Andrew Winhold

Frans, Ah, this target is great to know about. I've used Peter's sfrmat3 a bit but will check out SFREdge. For a quick A:B of the linear gamma option in MTF Mapper I've attached an example result from one of the edges imaged. Selected the ROI in the GUI, set it to reuse ROI on next image, ran results. Blue is with linear gamma selected. Green is with linear gamma not selected. Reduced smoothed is selected for both runs. The manual for the camera mentions white balancing not being available for monochrome...

Frans, Thank you for a great explanation! produce annotated output from raw grayscale 8-bit linear camera capture (e.g., PGM file): by default, 8-bit input images are assumed to be sRGB encoded, so TRC-to-linear is applied by default. If you add the "--linear" flag (or GUI option), you are telling MTF Mapper to skip TRC-to-linear, and instead values are just multiplied by 256 to obtain internal 16-bit encoding. Ok, our monochrome CCD obviously gives grayscale images but I'm working with .png at the...

Hi Frans, I've really enjoyed this tool and I appreciate the new (to me) ability to select ROIs and measure LSF and ESF. I don't see this asked in other posts so I'm doing so in this one: When I open a .jpg image taken with a monochrome CCD, the image seems to have a higher contrast in MTF Mapper than the original. The resulting saved annotated image shows this to be the case. The only setting I can think to change is the Linear gamma flag but that seems to make thing worse. Thank you for any time...

Thanks, Frans, for pointing this out. I'll need to get some specifics from the ones who work on the processing pipeline of the images. They've provided a mix of demosaicked and undemosaicked images but not to a pattern I can understand. Knowing they are likely removing rows is very useful as well.

Frans, This is really good to know. I've been using the documentation but haven't made it through all parts yet. I believe our CCD (https://www.onsemi.com/pub/Collateral/KAI-2020-D.PDF) is grgb so we'll want to include this argument. Thanks! Andy PS: Should there be any consideration for a monochrome sensor? We'll be using one in the near future that does not have a bayer patter on it.

Frans, This is really good to know. I've been using the documentation but haven't made it through all parts yet. I believe our CCD (https://www.onsemi.com/pub/Collateral/KAI-2020-D.PDF) is grgb so we'll want to include this argument. Thanks! Andy

Frans, Thanks for this detail! Yes we're using the raw Bayer images from our camera so this information is very useful. It sounds like we'll consider selecting the green channel as it's twice as prevalent vs the blue or red. Thank you again for your assistance on this. Best, Andy

Hi Frans, Thanks for the great answer and detail. I'll follow suit when referencing values in the future. I've attached two test images for reference and two screenshots using those images to show points where the MTF50 value in the annotated image and it's related SFR curve window differ by >0.1. From your response I think this difference is related to the the fact that the SFR curve is not 'reasonably smooth'. Most of the points match very well so I'm guessing this is why I'm finding a few outliers....