qlcplus icon indicating copy to clipboard operation
qlcplus copied to clipboard

Published 20 hours ago • verified publisher icon mcallegari

16bit movement issues

Open mcallegari opened this issue 2 years ago • 3 comments
trafficstars

Issue 1 https://www.qlcplus.org/forum/viewtopic.php?t=16788

Issue 2 https://www.qlcplus.org/forum/viewtopic.php?t=16789

Issue 3 https://www.qlcplus.org/forum/viewtopic.php?t=16790

mcallegari avatar Nov 18 '23 11:11 mcallegari

16 bit fade is being worked on in the following branch https://github.com/mcallegari/qlcplus/tree/16bitfade

mcallegari avatar Jan 08 '24 20:01 mcallegari

Test versions available for every platform https://www.qlcplus.org/downloads/test/index.php?p=

mcallegari avatar Feb 03 '24 15:02 mcallegari

As a kind reminder - the HTP mechanism does not work correctly with 16bit values. Say a scene has [MSB,LSB] 12,185 and another scene 13,22. Both scenes are active to trigger HTP. Doing the maths gives us 0x0CB9 and 0x0D16. The engine will spit out 0x0DB9 as a result. It should be 0x0D16.

GGGss avatar Feb 16 '24 07:02 GGGss

@GGGss I worked some more on 16bit. Can you please check the latest build from the 16bitfade branch?

Screenshot_20240223_232530

mcallegari avatar Feb 23 '24 22:02 mcallegari

I can confirm the HTP logic in the engine is valid. The LTP logic is - as expected - valid. Even in conflicting situations, the latest value wins. Tested with the QLC+-4.12.8_GIT-20240223-3188c91-16bit-FADE

GGGss avatar Feb 28 '24 12:02 GGGss

Adding a case report https://www.qlcplus.org/forum/viewtopic.php?t=17078

[EDIT] just checked and this was present pre-PR so not a regression but still to be fixed

mcallegari avatar Feb 28 '24 13:02 mcallegari

@GGGss another round of fixes. Apparently, relative EFX calculation has never worked as expected. Please check now.

mcallegari avatar Mar 03 '24 20:03 mcallegari

@mcallegari Checked and approved ... LSB values play nicely now ... I still need to test the outcome to the extremes (>255,255 <0,0 in what the EFX will produce. (I noticed you hard limited both limits into the code ... all well to me - because these are cases of operator-mind's errors!)

Let's use the sparse days to the public appearance of the v4.13 to debug a little more... - HTP & LTP couples now well-behave in my tests. A True 16b engine is born.

To me, this introduced a welcome add-on for the next version? True-relative-mode https://www.qlcplus.org/forum/viewtopic.php?p=72057#p72057 - But please accept my gratitude for your efforts

GGGss avatar Mar 04 '24 20:03 GGGss

@GGGss many thanks to you for testing! I will now consider this task as closed.

mcallegari avatar Mar 05 '24 07:03 mcallegari