RF 14-35mm f/4L IS – Distortion Correction Testing

[LogicExtremist]

Just wanted to add my thanks for the great review.

I’m a LR/CR user, and in comparing the FOV at 14mm between ACR and DPP, the ACR FOV is definitely wider to me. Your results appear to confirm this, which makes me happy.

I bought it to complement my RF 15-35, but now I’m not sure if there is much reason to use the 15-35 unless one needs the F2.8. I’m curious as to how you find the IQ on the 14-35 compared to the 15-35 other than the wider FOV.

Thanks again

Looking at Gordon Laing's review, where he compares the RF 16mm f/2.8, RF 14-35mm f/4 and RF 15-35mm f/2.8, the RF 15-35mm f/2.8 lens is much sharper in the corners at all apertures, and in the centre it's visibly a little sharper with more contrast when stopped down.

 

[docsmith]

Great tests!

 

[tron]

Gordon's review was very interesting and informative. I must say however that I am disappointed at his 14-35 4L IS findings.

Regarding Gordon's tests 14-35 4L:

For that amount of money it should be better at the edges. Also as an expensive optic it should be optically corrected just like the 2.8 version which is not disproportionately more expensive that the 4L. Their price difference justifies the difference in aperture but not the lack of optical corrections.

Neuro's tests:
On the other hand neuro's tests show that edges of these look really good at f/4. At least in my laptop and to me.

So

Of course I prefer the lens to be good at the edges. Its small size is a big advantage so maybe I will want to buy it in the future...

 

[LogicExtremist]

Gordon's review was very interesting and informative. I must say however that I am disappointed at his 14-35 4L IS findings.

Regarding Gordon's tests 14-35 4L:

For that amount of money it should be better at the edges. Also as an expensive optic it should be optically corrected just like the 2.8 version which is not disproportionately more expensive that the 4L. Their price difference justifies the difference in aperture but not the lack of optical corrections.

Neuro's tests:
On the other hand neuro's tests show that edges of these look really good at f/4. At least in my laptop and to me.

So

Of course I prefer the lens to be good at the edges. Its small size is a big advantage...

In Gordon's test, the RF 16mm IQ matches the RF 14-35 in the centre of the image when both are wide open or at F/4, and when stopped down to f/5.6 the RF 14-35mm is slightly sharper. When it comes to corner sharpness, wide open they're quite similar, and when stopped down to f/4 and f/5.6 the 16mm is very slightly sharper when pixel peeping.

Never thought I'd hear myself say this, but unless a 14-35 zoom range is required, the RF 16mm f/2.8 has similar IQ in centre and corners, is smaller, lighter, much cheaper and is a whole stop brighter! Perhaps, with all that considered, we've judged this little 16mm UWA prime lens a bit harshly.

 

[neuroanatomist]

What I still don't quite understand is that if there's mechanical vignetting, and the 13mm image across the full sensor visible in the RAW file is cropped down to 14mm, how is the image not being upscaled like you first theorised if the final corrected image is still the full sensor size? Or is that not the case

Because there’s no cropping. Rather, distortion correction stretches (warps) the periphery of the image out to the corners to fill the frame. There’s still interpolation going on but it’s only at the periphery, rather than the whole image being cropped and then interpolated up to the camera’s full resolution.

DxO’s optical corrections only, where the distortion-corrected image is not cropped to the normal 3:2 aspect ratio, confirms that. The lens projects a circular image into the camera, so the distortion is radially symmetrical. Because the sensor is a rectangular sample of that circle, the sides of the image are closer to the edge of the circle than the top and bottom. Therefore, the sides will be more distorted than the top and bottom, and require relatively more stretching to correct. That’s why the DxO optical corrections only is ~300 pixels wider.

 

[docsmith]

It seems like the EF 15-35 wins at the edges. But EF 14-35 is certainly "good enough". In addition to Gordon Liang's RF 16 f/2.8 video where he compares all there, there is this one showing the similar results between the RF 15-35 v RF 14-35.

But, looking at Neuro's and the images in this video, it isn't that the 14-35 is bad in the corners at 14-16 mm, just that the 15-35 is better.

 

[LogicExtremist]

Because there’s no cropping. Rather, distortion correction stretches (warps) the periphery of the image out to the corners to fill the frame. There’s still interpolation going on but it’s only at the periphery, rather than the whole image being cropped and then interpolated up to the camera’s full resolution.

