[Baltram]

Release v1.0

New features:

• Printer and Parser for Risen 3 Templates (.r3tpl <-> .r3tpldoc)
• Printer and Parser for Risen 3 Sectors (.r3sec <-> .r3secdoc)
• Ability to propagate changes in templates to the sectors
• Ability to recalculate bounding boxes, bounding spheres and matrixes of entities that have been moved/newly inserted
• Ability to change the archive generation of a .pak file (without repacking it)

This makes it finally possible to edit the world files (sectors) and the templates (e.g. items).
See the newly opened tutorials section for more infos.

As a small demonstration of the new possibilities I released the Food Diversity modification.

--

Neue Features:

• Printer und Parser für Risen 3-Templates (.r3tpl <-> .r3tpldoc)
• Printer und Parser für Risen 3-Sektoren (.r3sec <-> .r3secdoc)
• Möglichkeit zur Übertragung von Änderungen an den Templates in die Sektoren
• Möglichkeit zur Neuberechnung von Bounding Boxes, Bounding Spheres und Matrizen von Entities, die bewegt oder neu eingefügt wurden
• Möglichkeit zur Änderung der Archivgeneration einer .pak-Datei (ohne sie neu zu packen)

Das macht es endlich möglich, die Weltdaten (Sektoren) und Templates (z.B. Items) zu bearbeiten.
Weiteres im neuen Tutorial-Bereich.

Als kleine Demonstration der neuen Möglichkeiten habe ich die Modifikation Food Diversity veröffentlicht.

[Dark_Bauer]

Sehr geile Sache, danke für das Programm!

Ich werde also meiner eigenen Bitte von einst nachgehen und dem Scavenger seinen alten, richtigen Schrei wiedergeben.

Und ich freue mich, das die Namen der Dialogdateien wieder alte Strukturen angenommen haben (so geordnet wie sie in G3 etwa waren), das war in R1 und 2 nicht so, wenn ich mich nicht irre.

[George]

I've added a Printer and Parser for Risen 3 Documents (.hdr <-> .hdrdoc) to the Risen 3 Resource Manager and I want to share it with you.

EDIT: Baltram recompiled the Resource Manager and included my patch: r3resman_George.zip.

My changes are based on rmtools repo revision 230.
The zip archive contains a diff file and as alternative the changed source files.

r3resmanHdrPrinterParser.zip (22.94 KiB)


Thanks to Baltram (Risen 3 Resource Manager) and ogurets (analysis of the DT01 container).

[Nate-Bonnett]

Vielen Dank für das Programm.
Läuft echt super und jetzt habe ich endlich den kompletten Soundtrack von Risen 3 und konnte schon eine Mod veröffentlichen.

[Nate-Bonnett]

Wäre es auch möglich, die r3msh Dateien umzuwandeln?

[Baltram]

Wäre es auch möglich, die r3msh Dateien umzuwandeln?

Nur wenn sich jemand die Zeit nimmt, das Dateiformat zu analysieren. Bei mir reicht die Freizeit im Moment nicht.

[Nate-Bonnett]

Nur wenn sich jemand die Zeit nimmt, das Dateiformat zu analysieren. Bei mir reicht die Freizeit im Moment nicht.

Schade. Ich kennen mich mit sowas leider nicht aus. Sonst wurde ichs machen.

[George]

.r3msh-files are resource files and therefore structured as described in Baltrams post.

1. The actual meshdata (vertices, normals, uvs, ...) is stored in ResourceData.Data1 (the other three data blocks seem to be empty).

// uint8_t Data1[ResourceOffsetTable.Data1Size];
MeshData1 {
	char8_t  MeshClass[4]; // "GMSH"
	uint32_t _unknown1; // 0x01000000
	uint32_t _unknown2; // 0x65000000
	uint32_t IndicesSize;
	uint32_t _unknown3; // 0x00000000
	uint16_t Indicies[IndicesSize/2];
	uint32_t VerticesSize;
	uint32_t VertexSize;
	uint32_t _unknown4; // 0x00000000
	uint32_t VertexElementsCount;
	D3DVERTEXELEMENT9 VertexElements[VertexElementsCount];
	uint8_t Vertices[VerticesSize];
	uint32_t _unknown5[6]; // 000...
};

2. ResourceHeader.ResourceObject is of the type eCMeshResource2, which structure is shown in the following example (took it from Obj_Bld_TUT_Ruin_Plaza_1_L1.r3msh).

struct eSSubMeshClassDataEntry {
	bCBox Unk1;
	unsigned int Unk2;
	unsigned int Unk3;
}

class eCMeshResource2 {
	Version = 1;
	Properties {
		class bTObjArray<class eCSubMesh> SubMeshes = [
			class eCSubMesh {
				Version = 1;
				Properties {
					class bCString MaterialName = "AS_Tpl_Ceiling_1_1";
					class bCBox Extends = (box -776.097229 333.270752 -909.524109 1099.568115 541.254333 598.567566);
					unsigned int FirstIndex = 0;
					unsigned int IndexCount = 8907;
					unsigned int FirstVertex = 0;
					unsigned int VertexCount = 2270;
					unsigned int FirstBone = 0;
					unsigned int BoneCount = 0;
				}
				ClassData {
					unsigned int UnkCount = 230;
					eSSubMeshClassDataEntry ClassDataEntries[UnkCount];
				}
			},
			class eCSubMesh {
				Version = 1;
				Properties {
					class bCString MaterialName = "AS_Tpl_Ground_2";
					class bCBox Extends = (box -769.448975 174.146820 -1536.206055 1557.998047 556.443359 614.167603);
					unsigned int FirstIndex = 8907;
					unsigned int IndexCount = 1176;
					unsigned int FirstVertex = 2270;
					unsigned int VertexCount = 605;
					unsigned int FirstBone = 0;
					unsigned int BoneCount = 0;
				}
				ClassData {
					unsigned int UnkCount = 22;
					eSSubMeshClassDataEntry ClassDataEntries[UnkCount];
				}
			},
			class eCSubMesh {
				Version = 1;
				Properties {
					class bCString MaterialName = "AS_Tpl_Ground_3";
					class bCBox Extends = (box -3115.483887 -323.264618 -5184.551758 2453.021973 640.101013 2476.504639);
					unsigned int FirstIndex = 10083;
					unsigned int IndexCount = 17049;
					unsigned int FirstVertex = 2875;
					unsigned int VertexCount = 3955;
					unsigned int FirstBone = 0;
					unsigned int BoneCount = 0;
				}
				ClassData {
					unsigned int UnkCount = 413;
					eSSubMeshClassDataEntry ClassDataEntries[UnkCount];
				}
			},
			class eCSubMesh {
				Version = 1;
				Properties {
					class bCString MaterialName = "AS_Tpl_Pillar_2_1";
					class bCBox Extends = (box -782.033142 -119.825836 -1446.053833 1098.564819 527.600342 596.820374);
					unsigned int FirstIndex = 27132;
					unsigned int IndexCount = 28086;
					unsigned int FirstVertex = 6830;
					unsigned int VertexCount = 6825;
					unsigned int FirstBone = 0;
					unsigned int BoneCount = 0;
				}
				ClassData {
					unsigned int UnkCount = 769;
					eSSubMeshClassDataEntry ClassDataEntries[UnkCount];
				}
			},
			class eCSubMesh {
				Version = 1;
				Properties {
					class bCString MaterialName = "AS_Tpl_Roof_1_1";
					class bCBox Extends = (box -305.965942 500.739990 -1538.125732 969.948730 609.416748 -1034.113770);
					unsigned int FirstIndex = 55218;
					unsigned int IndexCount = 108;
					unsigned int FirstVertex = 13655;
					unsigned int VertexCount = 40;
					unsigned int FirstBone = 0;
					unsigned int BoneCount = 0;
				}
				ClassData {
					unsigned int UnkCount = 1;
					eSSubMeshClassDataEntry ClassDataEntries[UnkCount];
				}
			},
			class eCSubMesh {
				Version = 1;
				Properties {
					class bCString MaterialName = "AS_Tpl_Stairs_3";
					class bCBox Extends = (box -2017.009766 -725.530762 -4531.818359 2420.778076 605.659302 2795.842041);
					unsigned int FirstIndex = 55326;
					unsigned int IndexCount = 20709;
					unsigned int FirstVertex = 13695;
					unsigned int VertexCount = 4990;
					unsigned int FirstBone = 0;
					unsigned int BoneCount = 0;
				}
				ClassData {
					unsigned int UnkCount = 543;
					eSSubMeshClassDataEntry ClassDataEntries[UnkCount];
				}
			},
			class eCSubMesh {
				Version = 1;
				Properties {
					class bCString MaterialName = "AS_Tpl_Wall_12_1";
					class bCBox Extends = (box -3169.463379 -649.364380 -5252.328125 1646.322876 604.308960 1432.055298);
					unsigned int FirstIndex = 76035;
					unsigned int IndexCount = 5853;
					unsigned int FirstVertex = 18685;
					unsigned int VertexCount = 1424;
					unsigned int FirstBone = 0;
					unsigned int BoneCount = 0;
				}
				ClassData {
					unsigned int UnkCount = 167;
					eSSubMeshClassDataEntry ClassDataEntries[UnkCount];
				}
			},
			class eCSubMesh {
				Version = 1;
				Properties {
					class bCString MaterialName = "AS_Tpl_Wall_Broken_1_1";
					class bCBox Extends = (box -2020.229126 -725.530762 -4456.557617 2364.066162 640.101013 1003.992798);
					unsigned int FirstIndex = 81888;
					unsigned int IndexCount = 43779;
					unsigned int FirstVertex = 20109;
					unsigned int VertexCount = 14736;
					unsigned int FirstBone = 0;
					unsigned int BoneCount = 0;
				}
				ClassData {
					unsigned int UnkCount = 1307;
					eSSubMeshClassDataEntry ClassDataEntries[UnkCount];
				}
			}
		];
		class bCBox Boundary = (box -3169.463379 -725.530762 -5252.328125 2453.021973 640.101013 2795.842041);
	}
	ClassData {
	}
}

I was not able to find a mesh with a BoneCount greater zero.


3. ResourceHeader.ReferencedFileCount seems to be always zero.


@Baltram: Without a look in your Risen XmshReader i would have never come to the D3DVERTEXELEMENT9 structure...

[George]

Based on Baltrams Risen mCXmshReader included in Rimy3D, i wrote a basic Risen 3 mesh reader.
For now it ignores the eCMeshResource2 class (contains info about materials and submeshes) and simply reads the data stored in ResourceOffsetTable.Data1 block.

[Bild: 2I3WYyXz_R3msh.png]

[Baltram]

Based on Baltrams Risen mCXmshReader included in Rimy3D, i wrote a basic Risen 3 mesh reader.
[...]

Cool!
Maybe you could include a small .r3msh -> .obj converter in r3resman?
(Because I am not sure when I will find the time to finally release Rimy3D to the public.)

@Baltram: Without a look in your Risen XmshReader i would have never come to the D3DVERTEXELEMENT9 structure...

It's Nico who found out.

[hargreave]

Hi!
I apologize if my question is silly. The unpacked Template and Sector files all have (I guess) hex codes in them. Like so in the 'It_Swd_Worldslayer.r3tpldoc' file:

class eCCollisionShapeList Shapes = <01 00 00 00 47 45 43 32 13 57 EC CC 01 00 72 01 00 00 11 00 C9 C8 59 E8 7A 9A E2 F2 1C 00 00 00 1A 00 49 74 5F 50 69 65 5F 43 61 70 74 61 69 6E 73 62 6C 61 64 65 5F 4C 31 5F 43 56 35 DA E8 98 47 64 73 A6 40 00 00 00 00 00 80 3F 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 80 3F 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 80 3F 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 80 3F 76 34 D6 79 74 B9 4B CC 08 00 00 00 10 A5 25 B5 00 00 00 00 91 B3 94 7C 95 B7 D9 9E 04 00 00 00 00 00 00 00 3A A6 34 E3 A7 D7 1D 73 08 00 00 00 D5 61 5B 62 00 00 00 00 91 B3 94 7C C5 18 CC A7 04 00 00 00 00 00 00 00 91 B3 94 7C A5 F4 09 C8 04 00 00 00 01 00 00 00 91 B3 94 7C 78 C3 E5 B1 04 00 00 00 00 00 00 00 91 B3 94 7C 37 12 11 C6 04 00 00 00 01 00 00 00 91 B3 94 7C A7 24 97 77 04 00 00 00 01 00 00 00 91 B3 94 7C DC CF D6 D6 04 00 00 00 00 00 00 00 91 B3 94 7C 2D 02 D4 1B 04 00 00 00 01 00 00 00 91 B3 94 7C C8 20 CC DC 04 00 00 00 00 00 00 00 91 B3 94 7C E0 65 7D B2 04 00 00 00 00 00 00 00 91 B3 94 7C 0B AC 1E D4 04 00 00 00 00 00 00 00 91 B3 94 7C 31 3C 78 61 04 00 00 00 00 00 00 00 91 B3 94 7C 30 D5 BE A3 04 00 00 00 01 00 00 00 00 00 00 00>;

What does it bear?

[Baltram]

Hi!
