Pondering on the wide angle options (already have mZ 9-18mm)

ogremage

Leading Member
Messages
959
Reaction score
54
Location
UK
I really like my m.Zuiko 9-18mm as it is light weight and provides a nice wide 9mm, except when I zoom into corners at original size, especially since moving from GF1 to E-M5, and then I start getting these ideas I should exchange it for something else.

Last time before the Nepal trip, I didn't.

Now I'm going to Alps in July to do some hiking, and I still have time to replace it, but I don't like either of the replacement options.

The Oly 12mm is perfect except it is not wide enough. I use the 9-18mm mostly at 9mm and the 12mm just doesn't seem to provide anything close (the last thing is likely subjective).

The Pana 7-14mm is heavy and doesn't take filters and is expensive and is not a prime (so I'm not sure if it would be that much better IQ).

There's the SLR Magic 12mm and CV 12mm but both are only 12mm and heavy and/or MF. I think I'd take the Oly 12mm over them.

Using the 43 9-18mm is expensive, gets quite heavy, and likely isn't that much better than the mZ 9-18mm.

I was also thinking of taking the 20mm f/1.7 too and then stitch panoramas for some shots, but plants in the foreground won't stitch well anyway.

A faster lens would help with astro-photography too.

I was even tempted to buy a Fuji X-E1 + 14mm as it seems like the sort of camera / lens combo I'd find ideal, but that gets very expensive, I like my other m43 lenses so don't want to switch systems fully, and also I like to use the Oly 40-150mm for hiking with its light weight and versatility, but the Fuji 55-200mm to use in its place is 2.6x heavier.

Anyway, here are some sample pics with my 9-18mm:

http://farm9.staticflickr.com/8054/8134829890_c47fb80f7d_o.jpg

http://farm7.staticflickr.com/6143/5929135979_9e85480daf_o.jpg

http://farm9.staticflickr.com/8038/7928997246_85b11691de_o.jpg

http://farm9.staticflickr.com/8473/8134808743_fd796675d7_o.jpg

http://farm9.staticflickr.com/8169/7928988810_91ca3db281_o.jpg

http://farm8.staticflickr.com/7292/8729037969_df91168622_o.jpg

http://farm9.staticflickr.com/8441/7928929040_1e00fcac5d_o.jpg

Would the corners be much better with 12mm or 7-14mm?

Do I have a defective lens (it has been knocked about in my travels) or is this normal for the mZ 9-18mm?

Any advice, besides to stop worrying about IQ and go out and take pictures?

--
blog: http://photographyandtravels.blogspot.co.uk/
flickr: http://ow.ly/eojki
K1000 -> A70 -> K100D -> GF1 -> E-M5
P 20mm + O 9-18mm + O 45mm + O 40-150mm + O 12-50mm
 
Last edited:
Pikme wrote:

The 7.5 looks good, but your 9-18 still looks bad to me, especially the left side. Have you tried the upside down test to check the lens? (http://www.dpreview.com/forums/post/51037446)


Results of upside-down test for 9-18mm at 9mm, f/5.6, OOC JPEGs opened in multiple IrfanViewers, top left corner in each (well, or bottom right in case of the upside-down pics):

8910232940_7d52562a83_o.png







What do you think?



--
blog: http://photographyandtravels.blogspot.co.uk/
flickr: http://ow.ly/eojki


K1000 -> A70 -> K100D -> GF1 -> E-M5
7.5mm FE + O 9-18mm + P 20mm + O 45mm + O 12-50mm + O 40-150mm
 
Anders W wrote:

Could you please post an example with the same image, before and after correction, as processed by LR as well as DxO, preferably with sharpening reduced to zero (so we can rule that factor out completely) and with the lens a bit stopped down (so we can see the remaining CA more clearly; it might be covered up by ordinary blur otherwise).
Finally was able to prepare some tests. I have noticed that in DxO I actually did have an option turned on, which is called "lens softness." I could not get a very clear explanation how it differs from the unsharp mask, which is a separate item, but I am getting an impression that it tries to treat coma. I have included conversions with and without his option. The first one is an overview, and the remaining images are 100% crops from the bottom right corner. Some additional comments under the crops.



