Add LOD management functions #3831

Fernando-A-Rocha · 2024-10-29T15:10:36Z

Important

HLOD = High Level of Detail (normal models the player sees when near)
LLOD = Low Level of Detail (models with less detail the player sees when far)

Details

Adds new clientside model utility & Level-Of-Detail(LOD) management functions

WIP

To be used in conjunction with:

This solves the problem of having to include a Lua LOD_TABLE in every map's script, which has been happening for years (the map editor's LOD table happens to be incomplete as of submitting this PR), providing a simple solution to obtain an object's LLOD model ID in the code.

Related to multitheftauto/mtasa-resources#556

Examples

LLOD model of HLOD model ganghous03_LAx (3655) is lodganghous03_lax (3656).
Practical code examples:
WIP

Client/mods/deathmatch/logic/lua/CLuaFunctionDefs.Util.cpp

Client/mods/deathmatch/logic/lua/CLuaFunctionDefs.h

Client/mods/deathmatch/logic/lua/CLuaManager.cpp

Server/mods/deathmatch/logic/luadefs/CLuaFunctionDefs.Server.cpp

Server/mods/deathmatch/logic/luadefs/CLuaFunctionDefs.cpp

Server/mods/deathmatch/logic/luadefs/CLuaFunctionDefs.h

Shared/mods/deathmatch/logic/CLodModels.cpp

Shared/mods/deathmatch/logic/CLodModels.h

FileEX · 2024-10-29T16:18:45Z

The lod arrays from the map editor are incomplete, so I assume this function is also incomplete. Therefore, a note should be added on the wiki once this PR is merged

Fernando-A-Rocha · 2024-10-29T16:22:56Z

The lod arrays from the map editor are incomplete, so I assume this function is also incomplete. Therefore, a note should be added on the wiki once this PR is merged

No. I completed the table. It's good now.
See the mtasa-resources issue I linked where I explain it.

Client/mods/deathmatch/logic/lua/CLuaFunctionDefs.h

Client/mods/deathmatch/logic/lua/CLuaManager.cpp

Client/mods/deathmatch/logic/luadefs/CLuaObjectDefs.cpp

Server/mods/deathmatch/logic/luadefs/CLuaObjectDefs.cpp

Server/mods/deathmatch/logic/luadefs/CLuaObjectDefs.h

Shared/mods/deathmatch/logic/CLodModels.h

Shared/mods/deathmatch/logic/CLodModels.cpp

Shared/mods/deathmatch/logic/CLodModels.h

Client/mods/deathmatch/logic/luadefs/CLuaObjectDefs.cpp

Client/mods/deathmatch/logic/luadefs/CLuaObjectDefs.h

Server/mods/deathmatch/logic/luadefs/CLuaObjectDefs.cpp

Server/mods/deathmatch/logic/luadefs/CLuaObjectDefs.h

Shared/mods/deathmatch/logic/CLodModels.h

Shared/mods/deathmatch/logic/CLodModels.cpp

lotsofs · 2024-11-01T01:55:15Z

The provided LOD model list not only is incomplete, it also contains plain incorrections. Examples are the LLODs for HLOD models 10428, 10369, 10439.

Fernando-A-Rocha · 2024-11-01T18:14:23Z

The provided LOD model list not only is incomplete, it also contains plain incorrections. Examples are the LLODs for HLOD models 10428, 10369, 10439.

I'm surprised, I thought the list was generated from the game files not missing anything.

Will you suggest a fix for the array?

derxgbb · 2024-11-01T20:29:35Z

I'm surprised, I thought the list was generated from the game files not missing anything.

Will you suggest a fix for the array?

Just a note:
Even the game .ipl files are different in other ports (ps2, xbox etc). For example the lanterns and special objects around four dragon casino is only persistent on ps2/xbox. https://imgur.com/MV61hvH
The objects does exist but never used in game.
I don't know if .ide files are different, they shouldn't but who knows.

Fernando-A-Rocha · 2024-11-02T11:38:30Z

I'm gonna investigate the HLOD => LLOD array. @lotsofs @spooky-spook

I'm surprised, I thought the list was generated from the game files not missing anything.
Will you suggest a fix for the array?