I apologize if my question is silly. The unpacked Template and Sector files all have (I guess) hex codes in them. Like so in the 'It_Swd_Worldslayer.r3tpldoc' file:

class eCCollisionShapeList Shapes = <01 00 00 00 47 45 43 32 13 57 EC CC 01 00 72 01 00 00 11 00 C9 C8 59 E8 7A 9A E2 F2 1C 00 00 00 1A 00 49 74 5F 50 69 65 5F 43 61 70 74 61 69 6E 73 62 6C 61 64 65 5F 4C 31 5F 43 56 35 DA E8 98 47 64 73 A6 40 00 00 00 00 00 80 3F 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 80 3F 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 80 3F 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 80 3F 76 34 D6 79 74 B9 4B CC 08 00 00 00 10 A5 25 B5 00 00 00 00 91 B3 94 7C 95 B7 D9 9E 04 00 00 00 00 00 00 00 3A A6 34 E3 A7 D7 1D 73 08 00 00 00 D5 61 5B 62 00 00 00 00 91 B3 94 7C C5 18 CC A7 04 00 00 00 00 00 00 00 91 B3 94 7C A5 F4 09 C8 04 00 00 00 01 00 00 00 91 B3 94 7C 78 C3 E5 B1 04 00 00 00 00 00 00 00 91 B3 94 7C 37 12 11 C6 04 00 00 00 01 00 00 00 91 B3 94 7C A7 24 97 77 04 00 00 00 01 00 00 00 91 B3 94 7C DC CF D6 D6 04 00 00 00 00 00 00 00 91 B3 94 7C 2D 02 D4 1B 04 00 00 00 01 00 00 00 91 B3 94 7C C8 20 CC DC 04 00 00 00 00 00 00 00 91 B3 94 7C E0 65 7D B2 04 00 00 00 00 00 00 00 91 B3 94 7C 0B AC 1E D4 04 00 00 00 00 00 00 00 91 B3 94 7C 31 3C 78 61 04 00 00 00 00 00 00 00 91 B3 94 7C 30 D5 BE A3 04 00 00 00 01 00 00 00 00 00 00 00>;

What does it bear?

There are hundreds of different data structures and it would be hundreds of hours of work to analyze them all so I did this only for the most common.
All the other not-so-important data structures are just printed out byte-by-byte by Risen 3 Resource Manager (each two-digit hexadecimal number representing one byte).

[Baltram]

Some .pak archives in Risen 3: Enhanced Edition have a _z suffix (e.g. 0_na_img_z.pak). Such archives cannot be generated using the original version of Risen 3 Resource Manager.

So here is an experimental version of this tool that makes it possible to convert normal archives into their _z counterpart:

r3resman_experimental.zip (279KB)

For example, if you have an archive named 9_na_img.pak, simply rename it to 9_na_img_z.pak.9 and drag it onto the new version of r3resman.exe. This gives you an archive named 9_na_img_z.pak that can be used with Risen 3: Enhanced Edition.

Also included in this experimental version of Risen 3 Resource Manager is the support for Risen 3 documents (.hdr) by George and the support for printing/parsing bTRefPtrArray objects.

