The Holding Tank

Home Page Photography

Photography RSS 2.0 Feed

| Photos & Themes RSS | Hardware | Technical | Links | Books and Magazines  | Downloads |

Hardware

| Lens Data and Reviews | Lens Selection | Long Lens Selection | Wide Range Zooms | Consumer Zooms |

|  Macro Equipment |  Tilt and Shift Lenses | My Bags | My Old Hardware |

| My New Hardware | Lens MTF Comparison | Lens Tests | Weather Protection | Canon AF System | 7D vs 20D ISO Performance |

MTF Study

| Lens MTF Area Summary and Figure of Merit |

| MTF Data Overview | Photodo / Canon MTF Cross Validation | 

| Contax, Nikkor and Leica R lenses vs Canon | 

| List of Canon Data Comparisons | List of Photodo Data Comparisons |

| Database Download |

MTF Data Overview

I am not an optics expert or camera guru, just someone who wants to get value for money when he purchases camera equipment, and is enough of a geek to dig into the technical side of it. Writing it up this way forces me to make sure it is reasonable, a by product is it may be of use to other photographers.

There are a number of factors affecting lens selection such as:

a)             Optical Quality

Here my guide is MTF (see an explanation here). However, this is not the only aspect as flare, bokeh, colour and many other factors contribute including various aberrations and distortions. MTF describes the sharpness and contrast of the lens and to some extent bokeh and is provided by Canon as theoretical plots. Photodo produce measured MTF plots and sometimes provided a measure of distortion. Treat the Photodo figure of merit with caution as my understanding is that these weight the lens performance at the centre and provide little seperation between lenses.

b)             Usability

This covers weight, size, auto-focus actuator type, if full time manual focus is available, hyperfocal or depth of field markings, maximum aperture and other factors.

c)             Cost

Fairly well defined this one!  

  d)       Special Features for Digital Use

There are number of special design aspects such as reduced internal reflection coatings for light reflected off of the sensor and telecentric like design (True telecentric lenses seem to be used in computer imaging). Joseph S. Wisniewski has a FAQ discussing some of these points. Interestingly the EF-S lenses may not be telecentric like as the closer to image plane distance seems mostly to be used for simplifying scaling of existing designs rather than increasing the distance of the exit pupil from the sensor as might be expected with a telecentric design. 

To understand my background and thus position you can see my hardware history here.

I have found in the past, with my film SLR system that prime lenses suited my style and typical subjects well, and are generally acknowledged to be of better quality, cost and maximum aperture than zooms. At that time MTF data was not available, at least to my knowledge, so this is a major advantage now.

However, these days some good zoom lenses are available, particularly the Canon L zooms, so this trade-off is no longer a “no contest”.

One bias to understand as you read this page, is in the past I tried zooms and moved to primes. Generally, you can accomplish most framing adjustments by walking a few paces. So for me a 16-35mm zoom replaces probably just two lenses, a 17mm and a 28 or 35mm, not 17, 20, 24, 28 and 35 as many might claim, this is my bias and arises from my way of working.

With my 35mm film system I found I tended to use the 28mm, 50mm and 85mm lenses the most, and generally the 28mm and 85mm more than the 50mm. I also had 17mm and 135mm lenses. So this is close to the classic square root of 2 focal length ratio system (square root of 2 is 1.414 so this is an approximate statement).

Now on the 20D, with its 1.6X crop factor sensor, these lenses are equivalent to 10mm, 17mm, 32mm, 53mm and 85mm. So if I would like to replace my old FD lens system, these are the lengths I need, clearly these round to 10, 17, 35, 50 and 85mm.

The situation I am in at the start of this study is having the 20D with the 18-55mm kit lens (it was so cheap why not, plus a normal zoom is always useful). In addition I have already plumped for the 100mm f2.8 Macro as this does the job of a medium telephoto and true macro lens, as so is quite enabling.

So the main thing to determine here is if I should go for a zoom based system or a prime based system.

Another guiding factor is EF-S lenses, I plan to avoid these as much as possible to leave my path open to a 1D or 1Ds type camera in the future. Additionally, I found Canon orphaned my FD glass last time, and I don’t want to be in the same situation again. There is some discussion of the likely future of APS-C and full frame 35mm DSLRs. The suggestion being the FF sensor may be affordable in the 5 year time frame.

All the EF-S lenses are optimised for a small image circle and have a noticeable fall-off in image sharpness beyond 10mm offset, in addition they are all limited aperture zooms. So with this in mind full frame lenses will generally be superior. An exception to this is the recently announced (Feb 05) EF-S 60mm f/2.8 Macro USM lens, so if Canon produce good quality large aperture prime lenses in the EF-S mount this comment will be less appropriate. Another reason to consider full frame in the future are the limits on wide angle lens resolution for APS-C sized sensors. 

Canon publish MTF data for all their lenses in a book “The eyes of EOS” and on the USA web site. Some of the EF-S lenses MTF charts are only available off the Japanese site. There are some inconsistencies with the Canon data as listed in the table below:

As it is difficult to compare the curves for two lenses with the data as presented, the approach was to transfer the graphical data to an Excel spreadsheet (download here you will need to enable macros when you open it) for each of the four curves for each lens, sampling the curves at the following offsets from the optical centre 0, 2.5, 5. 7.5, 10, 13, 15, 17.5, 20 & 22mm.

Naturally the EF-S data is only provided up to 13mm. So clearly, MTF data beyond 13mm is only significant if you expect to use a camera with smaller than a 1.6X crop factor in the future.

Once in electronic form, the data can be indexed and interpolated as needed.

This data transfer does not need to be blisteringly accurate, so reading the values off by eye is OK. The justification being there will be production tolerance variations, plus if the lens performance is that close there are other factors to consider in the selection.

Canon provide data for 10 lp/mm and 30 lp/mm, these are often referred to as contrast and detail respectively in this text, taking this interpretation from Canon.

In the case of MTF curves from Photodo, curves are provided for 10, 20 and 40 lp/mm. For the few lenses that I have used Photodo data in this study (the trace title will be prefixed PD) I have derived the Photodo 30 lp/mm curves by linear interpolation between the 20 and 40 lp/mm curves at each offset point. This seems to work fairly well. An important thing to understand is that Canon’s data are computer simulation (i.e. theoretical) and Photodo’s are measurements on real physical lenses.

When dealing with zooms, Canon only provide MTF data for the extremes of the zoom range, what happens in the middle is left undefined. The approach in this study has been to linear interpolate the MTF data from the wide and tele plots to provide a third estimate plot at the focal length of interest. Now many say that the ends of the zoom range represent the worst case and the zoom should be a lot better in the middle of the range. If this is true then I am erring on the side of pessimism. However, study of Photdo data (eg), who do provide a mid focal length set of curves suggests that the wide and tele are not always the worst case, although this may depend on aperture.

The data can be looked at a number of ways, in this write-up I only present the worst case of the MTF for the Sagittal and Meridional directions, mostly for simplicity. Looking at the average data does not tend to change the conclusions in most cases. Clearly comparing maximum aperture data for lenses of different max apertures is of limited significance and must be read with caution.

The methodology is to examine the performance of a set of lenses that cover a particular focal length, interpolating the zoom performance as noted above. A selection of f2.8L and f4L zooms are considered.

In the following section all performance comments are drawn on a comparative basis. It should be noted that although one lens may be better than another, it may not be significantly better. Rather like having a 500bhp sports car and a 200bhp family car. The difference does not matter if you only drive around gently in town.

So these comments need to be treated with some caution. Canon’s own guidance in this respect is quoted: “A MTF characteristic of 0.6 and above indicates a satisfactory lens. A MTF characteristic of 0.8 and above indicates a superior lens.” No guidance is given for the 30lp/mm characteristic however, unless this is meant to apply to both spatial frequencies.  Judging from typical curves then, 30lp/mm limits of 0.5 and 0.7 would seem consistent. See my lens image simulation page to get a feel for the effects of different MTF.

The offset from which these MTF characteristics first fall below the limit are tabulated below. Where N/A is marked it indicates the characteristic never falls below the limit line. This helps give an indication over how much of the frame exceptional and satisfactory contrast and detail can be expected in each case.

See the diagram below which indicates the coverage circles against the full 35mm frame (1Ds) and 1.6X crop factor APS-C frame (20D).

Arguably, detail at 30 lp/mm is more important for the smaller sensor of a 1.6X crop factor camera like the 20D.

Generally wider angle lenses are harder to design and make than normal and telephoto, so as the focal length increases the sharpness expectations should also.  

Last Updated 05/06/2008

All Content © 2005-09 Lester Wareham All Rights Reserved     

All material is supplied  as is and without warranty, use at your own risk.  

All opinions stated are the authors own.

All quoted information remains the copyright of the respective authors.

Contact: (anti spam remove underscores)

Note this email gets spammed a lot so if you want to get your message through please prefix "Holding Tank" in the title

f_ w_ a_ r_ e_ 9_ 0_ @_l_i_n_e_o_n_e_._n_e_t