Just a note: Even the game .ipl files are different in other ports (ps2, xbox etc). For example the lanterns and special objects around four dragon casino is only persistent on ps2/xbox. https://imgur.com/MV61hvH The objects does exist but never used in game. I don't know if .ide files are different, they shouldn't but who knows.

It's interesting that PS2 and XBOX versions of the games have different world mapping, but I think this is out of the scope of this LLOD-improvements PR.

https://tcrf.net/Grand_Theft_Auto:_San_Andreas/Version_and_Platform_Differences
There are a large number of objects in Las Venturas missing from the PC version, mainly outside the Four Dragons Casino and The Visage. This seems to be a simple omission from the map as they are still in the game.

Fernando-A-Rocha · 2024-11-26T02:28:03Z

Check it out so far, ive redone some functions. Moved things to model related files. Added isValidModel (general purpose func)

@TheNormalnij I added model id checks, and also made it reset in engineFreeModel and when model manager is unloaded. Is this ok?

TracerDS · 2024-12-03T06:31:11Z

Client/mods/deathmatch/logic/luadefs/CLuaModelDefs.cpp

+        CLuaCFunctions::AddFunction(name, func);
+}
+
+bool CLuaModelDefs::IsValidModel(std::string modelPurpose, std::uint32_t id)


what about an enum instead of string?

TracerDS · 2024-12-03T06:31:50Z

Client/mods/deathmatch/logic/luadefs/CLuaModelDefs.cpp

+
+bool CLuaModelDefs::SetModelLOD(std::string modelPurpose, std::uint32_t hLODModel, std::uint32_t lLODModel)
+{
+    if (!(modelPurpose == "object" || modelPurpose == "building"))


if(modelPurpose != "object" && modelPurpose != "building")

TracerDS · 2024-12-03T06:33:15Z

Shared/mods/deathmatch/logic/CLodModels.cpp

+
+constexpr std::size_t OBJ_LOD_MODELS_COUNT = 4289;
+constexpr std::pair<std::uint32_t, std::uint32_t> OBJ_LOD_MODELS_ARRAY[] = {
+    {694, 784},                // sm_redwoodgrp => lod_redwoodgrp (countryS)


maybe too much spacing before the comment?
{X, Y}, // comment
instead of
{X, Y}, // comment

TracerDS · 2024-12-03T06:34:43Z

Shared/mods/deathmatch/logic/CLodModels.cpp

+    {13804, 13808},            // cuntelandf4 => lodcuntelandf4 (LAhills)
+    {13805, 13807},            // celalandbiv => lodcelalandbiv (LAhills)
+    {13809, 13811},            // ce_grndpalcst02 => lodcepalcst02 (LAhills)
+    {13810, 13812},            // ce_grndpalcst05 => lodce_grndpalcst05 (LAhills)
+    {13813, 13815},            // cegraveblok03e => lodcegraveblok03e (LAhills)
+    {13814, 13787},            // ceroadn => lodceroadn (LAhills)
+    {13816, 13877},            // ce_safeground => ce_s


do we really need that?

lynconsix · 2024-12-14T16:09:06Z

Any other news? This PR is interesting!

ds1-e · 2024-12-14T20:13:12Z

These functions should be shared, like setLowLODElement is. e.g: if you would like to create object on server-side, and get it's LOD model directly.

lynconsix · 2024-12-24T19:41:56Z

What is missing for this PR to be completed?

Nico8340 · 2024-12-26T00:37:19Z

Any updates on this? @Fernando-A-Rocha

Fernando-A-Rocha · 2025-01-01T11:53:08Z

I'll get back to this PR soon.

There are several ways to approach the problem of LODs. We certainly need something that is both clientside and serverside due to createObject, createBuilding (soon available serverside) and setLowLodElement being shared functions...

lotsofs · 2025-01-06T19:19:10Z

I am back with my manually built list of HLOD vs LLODs, using the removeWorldObject feature from the editor on every object in the map, and where that failed finding a manual match using the IDE list on the wiki or similar.
MTASA Maps - Sheet17.csv

Here are my conclusions, and more importantly, things that I think should be changed in the big table in CLodModels.cpp:

Missing entries
These are all objects that gave me trouble deleting them in the editor, their LLOD would still show after deletion, and I had to manually find their LLODs in the IDE list. They are also missing from the table in CLodModels.cpp

    {10289, 10156},            // tempsf_3_sfe => LODA27 (???)
    {10561, 10681},            // bbgroundbitc_SFS => LODroundbitc_SFS (???)
    {10933, 11275},            // traintrax01_SFS => LOD_trax_SFSe07 (???)
    {10973, 11025},            // mall_03_SFS => LOD_sfs014 (???)
    {11228, 11277},            // traintrax01b_SFS => LOD_trax_SFSe08 (???)
    {11231, 11273},            // traintrax03b_SFS => LOD_trax_SFSe01 (???)
    {11232, 11274},            // traintrax03c_SFS => LOD_trax_SFSe06 (???)
    {11409, 11347},            // roadsSFSE34 => LODroadsSFSE34 (???)
    {17614, 17843},            // Lae2_landHUB02 => LODLae2_lndHUB02 (???)
    {4826, 4987},              // grifftop2 => LODfftop03 (???)
    {7498, 7843},              // vegaswrailroad01 => lodaswrailroad01 (???)
    {7582, 7951},              // miragebuild04 => lodagebuild04 (???)
    {7625, 7957},              // vgnhseing129 => lodhseing129 (???)
    {7878, 7879},              // vegasNroad242 => hubst4alpha (???)
    {8386, 8387},              // vgsSspagjun09b => lodSspagjun09b (???)
    {8388, 8389},              // vegasSedge29b => lodasSedge29b (???)
    {9310, 9390},              // chapel_SFN => LODpel_SFN (???)
    {9889, 9890},              // park3a_sfw => lod_park3a_sfw (???)
    {9239, 9405},              // track01_SFN => LODck01_SFN (???)
    {9240, 9404},              // track02_SFN => LODck02_SFN (???)

HLODs using the same LLOD
These objects all use LLOD 17749, in addition to the already included HLOD 17559:

    {17556, 17749},            // MStorCP1_LAe2 => LOD1MStorCP_LAe (???)
    {17557, 17749},            // MStorCP2_LAe2 => LOD1MStorCP_LAe (???)
    {17558, 17749},            // MStorCP4_LAe2 => LOD1MStorCP_LAe (???)

multiple LLODs belonging to the same HLOD
These objects all appear the same HLOD, but have different LLODs for some reason:
Inclusion recommended if it is desired that a user can request a HLOD belonging to a specific LLOD.
(not including entry oilderricklod01 that is already in the table)

    {3427, 16274},             // derrick01 => oilderricklod02 (???)
    {3427, 16275},             // derrick01 => oilderricklod03 (???)
    {3427, 16276},             // derrick01 => oilderricklod04 (???)
    {3427, 16277},             // derrick01 => oilderricklod05 (???)
    {3427, 16278},             // derrick01 => oilderricklod06 (???)
    {3427, 16279},             // derrick01 => oilderricklod07 (???)

Furthermore there is this entry: but it has other issues (read below)

    {3724, 5352},              // laxrf_cargotop => LODcargoshp03f (???)

LLODs that cover multiple HLODs
Existing entry:
{7243, 6996}, // vgncircus1b => lodcircus1 (vegasN)
is correct, but LLOD 6996 covers both HLODs 6994 and 7243 combined. The former is an entire casino building, and the latter is only the hedges surrounding its parking lot. A second opinion might be needed on what to do with this, but I'd heavily recommend changing the line to
{6994, 6996}, // vgncircus1b => lodcircus1 (vegasN)
as it makes much more sense after a visual inspectation.

Existing entry:
{9504, 9634}, // sboxbld4_sfw73 => lodxbld4_sfw73 (SFw)
is correct, but LLOD 9634 covers both HLODs 9504 and 9505 combined. The former is a big hill with houses on top, the latter is only the grass textures of the tiny gardens belonging to these houses. The line is fine as it is, but in the interest of completeness I thought it worth mentioning here.

