Were picture gamma correct calculations done?

[hhuter]

Hi Janspiry,
As an IT- and algorithm engineer, I understand the "directional cubic interpolation" very well.

So, when I saw the examples "zoomed pictures" what you thankfully also published, I could not believe, that the algorithm does really work so distractingly bad: darker lines (between hairs strands e.g.) seems darkend jaggy staircaselike and diagonalcaselike far too much and additional extra jaggy "staircases" seems to be done amplified imprinted in the resulted demo-pictures. Thats what I saw.
...
My first tough was:

http://www.ericbrasseur.org/gamma.html

problem is that "the computations are performed as if the scale of brightnesses was linear while in fact it is a power scale." In mathematical terms: "a gamma of 1.0 is assumed while it is 2.2." Lots of filters, plug-ins and scripts make the same error.

Awaiting your thoughs
Johannes

I did not check the original Matlab code, but I think, the inventors of this "zooming" algorithm had fooled themself with not ensuring to have linear brightness scalings for their pixel in the loaded images, which by definition have nonlinear scaled pixelbrightness numbers in their imagefiles.

AND ! ?
maybe, this would explain, why the proposed "directional cubic interpolation" had no successful recognition in this 11 years since publication:
Because wrong implemented and not considering the nonlinear brightness to pixelvalue scalings (again: which are inherent in nearly every picture format definitions).

Note:
I also contacted one of the papers authors and the Matlab implempentation providers
Dengwen Zhou via [email protected]
https://de.mathworks.com/matlabcentral/fileexchange/38570-image-zooming-using-directional-cubic-convolution-interpolation
about this subject.
Hoping for some response.