Posts: 3,978
Threads: 177
Joined: Apr 2022
Reputation:
220
The skew problem arose from using different scales for x and y. If you (Steve) aren't doing that your code should be fine. Rotozoom2 went for years without detection of problem.
b = b + ...
Posts: 660
Threads: 142
Joined: Apr 2022
Reputation:
58
(01-21-2023, 07:10 PM)SMcNeill Wrote: I'll have to check later to see if my DisplayImage suffers from this same skew problem. https://qb64phoenix.com/forum/showthread.php?tid=133
Yup, just checked it shows the skew on rotating when x and y aren't the same scale.
Posts: 2,698
Threads: 327
Joined: Apr 2022
Reputation:
217
(01-21-2023, 07:27 PM)James D Jarvis Wrote: (01-21-2023, 07:10 PM)SMcNeill Wrote: I'll have to check later to see if my DisplayImage suffers from this same skew problem. https://qb64phoenix.com/forum/showthread.php?tid=133
Yup, just checked it shows the skew on rotating when x and y aren't the same scale.
Hopefully I've fixed that. (I followed your changes to rotozoom as a basic roadmap to my own changes, so I hope they're right now.) At least, it seems to be working to me, but heck; I never noticed the problem to begin with! I'ma gonna pretend that it's working now, until someone tells me that it isn't. LOL!
Posts: 2,698
Threads: 327
Joined: Apr 2022
Reputation:
217
One thing I noticed tweaking DisplayImage, which you guys might want to go back and incorporate into your Rotozoom changes:
W& = _Width(Image&): H& = _Height(Image&)
Add a line below this one and make it:
W1& = W& *xScale: H1& = H& *yScale
Then when you do all the P(x) and P(y) calculations, you can use W1 and H1 and reduce the overall math in your routine, which would end up making it a little bit faster in the long run. And, when it comes to graphics, any little speed boost (which as simple to implement as this one), is one that a programmer should always think about making use of.
Posts: 660
Threads: 142
Joined: Apr 2022
Reputation:
58
(01-21-2023, 09:22 PM)SMcNeill Wrote: One thing I noticed tweaking DisplayImage, which you guys might want to go back and incorporate into your Rotozoom changes:
W& = _Width(Image&): H& = _Height(Image&)
Add a line below this one and make it:
W1& = W& *xScale: H1& = H& *yScale
Then when you do all the P(x) and P(y) calculations, you can use W1 and H1 and reduce the overall math in your routine, which would end up making it a little bit faster in the long run. And, when it comes to graphics, any little speed boost (which as simple to implement as this one), is one that a programmer should always think about making use of.
Thank you for pointing that out.
Posts: 660
Threads: 142
Joined: Apr 2022
Reputation:
58
updated routine(s)
Code: (Select All) Sub RotoZoom23d (centerX As Long, centerY As Long, Image As Long, xScale As Single, yScale As Single, Rotation As Single)
Dim px(3) As Single: Dim py(3) As Single
Wi& = _Width(Image&): Hi& = _Height(Image&)
W& = Wi& / 2 * xScale
H& = Hi& / 2 * yScale
px(0) = -W&: py(0) = -H&: px(1) = -W&: py(1) = H&
px(2) = W&: py(2) = H&: px(3) = W&: py(3) = -H&
sinr! = Sin(-0.01745329 * Rotation): cosr! = Cos(-0.01745329 * Rotation)
For i& = 0 To 3
x2& = (px(i&) * cosr! + sinr! * py(i&)) + centerX: y2& = (py(i&) * cosr! - px(i&) * sinr!) + centerY
px(i&) = x2&: py(i&) = y2&
Next
_MapTriangle (0, 0)-(0, Hi& - 1)-(Wi& - 1, Hi& - 1), Image& To(px(0), py(0))-(px(1), py(1))-(px(2), py(2))
_MapTriangle (0, 0)-(Wi& - 1, 0)-(Wi& - 1, Hi& - 1), Image& To(px(0), py(0))-(px(3), py(3))-(px(2), py(2))
End Sub
Sub RotoZoom23ds (centerX As Long, centerY As Long, Image As Long, xScale As Single, yScale As Single, Rotation As Single)
Dim px(3) As Single: Dim py(3) As Single
Wi& = _Width(Image&): Hi& = _Height(Image&)
W& = Wi& / 2 * xScale
H& = Hi& / 2 * yScale
px(0) = -W&: py(0) = -H&: px(1) = -W&: py(1) = H&
px(2) = W&: py(2) = H&: px(3) = W&: py(3) = -H&
sinr! = Sin(-0.01745329 * Rotation): cosr! = Cos(-0.01745329 * Rotation)
For i& = 0 To 3
x2& = (px(i&) * cosr! + sinr! * py(i&)) + centerX: y2& = (py(i&) * cosr! - px(i&) * sinr!) + centerY
px(i&) = x2&: py(i&) = y2&
Next
_MapTriangle _Seamless(0, 0)-(0, Hi& - 1)-(Wi& - 1, Hi& - 1), Image& To(px(0), py(0))-(px(1), py(1))-(px(2), py(2))
_MapTriangle _Seamless(0, 0)-(Wi& - 1, 0)-(Wi& - 1, Hi& - 1), Image& To(px(0), py(0))-(px(3), py(3))-(px(2), py(2))
End Sub
Code: (Select All) Sub RotoZoom_23r (centerX As Long, centerY As Long, Image As Long, xScale As Single, yScale As Single, Rotation As Single)
'uses radians
Dim px(3) As Single: Dim py(3) As Single
W& = _Width(Image&): H& = _Height(Image&)
Wp& = W& / 2 * xScale
Hp& = H& / 2 * yScale
px(0) = -Wp&: py(0) = -Hp&: px(1) = -Wp&: py(1) = Hp&
px(2) = Wp&: py(2) = Hp&: px(3) = Wp&: py(3) = -Hp&
sinr! = Sin(-Rotation): cosr! = Cos(Rotation)
For i& = 0 To 3
' x2& = (px(i&) * cosr! + sinr! * py(i&)) * xScale + centerX: y2& = (py(i&) * cosr! - px(i&) * sinr!) * yScale + centerY
x2& = (px(i&) * cosr! + sinr! * py(i&)) + centerX: y2& = (py(i&) * cosr! - px(i&) * sinr!) + centerY
px(i&) = x2&: py(i&) = y2&
Next
_MapTriangle (0, 0)-(0, H& - 1)-(W& - 1, H& - 1), Image& To(px(0), py(0))-(px(1), py(1))-(px(2), py(2))
_MapTriangle (0, 0)-(W& - 1, 0)-(W& - 1, H& - 1), Image& To(px(0), py(0))-(px(3), py(3))-(px(2), py(2))
End Sub
Posts: 3,978
Threads: 177
Joined: Apr 2022
Reputation:
220
01-22-2023, 03:33 PM
(This post was last modified: 01-22-2023, 04:08 PM by bplus.)
(01-21-2023, 11:09 PM)James D Jarvis Wrote: updated routine(s)
Code: (Select All) Sub RotoZoom23d (centerX As Long, centerY As Long, Image As Long, xScale As Single, yScale As Single, Rotation As Single)
Dim px(3) As Single: Dim py(3) As Single
Wi& = _Width(Image&): Hi& = _Height(Image&)
W& = Wi& / 2 * xScale
H& = Hi& / 2 * yScale
px(0) = -W&: py(0) = -H&: px(1) = -W&: py(1) = H&
px(2) = W&: py(2) = H&: px(3) = W&: py(3) = -H&
sinr! = Sin(-0.01745329 * Rotation): cosr! = Cos(-0.01745329 * Rotation)
For i& = 0 To 3
x2& = (px(i&) * cosr! + sinr! * py(i&)) + centerX: y2& = (py(i&) * cosr! - px(i&) * sinr!) + centerY
px(i&) = x2&: py(i&) = y2&
Next
_MapTriangle (0, 0)-(0, Hi& - 1)-(Wi& - 1, Hi& - 1), Image& To(px(0), py(0))-(px(1), py(1))-(px(2), py(2))
_MapTriangle (0, 0)-(Wi& - 1, 0)-(Wi& - 1, Hi& - 1), Image& To(px(0), py(0))-(px(3), py(3))-(px(2), py(2))
End Sub
Sub RotoZoom23ds (centerX As Long, centerY As Long, Image As Long, xScale As Single, yScale As Single, Rotation As Single)
Dim px(3) As Single: Dim py(3) As Single
Wi& = _Width(Image&): Hi& = _Height(Image&)
W& = Wi& / 2 * xScale
H& = Hi& / 2 * yScale
px(0) = -W&: py(0) = -H&: px(1) = -W&: py(1) = H&
px(2) = W&: py(2) = H&: px(3) = W&: py(3) = -H&
sinr! = Sin(-0.01745329 * Rotation): cosr! = Cos(-0.01745329 * Rotation)
For i& = 0 To 3
x2& = (px(i&) * cosr! + sinr! * py(i&)) + centerX: y2& = (py(i&) * cosr! - px(i&) * sinr!) + centerY
px(i&) = x2&: py(i&) = y2&
Next
_MapTriangle _Seamless(0, 0)-(0, Hi& - 1)-(Wi& - 1, Hi& - 1), Image& To(px(0), py(0))-(px(1), py(1))-(px(2), py(2))
_MapTriangle _Seamless(0, 0)-(Wi& - 1, 0)-(Wi& - 1, Hi& - 1), Image& To(px(0), py(0))-(px(3), py(3))-(px(2), py(2))
End Sub
Code: (Select All) Sub RotoZoom_23r (centerX As Long, centerY As Long, Image As Long, xScale As Single, yScale As Single, Rotation As Single)
'uses radians
Dim px(3) As Single: Dim py(3) As Single
W& = _Width(Image&): H& = _Height(Image&)
Wp& = W& / 2 * xScale
Hp& = H& / 2 * yScale
px(0) = -Wp&: py(0) = -Hp&: px(1) = -Wp&: py(1) = Hp&
px(2) = Wp&: py(2) = Hp&: px(3) = Wp&: py(3) = -Hp&
sinr! = Sin(-Rotation): cosr! = Cos(Rotation)
For i& = 0 To 3
' x2& = (px(i&) * cosr! + sinr! * py(i&)) * xScale + centerX: y2& = (py(i&) * cosr! - px(i&) * sinr!) * yScale + centerY
x2& = (px(i&) * cosr! + sinr! * py(i&)) + centerX: y2& = (py(i&) * cosr! - px(i&) * sinr!) + centerY
px(i&) = x2&: py(i&) = y2&
Next
_MapTriangle (0, 0)-(0, H& - 1)-(W& - 1, H& - 1), Image& To(px(0), py(0))-(px(1), py(1))-(px(2), py(2))
_MapTriangle (0, 0)-(W& - 1, 0)-(W& - 1, H& - 1), Image& To(px(0), py(0))-(px(3), py(3))-(px(2), py(2))
End Sub
+1 James and Steve I like the degree conversions washed out. Luv it when list of contributors gets to 4!
Thanks for updating this with Radians, I hope it's tested because I just added to my Tool box because I am betting it is!
Update: now tested and the blue grid edges are better without _Seamless, good!
b = b + ...
|