Diffraction Calculator Not Matching Experiences with Panasonic Cameras

I was recently discussing how this website's diffraction limit calculator has not matched my own experiences...

For the Panasonic FZ30 (1/1.8, 8.32mp, 2.2 micron) you're calculator suggests around F4 to be the diffraction limit... in practical use most people seem to get their best results at F6.3 to F7.1... this is odd as it is well into the area of diffraction. Furthermore, we've noticed that when adding the Olympus Tcon17 (1.7x teleconvertor) in front of the camera we get our best results at F5.0... now before i go on let me explain what I tend to call best results

Best results
- Sharpest clearest images
- least CA and PF
- appearance of the best resolution *i say appearance as we have no numbers to judge it on

I manually did the math for the FZ20 (1/2.5, 5mp, 2.2 micron) and came out to a diffraction F-stop in the ball park of F4.4-4.5... this is closer to the forums experience as we figured best results were at F5.2 to F5.6... now i know the deterioration is gradual and not immediate but can you better explain the difference between our real world experiences versus the equations?

Does focal length play a role in diffraction? or what about glass quality? might aperture design have an effect on it as well (number of blades, or Nikons rounded blade system)?

One more point of reference, the Panasonic FZ50 (1/1.8, 10mp, 2.0micron) the DPreview review on it says noticeable diffraction past F5.0 and some users tend to agree that F4.6 top F5.0 is a sweet spot but that hasn't been confirmed as much as the other 2 older cameras.

I'd sure like to hear anyone's thoughts on this.

I really appreciate your thorough look into this site's diffraction calculator.

The diffraction calculator is designed to be as accurate as possible while still keeping in universal, however as you and others have noticed, often times other factors in the type of lens-cmos/ccd optical system can complicate matters.

One particularly important concept is the difference between the diffraction limit and a lens's "sweet spot", which can be very crudely described as the optimum aperture which minimizes the sum total contribution of both diffraction and aberrations. Unless your lens is a perfectly designed lens (aka diffraction limited lens), often times aberrations play more of a role than diffraction when determining your sweet spot. This effectively means that the sweet spot is located at a narrower aperture than the diffraction limit. CA and PF, as you mention, are key characteristics which worsen as the aperture gets wider.

This does not, however, necessarily mean that the diffraction calculator is inaccurate; the diffraction limit should not necessarily correspond with what aperture looks best-- all it tells you is a theoretical maximum resolving ability (which you say is not being measured). An example of this is the concept of local contrast; often times the appearance of a sharp, clear image is more dictated by small-scale local contrast (MTF50) than by absolute resolution. Local contrast can in fact improve slightly beyond the diffraction limit, even if absolute resolution decreases. Further, in-camera sharpening tricks, Bayer conversion and other post-processing in-camera can mean that having an airy disc slightly larger than the pixel size will not be as noticeable, and therefore one can push the aperture narrower than one would ordinarily calculate.

Finally, by sharp, clear image one needs to specify "where". There is always a trade-off between optimum sharpness in the center of the image and sharpness towards the edges. Usually far in excess of the diffraction limit one often finds this an acceptable trade-off because of the huge improvement in edge (often foreground) sharpness (not to mention sharpness from a wider depth of field). This is particularly true when the image sensor is designed to use the entire lens projection, such as in the Panasonic FZ30 (and unlike cropped 35mm cameras).

Overall, I suspect that the lens on the Panasonic, being a fixed 12X zoom, probably needs to be stopped down significantly to minimize aberrations. If one were able to replace this lens with a fixed focal length prime lens, I would expect that the sweet spot would more closely correspond with the diffraction limit.

Hope this has answered your query. Please let me know if you discuss this any further with any colleagues also using the Panasonic digital cameras.

---
PS: F-number, not focal length, is the most important factor for diffraction. See the end of my page on digital camera sensor sizes for more on this. The focal length and absolute aperture size effectively cancel. This being said, longer focal length lenses (if not extreme zoom lenses) are often easier to design so that they are nearer their diffraction limit.

Thanks for the detailed explanation... do you mind if i post my email and your response to some of my colleagues? One thing however, the FZ30 and 50 may cover an extensive range but as for CA & PF which i have mentioned its not the CA & PF that you'd expect from a kit lens and its even lower then some of the Canon L glass primes I've seen, there are 2 factors that make this possible in such a long range of focal lengths,

1) Leica Designed lens (built at the very new Panasonic glass labratory which was able to be 10x more precise then Leica's best glass laboratory)
2) there are routines in the firmware that do get rid of CA and PF (i know its cheating) :)

put both of those together and you get a beautiful lens unlike all the other 12x zoom cameras out there.