PHOTOOG Photography writings by Olivier Giroux


Commentary: 24G's MTF

Here are super-imposed the charts for the 50mm f/1.4G (lower red & blue) and 24mm f/1.4G (higher red & blue) lenses. There isn’t a single point where the 50mm beats the 24mm.

Most satisfactory, I suspect this lens is quite special now.

Let’s see what happens when I super-impose the charts for the 24mm f/1.4G and Leica’s 24mm Summilux-M ASPH. The faint/light red curves below are Leica’s; I have drawn a yellow curve between the lowest pair and the second lowest pair of these that is comparable in metrics to Nikon’s blue curve.

Oooooh, it is special.

To the exception of the extreme corners, the theory is that Nikon’s implementation surpasses Leica’s.  The Leica is almost assuredly smaller and less conspicuous but it doesn't eclipse the competition with its performance.

Finally, it's relevat to know how it compares to the Canon.  The lower set of curves, the fainter ones, are comparable to Nikon's.


The Nikkor is better over the whole frame, but not by much.



Comments (4) Trackbacks (0)
  1. You do realize that (1) Leica and Zeiss provide measured MTFs for actual lenses, while Canon and [it is widely believed] Nikon provide MTFs for computed designs; (2) Leica provides curves for 5, 10, 20, and 40 cycles, while Nikon and Canon provide MTFs for 10 and 30 cycles; and most importantly (3) none of these manufacturers specifies the wavelengths of light used to generate their MTFs or what happens at middle or near focus rather than infinity… never mind with real sensors.

    The result? Attempts at a rank comparison of the lenses based on these MTFs are nothing more than pseudoscience.

    (For what it’s worth I shoot Nikon, Leica, Zeiss, and Olympus glass and no, I don’t own any of the lenses in question.)

  2. Yes, I realize these things. I’m reasonably well read on the subject actually.

    In response to your inquiries: (1) Leica’s current MTFs are actually computed by software rather than measured in the lab, (2) the chart above has interpolated data for 30lppm for the purpose comparison, (3) Leica and Zeiss use a broad spectrum of light (or white light) while Nikon uses a single wavelength of green light and that clearly favors Nikon because then chromatic aberrations don’t lower the MTF.

    The result? Indeed, pseudoscience.

    HOWEVER, the mere fact that manufacturers provide this data is an invitation to comparison. The data is the manufacturer way of saying “my product is better than the other product, so pick mine”. There is no consumer-facing documentation that is not intended for comparison to the competition.

    The fact that this documentation is provided in a way that’s not well organized across the industry is something to keep in mind, but whether it’s done closely or at a glance, people are still comparing them. I just went a bit overboard by folding all the data into a single graph.

    Alright. Your comment was pretty confrontational in tone. Please stay curteous, and have a nice evening.

    – Olivier

  3. Sorry if I seemed confrontational, and thanks for the reply. I knew that Zeiss and Leica provide white light MTFs; I was not aware that Nikon used green light or, apparently, that Leica had switched to providing computed MTFs. Zeiss, of course, states that *all* their MTFs are data from real lenses.

    Now, that said: The combination of white and green light MTFs is probably the single biggest problem in your comparison. This can make a HUGE difference. Your statement that “To the exception of the extreme corners, the theory is that Nikon’s implementation surpasses Leica’s” might in fact be true, but there is absolutely nothing in the posted data to support that conclusion. Indeed, I’d go a step further: If the curves are that close under that particular condition, it’s likely that the Leica lens absolutely stomps the Nikon — in principle.

    In practice, of course, we have to also think about the iaging system, and not the optical bench. The Nikon is a telecentric design that probably plays very well with the digital sensors it was designed for, while the Leica probably does a heck of a lot less well on that count…

    • My statement you quote was probably overreach, correct.

      I’m going to throw another caveat on top of the good ones you already mention: focusing the Leica accurately is probably easier to do consistently than the Nikon. If you’ve followed the story of this lens so far, you’ll know that all cameras (except the D3S apparently) have difficulty focusing with it.

      Bottom line: comparison is difficult.

      I can offer some first-hand experience however, with my review that I posted later:

      I doubt I will ever review the Leica unfortunately.

Leave a comment

No trackbacks yet.