DxO’s optical corrections only, where the distortion-corrected image is not cropped to the normal 3:2 aspect ratio, confirms that. The lens projects a circular image into the camera, so the distortion is radially symmetrical. Because the sensor is a rectangular sample of that circle, the sides of the image are closer to the edge of the circle than the top and bottom. Therefore, the sides will be more distorted than the top and bottom, and require relatively more stretching to correct. That’s why the DxO optical corrections only is ~300 pixels wider.

Thanks for the explanation, that makes sense now! It's fortunate that there's no cropping and upscaling, and only stretching/interpolation of the periphery, as that retains the centre sharpness!

 

[gruhl28]

In Gordon's test, the RF 16mm IQ matches the RF 14-35 in the centre of the image when both are wide open or at F/4, and when stopped down to f/5.6 the RF 14-35mm is slightly sharper. When it comes to corner sharpness, wide open they're quite similar, and when stopped down to f/4 and f/5.6 the 16mm is very slightly sharper when pixel peeping.

Never thought I'd hear myself say this, but unless a 14-35 zoom range is required, the RF 16mm f/2.8 has similar IQ in centre and corners, is smaller, lighter, much cheaper and is a whole stop brighter! Perhaps, with all that considered, we've judged this little 16mm UWA prime lens a bit harshly.

Yeah, all the previous reviews I've seen gave me the impression that the 14 - 35 gave much better image quality, at least in the corners, but this review doesn't seem to support that. I'm not really sure what to make of this. The 14-35 might be better at close distances where a lot of testing is done, but I'm not sure that all the reviews were based only on close distance.

 

[LogicExtremist]

It seems like the EF 15-35 wins at the edges. But EF 14-35 is certainly "good enough". In addition to Gordon Liang's RF 16 f/2.8 video where he compares all there, there is this one showing the similar results between the RF 15-35 v RF 14-35.

But, looking at Neuro's and the images in this video, it isn't that the 14-35 is bad in the corners at 14-16 mm, just that the 15-35 is better.

Just be aware when looking at a comparison that the quality observed is relative to the quality of the other lens lens, and not necessarily an absolute perfect standard. Remember that the rating of the centre, mid, periphery and corners of a lens can be rated as excellent/outstanding, very good, good, fair and poor. We first need to know where the other lens sits on this scale to assess where its IQ sits on the scale. We can do this a bit more objectively than just say it looks fairly okay in the corners.

In Neuro's test, to my eyesight, the very the extreme corners of the RF 14-35 look soft compared to the EF 11-24 at 14mm f/4 and f/8 which looks quite clear. The difference can be seen in the edges of the bricks and mortar, which are more sharply and clearly defined, with better contrast in the EF lens images.

How clear is the EF 11-24 in absolute terms?

According to the extract from the Optical Limits review, at 14mm the center is 'outstanding' but the borders are just 'good' to 'very good', and the corners are only 'fair' at f/4, but when stopped down to f/5.6 the corners become sharp.

MTF (resolution) at 50 megapixels
The resolution characteristic varies substantially across the zoom range. Of course, the most difficult setting remains the 11mm setting. The center quality is absolutely stellar even at f/4. However, the borders/corners are soft here. The extreme corners show CAs with a width of up to 8 (!) pixels which has certainly also affected our tool here . Stopping down lifts the quality in the outer image region but not quite as much as you would have hoped. The best quality is reached at f/8 with good to very good borders and fair corners. At f/11 diffraction is already limiting the results.
The situation changes completely at 14mm and 18mm. The center remains outstanding but the borders/corners are MUCH better. The borders are good to very good at f/4 whereas the corners are fair here. However, the corners are boosted to sharp quality at f/5.6 already.
The performance goes a little downhill at 24mm again with softer results at f/4 and f/5.6 but f/8 is decent actually. The centering quality of the tested sample was good but not perfect.

Canon EF 11-24mm f/4 USM L - Review / Lens Test Report - Analysis

Canon EF 11-24mm f/4 USM L - Review / Lens Test Report

www.opticallimits.com

If this evaluation is correct, then the comparison against the 11-24mm @ 14mm and f/4, which is only rated as 'fair' in the corners at these settings, shows that the 14-35mm can't match that, and in absolute terms, must be producing corners that would be rated less than 'fair'.

Whether that performance rates as 'good enough' for a particular individual's needs is a completely different matter.

Hope that provides some clarity in terms of absolute vs relative performance.