I have some doubts during the testing, so I'm sharing my test cases hoping to get more detailed guidance.

[Leopold Fizz]

Dear Author,

First and foremost, thank you for your invaluable contribution. Your software has greatly simplified lens resolution testing for individual users like myself. However, I’ve encountered some inconsistencies in my tests and would appreciate your expertise.

Test Setup:

Equipment: Nikon Z6 with Z 24-70mm f/4 lens.

Charts:

A commercially printed ISO 12233-2000 chart (high-quality, slightly smaller than A1).

A self-printed lengrid-A1 chart (provided by you). While its slanted edges appear slightly coarser than the commercial chart under close inspection, I find it sufficiently sharp for testing purposes.

Observations:

When photographing the ISO 12233-2000 chart to fully fill the sensor height, MTF Mapper reports ~60 lp/mm at the central slanted edges.

With the lengrid-A1 chart under the same conditions (full sensor height), MTF Mapper yields <45 lp/mm at the center—a significant discrepancy.

Adjusting the distance to fill only half the sensor height with the lengrid-A1 chart improved results: ~50 lp/mm centrally (some edges nearing 60 lp/mm), with overall resolution higher than the full-frame attempt.

Questions:

What might explain these differences? Which results are more reliable?

For self-printing the lengrid-A1 chart:

Can the provided PDF be printed directly?

What is the minimum printer DPI required to ensure accuracy?

Your insights would be immensely helpful. Thank you again for your work and for any guidance you can offer.

Here's the raw file: https://drive.google.com/drive/folders/1VuYKd27P-26HJ5DlO7xLeSjjI8kFoLKT?usp=sharing

Best regards,
Fizz

[Frans van den Bergh]

Hi Fizz,

You are already on the right track: what you are observing, namely that the results improve when you move further away from the test chart, is what I would expect if print quality was a limiting factor.

To answer your questions:

1. Slanted-edge measurements are the product of three parts: MTF_measured = MTF_target * MTF_lens * MTF_sensor. At some point your chart magnification is low enough that the MTF_target part no longer has a measurable impact. If you start with a very high quality target (think chrome-on-glass targets) then MTF_target is essentially negligible even close up, but for most printed targets you will see the impact of MTF_target if you get close enough. There's even a method to measure the target MTF, and to correct for it: see this Imatest article
2. A better quality chart will typically give you a higher measured MTF result, meaning that the higher result is the more reliable one if you compare two charts. There are a lot of caveats that go along with this statement, e.g., I assume exposure is correct, and so on, but most defects will tend to decrease your measured result (with a few exceptions such as over-exposing the target).
3. The PDF files I have made available can be printed directly. They do not have any associated resolution; they are vector-graphics PDF files. The printer and/or printer driver may choose to rasterize the document too soon or at a coarser resolution, but the documents themselves can be used to produce very high quality prints.
4. The minimum DPI required to obtain accurate measurements depend a little bit on the printing process, e.g., an Inkjet print at 1200 DPI is usually less sharp than a laser print at 1200 DPI because of ink spread. The smoothness of the paper plays a role too (texture effectively adding noise), so plain office paper is too coarse for really high-quality work. I usually recommend the 8x rule: If your image has 6000 pixels across, then you want at least 8* 6000 = 48000 dots across the width of your printed target. For an A1 size target, that means around 1450 DPI. But keep in mind what I said about the printing process itself, since it is possible to have high DPI with bleeding in an Inkjet print. In this case, 14500 DPI at A1 size is a bit of an odd resolution, but if the same printer can do 1200 DPI at A0 size you may be good.

But I find that often I only require relative MTF comparisons, so as long as the chart is good enough to respond to your changing test conditions, e.g., two lenses tested on the same chart at approximately the same magnification, then I do not have to worry about the contribution of MTF_target, within reason.

Hope this helps!

Regards,
Frans

ps: for the record, I am still searching for an affordable printing process that produces really high quality targets, especially at larger sizes