Question about QuickMTF settings

[xpatUSA]

Hello QuickMTF Users . .

I use QuickMTF Light V1.14 and don't fully understand two settings:

A checkbox allows selection to display "peak MTF", e.g. MTF50P instead of MTF50. The graph changes only slightly between the two. While I understand the difference in MTF values obtained from bar and sine-wave targets, what is this "peak" number actually telling us? (bearing in mind that the target is a slant edge, i.e. it is neither a bar pattern nor is it a sine-wave).

A field allows entry of various numbers for "gamma" but I am not certain that it is referring to the gamma as in e.g. the linear to sRGB conversion. Again, the graph changes only slightly with widely different values. I glanced at Imatest's stuff and didn't find a clear answer there, although they do mention a similar type of selection. So, if I analyze a standard JPEG file, do I put 1 or 0.45? 2.2? 1.8? Is a TIFF or a PNG any different as far as QuickMTF is concerned?

I asked the author about the correct gamma setting for a TIFF file but he didn't seem to understand my question because the answer was written in "terse" and made no sense no me 

Old and confused and not liking it . . .

Ted

[Bart_van_der_Wolf]

Hello QuickMTF Users . .

I use QuickMTF Light V1.14 and don't fully understand two settings:

A checkbox allows selection to display "peak MTF", e.g. MTF50P instead of MTF50. The graph changes only slightly between the two. While I understand the difference in MTF values obtained from bar and sine-wave targets, what is this "peak" number actually telling us? (bearing in mind that the target is a slant edge, i.e. it is neither a bar pattern nor is it a sine-wave).

Hi Ted,

I'm just guessing, but could it be that the peak MTF50 is the highest MTF50 amongst those from a number of ROIs chosen on the same edge?

A field allows entry of various numbers for "gamma" but I am not certain that it is referring to the gamma as in e.g. the linear to sRGB conversion. Again, the graph changes only slightly with widely different values. I glanced at Imatest's stuff and didn't find a clear answer there, although they do mention a similar type of selection. So, if I analyze a standard JPEG file, do I put 1 or 0.45? 2.2? 1.8? Is a TIFF or a PNG any different as far as QuickMTF is concerned?

If the settings are anything like those in Imatest, then you should use 0.5 or 0.454545 for gamma adjusted output files, and Gamma 1.0 for linear Raw data.

Cheers,
Bart

[xpatUSA]

Hello Bart, thanks for the quick response!

I'm just guessing, but could it be that the peak MTF50 is the highest MTF50 amongst those from a number of ROIs chosen on the same edge?

Basically, the program only allows one ROI to be selected at a time and does not remember previous ROI's as far as I know. I think it has to do with bar vs. sine patterns of which I know you are well aware. Perhaps it is saying "here's what the MTF would be if this system were to measure a bar pattern"? But that is just my guess - and I was looking for something definitive from a QuickMTF user.

If the settings are anything like those in Imatest, then you should use 0.5 or 0.454545 for gamma adjusted output files, and Gamma 1.0 for linear Raw data.

Thanks, I've been using 0.45 up till now, I'll see if it'll take those extra decimals  

OT but I've been having fun with your lens test target (the 130mm-size 144-cycle with the added slant-edge) this past week. However, I do wish my Epson R260 would do a better job in the middle!

Ted

[Bart_van_der_Wolf]

OT but I've been having fun with your lens test target (the 130mm-size 144-cycle with the added slant-edge) this past week. However, I do wish my Epson R260 would do a better job in the middle!

Hi Ted,

My test target is also a torture test for the printer, any printer, not only the lens+sensor. But when you shoot if from a long enough distance, the printer shortcomings should become too small to have an effect on the lens test. Do make sure that the 1, 2, and 3 line wide patterns in the corners of the neutral mid gray background are fully resolved (requires 'finest detail' to be checked on, and probably a loupe to see that detail in the first place).

Cheers,
Bart

[xpatUSA]

Thanks again Bart,

I'll give it a shot. Not wishing to hijack my own thread, I'll leave target matters for another time . . .

Ted

P.S. long long ago, I visited the DOMO milk factory in Beilen on technical business.