Incorrect entries
Existing entry:
{3412, 3413}, // cunterb03 => lodcunterb07 (countrye)
is incorrect. HLOD 3412 is only the orange upper parts of a bridge, whereas HLOD 3411 is the entire bridge, including yet even more orange upper parts. LLOD 3413 matches HLOD 3412 much more closely
{3411, 3413}, // cunteRB01 => lodcunterb07 (countrye)

Existing entry:
{5786, 5945}, // shutters01_lawn => lodblok08_lawn (LaWn)
is incorrect. HLOD 5786 comprises just a set of door that go on entire building HLOD 5784 that LLOD 5945 matches much more closely.
{5784, 5945}, // MelBlok08_LAwN => lodblok08_lawn (LaWn)
However, the LLOD seems to be rotated 180 degrees compared to the HLOD 5784. Someone else besides me should look at this and make a judgement call.

Actually upon second thought, these are probably yet another cases of LLODs that cover multiple HLODs, but I already marked them as plainly incorrect and I can't be bothered rewriting.

LLOD to a LLOD
Existing entries:
{7174, 7366}, // lod_corpbuild31 => sham_superlod (vegasN)
{6971, 7174}, // vgn_corpbuild31 => lod_corpbuild31 (vegasN)
7174 is the LLOD for HLOD 6971, with 7366 in turn being a LLOD (SuperLOD) to 7174. Odd case. both cases are already in the list, so no action is needed I guess, but I had this in my notes and maybe someone has opinions.

Oddities that require subjective review
Non exhaustive list of objects that might give issues:

Existing entry
{3724, 5164}, // laxrf_cargotop => lodcargoshp03d (LAs2)
is correct, but the LLOD is rotated 90 degrees compared to its HLOD counterpart. This might give unwanted behavior. This also applies to the LODcargoshp03f I mentioned above, it is also rotated 90 degrees.

I dont have time to write the last entry, so here's my raw notes:
HLOD 9690, LLOD 9686
HLOD 9694, LLOD 9692
This concerns the pillars of Gant bridge. The HLOD 9690 and 9694 are each half a pillar (one the southern, one the northern), placed twice with one rotated 180 degrees so it connects. LLOD 9686 and 9692 is a single model representing an entire pillar. A subjective judgement call is needed for this. Furthermore: Deleting the pillar in the map editor reveals yet another model, which looks somewhat like a skeletal structure of like a character model. Someone should look at what the hell is going on here.

I'd submit my suggested changes myself, but I have no idea how to do a pull request to a pull request.

Fernando-A-Rocha · 2025-01-07T15:17:17Z

Wow thank you for your detailed review @lotsofs
I'll definitely take a deeper look
I've added you to my repository Fernando-A-Rocha/mtasa-blue so you can contribute to this branch directly!! 😃

These are all objects that gave me trouble deleting them in the editor, their LLOD would still show after deletion, and I had to manually find their LLODs in the IDE list. Assuming for this same reason they were also missing from the table in CLodModels.cpp

These objects all use LLOD 17749, in addition to the already included HLOD 17559

These objects all appear the same HLOD, but have different LLODs for some reason: Inclusion recommended if it is desired that a user can request a HLOD belonging to a specific LLOD. (not including entry oilderricklod01 that is already in the table) !!!! This will complicate lookups for a HLOD to a LLOD. Might need to adjust some code to return a table instead, or just have it pick one.

7243 => 6996 was correct, but LLOD 6996 covers both HLODs 6994 and 7243 combined. The former is an entire casino building, and the latter is only the hedges surrounding its parking lot. A second opinion might be needed on what to do with this, but I'd heavily recommend changing the line as it makes much more sense after a visual inspectation.

HLOD 3412 is only the orange upper parts of a bridge, whereas HLOD 3411 is the entire bridge, including yet even more orange upper parts. LLOD 3413 matches HLOD 3411 much more closely

HLOD 5786 comprises just a set of door that go on entire building HLOD 5784 that LLOD 5945 matches much more closely. !!!! However, the LLOD seems to be rotated 180 degrees compared to the HLOD 5784. Someone else besides me should look at this and make a judgement call.

lotsofs · 2025-01-24T13:28:02Z

I let the invitation expire by accident, but I have managed to submit PRs to your repo for each of my proposed changes.

=======================================

There are a bunch of LLODs that fit their HLOD rather awkwardly, such as being rotated in some way, or having additional structures like fences etc that are separated from the HLOD counterpart. I named like four above, but out of the thousands other entries there likely are more.

This might lead to some minor visual impurities, but I personally don't think it significant enough to warrant investigation, as LLODs are not meant to be seen from up close either way. However it is something to make note of.

Add missing entries to CLodModels.cpp

Add additional HLODs that each map to the same LLOD CLodModels.cpp

Add multiple LLODs for 1 HLOD to CLodModels.cpp

Update LLOD that covers multiple HLODs in CLodModels.cpp

Fix one entry in CLodModels.cpp

Fix entry 5945 in CLodModels.cpp

Fernando-A-Rocha · 2025-01-25T01:25:40Z

Thank you for the contributions. I'll take a look!

add shared func getObjectLODModel(int objectModel)

6f7c6c6

TracerDS suggested changes Oct 29, 2024

View reviewed changes

Fernando-A-Rocha and others added 2 commits October 29, 2024 15:13

Merge branch 'master' into add-lod-model-func

f182db2

refactor 1

cb928f1

Fernando-A-Rocha and others added 2 commits October 30, 2024 09:33

new arg parser

93251bd

Merge branch 'master' into add-lod-model-func

d0419a3

Fernando-A-Rocha changed the title ~~Add getObjectLODModel(int objectModel)~~ Add getObjectLODModel and getObjectModelOfLOD Oct 30, 2024

rename to GetObjectLODOfModel

62cb6e8

Fernando-A-Rocha changed the title ~~Add getObjectLODModel and getObjectModelOfLOD~~ Add getObjectLODOfModel and getObjectModelOfLOD Oct 30, 2024

add getObjectModelOfLOD

c648cbb

Fernando-A-Rocha requested a review from TracerDS October 30, 2024 09:47

TracerDS suggested changes Oct 30, 2024

View reviewed changes

FileEX suggested changes Oct 30, 2024

View reviewed changes

Fernando-A-Rocha added 2 commits October 30, 2024 10:41

refactors

405ad8d

refactor

51e8967

Fernando-A-Rocha requested review from FileEX and TracerDS October 30, 2024 10:48

Fernando-A-Rocha added 3 commits October 30, 2024 12:27

test @TracerDS

21e7b28

fix OBJ_LOD_MODELS array @TracerDS

6244ebb

constexpr

73a4252

TracerDS suggested changes Oct 31, 2024

View reviewed changes

Shared/mods/deathmatch/logic/CLodModels.cpp Outdated Show resolved Hide resolved

Fernando-A-Rocha added 2 commits November 1, 2024 18:15

Update CLodModels.cpp

832c6bd

Update CLodModels.cpp

be36dc5

rename GetObjectLowLODOfModel and GetObjectHighLODOfModel

765f7aa

wip

b03ea49

Fernando-A-Rocha requested review from TheNormalnij and TracerDS November 26, 2024 02:27

TracerDS suggested changes Dec 3, 2024

View reviewed changes

ds1-e mentioned this pull request Dec 14, 2024

Ditch editor's mapEditorScriptingExtension file or cut down on its size multitheftauto/mtasa-resources#558

Open

Merge branch 'master' into add-lod-model-func

e8150c1

Fernando-A-Rocha and others added 7 commits January 7, 2025 15:17

Merge branch 'master' into add-lod-model-func

80604f0

Add missing entries

ba04a9c

These are all objects that gave me trouble deleting them in the editor, their LLOD would still show after deletion, and I had to manually find their LLODs in the IDE list. Assuming for this same reason they were also missing from the table in CLodModels.cpp

Add additional HLODs that each map to the same LLOD CLodModels.cpp

6168d02

These objects all use LLOD 17749, in addition to the already included HLOD 17559

Fix one entry in CLodModels.cpp

d9d5d51

HLOD 3412 is only the orange upper parts of a bridge, whereas HLOD 3411 is the entire bridge, including yet even more orange upper parts. LLOD 3413 matches HLOD 3411 much more closely

