DeNoise AI - Automatic mode?

Do the default settings represent the software’s best guess for noise reduction on an image? If not, would love to see the software provide a starting point.


Yeo Juliber - welcome to our Community :slight_smile:

They do not, because they are the same for every image. I think the issue would be that in order to provide a best guess the software would have to do a full analysis of the image when it opens which can take some time. I think a good solution would be to have an auto-button that would preform the analysis if the user wishes.

1 Like

Definitely would like the option of an auto button!

1 Like

As in AI Clear, Denoise AI only looks at a small section of the image (the preview) so analyzing it would not increase the time a lot. In fact, if it did a good job in auto it would be faster since moving the slider and testing several times would take more time.

But if you wanted auto values, you’d want it to examine the entire image to get a proper noise profile. For instance what if you are zoomed in on a bright area but most of your image is in shadow or vice versa…

If I asked the program for it’s best guess at settings I’d want it to do so based on the entire image.

1 Like

I see your point and I think any auto mode (I agree it should be switchable) will choose a mid-tone area where noise is most noticeable. I think noise is fairly evenly distributed over a picture.

This option is under heavy research right now. Bowen and team have been testing several versions of this to see if it can really predict quality settings based on the image. For now they are just defaults with no per image smarts behind them


Didn’t Noise Ninja do this?

Ideally the Auto Settings Engine can learn from the preferred settings of the user much like junk mail software has a corpus that gets better with use . That way personal info such as camera serial #s lenses and particular ISO/noise levels can be taken account for a much improved result. Then batch processing will take on a whole new level of utility.

something this way comes…