Started 7 months ago | Discussions thread
Joofa Senior Member • Posts: 2,655
Re: BxU, JBM and CRB / FI

fvdbergh2501 wrote:

Joofa wrote:

Many of the stuff thrown around in this thread such as MTF50, MTFxx, or tools such as MTFMapper, or even ImaTest, are not designed to work with natural images, in the way as asked in this thread.

Unfortunately, the discussions involving them, such as on this thread, proceed into the domain of natural images without even realizing that they are not applicable for such experiments.

Very few metrics address such issues for natural images. For e.g., JIDM / JBM is one of the few that does operate on natural images. This is an important consideration, regarding suitability of techniques or tools to the problem at hand, that is almost invariably overlooked on this forum.

Here we have some possible scenarios (not exhaustive, mind you) :

1. Microcontrast is a property of an optical system that is observable through the transfer function (i.e., it shows up in the MTF), and therefore it is measurable using the slanted-edge method, for example;

2. Microcontrast is not reflected in the optical system transfer function, but it is readily measurable in images under controlled circumstances. This puts microcontrast into the same category as image noise (photon noise, read noise, etc.), which ultimately does not bias the MTF, but can quite clearly reduce perceived sharpness of an image;

IMHO, Microcontrast is not a property of an optical system. MTF is a property of an optical system, for e.g. Microcontrast, whatever that is, is a property of an image - imaged object that is. MTF can alter the final image reproduced, and any associated microcontrast, as the 'original' object being imaged passes through the optical system.

3. Microcontrast is a perceptual phenomenon that strongly depends on context, meaning it cannot be perceived without some understanding of what the subject of the image is. To me, this sounds rather far-fetched, but I am leaving this scenario here to cover the "I know it when I see it" case.

So I can readily agree that there are phenomena not captured by the MTF. What remains to be shown is that microcontrast is indeed such a phenomenon.

Firstly, let me say that I'm not against MTF for measuring optical system properties associated with a lens. And, for that purpose I think your MTFMapper is a fantastic tool to have in one's arsenal. You have put in a lot of effort and it is helpful for many people as I see it.

Secondly, there are different ways of looking at the same optical system. MTF is one. However, the same imaging plane intensity pattern could be obtained by 'replacing' the lens (with traditional notions of MTF) with another lens which intensity point image is a different function than the original PSF and replacing the original continuous input object structure with a set of equally spaced incoherently radiating points. What is the 'MTF' in this case, then? BTW, this phenomenon is called the matrix theory of optics, not to be confused with matrix theories of geometric optics or ray optics, etc.

Whatever approach one takes, at the end of the day one needs to analyze image content - natural image content in our case. And, not images of bar charts, which are fine for determining the MTF of a lens. For defining 'microcontrast' one doesn't need MTF, which is a property of the optical system. One can study the effects of MTF on microcontrast, but that is not the definition of microcontrast. Hence, that is why I consider discussions of micorcontrast limited to just MTF to be lacking.

Post (hide subjects) Posted by
(unknown member)
(unknown member)
(unknown member)
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow