Notes |
|
(0004444)
|
Qent
|
2012-08-25 06:01
|
|
It would help me to know the exact mechanics of the phenomenon. From what I can tell, Chocolate Doom and Zandronum are very similar, while Odamex's is many times stronger. |
|
|
|
I want it so we can play rocket jump wads (: |
|
|
|
I surely need more technical details on what exactly you would like to have changed how. |
|
|
|
I think it's because of the greater aircontrol skulltag has.
'http://zdoom.org/wiki/CVARs:Configuration [^]' says "Note: Skulltag ignores this value unless its compat_limited_airmovement flag is set. The default behavior in that port corresponds to an aircontrol value of 0.25." at sv_aircontrol and is by "Default: 0.00390625 " |
|
|
(0004456)
|
Qent
|
2012-08-25 23:34
(edited on: 2012-08-25 23:39) |
|
It's not. I tested with compat_limited_airmovement and sv_aircontrol 0.
BTW the wiki is incorrect. sv_aircontrol does have an effect when compat_limited_airmovement is false, albeit a strange one. I'll fix the wiki sometime if no one does.
|
|
|
(0004607)
|
unknownna
|
2012-09-07 08:42
(edited on: 2012-09-07 08:45) |
|
|
|
(0004675)
|
Torr Samaho
|
2012-09-15 19:46
(edited on: 2012-09-15 19:48) |
|
Quote from unknownna In 1.23, rocket blasts were a lot more powerful and sent the players flying far away.
Watermelon, can you confirm that ZDoom 1.23 behaves exactly as you would like it to in this regard? If so, what's the first ZDoom version where the behavior was changed? Here you'll hopefully find enough old ZDoom versions to pinpoint where the behavior was changed.
|
|
|
(0004710)
|
unknownna
|
2012-09-16 17:12
(edited on: 2012-09-16 17:22) |
|
Quote from Torr Samaho Watermelon, can you confirm that ZDoom 1.23 behaves exactly as you would like it to in this regard?
ZDaemon 1.09 seems to behave like ZDoom 1.23.
Quote from Torr Samaho If so, what's the first ZDoom version where the behavior was changed? Here you'll hopefully find enough old ZDoom versions to pinpoint where the behavior was changed.
From what I can tell, it seems that ZDoom changed the behavior somewhere between 2.0.42 and 2.0.43.
|
|
|
|
I found a difference regarding explosion trust between ZDoom 2.0.42 and 2.0.43 and ported it over. Please try if this binary behaves as desired. If so, I can add a new flag to toggle this behavior. |
|
|
(0004714)
|
unknownna
|
2012-09-16 22:23
(edited on: 2012-09-16 22:49) |
|
Very nice, Torr. Thank you very much for this. It works well, but it seems that ZDaemon 1.09 and ZDoom 1.23 sends you approximately 2-4 map units further on average. I think it has something to do with the damage calculation since the rockets seem to consistently deal different amounts of damage in each port.
ZDoom 2.0.21-2.0.42 is the strongest in terms of thrusting, but it doesn't deal the most damage.
ZDaemon 1.09 and ZDoom 1.23 are in the middle in terms of thrusting, but they deal the most damage it seems.
Zandronum is the weakest, even with compat_oldrandom set to 1.
I think that this bug could be related the issue.
Most people won't notice the differences, but they are there. The only thing missing now is some damage prediction for rocket thrusts online and we're set.
Note: For those of you who wish to test these binaries, set sv_aircontrol to 0 and compat_limited_airmovement to 1.
|
|
|
|
Quote from unknownna ZDoom 2.0.21-2.0.42 is the strongest in terms of thrusting, but it doesn't deal the most damage.
ZDaemon 1.09 and ZDoom 1.23 are in the middle in terms of thrusting, but they deal the most damage it seems. If we want to refine this further, we need to find out which ZDoom versions changed this behavior.
Quote from unknownna The only thing missing now is some damage prediction for rocket thrusts online and we're set. You mean that the clients should explode the rockets themselves (something we explicitly turned off a while ago because people don't like mispredictions in this regard) and estimate how much they will get thrusted? Does ZDaemon have anything like this? |
|
|
(0004721)
|
unknownna
|
2012-09-20 16:28
(edited on: 2012-09-20 16:41) |
|
Quote from Torr Samaho You mean that the clients should explode the rockets themselves (something we explicitly turned off a while ago because people don't like mispredictions in this regard) and estimate how much they will get thrusted? Does ZDaemon have anything like this?