--

Einige .pak-Archive in Risen 3: Enhanced Edition haben ein _z-Suffix im Dateinamen (z.B. 0_na_img_z.pak). Solche Archive können nicht mit der originalen Version des Risen 3 Resource Managers erstellt werden.

Hier nun eine experimentelle Version des Tools die es ermöglicht, normale Archive in ihre _z-Variante umzuwandeln:

r3resman_experimental.zip (279KB)

Hat man etwa ein Archiv mit dem Namen 9_na_img.pak, muss man es lediglich umbenennen in 9_na_img_z.pak.9 und es auf die neue r3resman.exe ziehen. Das Ergebnis ist ein Archiv mit dem Namen 9_na_img_z.pak, das für Risen 3: Enhanced Edition verwendet werden kann.

Zusätzlich unterstützt diese experimentelle Version des Risen 3 Resource Managers dank George Risen 3-Dokumenten (.hdr), außerdem enthalten ist print/parse-Unterstützung für bTRefPtrArray-Objekte.

--

source

[hargreave]

There are hundreds of different data structures and it would be hundreds of hours of work to analyze them all so I did this only for the most common.
All the other not-so-important data structures are just printed out byte-by-byte by Risen 3 Resource Manager (each two-digit hexadecimal number representing one byte).

Oh, ok. Thank you!

[Kuchenschlachter]

As promised I tried to implement FloatScale and ColorScale.
r3resman_ColorScaleFloatScale.zip
source_r3resman_ColorScaleFloatScale.zip

I'm glad IF it works but frankly if it doesn't I don't care...
There are a lot of issues. The only one that concerns useres is that my version isn't "precise". It means converting a file to doc and back results in a slightly different file. That really really really sucks.
I'm sorry but I can't do more work on this.
Good luck.

Edit: I take it back, there is a second problem concerning users... really really really awful float values, like negative zero, numbers that are nearly zero but are shown as exactly zero and gigantic numbers... I'm still certain those are floats.

[MN32410]

As promised I tried to implement FloatScale and ColorScale.
r3resman_ColorScaleFloatScale.zip
source_r3resman_ColorScaleFloatScale.zip

I'm glad IF it works but frankly if it doesn't I don't care...
There are a lot of issues. The only one that concerns useres is that my version isn't "precise". It means converting a file to doc and back results in a slightly different file. That really really really sucks.
I'm sorry but I can't do more work on this.
Good luck.

Edit: I take it back, there is a second problem concerning users... really really really awful float values, like negative zero, numbers that are nearly zero but are shown as exactly zero and gigantic numbers... I'm still certain those are floats.

Thank you for your hard work. I will test it at the weekend.

[tombom81]

is that my version isn't "precise".

thanks for your work!
FB 8A AF 3D -> 0.085714 -> d2 8A AF 3D
Guess the Piranhas calculated using "double" then cut it to "single".
So there's no change to get rid of the difference (FB<>D2).

and gigantic numbers... I'm still certain those are floats.

talking about class bCColorScale NightSkyColor
0x846: 71 2C C6 6D = 7666469788512741400000000000.000000
and class bCColorScale NightSkyHazeColor?
0xA050: c2 5D 62 EC = -1094639975936653500000000000.000000
I converted the hdrdoc back to hdr and at 0x846 71 2C C6 6D to be found and at 0xA050 there was c2 5d 62 ec again.
So hopefully that should not be a problem (though it's strange).

[George]

You could create a pull request containing your changes. I've done that with my changes, but I'm still waiting for a reaction from Baltram.
Repository of baltrams rmtools: https://github.com/Baltram/rmtools

[Baltram]

I've done that with my changes, but I'm still waiting for a reaction from Baltram.

Well you would have waited forever if you didn't post this here. The only time I logged into GitHub so far was when I opened the repository.

So I merged the two pull requests and I am very happy Rimy3D can read Risen 3 meshes now - thanks!

[George]

Well you would have waited forever if you didn't post this here. The only time I logged into GitHub so far was when I opened the repository.

I assumed that github sent a notification mail to you, since the pull request affects your repository.