes, vignetting is stronger compared to my initial 1.5EV statement; now it is close to 2.0EV.

I am sure the guys at Optical Limits have more sophisticated measurement methods and tools at hand, though. OR -- maybe the R and RP exhibit different vignetting thanks to different micro lens design. Don't have an EOS R at hand, though.

Cheers!

This is interesting. Would you share that RAW file with me? I can share some RAWs of my 50L with the R and the 5D4 if you want.

I found it tricky to measure (and even define) vignetting. It depends how directional your source is. Walls/ceilings are good targets but they can be a bit too directional. I tried putting a napkin over the lens and when I point it to a light source, for example, I still see a hot spot.

I measured about 2.2 stops for my 50L. If I choose the center to be strictly the center, and measure the corners in the extreme corners with a small box, I get a bit more.

medon78 wrote:

... sure, let me know what your results are!

Cheers, Medon

https://www.dropbox.com/sh/uqd1pqbh82tq8c6/AAA5V2d8HQDZj_vU_ZxLE73ha?dl=0

Would you share that RAW file with me?

Thanks. I will post my findings. I may start a new thread in the PST forum with a link here.

Here is a curve derived from one of you RAW files:

It goes to -2.1 or -2.2 in the extreme corners.

I tested my EF 50L on the R a bit unscrewed and screwed (with fewer samples).

Yellow curve: unscrewed (the camera cannot recognize the lens). The red curve: the camera recognizes the lens and "cheats".

medon78 wrote:

J A C S wrote:

Here is a curve derived from one of you RAW files:

It goes to -2.1 or -2.2 in the extreme corners.

OK! Thank you very much. So... no -3EV or more as the Optical Limits guys sampled. Then there must be either something wrong with my uncalibrated ceiling ;)... hmmm.

I do not think that you can make the vignetting weaker by doing a bad test, this is very hard. On the other hand, a poor test can easily make it look stronger, if the light is more directional than one would think. Taking a shot of a wall pained with a glossy paint is an example. This is especially tricky with wide lenses.

If you would like to compare other lenses, I would have the EF 85/1.8 and EF 50/1.4 at hand.

Since I have the spreadsheet ready, yes, please. You can try to unscrew the lens a bit so that the camera does not cheat. You need to tell the camera to take shots without a lens first (it is somewhere in the menu).

You can see my thread in the PST forum as well, where I tried to catch my R cheating. 

Vignetting is not much different from 35/2 IS lens. I used the latter profile in LR and it looks reasonable.

https://www.the-digital-picture.com/Reviews/Lens-Vignetting-Test-Results.aspx?FLI=0&API=0&FLIComp=0&APIComp=0&Lens=1224&Camera=1221&LensComp=824

medon78 wrote:

forpetessake wrote:

Vignetting is not much different from 35/2 IS lens. I used the latter profile in LR and it looks reasonable.

https://www.the-digital-picture.com/Reviews/Lens-Vignetting-Test-Results.aspx?FLI=0&API=0&FLIComp=0&APIComp=0&Lens=1224&Camera=1221&LensComp=824

In

https://www.dpreview.com/forums/post/63469698

... J A C S analyzed some RAW files. Somehow it is hard to reproduce those -3EV results.

Overall, as you say, vignetting in real life use seems not to be excessive.

I don't know the reason, most sites measured over 3 stops. Could it be possible that vignetting is corrected in RAW?

forpetessake wrote:

medon78 wrote:

forpetessake wrote:

Vignetting is not much different from 35/2 IS lens. I used the latter profile in LR and it looks reasonable.

https://www.the-digital-picture.com/Reviews/Lens-Vignetting-Test-Results.aspx?FLI=0&API=0&FLIComp=0&APIComp=0&Lens=1224&Camera=1221&LensComp=824

In

https://www.dpreview.com/forums/post/63469698

... J A C S analyzed some RAW files. Somehow it is hard to reproduce those -3EV results.

Overall, as you say, vignetting in real life use seems not to be excessive.

I don't know the reason, most sites measured over 3 stops. Could it be possible that vignetting is corrected in RAW?

No, I tested for this in the other thread.

J A C S wrote:

medon78 wrote:

J A C S wrote:

Here is a curve derived from one of you RAW files:

It goes to -2.1 or -2.2 in the extreme corners.

OK! Thank you very much. So... no -3EV or more as the Optical Limits guys sampled. Then there must be either something wrong with my uncalibrated ceiling ;)... hmmm.

I do not think that you can make the vignetting weaker by doing a bad test, this is very hard. On the other hand, a poor test can easily make it look stronger, if the light is more directional than one would think. Taking a shot of a wall pained with a glossy paint is an example. This is especially tricky with wide lenses.

Are there methodologies that can eliminate focusing distance or colour profiles as a variable when expressing vignetting in EV ? Differences between various websites' results in EV may just come from that, no ?

MayaTlab0 wrote:

J A C S wrote:

medon78 wrote:

J A C S wrote:

Here is a curve derived from one of you RAW files:

It goes to -2.1 or -2.2 in the extreme corners.

OK! Thank you very much. So... no -3EV or more as the Optical Limits guys sampled. Then there must be either something wrong with my uncalibrated ceiling ;)... hmmm.

I do not think that you can make the vignetting weaker by doing a bad test, this is very hard. On the other hand, a poor test can easily make it look stronger, if the light is more directional than one would think. Taking a shot of a wall pained with a glossy paint is an example. This is especially tricky with wide lenses.

Are there methodologies that can eliminate focusing distance

One usually tests with focusing distance infinity.

or colour profiles as a variable when expressing vignetting in EV ?

I measure RAW files. Most websites however (except DXO, I guess) measure JPEGs, and the tone curve would affect what they report.

Differences between various websites' results in EV may just come from that, no ?

J A C S wrote:

MayaTlab0 wrote:

J A C S wrote:

medon78 wrote:

J A C S wrote:

Here is a curve derived from one of you RAW files:

It goes to -2.1 or -2.2 in the extreme corners.

OK! Thank you very much. So... no -3EV or more as the Optical Limits guys sampled. Then there must be either something wrong with my uncalibrated ceiling ;)... hmmm.

I do not think that you can make the vignetting weaker by doing a bad test, this is very hard. On the other hand, a poor test can easily make it look stronger, if the light is more directional than one would think. Taking a shot of a wall pained with a glossy paint is an example. This is especially tricky with wide lenses.

Are there methodologies that can eliminate focusing distance

One usually tests with focusing distance infinity.

One thing I like with Photographylife's tests is that they do it at two focusing distances.

I'm not sure that all lens testing websites are quite rigorous enough to bother with that or many other variables though.

I mean, this is Les Numériques's latest chart concerning the Nikon 58 0.95 :

According to them, the vignetting actually increases past f2 when closing the lens down... ???

or colour profiles as a variable when expressing vignetting in EV ?

I measure RAW files.

Gotcha.

Most websites however (except DXO, I guess) measure JPEGs, and the tone curve would affect what they report.

That could be where some of the differences are seen, no ? Tone curves can dramatically affect pictures.

DXO might be evaluating raw files, but sometimes they still managed to produce some strange results :

A lot has been said about the way they aggregate their data into meaningless scores, but sometimes the data itself - or at least the way it's presented - looks a bit weird.

Honestly these days, as far as vignetting is concerned, I find its expression in EV at infinity, provided it's actually properly done (and I don't trust a lot of websites in that regard), quite insufficient to actually describe how a lens is affected by vignetting.

MayaTlab0 wrote:

Honestly these days, as far as vignetting is concerned, I find its expression in EV at infinity, provided it's actually properly done (and I don't trust a lot of websites in that regard), quite insufficient to actually describe how a lens is affected by vignetting.

All I care is how it is affected the way I use my lenses. I never focus wide open too close.

Anyway, I will do more tests when I have the time. I can also shoot a short video with varying the focal plane. I pretty much determined that the camera does not correct for vignetting unless it is told so, so the video would be informative. I will run RAW tests as well.