Replies: 15 comments 39 replies
-
|
In fact, since I referenced #25814, might as well show how fickle this is on the examples from that issue: Creature 19163 has no movementId, so let's look at the models used: And yet, only one sniffed case matched these numbers. But it did match them, 6.9444 is not a number you just randomly decide to type. But why the other ones didn't?.. Do Blizzard also have a speed override in CreatureModelData?.. In CreatureDisplayInfo?.. It's so strange... |
Beta Was this translation helpful? Give feedback.
-
After all 2.5 and 8 are indeed default values, we need to expect they will be used in thousands of cases but also expect that they are wrong because Let's discuss that c) case. Why don't you think those Walk 0 and Run 0 are not for case I mentioned in #25814 (comment)? Fuck. SELECT entry,modelid1,modelid2,modelid3,modelid4,speed_walk,speed_run,movementid FROM creature_Template WHERE movementid=150To many uses different speed than 1.6 & 1.42857 but that's ok. They're mostly heroic entries(heroic entries mostly are not updated) or TBC or not updated or unused \ not spawned transform templates, etc. We can update lots of them from sniffs, not all but more than half of those 150 SELECT entry,NAME,modelid1,modelid2,modelid3,modelid4,speed_walk,speed_run,movementid FROM creature_Template WHERE movementid=150 AND (speed_walk BETWEEN 0.9 AND 1.1 OR speed_run BETWEEN 1.13 AND 1.15)#entry iN (25618)#name like '%evidence marker%'#modelid4 between 18579 and 18582--
I probably miss something but from where those speeds are taken? I checked that case, it's interesting because one creature uses 4 models and 3 of them uses different speed. I checked all creatures with those models and turned out speed matches. Of course some of those creatures were not updated from sniffs but mostly few or more creatures with the same model used same speed. So speed is linked directly to model id and also to movementId and not to creature_template and model ids as I was thinking before(if creature_template speed is 0 = take speed from model, if !=0 = take speed from creature_template) What do you think? |
Beta Was this translation helpful? Give feedback.
-
What do you mean? The second one clearly has movementId, therefore its speed, wherever it comes from, get overridden by the speed from CreatureMovementInfo, which for ID=150 is indeed 4 walk, 10 run. If a creature has movementId with both non-zero overrides - it no longer matters which model does it uses. But, of course, the first one doesn't have movementId, so according to my theory it should be using the values from its model. But its model has 2.5 walk, 6.944445 run, which doesn't match the sniff, so go figure... :( BTW, you can use
Ideally, you should be able update absolutely all of those that have movementId=150 to 4 walk 10 run speeds. But before I can guarantee that - I need to figure out the reason behind cases like the ones I showcased at the end of my original post. I would need to manually scour these logs to see if they don't have any auras applied to them that would mess with the numbers and prevent them from matching CreatureMovementInfo data.
They are taken from the model files themselves. You need some way to examine them. I've been using some kind of a model viewer that looks like this, I have no idea where to even download it from these days: Of course it takes too long to do for every single model, so I made for myself a basic property tree viewer
I've been relying purely on sniffs when performing those tests to see if my theories match reality. I only used TDB speed_walk/speed_run to fill the CreatureMovementInfo data, and only when there was a clear domination of one value over the others we have in TDB, and only when it fit the pattern I mentioned of values in CreatureMovementInfo being sorted.
creature_template speed_walk/run don't come from cache. They're filled manually, hopefully using sniffs as a basis, they will never be 0. But it's not safe from errors. What if the creature is first sniffed when it's in combat with someone, and has a bunch of slowing debuffs applied? If WPP would spew out .sql with the speeds of that instance of the creature - it would be all wrong. Or do you mean this is how you envision it should be implemented in TC? Basically treating speed_walk and speed_run columns as overrides? In this case you'll essentially be implementing it as it existed prior to WotLK, prior to CreatureMovementInfo's creation, when speed overrides were stored on per-Creature basis. But since we now have CreatureMovementInfo now - why not just skip the middle-man and use values directly from it? Again, I cannot stress this enough, movementId comes from cache, so it will be correct, unlike whatever data we currently have in speed_walk/run. What I currently view as the ideal solution:
Then the process of selecting a speed would look sort of like:
But there are still too many unknowns to definitely say if it should work like this or not... |
Beta Was this translation helpful? Give feedback.
-
Nothing, was just suprised by how many uses valid speed rates with that movementId
If you meant those kids with 4/4 models, I doubt they uses auras. Even if they did, how auras can change speed so accurately? From one commonly used speed to another commonly used. Usually if speed is changed by auras, you will notice it literally in seconds if you know which values are commonly used
But why we never used them & how can we treat them as correct if sniffs shows different values? Did you found models with correct values which matches sniffs?
Maybe, but again, how many cases matches sniffs in general?
Yeah, that was the plan I described in my issue. To set speed in creature_template to 0 and pick speed from models directly. Or use speed in creature_template and not from model (one model, two entries, two different speed rate cases). Then parse as much sniffs as possible and set speed in creature_template to 0 if not needed. For example if you drop speed from creature_template, you will be not able to parse them anymore. You can change parser of course. I understand what we can link speed to movementId but I don't understand how then we can change speed all for models. It requires parsing for model and its speed, perhaps?
Ouch. It requires massparsing. Massparsing requires a man who can parse all sniffs. Actually it doesn't look like all sniffs were massparsed at some point(thousands of entries were not updated). There were some small massparsings probably. I wanted to create VerifiedBuildOld in creature_template, migrate all current values to it from VerifiedBuild and start to use that nice field correctly. Because that field currently tells nothing. It doesn't tell all fields were updated because they were not updated but VerifiedBuild was. |
Beta Was this translation helpful? Give feedback.
-
Yes. Look at the first two images in "The Problem" section of my original post. Or remember your example with kids and 6.9444 speed. Or, here's more: These numbers are way too specific to NOT be coming from models. But yeah, there aren't that many cases like that. Most just seem to use 2.5/8 for god knows what reason. 😐
Your kids case does. Different models -> different speeds. This confirms that speeds are set either per-CreatureDisplayInfo or per-CreatureModelData. CreatureMovementInfo override is set per-Creature, but the rest seem to be dependent on displays/models, not creature templates.
Yes. It already generates queries to insert/replace bounding_radius and combat_reach in creature_model_info table instead of creature_template. All that needs to be changed is for it to do the same for speeds.
Yeah, it would require loading and parsing them at least once to assemble all that data. The way I see it, it's that there are 3 layers that can override each other.
This is the only explanation I can come up with. We have full access to models, and partial access to CreatureMovementInfo data with my research. However that middle layer, sadly, remains a gray box for us, and we can only fill it from sniffs. There's already a precedent that they have a lot of data duplication happening in these levels. For example, both CreatureModelData and CreatureDisplayInfo have a field for "SizeClass". Model files contain "bounding sphere radius", and yet it rarely (if ever) matches bounding_radius that we're collecting from sniffs inside our creature_model_info, which attributes it to each CreatureDisplayInfo. Who's to say that speeds cannot be duplicated in 3 places in a similar fashion...
Yeah. What you want is to have a VerifiedBuild for EVERY COLUMN in the table, in order to reliably keep track of what's updated when. Obviously it would be an overkill. Initially that field was called WDBVerified, and it was used only for data that came from cache (its name coming from its namesake .wdb cache files). For all its worth, speeds shouldn't even be in creature_template to begin with, but instead in creature_template_addon, because they don't come from cache. As should more than half of all other columns in that table. Nobody will want to deal with that. |
Beta Was this translation helpful? Give feedback.
-
I think it's possible to have testing fields
Not for every column. It's useless even for every sniffable column. Actually people here wants to split creature_template into multiple tables. So maybe it's the perfect opportunity to split them to include in one table all easily sniffable fields. Excluding stuff like gossip_menu_id. Parser tries to remove it always. So even if whole template is sniffed, gossip_menu_id mostly will be not. I doubt there is a way to update all stats correctly and always but you can at least mark template as 'updated at some point', this is what I wanted. I'm not satisfied to not know was template sniffed or absolutely not |
Beta Was this translation helpful? Give feedback.
-
Current values are human-readable because those are real values(1.28 is just slighly faster than normal run speed, 1.7 is pretty fast, 4 is rocket, 15 is space shuttle) |
Beta Was this translation helpful? Give feedback.
-
|
Ugh I checked them from sniffs |
Beta Was this translation helpful? Give feedback.
-
|
That's fine. CreatureMovementInfo ID 100 has 0 walk, 6 run. So it doesn't override the walk speed, and it gets picked from whatever it would normally get picked if the creature had no movementId specified. 22989's model Lobstrok.mdx has speeds: walk 2.5, run 6.944445
1086's model BogBeast.mdx has speeds: walk 1.666667, run 7.777778
Matches perfectly. |
Beta Was this translation helpful? Give feedback.
-
|
I forgot about 0 |
Beta Was this translation helpful? Give feedback.
-
|
Nice. Sad thing is movementId is used only by 5k creatures (1 / 6) SELECT entry,NAME,modelid1,speed_walk,speed_run,movementid FROM creature_Template WHERE movementid=60 |
Beta Was this translation helpful? Give feedback.
-
|
(87, 4, 4.75 |
Beta Was this translation helpful? Give feedback.
-
|
So what exactly is the deal with the facing stuff there. and from what dbc does that value come from exactly and it is a override value for the TurnRate field in update object? |
Beta Was this translation helpful? Give feedback.
-
|
It's purely for visual fanciness. It affects how fast does the model facing adjust to match the real facing. Usually facing changes are smoothed out over a tiny period of time by the client, and that value allows them to control how fast does that happen. Server doesn't need it, I added it to the table just for completeness sake. Here's an example: https://imgur.com/y6KYk6O
Huh? It's from CreatureMovementInfo.dbc, it's the only column there that's shipped with the client.
No idea how it's implemented, and don't care to find out - it's purely a visual thing for the client. But I doubt it's related to TurnRate, I think that speed only controls the keyboard turning rate. |
Beta Was this translation helpful? Give feedback.
-
|
Is it related to bug when pathing creatures instantly changes orientation when moves to next point? |
Beta Was this translation helpful? Give feedback.
-
|
I'm not aware of such bug. An example would be nice. But I doubt it's related, it's more likely that waypoints have points too close to each other and could use some cleaning up. Unless you're talking about creatures turning smoothly around each point of the path, in which case it's a Cata+ feature. |
Beta Was this translation helpful? Give feedback.
-
|
Any creature, heh |
Beta Was this translation helpful? Give feedback.
-
|
Dunno, it's smooth on my server. The smoothing is very short, takes place over just a few frames. Same for opening interaction with NPCs. Retail might've just increased the duration of the interpolation. |
Beta Was this translation helpful? Give feedback.
-
|
They changed speed at least in 121 |
Beta Was this translation helpful? Give feedback.
-
|
Ok, so that answers the few oddities I depicted on screenshots in the original post. Those that have "Server Speed: 6, Override Speed: 8". I wonder if whoever did that even knew what they did. :D I edited the post and added a comment about the speed change to the SQL. |
Beta Was this translation helpful? Give feedback.
-
|
I think I need to re-check your table(should be fine in general since values are ordered and they just visually looks correct, probably can find unknown values, who knows) |
Beta Was this translation helpful? Give feedback.
-
|
I don't want to tell you what to do. I mean, yeah, putting movementId's speeds into creature's walk_speed/run_speed sounds like a reasonable temporary solution. Those aren't affected by randomized models, so we can disregard that feature. Mounts are still an issue, tho. You would be filling walk_speed/run_speed with creature's own speeds, but if the creature is permanently mounted (c_t_a.mount) - it should be using those of the mount. So maybe exclude mounted creatures for the time being (which is easier said than done, because I imagine there are some cases where the mount is missing from c_t_a, but every existing spawn has a mount in c_a). |
Beta Was this translation helpful? Give feedback.
-
|
I somehow missed this comment. Currently we load mountDisplayId in c_a \ c_t_a but I think Blizz literally mounts creature to mount entry in every case or in multiple cases This way after mounting we temporary change speed of creature and after dismounting we change speed to normal |
Beta Was this translation helpful? Give feedback.
-
|
(64, 1, 1, 0, ''), -> (64, 1, 0, 0, ''), |
Beta Was this translation helpful? Give feedback.
-
|
What were the sources? Please don't look at Soldier of the Frozen Wastes, Unstoppable Abomination, Soul Weaver - their speeds are randomized by a script when they spawn. |
Beta Was this translation helpful? Give feedback.
-
|
You're right |
Beta Was this translation helpful? Give feedback.
-
|
Auras? (76, 3, 3, 0, ''), [6] UpdateType: CreateObject1 ServerToClient: SMSG_AURA_UPDATE_ALL (0x7A25) Length: 35 ConnIdx: 2 Time: 03/16/2012 17:55:48.794 Number: 1026276 What do you think about #25964 (reply in thread)? |
Beta Was this translation helpful? Give feedback.
-
|
They're mounted, aren't they. So they use the speed of the mount. CreatureMovementInfo 140's RunSpeedOverride is 9.7, which matches your sniff. Conclusion: they're using creature 34558 as the mount. As for the walk speed, I suppose it's coming from Valiant's own movementId (76, walk 3, run 3), because the mount's movementId doesn't override walk speed. Although that does put a hamper on the theory that they should be using the mount model's speeds instead. |
Beta Was this translation helpful? Give feedback.
-
|
Lol, that thing is funny |
Beta Was this translation helpful? Give feedback.
-
|
Have you heard something about scaling with level bounding radius + combat reach and linked to entry combat reach and bounding radius for tameable(not yet tamed) creatures? |
Beta Was this translation helpful? Give feedback.
-
|
Nope. I heard about hunter pets increasing scale over time, but I don't know if that affects bounding radius and combat reach. Would be rather unbalanced if it did, in my personal opinion. |
Beta Was this translation helpful? Give feedback.
-
|
I checked hunter pet with the same model and its scale was 1 but bounding radius and combat reach were completely different from values of untamed creature |
Beta Was this translation helpful? Give feedback.
-
|
I checked values of 335 using both current 335 and master DB, sometimes classic+ sniffs (119, 2.7, 7.2, 0, '2.5 7.2 in classic sniffs, don't have 3xx sniff atm, still used only by one creature, why 2.7?'), [2] UpdateType: CreateObject2 Maybe walk 2.5? entry name movementid speed_walk speed_run
------ ---------- ---------- ---------- -----------
33651 VX-001 421 1 1.14286
34108 VX-001 (1) 421 1 1.14286Quite possible was changed at some point or DB value is initially wrong (254, 10, 28, 0, ''), entry name movementid speed_walk speed_run
------ ------ ---------- ---------- -----------
29884 Gymer 254 4 4 entry name movementid speed_walk speed_run
------ -------------------------- ---------- ---------- -----------
161268 Jrokgar 254 1 1.14286
47467 Brotherhood Flying Machine 254 1 1.14286
52597 Wings of Aviana 254 1 1.14286
52621 Wings of Aviana 254 1 1.14286
148117 Grong 254 1 1.14286
155328 Spud 254 1 1.14286
144638 Grong the Revenant 254 4.8 4
147550 Harvester 254 4.8 4
143579 Spud 254 4.8 4
153989 Model W 254 4.8 4
144637 Grong 254 4.8 4
143590 Donk 254 4.8 4
29884 Gymer 254 4 4ServerToClient: SMSG_UPDATE_OBJECT (0x4715) Length: 233 ConnIdx: 2 Time: 06/17/2012 00:50:35.090 Number: 182401 Few new (238, NULL, 22, 0, 'Walk is unk, run speed taken from creature which uses mount with this movementId'), |
Beta Was this translation helpful? Give feedback.
-
|
If only I could remember what I was doing 10 months ago... :D
I don't know, could've been a typo on my part. I checked every DB I have and they all have walk_speed=1 for him. In Classic I only encountered him once and he was running at that time, so I recorded the 7.2 speed. I can see it being typo, since "7.2" was the number I wanted to type, so I probably subconsciously pressed "7" ahead of time. So yeah, I agree with using either 0 or 2.5 there, since there's no way to tell for sure.
Yeeeah, this I can't even explain like the above. No idea what happened there. Yes, if anything, I think it should be 2.5, because the (0, 7, 0) case already exists, and it doesn't seem like they were duplicating records there.
Maybe changed. As you can see, all the ones with speed_walk=4.8 are waaaay newer records, with IDs>100000. In the 2013 sniffs that I have both 29884 and 47467 have (12, 28) speeds. So I guess whether it's 10 or 12 depends on whether you trust in that patch that changed speed_walk to 4. To assemble all that data I was using a combination of 3.3.5 TDB, 7.3.5 TDB, and the DB of the server I was working on, which, I think, initially sourced it from YTDB. So my (10, 28) were most likely based on the data that came from that patch. I'll edit my OP with these corrections. |
Beta Was this translation helpful? Give feedback.
-
|
You also just made me realize that some run speeds could be guessed if they belong to the part of the table that's sorted and if the preceding record and the succeeding record have the same run speeds - so the record in question will also have that same run speed too. Filled a few of those. Not like it makes much difference. |
Beta Was this translation helpful? Give feedback.
-
Unfortunately, no 3xx sniff
I didn't stored even a single example of changed value except the one I added months ago, not enough data in master DB to be sure something was changed |
Beta Was this translation helpful? Give feedback.
-
|
Haha entry name walkspeed runspeed movementid
------ ---------------------------------------- --------- -------- ------------
25189 Riding Horse (Brown) (16 Speed) 0 16 199
25191 Riding Horse (White Stallion) (16 Speed) 0 16 199 |
Beta Was this translation helpful? Give feedback.
-
|
Looks like creatures does not always take run_speed from mount's movementId I didn't stored more examples, will add later This case is really weird, I have only 1 4xx sniff with Garadar Wolf Rider. Speed of Telaari Elekk Rider is 1.38571 in several sniffs entry modelid1 modelid2 modelid3 modelid4 speed_walk speed_run WalkSpeed RunSpeed WalkSpeed RunSpeed movementid NAME
------ -------- -------- -------- -------- ---------- --------- --------- -------- --------- -------- ---------- ---------------------
19069 18509 0 0 0 1.2 1.14286 (NULL) (NULL) (NULL) (NULL) 0 Garadar Riding Wolf
19072 18511 0 0 0 1.2 1.14286 0 1.38571 0 9.7 140 Telaar Elekk Mount entry mount MountCreatureID speed_walk speed_run WalkSpeed RunSpeed WalkSpeed RunSpeed movementid name
------ ------ --------------- ---------- --------- --------- -------- --------- -------- ---------- ---------------------
19068 18509 0 1 1.38571 (NULL) (NULL) (NULL) (NULL) 0 Garadar Wolf Rider
19071 18511 0 1 1.38571 (NULL) (NULL) (NULL) (NULL) 0 Telaari Elekk Rider Defenders and riders uses same models entry modelid1 modelid2 modelid3 modelid4 speed_walk speed_run WalkSpeed RunSpeed WalkSpeed RunSpeed movementid NAME
------ -------- -------- -------- -------- ---------- --------- --------- -------- --------- -------- ---------- --------------------
18489 17971 18506 18507 18508 1 1.14286 (NULL) (NULL) (NULL) (NULL) 0 Garadar Defender
19068 17971 18506 18507 18508 1 1.38571 (NULL) (NULL) (NULL) (NULL) 0 Garadar Wolf Rider Some of Kor'kron Vanguards are obviously mounted to the second mount because there's no other mount with model 14336. Their mount model is randomized but speed does not match. 25194 also is spawned in game as vehicle, with speed from movementId entry modelid1 modelid2 modelid3 modelid4 speed_walk speed_run WalkSpeed RunSpeed WalkSpeed RunSpeed movementid NAME
------ -------- -------- -------- -------- ---------- --------- --------- -------- --------- -------- ---------- ----------------------
14539 14575 0 0 0 1 1.38571 0 1.38571 0 9.7 140 Swift Timber Wolf
25194 14336 14334 14574 14575 1 2.28571 0 2.28571 0 16 199 Kor'kron Riding Wolf entry mount MountCreatureID speed_walk speed_run WalkSpeed RunSpeed WalkSpeed RunSpeed movementid name
------ ------ --------------- ---------- --------- --------- -------- --------- -------- ---------- -------------------
27553 0 0 1 1.38571 (NULL) (NULL) (NULL) (NULL) 0 Kor'kron Vanguard entry guid path_id mount MountCreatureID bytes1 bytes2 emote visibilityDistanceType auras name
------ ------ ------- ------ --------------- ------ ------ ------ ---------------------- ------ -------------------
27553 111212 0 14574 0 0 1 0 0 (NULL) Kor'kron Vanguard
27553 111218 1112180 14574 0 0 1 0 0 Kor'kron Vanguard [13] Walk Speed: 2.5 [14] Walk Speed: 2.5 [9] Walk Speed: 2.5 [13] UpdateType: CreateObject1
[13] GUID: Full: 0xF1306BA100002320 Type: Creature Entry: 27553 Low: 8992
[13] Object Type: 3 (Unit)
[13] Has GameObject Rotation: False
[13] Has AnimKits: False
[13] Has Attacking Target: False
[13] Self: False
[13] Has Vehicle Data: False
[13] Living: True
[13] Unknown array size: 0
[13] Has GameObject Position: False
[13] Has Stationary Position: False
[13] Transport: False
[13] Lacks orientation: False
[13] Movement Flags: 1 (Forward)
[13] Has MovementInfo spline: True
[13] Lacks pitch: True
[13] Has Spline Data: True
[13] Has Fall Data: False
[13] Lacks spline elevation: True
[13] Has Transport Data: False
[13] Lacks timestamp: False
[13] Has extended spline data: True
[13] Spline Mode: 0 (Linear)
[13] Has spline start time: False
[13] Spline Waypoints: 6
[13] Has spline vertical acceleration: False
[13] Spline type: Normal
[13] Spline flags: 3145728 (Unknown5, Walkmode)
[13] RunBack Speed: 4.5
[13] SwimBack Speed: 2.5
[13] Spline Time: 20
[13] [0] Spline Waypoint: X: 4915.338 Y: 1279.6483 Z: 225.59161
[13] [1] Spline Waypoint: X: 4915.9707 Y: 1278.874 Z: 225.59161
[13] [2] Spline Waypoint: X: 4920.2617 Y: 1273.623 Z: 225.23846
[13] [3] Spline Waypoint: X: 4921.3716 Y: 1272.2559 Z: 225.22162
[13] [4] Spline Waypoint: X: 4932.0703 Y: 1265.8212 Z: 225.39421
[13] [5] Spline Waypoint: X: 4932.0703 Y: 1265.8212 Z: 225.39421
[13] Spline Duration Multiplier Next: 1
[13] Spline Full Time: 2171
[13] Spline Duration Multiplier: 1
[13] Spline Id: 76443649
[13] Spline Endpoint:: X: 4932.0703 Y: 1265.8212 Z: 225.39421
[13] Pitch Speed: 3.141592741012573242
[13] Swim Speed: 4.72221994400024414
[13] Walk Speed: 2.5
[13] Time: 3362060837
[13] FlyBack Speed: 3.141592741012573242
[13] Turn Speed: 7
[13] Run Speed: 9.699999809265136718
[13] Fly Speed: 4.5
GUID 2: Full: 0xF1306BA100002320 Type: Creature Entry: 27553 Low: 8992
[13] Position: X: 4915.9707 Y: 1278.874 Z: 225.59161
[13] Orientation: 5.3975
[13] OBJECT_FIELD_GUID: Full: 0xF1306BA100002320 Type: Creature Entry: 27553 Low: 8992
[13] OBJECT_FIELD_TYPE: 9
[13] OBJECT_FIELD_ENTRY: 27553
[13] OBJECT_FIELD_SCALE_X: 1
[13] UNIT_FIELD_BYTES_0: 0/1/0/1
[13] UNIT_FIELD_HEALTH: 42540
[13] UNIT_FIELD_MAXHEALTH: 42540/5.9611E-41
[13] UNIT_FIELD_LEVEL: 75/1.05E-43
[13] UNIT_FIELD_FACTIONTEMPLATE: 2054/2.878E-42 (2054)
[13] UNIT_VIRTUAL_ITEM_SLOT_ID: 34682/4.86E-41
[13] UNIT_VIRTUAL_ITEM_SLOT_ID + 1: 12452/1.7449E-41
[13] UNIT_FIELD_FLAGS: 32768
[13] UNIT_FIELD_FLAGS_2: 2048
[13] UNIT_FIELD_BASEATTACKTIME: 2000/2.803E-42
[13] UNIT_FIELD_BASEATTACKTIME + 1: 2000/2.803E-42
[13] UNIT_FIELD_BOUNDINGRADIUS: 0.3906
[13] UNIT_FIELD_COMBATREACH: 1.5749999
[13] UNIT_FIELD_DISPLAYID: 23864/3.344E-41
[13] UNIT_FIELD_NATIVEDISPLAYID: 23864/3.344E-41
[13] UNIT_FIELD_MOUNTDISPLAYID: 14336/2.0089E-41
[13] UNIT_FIELD_BYTES_1: 0/0/1/0
[13] UNIT_MOD_CAST_SPEED: 1065353216/1
[13] UNIT_MOD_CAST_HASTE: 1065353216/1
[13] UNIT_FIELD_BYTES_2: 1/1/0/0
[13] UNIT_FIELD_HOVERHEIGHT: 1
[14] UpdateType: CreateObject1
[14] GUID: Full: 0xF1306BA100002321 Type: Creature Entry: 27553 Low: 8993
[14] Object Type: 3 (Unit)
[14] Has GameObject Rotation: False
[14] Has AnimKits: False
[14] Has Attacking Target: False
[14] Self: False
[14] Has Vehicle Data: False
[14] Living: True
[14] Unknown array size: 0
[14] Has GameObject Position: False
[14] Has Stationary Position: False
[14] Transport: False
[14] Lacks orientation: False
[14] Movement Flags: 1 (Forward)
[14] Has MovementInfo spline: True
[14] Lacks pitch: True
[14] Has Spline Data: True
[14] Has Fall Data: False
[14] Lacks spline elevation: True
[14] Has Transport Data: False
[14] Lacks timestamp: False
[14] Has extended spline data: True
[14] Spline Mode: 0 (Linear)
[14] Has spline start time: False
[14] Spline Waypoints: 4
[14] Has spline vertical acceleration: False
[14] Spline type: Normal
[14] Spline flags: 3145728 (Unknown5, Walkmode)
[14] RunBack Speed: 4.5
[14] SwimBack Speed: 2.5
[14] Spline Time: 20
[14] [0] Spline Waypoint: X: 4919.812 Y: 1282.776 Z: 226.48846
[14] [1] Spline Waypoint: X: 4920.373 Y: 1281.9482 Z: 226.48846
[14] [2] Spline Waypoint: X: 4930.3213 Y: 1272.7078 Z: 226.13947
[14] [3] Spline Waypoint: X: 4930.3213 Y: 1272.7078 Z: 226.13947
[14] Spline Duration Multiplier Next: 1
[14] Spline Full Time: 1650
[14] Spline Duration Multiplier: 1
[14] Spline Id: 76443650
[14] Spline Endpoint:: X: 4930.3213 Y: 1272.7078 Z: 226.13947
[14] Pitch Speed: 3.141592741012573242
[14] Swim Speed: 4.72221994400024414
[14] Walk Speed: 2.5
[14] Time: 3362060837
[14] FlyBack Speed: 3.141592741012573242
[14] Turn Speed: 7
[14] Run Speed: 9.699999809265136718
[14] Fly Speed: 4.5
GUID 2: Full: 0xF1306BA100002321 Type: Creature Entry: 27553 Low: 8993
[14] Position: X: 4920.3735 Y: 1281.9485 Z: 226.51437
[14] Orientation: 5.3079786
[14] OBJECT_FIELD_GUID: Full: 0xF1306BA100002321 Type: Creature Entry: 27553 Low: 8993
[14] OBJECT_FIELD_TYPE: 9
[14] OBJECT_FIELD_ENTRY: 27553
[14] OBJECT_FIELD_SCALE_X: 1
[14] UNIT_FIELD_BYTES_0: 0/1/0/1
[14] UNIT_FIELD_HEALTH: 42540
[14] UNIT_FIELD_MAXHEALTH: 42540/5.9611E-41
[14] UNIT_FIELD_LEVEL: 75/1.05E-43
[14] UNIT_FIELD_FACTIONTEMPLATE: 2054/2.878E-42 (2054)
[14] UNIT_VIRTUAL_ITEM_SLOT_ID: 30181/4.2293E-41
[14] UNIT_VIRTUAL_ITEM_SLOT_ID + 1: 12452/1.7449E-41
[14] UNIT_FIELD_FLAGS: 32768
[14] UNIT_FIELD_FLAGS_2: 2048
[14] UNIT_FIELD_BASEATTACKTIME: 2000/2.803E-42
[14] UNIT_FIELD_BASEATTACKTIME + 1: 2000/2.803E-42
[14] UNIT_FIELD_BOUNDINGRADIUS: 0.3906
[14] UNIT_FIELD_COMBATREACH: 1.5749999
[14] UNIT_FIELD_DISPLAYID: 23864/3.344E-41
[14] UNIT_FIELD_NATIVEDISPLAYID: 23864/3.344E-41
[14] UNIT_FIELD_MOUNTDISPLAYID: 14575/2.0424E-41
[14] UNIT_FIELD_BYTES_1: 0/0/1/0
[14] UNIT_MOD_CAST_SPEED: 1065353216/1
[14] UNIT_MOD_CAST_HASTE: 1065353216/1
[14] UNIT_FIELD_BYTES_2: 1/1/0/0
[14] UNIT_FIELD_HOVERHEIGHT: 1
[9] UpdateType: CreateObject1
[9] GUID: Full: 0xF1306BA10043EEBE Type: Creature Entry: 27553 Low: 4452030
[9] Object Type: 3 (Unit)
[9] Has GameObject Rotation: False
[9] Has AnimKits: False
[9] Has Attacking Target: False
[9] Self: False
[9] Has Vehicle Data: False
[9] Living: True
[9] Unknown array size: 0
[9] Has GameObject Position: False
[9] Has Stationary Position: False
[9] Transport: False
[9] Lacks orientation: False
[9] Movement Flags: 1 (Forward)
[9] Has MovementInfo spline: True
[9] Lacks pitch: True
[9] Has Spline Data: True
[9] Has Fall Data: False
[9] Lacks spline elevation: True
[9] Has Transport Data: False
[9] Lacks timestamp: False
[9] Has extended spline data: True
[9] Spline Mode: 0 (Linear)
[9] Has spline start time: False
[9] Spline Waypoints: 4
[9] Has spline vertical acceleration: False
[9] Spline type: Normal
[9] Spline flags: 3145728 (Unknown5, Walkmode)
[9] RunBack Speed: 4.5
[9] SwimBack Speed: 2.5
[9] Spline Time: 824
[9] [0] Spline Waypoint: X: 4937.2065 Y: 1258.5166 Z: 225.79083
[9] [1] Spline Waypoint: X: 4937.134 Y: 1259.5139 Z: 225.79083
[9] [2] Spline Waypoint: X: 4938.4463 Y: 1241.5406 Z: 226.04083
[9] [3] Spline Waypoint: X: 4938.4463 Y: 1241.5406 Z: 226.04083
[9] Spline Duration Multiplier Next: 1
[9] Spline Full Time: 1859
[9] Spline Duration Multiplier: 1
[9] Spline Id: 213233153
[9] Spline Endpoint:: X: 4938.4463 Y: 1241.5406 Z: 226.04083
[9] Pitch Speed: 3.141592741012573242
[9] Swim Speed: 4.72221994400024414
[9] Walk Speed: 2.5
[9] Time: 4216388466
[9] FlyBack Speed: 3.141592741012573242
[9] Turn Speed: 7
[9] Run Speed: 9.699999809265136718
[9] Fly Speed: 4.5
GUID 2: Full: 0xF1306BA10043EEBE Type: Creature Entry: 27553 Low: 4452030
[9] Position: X: 4937.6973 Y: 1251.7987 Z: 225.89815
[9] Orientation: 4.785285
[9] OBJECT_FIELD_GUID: Full: 0xF1306BA10043EEBE Type: Creature Entry: 27553 Low: 4452030
[9] OBJECT_FIELD_TYPE: 9
[9] OBJECT_FIELD_ENTRY: 27553
[9] OBJECT_FIELD_SCALE_X: 1
[9] UNIT_FIELD_BYTES_0: 0/1/0/1
[9] UNIT_FIELD_HEALTH: 39760
[9] UNIT_FIELD_MAXHEALTH: 39760/5.5716E-41
[9] UNIT_FIELD_LEVEL: 73/1.02E-43
[9] UNIT_FIELD_FACTIONTEMPLATE: 2054/2.878E-42 (2054)
[9] UNIT_VIRTUAL_ITEM_SLOT_ID: 30181/4.2293E-41
[9] UNIT_VIRTUAL_ITEM_SLOT_ID + 1: 12452/1.7449E-41
[9] UNIT_FIELD_FLAGS: 32768
[9] UNIT_FIELD_FLAGS_2: 2048
[9] UNIT_FIELD_BASEATTACKTIME: 2000/2.803E-42
[9] UNIT_FIELD_BASEATTACKTIME + 1: 2000/2.803E-42
[9] UNIT_FIELD_BOUNDINGRADIUS: 0.3906
[9] UNIT_FIELD_COMBATREACH: 1.5749999
[9] UNIT_FIELD_DISPLAYID: 23867/3.3445E-41
[9] UNIT_FIELD_NATIVEDISPLAYID: 23867/3.3445E-41
[9] UNIT_FIELD_MOUNTDISPLAYID: 14574/2.0423E-41
[9] UNIT_FIELD_BYTES_1: 0/0/1/0
[9] UNIT_MOD_CAST_SPEED: 1065353216/1
[9] UNIT_MOD_CAST_HASTE: 1065353216/1
[9] UNIT_FIELD_BYTES_2: 1/1/0/0
[9] UNIT_FIELD_HOVERHEIGHT: 1 |
Beta Was this translation helpful? Give feedback.
-
|
I forgot about one thing... in classic they simply added movementId to all creatures In TBC they added more entries but I don't have wdb data to confirm if movementId again is used by all creatures, probably yes |
Beta Was this translation helpful? Give feedback.
-
|
Heh, maybe because their archived classic DB didn't have CreatureMovementInfo assignments (because it didn't exist back then) so they just massconverted all the unique speed combinations to CreatureMovementInfo AGAIN, instead of using WotLK data for that. |
Beta Was this translation helpful? Give feedback.
-
|
In Classic Vanilla 1.13, the movement ids used in creature template go from 1693 to 1778. In Classic TBC 2.5.1, they go from 1693 to 2074. There are no creatures with 0 as the movement id. |
Beta Was this translation helpful? Give feedback.
-
|
Hope he's still somewhere around & can share it |
Beta Was this translation helpful? Give feedback.
-
|
He told to share it if it can help improving TC or something like that but I think that way it will never be used by anyone and anywhere |
Beta Was this translation helpful? Give feedback.
-
|
10.1.7 added WalkSpeed and RunSpeed to CreatureModelData.db2 |
Beta Was this translation helpful? Give feedback.
-
|
wait that's good news, does this mean we can finally implement correct walk/run speeds attached to MovementID? |
Beta Was this translation helpful? Give feedback.
-
I've been sitting on this info for a long while, but wasn't posting it because it wasn't handling 100% of the cases. Over time I made some other discoveries, which brought it closer to that goal, but still not quite there. So take this with a small grain of salt.
CreatureMovementInfo
This DBC appeared in WotLK. If you were to look at it, you would discover that it has hundreds of rows with the same single value of 0. This is a clear sign that some columns of this table are "serverside" - hidden from the DBCs because the client doesn't need them, but nevertheless present and used by the server. The only value that's in there for the client is the rate of smooth facing changes, added mainly to let the client draw the creature's model sloooooowly turning around instead of changing the orientation almost instantaneously, which looks especially weird with large enemies (Flame Leviathan from Ulduar, for example).
So what columns are we missing? The answer can yet again be gleamed from the few WowEdit screenshots we have:
Let me bring your attention to the "Movement Override" groupbox. You can see that the "Walk" and "Run" textboxes are readonly there. That means you can't edit them directly, their data comes from elsewhere, and is there just for display, just like with "Aggro Radius" groupbox directly above it - its values come from the Radius (CreatureRadius?) record that's selected from the dropdown combobox, and the textboxes are there just for convenient display, but if you want to actually edit them - you press the "Copy" or "Edit" buttons.
So where does the data for "Walk" and "Run" "overrides" come from? From the CreatureMovementInfo record, which is selectable by clicking the button directly above them.
So, the two columns of data that we're missing in CreatureMovementInfo are the walk and run speed overrides.
But what does this give us, you might ask, this data is serverside and we don't have access to it. Which is true... But we can extrapolate it.
Filling CreatureMovementInfo
So here's what I did. I went through every CreatureMovementInfo record, I selected the rows from our `creature_template` that had that that CreatureMovementInfo ID in `movementId`. I filtered out all the rows that had blatantly default values (2.5 walk, 8 run, aka speed_walk=1, speed_run=1.14286), which most likely means they were never filled from sniffs. And then I looked at the speed_walk and speed_run values in the remaining rows, and almost always there were a few non-default values that were dominating the results.
Sometimes there was only one value dominating - either the walk or the run speed, and the other one was varying. But it's easily explainable: the values in the screenshot are "overrides", so they were using the value of 0 to specify that that speed (either walk or run) should not be overridden.
So I started writing them down. And before long, a pattern emerged:
Do you see what's happening here? They're SORTED. First by run speed, then by walk speed, in ascending order.
How did this happen? The most likely explainable would be that until WotLK they had the ability to specify these speed overrides on every creature separately. With WotLK (or whenever they decided to create that CreatureMovementInfo table), they wanted to organize it a bit better, and also add the turn rate modifier there (which got bundled with the client), so they took all the existing data, grouped it up, sorted it by run speed, then by walk speed, and generated the contents for CreatureMovementInfo table, and every Creature record got the respective MovementInfo assigned via ID instead of raw speed values.
So I kept going. Knowing that there's a pattern considerably sped up the process.
And so, without further ado, here's the data:
`SmoothFacingChaseRate` column is filled directly from .dbc/.db2. This table is meant as a direct replacement to that DBC.
The pattern of rows being ordered only holds up until `SmoothFacingChaseRate` starts changing. This is to be expected, the aforementioned mass-conversion happened before this new column saw any use by designers. All the new rows must either be combinations of walk/run speeds that were never encountered before, or already existing walk/run speeds but now with a non-default `SmoothFacingChaseRate`. Or, you know, it's also possible that someone could've created a duplicate row if they failed to find the one they wanted in the list.
NULLmeans I couldn't definitively source that value from anywhere. The new records that got added after the mass-conversion are sparsely used, so there isn't much examples to base the data upon. And those creatures that do happen to use them often don't have the speeds sniffed. So NULLs get more and more prevalent the further it goes. I added comments to the table to describe why the certain values are missing.However the vast majority of creatures use the early rows, for which the data is abundant.
WotLK uses only rows up until and including ID 821.
Why is this info valuable?
For one simple reason: `creature_template`.`movementId` comes from cache. Therefore, unless someone manually tampered with it, it's 100% correct. And so with this data we can correct the speeds of thousands of creatures at once.
So how is a creature's speed determined?
(auras are completely neglected here, this thread is solely about base speeds)
It's based on 3 factors:
1. Whether the creature is mounted
Believe it or not - when a creature mounts up - its speed changes to that of the mount. Just like that, without any auras. Examine sniffs to find this out for yourself.
Mount's speed is determined in the same way. If the mount's `creature_template` record has `movementId` - it uses the speed from CreatureMovementInfo, unless it's 0 ("don't override"). Otherwise - the (sort of) model speed is used.
Hold up. Did I say "mount's creature_template record"? What? Don't we use CreatureDisplayInfo for mounts?
Yup, my dear viewer, we are doing it wrong. Mounts aren't just models, mounts are creatures in and of itself. They're already used as such for mounting auras, which reference the Creature record via MiscValue. The rest of the mounted creatures in the game should do the same. Here, in their scripting system, you can see that the parameter for "Unit(s) mount a creature" is a "Creature" record.
This means that, yes, all the data we have for creature mounts in `*_addon*` tables is essentially wrong, and needs to be switched from CreatureDisplayInfo ID to the appropriate mount's `creature_template`.`entry`. This is further complicated by the fact that there can be multiple Creature records using the same mount model, and they can have different CreatureMovementInfo assigned, and each would need to be examined separately to determine which one should be chosen for the speeds to match.
This also means that if a mount has multiple models - a random one should be chosen for the creature, much like the way mounting auras work.
2. Whether the creature has a CreatureMovementInfo record assigned
This is fairly straightforward. If the creature is not mounted - it uses its own `creature_template`.`movementId`. If that CreatureMovementInfo's speed override is 0 - this step is skipped. If it's not 0 - that's the speed the creature will have.
3. Which model is used by the creature (sort of)
And this is where the main difficulty comes. And the problem I described in the intro. That "(sort of)" part. But we'll get to it later.
For now, let's figure out how the fuck does the speed depend on the model.
Here's the thing... Models have movement speeds embedded in them.
Models have animations. Each animation has a field that contains the movement speed it's supposed to be played at. I.e. the speed the animation was created for. If the creature's real speed is higher or lower - it needs to be appropriately slowed down or sped up.
So, what we can do, is load the model, find the animations sequences corresponding to "walk" and "run" animations (or "fly walk", "fly run", or maybe there are more. Or maybe they even depend on animation tier. I haven't tried digging that deep), and look up their movement speeds. Loading the models every time is expensive, so it would be easier to load this data once and then fill `creature_model_info` with that data in the new columns.
And... this works. In some cases.
The problem
This. This is the main hurdle that I failed to overcome. I could not figure out why do SOME sniffed creatures match the speed of their model, and SOME sniffed creatures - did not.
Those that did not - most often had the default 2.5 walk, 8 run speeds. As if it was forced to be default somehow. But I just don't know how.
But those that did - matched the model speeds exactly. And those speeds are often obscure ones, like 2.2222 or 6.94444. It's not the kind of speed you would enter in a textbox for no reason. But sometimes only one of the two speeds (walk and run) would match:
And sometimes it would be all over the place, like this example, where the CreatureMovementInfo override doesn't match at all, but the model's walk speed matches, but run speed doesn't:
Here's an example of a mounted creature (the last column is meant to be named "Mount Override Speed:" and show the mount's CreatureMovementInfo's values, but for that I need to select a Creature from which to take that info from, so it became a combobox):
Nothing matches the run speed, right? But we have several options for the mount's Creature record:
If we were to pick Naxxramas Deathcharger, then the run speed would match, but the walk speed wouldn't:
And here's another mounted example from the same Naxxramas, but here the RIDER's CreatureMovementInfo override matches perfectly, but the mount's one doesn't (for walk):
So it's basically a mess. It does work for most cases, tho:
But then sometimes it randomly doesn't:
Outdated
Outdated
Outdated
Perhaps the application I wrote for this doesn't consider something, like the auras being applied in the middle of a fight. Needs laborious manual checking.
Or perhaps the problem is in that I only have MoP+ sniffs, but I'm testing it against WotLK DBCs and creature_templates. I wish I had WotLK sniffs.
Epilogue
For now, this is all I have. I'm at an impasse. At first glance it looks like it should work perfectly, but as you start examining it in details - all kinds of imperfections start sticking out. But where it does work - it works beautifully. As that quote goes: "60% of the time, it works every time"...
Maybe you can test it better, given that you have WotLK sniffs. I only have a bunch of data grabbed with Lua's GetUnitSpeed from Classic, which needs a lot of cleanup to be useful (while run/flight/swim speeds were nice and tidy, "current" speed was all over the place, as it's calculated based on spline's distance divided by duration).
Even if my theories about how final movement speed is determined are ultimately incorrect, you can still extract a lot of usefulness from my CreatureMovementInfo data research.
P.S. Please, for the love of all that is holy, stop putting speed multipliers into `creature_template` and just store the data there in yards/sec. It looks SO MUCH BETTER. "8" is a much more sensible number than "1.14286", "30" than "4.28571", etc. You can still multiply then by speed rates from config, nothing inhibits that.
Ref #25814
Beta Was this translation helpful? Give feedback.
All reactions