Sigma 50-150 OS and MTF chart comparison

Started May 14, 2012 | Discussions thread
Flat view
motobloat Regular Member • Posts: 497
Sigma 50-150 OS and MTF chart comparison

So I spent some time comparing the MTF charts of the Sigma 50-150 OS with a few other Sigma lenses. Several reviews of their lenses have mentioned that Sigma's MTF charts are good indicators of the lenses' actual performance, so I thought I'd take a look. I had to carefully stretch the graphs in GIMP in order to line up the full-frame lens charts with the APSC charts. All of the grid lines (horizontal and vertical) were perfectly overlapping, so I am very confident in the comparability here.

If you want to do some reading on MTF charts, there is a quick explanation here: and here and here

All of Sigma's charts are at the widest aperture and are displayed in the exact same way as Nikon's and Canon's MTF charts. However, their MTF testing procedures may be different so I did not bother to directly compare them.

The most important set to compare is the green vs. the magenta lines that test resolution (MTF 30) because that is where the big differences lie. Almost all of the lenses perform about the same in terms of contrast (MTF 10). Also, the key test points on the x-axis are at 8mm and 12mm, because these are the edges of the APS-C frame (which is approximately 16mm x 24mm). Another key point is of course at 0mm, which is the center performance. When a lens has divergent solid (sagittal) and dashed (meridional) lines, a sensible approach is to draw a line halfway between the two lines to evaluate the lens.

So first, tests at 50mm f/2.8

The lens to beat here is probably the 17-50 f/2.8 OS , which is widely noted to have excellent wide-open performance at 50mm.

Here we see very good results indeed. Very slightly less sharp at the center (green line below the purple line at 0mm on the x-axis) but about the same at 8mm and 12mm (for the lower line). Overall, when comparing the 17-50 @ 50mm versus the 50-150 @ 50mm, you won't see much difference. The 50-150 will be a tiny bit worse in the center but a bit better in the corners.

Since we know the new 50-150 lens is based on the 70-200 f/2.8 OS , we should compare them.

Looking at the graph below, we can see that the 50-150 @ 50mm handily beats the 70-200 @ 70mm, at 0mm and 8mm, with about the same performance at 12mm. By 12.5mm, things start to fall apart, but that is also PAST the edge of the APS-C frame anyway (which is 24mm wide, or 12mm from the center). So the 50-150 f/2.8 @50mm should be a bit better than the 70-200 f/2.8 @70mm, besting an already good performance.

So next, tests at 150mm f/2.8

The lens to beat here is probably the 150mm f/2.8 OS Macro , which is widely noted to have astoundingly good wide-open performance.

Looking at the graph below, we see that the prime lens does indeed best the 50-150, but only in one metric - the Meridional line test. While many pages say that "The closer the dotted line and thick lines are on the chart, the more pleasing the bokeh will be" this is not really true at all.* Just because the lines are the divergent doesn't mean the lens will have bad bokeh. The best way to judge bokeh quality is from sample photos.

In fact, what actually does tell you is that there is some astigmatism going on, which will reduce resolution. Taking the midpoint of the two green lines, however, we can see that it doesn't suffer much from the 150mm macro, and it is a very good result considering that the 150mm is a prime lens and the 50-150mm is a zoom.

Overall, the 50-150 is about the same in the center as the 150mm macro, but will suffer a bit at the edges due to astigmatism. Strong results though

Finally, let's compare it to its big brother, the 70-200 f/2.8 OS at 200mm.

Looking at this chart, we can see they start out the same in the center, but the 70-200 rapidly drops off in sharpness, and the 50-150 @ 150mm easily bests the 70-200 @ 200mm.

So what are the conclusions?

Overall, I was really hoping for an improvement over the 70-200 OS, which had some issues in sharpness at the corners of the frame, and it looks like this lens delivers. In fact, it appears to compete well with some of Sigma's best performers at 50mm f/2.8 and 150mm f/2.8. If you want to see even more test charts , including some comparisons against some rather low-quality Sigma lenses, they are in my gallery.

Here is also some food for thought, although not directly comparable. The highly-regarded Nikon 70-200 VRII has similar MTF patterns as the Sigma 50-150, with a worse, diverging meridional (dashed) line.

* See page 33 here for an explanation (highly technical) of why bokeh can't be determined from MTF graphs: $File/CLN_MTF_Kurven_EN.pdf

Flat view
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow