DIY Lens MTF Testing

Following on in the tradition of various test sites (and because I can do what I please), I'll categorise the numbers on the above charts as:

>3000 : Excellent
2500 - 3000: Very Good
2000 - 2500: Good
1500 - 2000: Just OK
1000 - 1500: Poor
500-1000: very poor
< 500: Terrible

This scale is relevant only to tests on APS-C 24 mp cameras.

Moredsir wrote: Hi Craig, that's just what I am after. Any chance I could make use of the scripts?

This little project is on the back burner for the moment as I was still somewhat uncertain about test charts. You do NEED quality professionally printed test charts.

The code and scripts are a bit cobbled together at the moment. I'll review matters over the weekend and send you a PM. It's all on Linux so if you are on Windows, you would have to at least rewrite the scripts as batch files and as far as I am concerned you would be mostly on your own as I spent most of my career avoiding Windows in favour of UNIX . My windows knowledge is err... limited. I'll also have to review the Java code for portability, though shouldn't be a big issue.

In the mean time you could get hold of MtfMapper and dcraw and whatever else is needed and just have a play around with the MtfMapper surface plots using a chart printed on any old ink jet printer. The absolute MTF figures won't be flattering but it will serve for familiarization purposes. You can probably get a professionally printed A1 test chart for $50 or so. It must be high quality - things like the cheap large canvas prints won't cut it.

Install inkscape for viewing and printing the SVG test charts. Using wide area metering if shooting the test charts. +1 EV is probably about right, but check your histogram.

The other use for MtfMapper is checking (and hence micro adjusting) AF. It's fairly easy and there is no requirement for a high quality test chart.

Thanks to all for this thread - as a result of it I have started to play with MTF Mapper which looks like being an excellent tool for simply comparing lenses. I am sure that my initial trials using an A4 'lensgrid' chart and a copy stand will not fit everyone's idea of precision testing - but one of my interests is in copying historical documents so I am very keen to see how my lenses perform in this sort of set up.

However, I would also like to see how they perform from further away - the copy stand is too short for the MFD of many lenses in my collection of old glass. My next step must be a larger test chart - I shall get some A3 prints and investigate my local print shop for larger sizes.

Bob J wrote: Phil Wood wrote: one of my interests is in copying historical documents so I am very keen to see how my lenses perform in this sort of set up. Visiting the National Archives at Kew?

I certainly spend time at Kew, easily the most photographer friendly archive I visit.

I have now tested a few lenses and am getting some surprising results - I fear I need to eradicate a few more variables from the set up before I will trust the results. The biggest surprise has been that an old Minolta 28-100 D (the silver plastic one) is out performing (by some margin) a modern Sony DT 18-55 SAM II and an earlier Sony DT 18-55 SAM at 50mm.

Phil Wood wrote: I certainly spend time at Kew, easily the most photographer friendly archive I visit.

They are a good bunch. I worked there for about 18 months back in the early 90s when it was still the Public Record Office - working on the specification for a new document ordering system and the first instance of their 'Guide' being available electronically... My system will have been replaced at least twice since then.

I felt terribly lowbrow, as it seemed like even the cleaners had history degrees, while I was just the token techie.

Bob J wrote: I felt terribly lowbrow, as it seemed like even the cleaners had history degrees, while I was just the token techie.

But could they speak SQL?

Miranda F wrote: Phil Wood wrote: The biggest surprise has been that an old Minolta 28-100 D (the silver plastic one) is out performing (by some margin) a modern Sony DT 18-55 SAM II and an earlier Sony DT 18-55 SAM at 50mm. I'm a little surprised at that. The ones I've tried have had quite poor corners anywhere near full aperture, but my Sony 18-55 is better. For best cheap sharpness at close-focus I would highly recommend the 35-70 F4 which is very good in close-up, where a lot of lenses go soft.

I have only scratched at the surface with my tests - and can see where there may be problems that may render comparisons between lenses tested at different times more than a little dubious. Notably I have tested in natural light (often obligatory when photographing historic documents) - but it must lead to inconsistencies in contrast. I shall try again with an LED ring flash (seems to me to be the easiest solution to hand).

I have tried a 35-70 f4, but it cannot focus (at f4) on the A4 target at 50mm and fill the frame. Obviously it can focus in macro, but then the lens is a 35mm. I chose to compare at 50mm because I have a 50mm prime to assess the various zooms against. I am finding that a few old zooms don't have the MFD to cope. Once I am happier with my test setup I shall try all my zooms capable of 50mm - perhaps a stepped down 35-70 f4 will outperform those that can focus wide open?