16 Nov 2006 Understanding MTF Charts (Podcast 62)
There’s always a lot of talk in the forums about how sharp this lens is, how sharp that lens is. Although getting advice from people that already own and are using a particular lens will always be useful and online lens reviews can really help in making up your mind whether or not to take the plunge, another way to judge for ourselves if a lens will really cut it is to view it’s MTF Chart. MTF charts are subject to some variance, so by all means don’t make them the single point of reference. As a great starting point though, we can compare any particular lenses MTF chart with that of a lens that we already own for example, to see if it would provide better or worse results. Having a baseline to compare with will hopefully help.
Another word of warning before we start is that how an MTF Chart is created varies between manufactures, so it is not advisable to compare say a Canon and a Nikon lenses MTF Chart. Also many manufacturers don’t even publish their MFT Charts. Canon for example releases an MTF chart for almost all of their lenses, while Nikon does not, at least not in the US. You can find MTF Charts for Nikon lenses on the Japanese Web site though. I’ll drop some links to the Canon and Nikon Japan Web sites into the show notes, and we can look at them to get the point. So let’s get started and try to make sense of all those squiggly lines.
So, if you are listening at a computer, please look at the show notes, either in iTunes or on the Podcast page at martinbaileyphotography.com and jump to either the Canon or the Nikon site and select one of the lenses, and take a look at the MTF chart at the bottom of the page. If the page you looked at doesn’t have the chart, go back to the lens list and keep looking until you find one that does. If you are listening while away from a computer, try to recall what an MTF chart looks like. It’s a small square graph with 0 to 1 in 0.1 steps going up the vertical axis. We also can see 0 to 20 in steps of 5 going along the horizontal axis, with a little more space after the 20. Along the top are a number of solid and dotted lines, some coloured and some black. We’ll get into the meaning in a moment, but if you aren’t already looking at a chart, hopefully now you’ll have recalled seeing one of these charts at some time in the past. It would also be great if you could take a look later when you can get to a computer.
In starting to understand an MTF chart, we should think of the vertical axis of the chart as the lens axis. That is, if we imagine a line drawn down the very centre of the lens’ cylinder from the front element to the back element it would coincide with the left hand side vertical axis of the MTF chart. The numbers 0 to 1 in 0.1 steps can be thought of as 0 to 100%. The 0 to 20 along the bottom is millimetres. As the chart actually extends a little further past the 20, we need to think of this as being roughly 22 millimetres. This is basically the number of millimetres from the centre of the image, which is about 22mm. Note that we are talking about the image on the film or digital sensor, not 20mm from the centre of any single lens element, which of course can be much larger. Remember that a piece of 35mm film or a full sized digital image sensor is 24mm high and 36mm wide. This means the diagonal line across the film or sensor is fractionally over 43mm, so a chart of the radius of the lens at 22mm would give us an image of 44mm that is necessary to create a 35mm camera’s image circle.
MTF stands for Modulation Transfer Function, which is the normalized spatial frequency response of the lens. Now before you switch me off and start listening to a nice easy to understand Podcast, let me explain spatial frequency and how it’s measured. It really isn’t that difficult so please bear with me. If it’s tough to understand, just rewind a little and listen again. Basically we’re dealing with measurement of contrast and resolution of patterns of pairs of lines per millimetre. Contrast is measured using patterns of 10 pairs of lines per millimetre and resolution is measured using patterns of 30 pairs of lines per millimetre. Now, by pairs of lines, I mean one white line and one black line. You could say that this is just one line, and you would not be wrong as a white line alone does not make a line. It needs the black line to exist, but let’s just understand that 30 pairs of lines means 30 black lines separated by 30 white lines. Anyway, I’m not sure if what I’m about to say is a very good scientific explanation, and to be honest if it helps to understand these charts I don’t really care, but let’s think of spatial frequency quite simply as how frequently (quote/unquote) a pair of lines appear in the space (quote/unquote) of one millimetre. I’m going to repeat that… How frequently a pair of lines appears in the space of one millimetre is the spatial frequency.
So, what do all the lines in the charts mean? Well, in the Canon MTF charts you’ll notice eight lines. There are thin and thick, black and blue solid lines and thin and thick, black and blue broken lines. The thick lines represent measurements of patterns of pairs of 10 line pairs per millimetre, or the contrast of the lens, and the thin lines represent measurements of 30 line pairs per millimetre, or the resolution of the lens. The black lines are measurements taken with the lens wide open at its widest aperture and the blue lines represent measurement taken with the lens at f/8. The solid lines represent Sagittal measurements, which are of sets of lines parallel to a line drawn from the centre to the edge of the lens, and the broken lines represent Meridonial measurements, which are of sets of lines running at 90 degrees to the lens axis. Multiple pairs of Meridonial and Sagittal line patterns at right angles to each other are measured diagonally across the 35mm frame.
On Nikon MTF charts, you’ll notice only thin blue and red lines, both solid and broken, making a total of four lines. From the Nikon Japan Web site I see that the red lines represent 10 line pairs per millimetre measurements, and the blue lines represent 30 line pairs per millimetre measurements. Unbroken lines represent the Sagitall line patterns, that is lines running parallel to a line drawn from the centre to the end of the lens, and the broken lines represent the Meridonial line patterns, which are lines running at 90 degrees to the lens axis.
So now that we now what the various lines mean, how do we make sense of the MTF charts? Basically the higher up the chart the 10 LP/mm line, that is the thick lines on the Canon charts or the red lines on the Nikon charts are, the higher the contrast reproduction capability of the lens will be. The higher up the chart the 30 LP/mm line, that is the thin lines on the Canon or the blue lines on the Nikon charts, the higher the resolving power, which means the lens will give sharper images. Remember that for Canon charts the black lines show the lens wide open while the blue lines show the lens stopped down to f/8, so we can compare how well the lens will perform wide open compared to f/8, which is usually considered the sharpest aperture for most lenses. Lenses that have black and the blue lines close together will perform pretty much the same wide open as they would at f/8. In one book I have from Canon it states that if the thick lines on a Canon chart that represent the 10 LP/mm are above 0.8 that lens will have excellent image quality. It also says that above 0.6 will be “satisfactory”.
Let’s take a look at Canon’s 50mm F1.4 lens’ MTF Chart to put this into practice. Hopefully you’re at a computer and can follow along. If you can see the show notes, find the first link which will take you to a Flash MTF chart on Canon Japan’s Web site. Because this chart is made with Flash, you can maximize your browser window to view the chart much larger than usual, which will make it easier to follow.
If we look first at the thick black solid line which represents a measurement of the 10 line pairs per millimetre with the lens’ aperture wide open at F1.4, we can see that the line is running from between the 0.7 and 0.8 lines on the left, right down to just over 0.2 on the right hand side of the chart. We learn from this that the lens has pretty good contrast in the middle, but that drops off to pretty low levels in the outer 5mm or so of the lens. If we look at the thick solid blue line though, which represents a measurement of the 10 line pairs per millimetre with the lens stopped down to F8, we can see that this lens has excellent contrast throughout. The thick blue line is only just below the top of the chart all the way to around the outer 3mm of the lens, when it drops to just below 0.9, which is still very good.
Let’s look at the thin solid lines now to see what we can learn. Firstly, again let’s look at the thin solid black line, which is a measurement of 30 line pairs per millimetre with the lens wide open at F1.4. We can see from this that in the center, the lens’ resolution is only 0.5 and it drops off to between 0.1 and 0.2 on the outer edge of the lens, the right side of the chart. This shows us that the resolution of the lens is really not that great wide open. I own one of these lenses and I can tell you that this is very true. This is a beautiful lens with a wide following, but anyone that has used it wide open at F1.4 will be able to tell you, it’s really quite soft. This is though exactly what most people buy the lens for, in addition to it just being an incredibly bright lens.
If we look at the solid thin blue line, we can see that the lens has excellent resolution when stopped down to F8, right to the outer 3mm of the lens. The line is meandering around the 0.9 line until around 17mm and drop below 0.8 at around 19mm, which really tells us this is a very sharp lens pretty much throughout at F8.
So far, we’ve look only at the solid lines, and I think you will now be getting the picture. The solid lines if you remember are for the Sagittal measurements, which are of sets of lines parallel to a line drawn from the centre to the edge of the lens. The broken lines are the Meridonial measurements, which are of sets of lines running at 90 degrees to the lens axis. If we look at the chart we can see that each pair of solid and the broken lines is running pretty closely together. I’m not going to work you through all of the broken lines as I just did for the solid lines, but the thing to note here is that the closer these solid and broken lines are to each other, the better the lens’ bokeh, or the areas not in focus, will be. That is another reason people really like this lens, because it has beautiful bokeh, and we’ve just confirmed that by looking at the chart. You’ll notice that this chart at first glance really not look that great, but with the knowledge to read it properly you can appreciate the benefits of such a lens. Also bear in mind that wide open for the lens that is represented by the black lines, is F1.4. We could close down a full two stops to sharpen up the image a little and we’d still be at F2.8, which is the starting point for most other lenses. So it’s really not surprising that the chart looks a little rough at first glance.
Most charts I’ve looked at start off relatively high on the left had side, which is the centre of the lens, and then drops off, sometimes gradually, and sometimes more sharply, to the right. This shows us that the lens is sharper in the middle than around the edges. Notice that for prime lenses there will only be one MTF chart, but for zoom lens, you’ll usually see two charts – one for the wide and one for the telephoto end of the zoom range. If we compare these pairs of charts for a zoom lens we can see how good a zoom lens will be at one end of its zoom range compared to the other.
The other age old topic that crops up in our forum and I’m sure many others is how much sharper prime lenses are compared to zoom lenses. One that came up recently in our forum at martinbaileyphotography.com was a comparison of the Canon EF 400mm F2.8 IS lens compared to the 100-400mm F4.5-5.6 IS lens. An interesting thing you might want to try here to go to the Canon Web site and open two windows, one showing the 400mm F2.8 and one showing the 100-400mm lens, and compare the same focal lengths. You’ll see that the 100-400mm at 400mm is a very respectable lens, though it drops of quite sharply towards the edge, but the 400mm F2.8 has five of the eight lines running almost across the top of the chart for the entire width of the lens, with just three lines dropping below 0.9 from around 10mm from the center of the lens, but staying above 0.8 to the very edge. This shows us that the 400mm F2.8 is one sharp lens.
So today’s episode might have been pretty heavy going, but hopefully if you weren’t up on this before, you should now be armed with the knowledge to make more educated decisions on any future lenses that you consider adding to your arsenal. Please do also make use of online lens reviews and others opinions too, as the MTF charts do not always tell the whole story. Also, please don’t just try to find a lens that has all the lines running along the top. The perfect lens does not exist, even if the 400mm F2.8 comes close.
So, have a great week, whatever you do. Bye bye.
Music from Music Alley: www.musicalley.com/
Let’s look at the Canon 50mm F1.4 lens as an example: http://cweb.canon.jp/ef/lineup/standard/ef50_f14/mtf.html
Canon provides MTF Charts for almost all of their lenses on all the country sites that I’ve seen to date. Here’s a link to lens pages on the Canon US Web site: http://www.usa.canon.com/eflenses/lineup/index.html
Here’s a few more examples from the Canon Japan Web site that use Flash so you can resize them to large sizes for easier viewing:
Nikon’s Japanese Web site does contain MTF Charts for their lenses, although the US site does not seem to. As the lens name is displayed mostly in English, it’s not difficult to browse from the below pages.
The music in this episode is from the PodShow Podsafe Music Network at http://www.music.podshow.com/
Subscribe in iTunes for Enhanced Podcasts delivered automatically to your computer.
Download this Podcast in MP3 format (Audio Only).
Download this Podcast in Enhanced Podcast M4A format. This requires Apple iTunes or Quicktime to view/listen.
Posted on behalf of Martin by Michael Rammell, a Wedding Photographer based in Berkshire, England. Michael also has a long-standing passion for Nature & Landscape photography. To catch up with Michael, visit his Web site, and follow him on the following social networking services.