[Medex]

I do monochrome conversions about a year - different Pentax models and Ricoh GR APS-C models. Finally K-1 monochrome was finished and this is 39th successful conversion.
I just want to share an example of high ISO picture (ISO 51200) shot on monochrome camera in Pixel Shift Resolution mode vs Normal mode. Not worth to make discuss about intact K-1 sensor (with Bayer CFA in place) PSR mode and monochrome sensor because they deliver practically same details and noise. But I was not expecting any additional improvement using PSR mode on monochrome sensor. Seems like monochrome PSR delivers perceptibly less noise (about 1,5-2 stops).
Here are comparison pictures (200 percent crops in Lightroom)



Shortly about Pentax K-1 monochrome resolution. The test was made at 1/3 of test chart height (12233). Number should be multiplied by 300 to get LW/PH value.
Lens - DFA* 50/1.4 @f1.4!
Result: sensor resolves max possible - about 16,4x300. Physically sensor has 4912 pixels in vertical dimension.
Now some calculations / approximations. As I see from tests (my own and from internet) demosaicing in RAW converter is the reason of resolution loss by about 1/3. For example Fujifilm X-H2 has 40 mpix sensor, it means about 5100 pixels in vertical dimension. Test shows that this sensor resolves 3400 LW/PH (https://www.digitalcameraworld.com/reviews/fujifilm-x-h2-review). Similarly 24 mpix sensor from GR III has 4000 pixels in vertical dimension and resolves about 2800 LW/PH (my own data). Same with Sony a6600 (https://www.digitalcameraworld.com/reviews/fujifilm-x-h2-review).
Normal sensor should have 4912 / 2 *3 = 7368 pixels in vertical dimension to resolve same amount of lines as K-1 monochrome. Then horizontal dimension should be 11052 pix and sensor should have 81,4 mpix. This is reasonable equivalent what K-1 pixel shift can deliver.

Attached Images

View Picture EXIF

File Date

 

---

Last edited by Medex; 10-08-2022 at 06:28 AM.