What is the "Levels>Auto Input Levels" operation actually doing?

[Ofnuts]

(01-31-2024, 04:34 PM)ajax Wrote: Well now look what I just did.  Consider the following:



After making the adjustments (manually), separately on the R&G&B channels that I previously mentioned, I then went and followed the advice regarding adjustments to the channel identified as "Value" (i.e., R/G/B combined) and produced the above result.  While NOT exactly the same it is pretty close.  Also, by fiddling with the black and white point, in the Value Channel, just a bit more I'm pretty sure I could have made one that is indistinguishable from the one made using "Auto Input Levels".  If left to choose I'd likely have picked the one created manually.

While it might be a bit off subject I'm a little curious about why this (Auto Input Levels) capability is NOT included by the developers of my scanner software.  The original image shown above is NOT peculiar to this shot or film type but rather very common across the board (i.e., all images have the same properties).  Given that this comes from a negative there is a fair amount of conversion associated with producing the positive image.  Seems like an Auto Level adjustment ought to be an option at minimum.

Anyway, thanks for the help.  I've definitely learned some important things about the Levels Tool.

Because "auto" anything isn't really predictable. Scanners are meant to mostly scan "normal" stuff, and so they aim for color accuracy.

Also "auto" isn't a good way to "fix" negatives. You don't want the result to look good, but to be accurate. There are better methods (CLUT based, with a CLUT for each film type).