| Summary: | Height Map to Normal Map creates a line at the top/bottom edge. | ||
|---|---|---|---|
| Product: | [Applications] krita | Reporter: | sidartalei |
| Component: | Filters | Assignee: | Krita Bugs <krita-bugs-null> |
| Status: | RESOLVED FIXED | ||
| Severity: | normal | CC: | dennis.ranke, halla |
| Priority: | NOR | ||
| Version: | 4.0 | ||
| Target Milestone: | --- | ||
| Platform: | unspecified | ||
| OS: | All | ||
| Latest Commit: | https://commits.kde.org/krita/a0e2a23f0e7ab1054b9fd49cad29c3c58590cacc | Version Fixed In: | |
| Sentry Crash Report: | |||
| Attachments: | The canvas in wrap mode, the lines become apparent. | ||
Yes, this is as far as I understand a problem with the convolution painter itself. I pass the wraparound variable correctly, as far as I know. I did a read through the code and I /think/ I identified two issues that relate to this bug:
1. In KisEdgeDetectionKernel::convertToNormalMap there is this code to apply the horizontal and vertical convolution kernels:
horizPainterLR.applyMatrix(kernelHorizLeftRight, device,
srcTopLeft - QPoint(0, ceil(horizontalCenter)),
srcTopLeft - QPoint(0, ceil(horizontalCenter)),
rect.size() + QSize(0, 2 * ceil(horizontalCenter)), BORDER_REPEAT);
and
verticalPainterTB.applyMatrix(kernelVerticalTopBottom, device,
srcTopLeft - QPoint(0, ceil(verticalCenter)),
srcTopLeft - QPoint(0, ceil(verticalCenter)),
rect.size() + QSize(0, 2 * ceil(verticalCenter)), BORDER_REPEAT);
If you look at the QPoints and the QSize, you see that in both cases the x is set to zero and the y to some value based on the horizontal/vertical center. I guess that in one of the cases, probably the vertical one, the two vector components should be flipped around.
2. The definition of BORDER_REPEAT which you pass into applyMatrix is defined as:
BORDER_REPEAT = 1 // Use the border for the missing pixels
Which is actually a clamp, not a wrap-around. So a wrapping normal map will have some artifacts on the borders. Ideally, the filter would gain an option to select a newly coded BORDER_WRAP mode, although that's really more of a feature request rather than a bug.
Ok, I guessed wrong, it's actually the horizontal applyMatrix that is wrong:
horizPainterLR.applyMatrix(kernelHorizLeftRight, device,
srcTopLeft - QPoint(ceil(horizontalCenter), 0),
srcTopLeft - QPoint(ceil(horizontalCenter), 0),
rect.size() + QSize(2 * ceil(horizontalCenter), 0), BORDER_REPEAT);
gets rid of the line at the top and bottom edges for me.
Git commit bc30d50e80024d7b2352fa0b3112ecff5e9c36aa by Boudewijn Rempt. Committed on 28/03/2018 at 14:08. Pushed by rempt into branch 'master'. Fix height map to normal map in wraparound mode patch by Dennis Ranke. Thanks! M +3 -3 libs/image/kis_edge_detection_kernel.cpp https://commits.kde.org/krita/bc30d50e80024d7b2352fa0b3112ecff5e9c36aa Git commit a0e2a23f0e7ab1054b9fd49cad29c3c58590cacc by Boudewijn Rempt. Committed on 03/04/2018 at 11:18. Pushed by rempt into branch 'krita/4.0'. Fix height map to normal map in wraparound mode patch by Dennis Ranke. Thanks! (cherry picked from commit bc30d50e80024d7b2352fa0b3112ecff5e9c36aa) M +3 -3 libs/image/kis_edge_detection_kernel.cpp https://commits.kde.org/krita/a0e2a23f0e7ab1054b9fd49cad29c3c58590cacc |
Created attachment 111558 [details] The canvas in wrap mode, the lines become apparent. The Height Map to Normal Map filter creates height information on the top and bottom of the image.