What I can tell you is that under the same circumstances (300 ping), ZDaemon 1.09 plays considerably better when it comes to rocket jumping. I don't notice any mispredictions there. The only thing I notice in ZDaemon is that the DSOOF sound is delayed when you land on a floor. In Zandronum, you seem to *jitter/glitch/teleport/warp* when you're thrusted, and if you don't press +jump, you won't even see any z thrusting locally, which in turn causes the client to mispredict its own z position, thinking that it's instantly located on the lowest floor when it actually is falling down. Zandronum feels very amateurish in comparison.
Quote from Torr Samaho If we want to refine this further, we need to find out which ZDoom versions changed this behavior.
I couldn't test ZDoom 2.0.13, the version after 1.23 beta 33, since the download only contains the source code. But there's a difference between 1.23 beta 7 and 1.23 beta 24. It seems that beta 24 behaves like beta 33, but beta 7 is even weaker than Zandronum. I couldn't test 1.23 beta 20 since the download only contains the source code.
|
|
|
(0004723)
|
Torr Samaho
|
2012-09-20 18:52
(edited on: 2012-09-20 18:53) |
|
Zandronum doesn't have any client side explosion prediction, so it's not surprising that it doesn't work well with a high ping in this regard. Adding such a kind of prediction is not trivial and may have adverse side effects. I'm not sure if it's worth the effort.
Quote from unknownna I couldn't test ZDoom 2.0.13, the version after 1.23 beta 33, since the download only contains the source code. But there's a difference between 1.23 beta 7 and 1.23 beta 24. It seems that beta 24 behaves like beta 33, but beta 7 is even weaker than Zandronum. I couldn't test 1.23 beta 20 since the download only contains the source code. Let me ask differently, is there any ZDoom version on Sourceforge (or elsewhere, it just has to include the full source) that behaves exactly as desired?
|
|
|
|
That binary is awesome. It works really well. Sorry I didn't get back to this ticket, the university year has started.
I tried it out and it worked really well (the binary that is).
COMPAT_OLDROCKETTHRUST maybe? |
|
|
(0004726)
|
unknownna
|
2012-09-20 19:49
(edited on: 2012-09-20 19:56) |
|
Quote from Torr Samaho I'm not sure if it's worth the effort.
Of course it's worth the effort in the long run. If anything, it should have been added ages ago. I already requested this a while ago. What's the point of playing seriously and competitively when you can't even maneuver around properly? Don't expect competitive players (IDL) to take this port seriously until we've fixed all of the evident prediction errors. We already fixed most of them in 1.0.
Quote from Torr Samaho Let me ask differently, is there any ZDoom version on Sourceforge (or elsewhere, it just has to include the full source) that behaves exactly as desired?
Well, as I said earlier, ZDaemon (and Odamex's "ZDoom physics" flag) is based on ZDoom 1.23 beta 33, so the source code of that version should give you what you need.
|
|
|
|
Quote from Watermelon COMPAT_OLDROCKETTHRUST maybe? I added the flag shortly after unknownna confirmed that the testing binary works :)'https://bitbucket.org/Torr_Samaho/zandronum/changeset/a90c6f1434bd6fe7ca4894afd6d1b3392ca76545 [^]'
Quote from unknownna Of course it's worth the effort in the long run.
Let me clarify my statement: Of course it won't hurt to have proper prediction for this, but there are countless of things that would be nice to have. I'm not sure that the effort is better spend on this than let's say upgrading to ZDoom 2.4.
Quote from unknownna If anything, it should have been added ages ago. I already requested this a while ago. And the statement I made there still holds: 0000517:0001902.
Quote from unknownna Well, as I said earlier, ZDaemon (and Odamex's "ZDoom physics" flag) is based on ZDoom 1.23 beta 33, so the source code of that version should give you what you need. So you can confirm that ZDoom 1.23 beta 33 behaves exactly as desired? I wasn't completely certain about this from the previous comments. |
|
|
(0004730)
|
unknownna
|
2012-09-20 20:34
(edited on: 2012-09-20 20:43) |
|
Quote from Torr Samaho I'm not sure that the effort is better spend on this than let's say upgrading to ZDoom 2.4.
