Here are links to Sony's own MTFs for the 50mm 1.2 GM and the 50mm 1.4 GM (go about halfway down the page at each link).
I am very far from expert at reading MTF charts, so I will readily accept correction from the more knowledgeable in how I understand these.
But my reading of these MTFs is that, contrary to Lenstip's results, there is little difference in sharpness between the two lenses and what there is slightly favors the 1.2 GM. I also interpret some of the divergences in the frequency lines to indicate that the 1.2 is a bit less corrected than the 1.4 which, assuming a design choice, may indicate that the 1.2 has better or more special rendering.
chiron wrote:
Here are links to Sony's own MTFs for the 50mm 1.2 GM and the 50mm 1.4 GM (go about halfway down the page at each link).
I am very far from expert at reading MTF charts, so I will readily accept correction from the more knowledgeable in how I understand these.
But my reading of these MTFs is that, contrary to Lenstip's results, there is little difference in sharpness between the two lenses and what there is slightly favors the 1.2 GM. I also interpret some of the divergences in the frequency line to indicate that the 1.2 is a bit less corrected than the 1.4 which, assuming a design choice, may indicate that the 1.2 has better or more special rendering.
So: Sony's charts are calculated, not measured. Since they're calculated the same way, though, you're still getting a generalized comparable chart between the two.
Lenstip's results are measured, but they are also only single instances of a lens. So you get "real world" results, but those results are not necessarily generalizable. You just know how two samples of two lenses compare.
MAubrey wrote:
So: Sony's charts are calculated, not measured. Since they're calculated the same way, though, you're still getting a generalized comparable chart between the two.
Lenstip's results are measured, but they are also only single instances of a lens. So you get "real world" results, but those results are not necessarily generalizable. You just know how two samples of two lenses compare.
Yes, this is an issue with almost all published lens reviews. Any two specific copies of a lens fall in a range of variation and most sites for very real practical reasons test only one copy of a lens. But with single copies the sample size is so small that it actually isn't generalizable at all with any confidence.
As far as I know, only Fred goes to some trouble to make sure he is testing and comparing well-centered copies and only Roger Cicala tests multiple copies (and his tests of multiple copies usually demonstrate how much variation there is among different copies and therefore how unreliable single copy comparisons can be).
Sony's calculated MTFs offer a more reliable guide to how average/good copies should perform and how one lens compares to another.
With Sigma's 1.4 50mm, now joined by Sony's 1.4 (and the existing 1.2) photographers have three lens with significantly different price points, and minutely different performances.
For still images, I don't think there is much to detrimentally affect anyone's photographs. Unless of course we are pixel peeping.
Take the same photograph will all three lenses at 1.4 and I seriously doubt that anyone could tell them apart without zooming in.
If either of these 1.4 lenses had been available when I purchased the 1.2, I would have seriously considered either of them.
Allowing 3rd party lenses is benefiting the e-mount system. Canon needs to take note.
From my take on the MTF: at f/8 the 1.2GM looks better, wide open I would prefer the 1.4GM. From my perspective higher lines and keeping the lines closer together (less astigmatism) is good.
The question about overcorrecting affecting the rendering sounds complex. The MTF describes the image that is in focus, but the rendering is as I understand it concerned with the optical behavior for out of focus sections. So I'm not sure MTF charts give any insight into this.
I have to say honestly, I don't get much of the discussion about rendering. I'm not saying it's not real, just that I'm not in tune enough to appreciate it. From what I can tell, a good sign is uniform bokeh balls with no outlining. The bokeh balls will give an indication how an out-of-focus point will be displayed, so it is to some extent a measure of rendering. I also prefer less cats eye as I don't want things to get swirly at the edges and I like less vignette.
The other thing seems to be how fast and nicely the out -of-focus section transitions. I think the transition is also where loca would be important. I would like the out of focus sections to still maintain an apochromatic character, just not be in focus. So no or low loca should give better rendering (I think). I don't believe you can see that from the MTF, but I would like to be corrected if I'm wrong.
They're probably both excellent if you need 1.2 get 1.2. It's difficult to get anything oof at this FL and the 1.4 isn't much smaller or less expensive imo.
If you are looking for f/8 you can get lighter options that might even be better. My opinion on the 50/1.2 nikkor I used to have is that it probably topped out by f/4, the lenstips seemed to say the same on these. MTF is contrast not resolution though. Not that I think anyone is going to listen to me on that
It's probably not going to be a good idea getting rendering out of MTF.
To answer the OP's question the MTF charts are pretty similar. Keep in mind that the left chart is at f/1.2 for the f/1.2 lens and f/1.4 for the f/1.4 lens. The right chart is at f/8 for both lenses. Although they are close the chart is slightly better for the f/1.2 lens but only a little better. At the wide open aperture it has a couple of percentage points higher contrast at 30 lp/mm than the f/1.4 in the center and in the corner the f/1.4 lens drops slightly below 60% contrast for the trans orientation and 30 lp/mm and neither orientation drops below 60% for the f/1.2 lens. This slightly higher contrast at 30 lp/mm has to be weighed against the lines generally be closer (probably suggesting less astigmatism) for the f/1.4 lens. The MTF suggests pretty low astigmatism for both lens, but the final point is the trans line at f/8 and 30 lp/mm is higher for the f/1.2 lens suggesting less astigmatism stopped down for the f/1.2 lens. So overall I would say the f/1.2 lens looks just a quite small amount better in the Sony MTF charts, but probably not enough better to really be able to see in real world images. I would expect really similar performance.