Overview of the scene. 9-18 @ 9 mm, f/5.6
Overview of the scene. 9-18 @ 9 mm, f/5.6



100% crop of the OOC JPEG. Sharpness=0 NR=low
100% crop of the OOC JPEG. Sharpness=0 NR=low



ACR 6.7 conversion. sharpening=off, NR=off, CA = check, defringe=all edges
ACR 6.7 conversion. sharpening=off, NR=off, CA = check, defringe=all edges







DxO Optics Pro 8, sharpening=off, NR=off, CA=check, Purple fringe=check
DxO Optics Pro 8, sharpening=off, NR=off, CA=check, Purple fringe=check






DxO Optics Pro 8, sharpening=off, NR=off, CA=check, PF=check, lens softness = check
DxO Optics Pro 8, sharpening=off, NR=off, CA=check, PF=check, lens softness = check

The sharpness of the OOC JPEG is not bad, but it does show some red-green CA. It also shows USM halos elsewhere.


I think DxO gives somewhat sharper results than ACR, but without the lens softness option the difference is not dramatic. When the lens softness correction is turned on, the difference is very significant. This is without any USM enabled, which could be done on top of lens softness correction. This option does not create outlines like the USM tends to, but it can create a different type of artifacts along the edges. It is easy to dial this correction down.

ACR can dial in some sharpness with USM, but so can DXO on top of the softening correction.

Overall, I think that the last crop shows a very satisfactory corner sharpness without any offensive artifacts, although obviously it is not CA alone that is being treated here. The main point though is that the 9-18 zoom can deliver sharp images with some postprocessing, but the great thing is that it's completely automatic, and does not require any tinkering.


Vlad
 
Vlad S wrote:
Anders W wrote:

Could you please post an example with the same image, before and after correction, as processed by LR as well as DxO, preferably with sharpening reduced to zero (so we can rule that factor out completely) and with the lens a bit stopped down (so we can see the remaining CA more clearly; it might be covered up by ordinary blur otherwise).
Finally was able to prepare some tests.
Thanks for getting back on this.
I have noticed that in DxO I actually did have an option turned on, which is called "lens softness." I could not get a very clear explanation how it differs from the unsharp mask, which is a separate item, but I am getting an impression that it tries to treat coma.
It's not only coma I think. Part of the reason why DxO test lenses, if I understand things correctly, is that they develop lens profiles specifically adapted to their optical properties in all relevant regards.

I have included conversions with and without his option. The first one is an overview, and the remaining images are 100% crops from the bottom right corner. Some additional comments under the crops.

Overview of the scene. 9-18 @ 9 mm, f/5.6
Overview of the scene. 9-18 @ 9 mm, f/5.6

100% crop of the OOC JPEG. Sharpness=0 NR=low
100% crop of the OOC JPEG. Sharpness=0 NR=low

ACR 6.7 conversion. sharpening=off, NR=off, CA = check, defringe=all edges
ACR 6.7 conversion. sharpening=off, NR=off, CA = check, defringe=all edges

DxO Optics Pro 8, sharpening=off, NR=off, CA=check, Purple fringe=check
DxO Optics Pro 8, sharpening=off, NR=off, CA=check, Purple fringe=check

DxO Optics Pro 8, sharpening=off, NR=off, CA=check, PF=check, lens softness = check
DxO Optics Pro 8, sharpening=off, NR=off, CA=check, PF=check, lens softness = check

The sharpness of the OOC JPEG is not bad, but it does show some red-green CA. It also shows USM halos elsewhere.

I think DxO gives somewhat sharper results than ACR, but without the lens softness option the difference is not dramatic. When the lens softness correction is turned on, the difference is very significant. This is without any USM enabled, which could be done on top of lens softness correction. This option does not create outlines like the USM tends to, but it can create a different type of artifacts along the edges. It is easy to dial this correction down.

ACR can dial in some sharpness with USM, but so can DXO on top of the softening correction.

Overall, I think that the last crop shows a very satisfactory corner sharpness without any offensive artifacts, although obviously it is not CA alone that is being treated here. The main point though is that the 9-18 zoom can deliver sharp images with some postprocessing, but the great thing is that it's completely automatic, and does not require any tinkering.
I have no difficulties believing that DxO is better than ACR as well as LR when it comes to sharpening, thanks to better algorithms in that specific regard as well as their work with the lens profiles. However, LR (and reasonably recent versions of ACR) do not use USM alone and can certainly to a better job than the one you exemplify here. So with everything properly adjusted, the likely pecking order with respect to sharpening is DxO, followed by LR/ACR, followed by OOC jpeg.

When it comes to CA, see the comments I just added in the thread where we are already discussing that subject.
 

Keyboard shortcuts

Back
Top