Fernando-A-Rocha added 7 commits January 24, 2025 13:35

Merge pull request #1 from lotsofs/patch-1

49259f7

Add missing entries to CLodModels.cpp

Merge pull request #2 from lotsofs/patch-2

9e4ca81

Add additional HLODs that each map to the same LLOD CLodModels.cpp

Merge pull request #3 from lotsofs/patch-3

5cd469f

Add multiple LLODs for 1 HLOD to CLodModels.cpp

Merge pull request #4 from lotsofs/patch-4

b7f0a9c

Update LLOD that covers multiple HLODs in CLodModels.cpp

Merge pull request #5 from lotsofs/patch-5

a1b3f86

Fix one entry in CLodModels.cpp

Merge pull request #6 from lotsofs/patch-6

404cd79

Fix entry 5945 in CLodModels.cpp

Merge branch 'master' into add-lod-model-func

d7eeea5

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add LOD management functions #3831

Add LOD management functions #3831

Fernando-A-Rocha commented Oct 29, 2024 •

edited

Loading

FileEX commented Oct 29, 2024

Fernando-A-Rocha commented Oct 29, 2024 •

edited

Loading

lotsofs commented Nov 1, 2024

Fernando-A-Rocha commented Nov 1, 2024

derxgbb commented Nov 1, 2024

Fernando-A-Rocha commented Nov 2, 2024 •

edited

Loading

Fernando-A-Rocha commented Nov 26, 2024 •

edited

Loading

TracerDS Dec 3, 2024

TracerDS Dec 3, 2024

TracerDS Dec 3, 2024

TracerDS Dec 3, 2024

lynconsix commented Dec 14, 2024

ds1-e commented Dec 14, 2024 •

edited

Loading

lynconsix commented Dec 24, 2024

Nico8340 commented Dec 26, 2024

Fernando-A-Rocha commented Jan 1, 2025

lotsofs commented Jan 6, 2025 •

edited

Loading

Fernando-A-Rocha commented Jan 7, 2025

lotsofs commented Jan 24, 2025

Fernando-A-Rocha commented Jan 25, 2025

Add LOD management functions #3831

Are you sure you want to change the base?

Add LOD management functions #3831

Conversation

Fernando-A-Rocha commented Oct 29, 2024 • edited Loading

Details

Examples

FileEX commented Oct 29, 2024

Fernando-A-Rocha commented Oct 29, 2024 • edited Loading

lotsofs commented Nov 1, 2024

Fernando-A-Rocha commented Nov 1, 2024

derxgbb commented Nov 1, 2024

Fernando-A-Rocha commented Nov 2, 2024 • edited Loading

Fernando-A-Rocha commented Nov 26, 2024 • edited Loading

TracerDS Dec 3, 2024

Choose a reason for hiding this comment

TracerDS Dec 3, 2024

Choose a reason for hiding this comment

TracerDS Dec 3, 2024

Choose a reason for hiding this comment

TracerDS Dec 3, 2024

Choose a reason for hiding this comment

lynconsix commented Dec 14, 2024

ds1-e commented Dec 14, 2024 • edited Loading

lynconsix commented Dec 24, 2024

Nico8340 commented Dec 26, 2024

Fernando-A-Rocha commented Jan 1, 2025

lotsofs commented Jan 6, 2025 • edited Loading

Fernando-A-Rocha commented Jan 7, 2025

lotsofs commented Jan 24, 2025

Fernando-A-Rocha commented Jan 25, 2025

Fernando-A-Rocha commented Oct 29, 2024 •

edited

Loading

Fernando-A-Rocha commented Oct 29, 2024 •

edited

Loading

Fernando-A-Rocha commented Nov 2, 2024 •

edited

Loading

Fernando-A-Rocha commented Nov 26, 2024 •

edited

Loading

ds1-e commented Dec 14, 2024 •

edited

Loading

lotsofs commented Jan 6, 2025 •

edited

Loading