Correction #4 - Vignetting

I purposely shot an image with an aperture of f/3.5 at maximum wide angle to maximize the amount of vignetting in the original. Of course, in normal daylight circumstances, one would rather shoot around f/8. Given the fact that DxO Optics Pro is camera- and lens-specific, I expected the vignetting to be at least, if not better, than what a general application like Photoshop CS achieves. Photoshop CS removes, just like Nikon Capture 4.1 virtually all vignetting.

Again, DxO reduced vignetting but did not completely eliminate it as shown in this example. This makes a strong case for an optional slider that can be adjusted by the user. The only benefit is that DxO works on JPEG in an automated way while Photoshop CS and Nikon Capture vignetting reduction only work in RAW and require user intervention. Vignetting can also be fixed in Photoshop on JPEGs via a levels adjustment layer with radial gradient mask.
Update (September 22, 2004): The just released PTLens 5 allows you to adjust vignetting via a slider.

The examples below are downsampled crops of the upper right hand corners of the originals.

Original	After DxO Optics Pro
850 KB	937 KB
Notice the vignetting to the right	DxO removed vignetting partially (the sky on the right side is still darker than the sky on the left side)

After Photoshop (JPEG)
850 KB
Vignetting is compeletely removed via a Levels Adjustment Layer with Radial Gradient Mask.

After Photoshop CS (RAW)	After Nikon Capture 4.1 (RAW)
1,007 KB	1,096 KB
Vignetting is compeletely removed	Vignetting is compeletely removed

Conclusion

DxO is very easy to use and allows for automation as it requires virtually no user intervention. This benefit is at the same time a disadvantage because the results are "as is" due to the lack of user adjustable parameters or sliders. In theory this should not be necessary because DxO uses tailored mathematical models which take into account specific camera-lens combinations. But in reality, the distortion*, chromatic aberration, and vignetting results were only partial.

Also, DxO is an expensive solution for just four JPEG-only corrections for a single camera-lens ombination. Additional lenses or cameras require the purchase of additional modules (if available). Of course, more important than price is the actual performance:

• The visible result of the Blur correction was very similar to the visible result of a Photoshop Unsharp Mask. However, it is hindered by the JPEG format and the lack of user controls. But the main drawback is that DxO Blur requires other image adjustments—such as contrast enhancements—to be applied on arleady sharpened images, which is certainly not the generally accepted worfklow order.
   
• Barrel Distortion* was only partially reduced compared to PTLens, a freeware based on Panorama Tools which has more functionality and fewer limitations.
   
• Chromatic Aberrations were only partially removed compared to Photoshop Camera RAW or manual JPEG techniques, and some types like purple fringing are not covered by DxO Optics Pro.
   
• Vignetting was only partially eliminated and the result was inferior to Photoshop (JPEG and RAW) and Nikon Capture (RAW), partly due to the lack of controls.

* Update (September 22, 2004): It seems that the D70 with the 18-70mm kit lens reports all focal lengths between 18 to 21mm as 18mm in the EXIF data. Because DxO is based on EXIF information, it is unable to properly correct this most critical part of the zoom range. In particular, the results at 18mm are not optimal. DxO Labs has informed us that this is because they made a compromise by averaging the settings between 18mm and 21mm. Our thoughts on this:

• DxO Labs should not have made this compromise, as most photographers will shoot at maximum wide angle at 18mm and not at 20 or 21mm. This looks more like a mathematical compromise rather than one made by photographers and it raises questions about how DxO Labs optimizes its parameters. DxO Labs has admitted the compromise made was not ideal and will correct it, thanks in part to our review.
• Since DxO Optics Pro is based on EXIF, other body and lens combinations may face similar issues. DXO Labs confirmed this assumption but did not specify which camera and lens combinations were affected.

Automation and ease-of-use often require compromises that lead to less than optimal image quality. I was hoping that this camera-lens customized automation would offer the best of both worlds. Unfortunately it does not achieve that. DxO Optics Pro is based on a great idea but it needs improvement in five areas:

• First and foremost, better than partial results for distortion, aberration, and vignetting
   
• Have the option of (adjustable) sharpening at the end of the workflow
   
• Optional user intervention (unless results are optimal in all situations)
   
• RAW support (we assume an upcoming version will address this)
   
• A more affordable pricing and pricing structure

A free DxO Optics Pro demo can be downloaded from the DxO Labs website.

This article is written by Vincent Bockaert, author of The 123 of digital imaging.