Modulation Transfer Function...What does it mean

FlowtographyBerlin

@people: What I really don't understand about forums is why it is so frequent that people tell eachother that they need to "go back to class", "do some more reading", "learn the basics before you post", all that. What the heck, seriously.

Everyone makes mistakes, is mistaken, some more, some less, some have a different first language than English, some people know about this a bit and not about that, some the other way around. There are many people who don't know jack about physics but take excellent photos, and then there are many people who talk about lens scores and rarely actually shoot. So? That's what a forum is for, to exchange information (& opinions).

Why, why in the world is it always that people must tell eachother that they're not expert enough on a topic?

Be nice to eachother!

Golf007sd

I, on behalf of the members, I want to take the time and thank both TTJ @ Ade for there input. They both have taken their time and energy in trying to bring this very complicated issue in reading/understand MTF charts into light for us all. Thanks guys....hats of to you both.

At NRF we are like a family; and like any other families, the individuals within the circle will, at times, have a way of expressing themselves in a fashion that is a bit sharp...nothing is wrong with two family members having a spirited conversation. I trust each pro-active member has respect for the other, even though they may not see eye-to-eye. So long as each address their point in relation to the topic being discussed, I find it all very enlightening and informative.

Cheers to you all....

FlowtographyBerlin

I, on behalf of the members, I want to take the time and thank both TTJ @ Ade for there input. They both have taken their time and energy in trying to bring this very complicated issue in reading/understand MTF charts into light for us all. Thanks guys....hats of to you both.

That, too. Thank you!

Ade

So now that we have some idea of what MTF is, how it is measured, and how it relates contrast to resolution, we can explore the many different ways MTF data is used.

Usually, when MTF is measured (or calculated) we end up with a lot of measurement points. So we produce different charts depending on which relationships we want to explore.

I've created a few "representative" charts. Please note that I made up the values and the curves in the charts (they don't pertain to any real lens). However the overall shape of the curves might be similar to what you'd see in a real chart.

Here's the most common type of chart, plotting MTF vs. Distance from the lens center:


Figure 1: MTF vs. Distance

This kind of MTF chart is useful to evaluate the overall performance and consistency (uniformity) of a lens. I think by now we're all familiar with this chart. I don't want this post to be about the mechanics of chart-reading, but the important point to note is with this type of chart we are analyzing MTF across the physical surface of the lens, from center to edge.

Here's another very common type of chart, plotting MTF vs. Frequency:


Figure 2: MTF vs. Frequency

This chart is very different from the first one. We are still measuring MTF (as can be seen on the Y-axis). But we are no longer measuring from center to edge. Rather, we are measuring the maximum resolution of the lens, at a particular location (usually center).

E.g., from this type of chart, you can tell at 50% contrast (MTF 50) this fictional lens has a resolution of around 1900 LP/PH.

The next chart is a very interesting one, we plot the effects of back focus (or front focus) on MTF:


Figure 3: MTF vs. Focus Shift

This chart examines how MTF is affected when the lens is not perfectly in focus. During MTF testing, it is common to use "focus bracketing". That is, we take pictures of the test chart at several different focus positions, then pick the best one. If we'd plot the results at different focus positions, we get the picture above.

By the way, this is also how "contrast detect" auto-focus systems work. The contrast detect AF algorithm measures contrast at slightly different focus positions, then moves the focus to the peak contrast position.

Now I'm going to present some variations of the above charts.

The first variant is pretty common, we plot MTF 50 resolution vs. aperture (f-stop):


Figure 4: MTF 50 vs. F-stop

Lenses are often limited by aberration at wide-open apertures and then by diffraction when stopped-down. By plotting the MTF 50 resolution vs. F-stop, we can find the "sweet-spot" aperture of the lens (which is f/4 for the above lens). This is the aperture where the lens is the "sharpest".

Another variant is plotting MTF vs. Frequency (as in Figure 2) but analyzing R, G, B wavelengths separately:


Figure 5: MTF at RGB wavelengths

This type of chart looks at how the lens design handles single colors (monochromatic light). When a lens is being designed, it would not be unusual to run simulations to predict how the lens will react to R, G, B wavelengths.

The last variant I'll illustrate today looks at MTF 50 resolution as a function of ISO:


Figure 6: MTF vs. ISO

This one is probably self-explanatory, but I included it here as a reminder that MTF measures the entire system (including the camera's sensor) -- not just the lens.

As you may appreciate, there are many ways to look at MTF data. There are actually many other types of MTF charts. For example, in biology and medicine, a researcher might plot our eye's MTF vs. Age to see how our vision deteriorates with old age.

With a better understanding of what MTF really is -- rather than how to decipher a particular chart -- it is possible for us to look at a variety of MTF charts and gain further insights about the equipment which we use everyday.

JJ_SO

Great work! Thank you very much for showing it!

Golf007sd

Good stuff. Thank for the info.