I think most developers and modders wish to see Zandronum upgrade to ZDoom 2.4.0.
Quote from Torr Samaho So you can confirm that ZDoom 1.23 beta 33 behaves exactly as desired? I wasn't completely certain about this from the previous comments.
They (ZDaemon 1.09 and ZDoom 1.23 beta 33) should pretty much be identical to one another in terms of thrusting and damage. When having a megasphere, holding +attack and firing two rockets near a wall (MAP01) should have you end up with:
Armor : 95
Health : 94
Y Postition : 1304
Currently in Zandronum you end up with:
Armor : 97
Health : 97
Y Position : 1305
ZDoom 1.23 beta 33 should behave as desired.
Quote from Torr Samaho I added the flag shortly after unknownna confirmed that the testing binary works :)https://bitbucket.org/Torr_Samaho/zandronum/changeset/a90c6f1434bd6fe7ca4894afd6d1b3392ca76545
There are some typos in the comments. Thurst should be thrust.
|
|
|
|
Could odamex/ZD rocket thrust be possibly tampered with further (which may give the extra 2 map unit differences)? They might have slightly tweaked the old ZD code further. |
|
|
|
Quote from Watermelon Could odamex/ZD rocket thrust be possibly tampered with further (which may give the extra 2 map unit differences)? They might have slightly tweaked the old ZD code further.
I don't think so.
Quote from Vulture When you enable ZDoom mode in Odamex, the ZDoom physics that is emulated is based on ZDoom 1.23 beta 33, which is the same version of ZDoom that ZDaemon is based on. The rocket jumping in Odamex should be exactly the same compared ZDaemon's.
Quote from EarthQuake
Quote from Galaxy_Stranger I've been poking around and I can't find an answer -
What version of ZDoom is the current version of ZDaemon based upon?
1.23 beta 33 |
|
|
|
Quote from unknownna ZDoom 2.0.21-2.0.42 is the strongest in terms of thrusting, but it doesn't deal the most damage.
ZDaemon 1.09 and ZDoom 1.23 are in the middle in terms of thrusting, but they deal the most damage it seems.
The PIT_RadiusAttack/P_RadiusAttack code in ZDoom 1.23b33 and 2.0.25 is identical, the damage difference has to be caused by something else (possibly the PNRG). Without more evidence it's very difficult to fine tune this further. |
|
|
|
Hmm, did you change something in this binary? It behaves differently now.
It should take 2 rockets to push you away from a wall with the 1.23b33 behavior. Right now it only takes one. But if you're invulnerable, it takes 2 rockets. |
|
|
|
I simplified the code before committing it. After looking into this more closely this may have changed the behavior. Please try if this binary works as the original testing binary. |
|
|
|
That took care of it. It seems that we now have several explosion-related flags:
sv_norocketjumping
compat_oldradiusdmg
compat_oldexplosionthrust
Which one of them has the final say in terms of behavior?
Quote from Torr Samaho The PIT_RadiusAttack/P_RadiusAttack code in ZDoom 1.23b33 and 2.0.25 is identical, the damage difference has to be caused by something else (possibly the PNRG). Without more evidence it's very difficult to fine tune this further.
Speaking of which, it seems that 1.23b33 explosions deal more damage than Vanilla Doom explosions. When standing close to a wall and firing a rocket while having 200 blue armor and 100 health, you should end up with 36 health and 136 armor. In 1.23b33 you end up with 35 health and 136 armor. |
|
|
|
Quote from unknownna Which one of them has the final say in terms of behavior? sv_norocketjumping and compat_oldexplosionthrust are complementary:
- sv_norocketjumping removes the z-momentum
- compat_oldexplosionthrust changes the xy-momentum
compat_oldexplosionthrust disables the two other ones though (at least for explosions that deal damage). If it's set, Vanilla Doom's radius damage code is used, i.e. explosions have infinite vertical range. |
|
|
(0005891)
|
Mobius
|
2013-01-31 07:01
|
|
Tested online and it works accordingly. |
|