I think Google maps just loads image tiles around the current projection
so when you translate, the images are already there and makes it look
smooth.
Here are few ideas you might try:
1.) cache the image buffer between panel resizes and reuse it
2.) create a layer or mousemode that takes a image snapshot of the
current map and then translate the image to the new locations during the
mouse drag and the notify the layers to update after the mouse button is
released or modify each layer to draw a cached offscreen image during
mouse movements and recalculate upon mouse up.
3.) If you don't like the the idea of translating an image with the
resulting unrendered gaps that show up, you could keep track of dirty
areas and the only recalulate those regions. NetBeans chart code had
some high performance ways of translating the chart in time and only
updating the new deltas which may be useful (with some surgery):
http://www.netbeans.org/source/browse/profiler/libs/jfluid-ui/src/org/netbeans/lib/profiler/ui/charts/SynchronousXYChart.java?rev=1.1&view=auto&content-type=text/vnd.viewcvs-markup
// draws scaled chart from provided viewOffsetX
private void drawChartFromOffset(Graphics2D g2, long offsetX) {
...
if ((copyWidth > 0) && (dirtyWidth > 0) && (chartHeight > 0)) {
// copy not-changed-areas to the left
if (copyAccel == COPY_ACCEL_RASTER) {
// BufferedImage.getRaster().setDataElements() used, seems to have better performance on Windows (HW acceleration)
int rasterWidth = offScreenImage.getRaster().getWidth();
int rasterHeight = offScreenImage.getRaster().getHeight();
int startX = chartInsets.left + dirtyWidth;
int startY = chartInsets.top;
if ((startX >= 0) && ((startX + copyWidth) <= rasterWidth) && (startY >= 0)
&& ((startY + chartHeight) <= rasterHeight) && (chartInsets.left >= 0)
&& ((chartInsets.left + copyWidth) <= rasterWidth) && (chartInsets.top >= 0)
&& ((chartInsets.top + chartHeight) <= rasterHeight)) {
Raster raster = offScreenImage.getRaster()
.createWritableChild(startX, startY, copyWidth, chartHeight, 0,
0, null);
offScreenImage.getRaster().setDataElements(chartInsets.left, chartInsets.top, raster);
}
} else {
// Graphics.copyArea() used, optimal for UNIXes, works well also for Windows (default)
g2.copyArea(chartInsets.left + dirtyWidth, chartInsets.top, copyWidth, chartHeight, -dirtyWidth, 0);
}
}
F.Bayliss wrote:
> Hello,
>
> I've been asked by my upper management to once again mimic the way
> Google allows the user to move the map. I've spent a lot of time on
> this, too much, and I don't see anyway it can work. I have several,
> several layers and the time it takes to process and draw a map
> movement event is slow and that results in a choppy map picture. I've
> tried several approaches, the best being drawing the map to an image
> buffer and replacing the map with that, letting them move what they
> think is the map. But this isn't acceptable because it takes time to
> create the image. I was hoping the experts had a better idea.
>
> Thanks
> Frank
>
> --
> [To unsubscribe to this list send an email to "majdart_at_bbn.com"
> with the following text in the BODY of the message "unsubscribe
> openmap-users"]
>
--
Rob Ratcliff
cell: 512-633-5751
office: 512-250-3444
fax: 512-233-2873
web: http://www.futuretek.com
--
[To unsubscribe to this list send an email to "majdart_at_bbn.com"
with the following text in the BODY of the message "unsubscribe openmap-users"]
Received on Thu Apr 10 2008 - 09:28:59 EDT