1 – close it after registering, and re-open it, because it does not work immediately after registering and leaving it open.

2 – Luminance, Chrominance, and Sharpness are the only controls, and that’s a good thing -ish.

3 – There are no [ Save As ] compression settings, that’s a bad thing -ish.

So the only way to control JPG or PNG compression settings appears to be to save as BMP, and then re-open the BMP in another program, and save as JPG or PNG with compressions choices.

Otherwise, WidsMob appears to save at JPEG quality 75, with 2×2 subsampling, and PNG ZIP, resulting in,
– if BMP is 100%,
– then PNG saved to 29% of the original file size,
– and JPG saved to 4% of the original file size
( yeah, it’s amazing how little information we need in order to resolve a visual image ).

4 – Luminance smoothed everything toward plasticy, losing subject details,
– Chrominance did nothing that I could see in a few samples,
– Sharpening made images look better to me.

5 – I just don’t have noisy images because modern imaging and image processing has gotten very sophisticated and intelligent, eliminating the perception of noise and compression artifacts.

So the need for this program is becoming less and less for new images.

Perhaps WidsMob Denoise is best thrown at our archive old original digital camera images, images we thought we might not use again because they look as dated as faded film prints look dated.

But this does not open raw or TIFF files, only JPG, PNG, and BMP, so WidsMob Denoise can’t even see the majority of my archive.

And WidsMob Denoise does not have any sophisticated intelligence or fuzzy logic to support an auto-analyzer mode, such as seen in Franzis Denoise Projects, so I cannot imagine processing images in a batch with all one setting and hope for the best.

Automated denoise wise, supposedly, taking a series of calibratable test images, such as total blackness ( with the lens cap on ), for each shutter speed, offers a subsequent noise-reduction program a reference for what should be black but if there’s noise, it can be analyzed and then auto-removed from real images captured at that same shutter speed, same for taking a series of calibrated test exposure values, analyzing the noise, then auto-eliminating that noise from real images which match those exposure values, and so on, and endless arduous task – who goes back and takes equivalent shots to everything they captured, to use as noise-reduction references, as test calibrations “… hmm, this was shot at 1/800th second at exposure value 14, so now I will take a test black field at 1/800 to use to subtract some noise from the real picture, and then I will take a test at exposure value 14 to find more noise to remove from the real picture … now, here’s my second picture from that day, what are the values for shutter speed and exposure value so I can create calibration images for subsequent noise reduction …” … no, that ain’t gonna happen, but they have now supposedly done all that at the camera factory to direct their in-camera noise reduction so we don’t have to.

And that’s the point.

Plus ( yes, there’s more ), printers do even more to smooth out the graininess, er, fix that, to smooth out the noisiness in noisy images, such that even 10 years ago, an 11×17 print from a 5 megapixel 2/3″ ~8.8×6.6mm sensor 2,655×1,996 pixel image resulting in an effective mere ~192 dots per inch, when printed by a printer laying down as much as 5,760 microdots across an inch of print using stochastic algorithms, interpolating 30 additional dots for every original image dot, that picture looked great, and still does, there is no way I would revisit it using WidsMob Denoise and hope for a better presentation.

But everyone reacts to noisiness differently, and your eye may beg to differ, so try WidsMob Denoise and you tell us what is your preference, what is your experience of WidsMob Denoise.

Thanks for exploring this and sharing.
.

[@Peter Blaise] My thanks as well for your thorough and insightful review. This might be useful on scanned images of old photos.

I never thought about capturing “noise” to have a reference to remove it later. I do something similar with audio files when needed/possible, but that’s a much simpler algorithm.