summaryrefslogtreecommitdiff
path: root/dix
diff options
context:
space:
mode:
authorSimon Thum <simon.thum@gmx.de>2008-08-07 11:06:21 +0200
committerPeter Hutterer <peter.hutterer@redhat.com>2008-08-08 16:19:20 +0930
commite00cd54c1ac57ea6e7767e90592200f343d9d08a (patch)
treefaffec13fc876cbc663737dfc31adc8214c2b472 /dix
parent6ca34549b89e73e3e0ec3cb7585686b880534cef (diff)
dix: remove misleading comment in ptrvelo.c
Signed-off-by: Peter Hutterer <peter.hutterer@redhat.com>
Diffstat (limited to 'dix')
-rw-r--r--dix/ptrveloc.c5
1 files changed, 2 insertions, 3 deletions
diff --git a/dix/ptrveloc.c b/dix/ptrveloc.c
index 330ab39a7..2406be8cd 100644
--- a/dix/ptrveloc.c
+++ b/dix/ptrveloc.c
@@ -152,9 +152,8 @@ Expected result is a series of filters, each progressively more integrating.
This allows for two strategies: Either you have one filter which is reasonable
and is being coupled to account for fast-changing input, or you have 'one for
-every situation'. You might want to have loose coupling then, i.e. > 1.
-E.g. you could start around 1/2 of your anticipated delta t and
-scale up until several motion deltas are 'averaged'.
+every situation'. You might want to have tighter coupling then, e.g. 0.1.
+In the filter stats, you can see if a reasonable filter useage emerges.
*/
void
InitFilterChain(DeviceVelocityPtr s, float rdecay, float progression, int stages, int lutsize)