-
Notifications
You must be signed in to change notification settings - Fork 2
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Corrupt merge of mvds #10
Comments
Hi @dsvensson, Please add those demos. they will help for sure... |
I noticed now that one of the demos is not playing back correctly when viewed separately. Playing back in slomo, and the -fps parameter didn't help. Might be that one is a regular demo recording, and the other one is a qizmo recording. Maybe that's what causes flag carrier to drift out of sync. I just googled some random file service, they should be available here for 7 days, |
@tcsabina did you manage to grab it or do I need to find some other place for it? |
Hi @dsvensson , Sorry, I completely forgot about this issue, and the demo you have saved. Could you add it again somewhere please? Sorry for the inconvenience. |
Nice, never seen that feature in GH before, here you go. iirc it's the final3.qwd that becomes slowed down when converting to mvd (before merge). Haven't measured exactly the ratio, but it can bee seen via |
perfect! |
Interesting. Both demos use the Pure CTF statusbar which shows the game clock. Comparing that game clock to wallclock shows that 1 minute of game clock is more than 1 minute wall clock in both demos - before converting. final3.qwd is about 70 seconds per 60 seconds game clock, and first3.qwd is 75 seconds per 60 seconds game clock modulo sloppy reactivity. Any idea why that might be? |
No longer able to reproduce the slowdowns in ezquake master, still some slowdowns in FTE. This is however already in the qwd's. What is true is that merging multiple POV's seems to have trouble clearing flag carrier status. first3.qwd and first3.mvd - [c]space pov are both fine, but when combining with final3, and jumping to space pov at for example Comparing time with centerprint status bar written into the demo and wallclock produces the following result, when considering final3.qwd/first3.qwd, so at least not an issue here, unless the drift seen in FTE is related, but maybe that's how FTE behaves in other demos as well, haven't had time to check yet. https://gist.github.com/dsvensson/a7d11ede880130bee40a2ccda145131c Going to have a look what mvdparser says about the converted mvd's, but they ofc contained unimplemented messages, the yak shaving continues 🔪. |
Ok, mvdparser shines a clear light on the problem, missing messages, and duplicated messages, inspection of first1.mvd, final3.mvd - both correct, and the combined mvd here: https://gist.github.com/dsvensson/b1612eee74e95cec2832fbbcb5c15eed It's very weird that the first 7 takes+drops for the observed player are lost, but it would be interesting to see if the merge algorithm works correctly if the drift was eliminated. |
Steps:
While the output is playable - very cool stuff! - and looks mostly correct, it seems like the flag carrier status is not correctly cleared leading to cases where flag is in the base position, while also on the back of an attacking enemy etc. I can provide source demos and merged demos if needed.
The text was updated successfully, but these errors were encountered: