Blogger: Whites in photos are grey and faded. How to turn off auto-photo enhance. & Google+ form settings analysis

In a series of articles, of this blog transitioning to the new Blogger theme and system, one of the issues that appeared was that uploaded pictures, even png's for that matter, with perfect whites before, looked greyed out or faded once they appeared in Blogger. Uploading pictures in Blogger with a white background and making them turn into grey is not what I had in mind.

To blame is the new photo enhance function, and an idiosyncratic implementation.Likely the alpha transparency would go missing as well, but this wasn't tested yet.

Update: After a quick test it shows that apparently Google does not enhance png's with an alpha transparency.

Example:
Take this example of Google+ Cheat Sheet:

lorenz-lo-sauer-google-plus-cheat-sheet-google+-formatting.jpg


Solution: As you can see the picture's background isn't white anymore, as it should be. That is owing to Blogger's integrated photo enhancement feature of Google+ called Auto Enhance and is turned on by default.
To turn it off go to your Google+ Preferences by clicking here pressing CTRL + F , then search for Auto Enhance and set it to Off.
As you can, see whites are white again.

Technical:

  • The settings are stored automatically, as any form changes are tied to an AJAX submit, containing the user's current session Id, along with an action-request Id of the Model-controller on the server-side offering additional security. The payload of the form is simply a variable 'at' being assigned the form Element in question, in this case AutoE, and a hashed value rather than a cleartext value, offering additional privacy protection.


AJAX Endpoint and query string
https://plus.google.com/u/0/_/photos/setsettings/?hl=en&f.sid=756568035&_reqid=669724&rt=j
 

  • A caveat regarding the idiosyncratic implementation: In a lossless format such as an png image, whites without noise are likely desired to remain so by the user. No noise would mean that all pixel-neighbors are of the same value in a say 8x8 raster or KernelThe measure of noise would depend on the probability distribution function, but should be able to qualify the state of being noiseless
  • in the context of the image algorithm.
LihatTutupKomentar