You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I have some demosaiced (aka Linear RAW) DNGs created by a rather oddball camera (I don't own it, but have put some effort into reverse engineering/supporting it).
These cause the asserts from 831a9bb to fail because rawdata.getWidth() returns a width that is three times the image width since it has three samples per pixel
Is the value of DNG samplesPerPixel readily available where these asserts are?
Alternatively, maybe asserting (rawData.getWidth() % W) == 0 would work here, under the assumption that getWidth() won't ever be exactly zero if getHeight is nonzero?
The text was updated successfully, but these errors were encountered:
Entropy512
added a commit
to Entropy512/RawTherapee
that referenced
this issue
Mar 17, 2024
Demosaiced DNGs (aka LinearRAW PhotometricInterpretation) have 3 samples per pixel, so rawData.getWidth() returns triple the width
Only throw an assert if getWidth is not a multiple of W. This new implementation assumes that getWidth() will never be zero
when getHeight() is nonzero
FixesBeep6581#6996
copyOriginalPixels is where rawData gets resized. If ri->getSensorType() == ST_BAYER || ri->getSensorType() == ST_FUJI_XTRANS || ri->get_colors() == 1, then its height and width are H and W. Otherwise, they are H and 3 * W.
I have some demosaiced (aka Linear RAW) DNGs created by a rather oddball camera (I don't own it, but have put some effort into reverse engineering/supporting it).
One example is at https://filebin.net/a8zb1jja5rwb8dk5
These cause the asserts from 831a9bb to fail because rawdata.getWidth() returns a width that is three times the image width since it has three samples per pixel
Is the value of DNG samplesPerPixel readily available where these asserts are?
Alternatively, maybe asserting (rawData.getWidth() % W) == 0 would work here, under the assumption that getWidth() won't ever be exactly zero if getHeight is nonzero?
The text was updated successfully, but these errors were encountered: