Originally posted by Pål Jensen Strange then that it doesn't happen when used on a tripod when not using 2s delay (all else equal)...
Things are being mixed up here.
1) only corner softness around horizon correction
That seems related to the timing of the horizon correction of the sensor.
If only the corners are soft this is about the sensor adjustment and definitely not vibration induced (definitely not "shutter shock").
2) general vibration induced softness when used on tripod
That is what I referred to mainly. This is the topic created by the interns at dptabloid.
Here the softness clearly is in the center of the image as well.
And this very likely is related to the resonance effects described if the lieklyhood of occurence is as low as it is when using the camera in a well dampenend scenario (e.g. handheld, proper dampened tripod).