Forza Motorsport Resource Extraction (.carbin)

[shakotay2]

but there are some cars that simply don't want to be converted, as for example the AMC Rebel "the machine", which make every H20 version crashing, and lead to an incomplete obj.

I'd need the modelbins in question to check what is causing the problem.

ahh c, and idk if you've seen my plugin (collects modelbin files from all dir and writes to respective modelbins.txt), but I wanted to do it where it uses exe to directly produce the obj's too
p.s. could you take name of folder in which modelbins.txt is in and have that as the obj name pls, i'm happy to help

I really appreciate this but after 30 min of weighting the pros and cons I found that handling a command line parameter would lead to further confusions atm.

I need a better understanding of the uv layering (i.e. a more or less fully textured model).

Atm in jm versions I use the BOOL CheckUV_offsets(HWND hwnd) function
to handle uv-offsets manually (line uv-offsets 4 4 4 4 4 4 4 4 (example) in the obj files).

This uv selection has to be done for each xxx_a.modelbin separately (I'm not even sure if once you found a suiting offset combination it would fit for all modelbins of a model).

The idea of your plugin processing modelbins automatically is good but it severely interferes with the manual uv selection (per model, or maybe even worse, per modelbin?). It would produce dozens of obj files containing many wrong uv maps which I fear could cause a shit storm.

I don't have the time for texturing so still waiting for a fully textured model, see above.

[SoniKalien]

but there are some cars that simply don't want to be converted, as for example the AMC Rebel "the machine", which make every H20 version crashing, and lead to an incomplete obj.

Does anybody here havec some insight wether its coming from the original modelbin and its extraction or from the conversion by H2O and so can be done at the end.

Thanks in advance!!

MakeH20-J5 gives me "more than one v block" error. No crash but mesh has "invalid vertex index" on import.

I will try to narrow down the offending part.

It is:

Code: Select all

amc_rebel_70\scene\Exterior\Hood\hood_a.modelbin

With that line removed I get no problems with MakeH20 but the resulting .obj is still slow to import but eventually does.

[Frazzlee]

but there are some cars that simply don't want to be converted, as for example the AMC Rebel "the machine", which make every H20 version crashing, and lead to an incomplete obj.

I'd need the modelbins in question to check what is causing the problem.

ahh c, and idk if you've seen my plugin (collects modelbin files from all dir and writes to respective modelbins.txt), but I wanted to do it where it uses exe to directly produce the obj's too
p.s. could you take name of folder in which modelbins.txt is in and have that as the obj name pls, i'm happy to help

I really appreciate this but after 30 min of weighting the pros and cons I found that handling a command line parameter would lead to further confusions atm.

I need a better understanding of the uv layering (i.e. a more or less fully textured model).

Atm in jm versions I use the BOOL CheckUV_offsets(HWND hwnd) function
to handle uv-offsets manually (line uv-offsets 4 4 4 4 4 4 4 4 (example) in the obj files).

This uv selection has to be done for each xxx_a.modelbin separately (I'm not even sure if once you found a suiting offset combination it would fit for all modelbins of a model).

The idea of your plugin processing modelbins automatically is good but it severely interferes with the manual uv selection (per model, or maybe even worse, per modelbin?). It would produce dozens of obj files containing many wrong uv maps which I fear could cause a shit storm.

I don't have the time for texturing so still waiting for a fully textured model, see above.

ahh yeah I see what you mean, thank you for explaining , I'll try to think of something to help with UV's

[Doc0]

but there are some cars that simply don't want to be converted, as for example the AMC Rebel "the machine", which make every H20 version crashing, and lead to an incomplete obj.

Does anybody here havec some insight wether its coming from the original modelbin and its extraction or from the conversion by H2O and so can be done at the end.

Thanks in advance!!

MakeH20-J5 gives me "more than one v block" error. No crash but mesh has "invalid vertex index" on import.

I will try to narrow down the offending part.

It is:

Code: Select all

amc_rebel_70\scene\Exterior\Hood\hood_a.modelbin

With that line removed I get no problems with MakeH20 but the resulting .obj is still slow to import but eventually does.

Thanks, you are my savior !!

It works smoothly, and i have a nice obj without the regular hood (whatever the LOD) and without error in Blender, but, wich is strange, is that if i use MakeH20-J5 on the modelbin.text generated by the dir_modelbin in the exterior/hood folder alone, i got the nice regular hood i need. Seems the error is just when MakeH20-J5 (and the others) run on the full modelbin text.

How did you find the line, by try and error or you read it in the matrix?

[SoniKalien]

It works smoothly, and i have a nice obj without the regular hood (whatever the LOD) and without error in Blender, but, wich is strange, is that if i use MakeH20-J5 on the modelbin.text generated by the dir_modelbin in the exterior/hood folder alone, i got the nice regular hood i need. Seems the error is just when MakeH20-J5 (and the others) run on the full modelbin text.

That is strange, as dir_modelbin.cmd generates absolute paths, and these paths have no influence on the generated obj. The only difference would be that you only have all the hoods listed instead of every single car part.

I did try hood_a alone and with the same results and errors.

How did you find the line, by try and error or you read it in the matrix?

I prayed, and Jesus showed me the correct path.

No really, if you watch the MakeH20 it will list each part after it has been processed. So if it pauses with an error message, the problem lies with the next file that is listed in modelbins.txt after the last part that is displayed.

[zerozone]

Things are going pretty well with the last JM, JM4 and 5 are O.K., i can do things like that:

or

Each model can take for me 30 mn to 2 hours to have something correct using Blender.

At the end models are pretty good looking :



but there are some cars that simply don't want to be converted, as for example the AMC Rebel "the machine", which make every H20 version crashing, and lead to an incomplete obj.

Does anybody here havec some insight wether its coming from the original modelbin and its extraction or from the conversion by H2O and so can be done at the end.

Thanks in advance!!

How do you know what parts goes on what bone? the names doesn't match and is quite difficult to identify some parts

[Doc0]

To Sonik, i will send you the files if you want, i used J5 and Blender for import (and for what is worth, on a 32 bits WIN 7). And now i have a full AMC rebel

How do you know what parts goes on what bone? the names doesn't match and is quite difficult to identify some parts

To be honest, i don't use the skeleton, it's the only file i remove from the modelbin_text generated. I run the H2O, now the J5, and i import the obj in Blender.
After, i separate first the lods in another layer, using the outliner.
after that, still with the outliner now in the layer with all the lods, i put doors, hood, trunk, wipers, suspension, wheel, dash, etc, each part that need to be adjust or rescaled in its layer.

Next, i just have to rescale the dash, steering wheel and wipers, an dmove the trunk and hood.
For the doors, there is always one which is good, and the other just need to be adjusted.

Now there is always some bad surprise, as having to look for some part in the lod1 (rims, pars of the dash, some hoods, etc. ) or dash so distorted that i take half an hour to adjust. And these fu...g wipers.

[Frazzlee]

To Sonik, i will send you the files if you want, i used J5 and Blender for import (and for what is worth, on a 32 bits WIN 7). And now i have a full AMC rebel

How do you know what parts goes on what bone? the names doesn't match and is quite difficult to identify some parts

To be honest, i don't use the skeleton, it's the only file i remove from the modelbin_text generated. I run the H2O, now the J5, and i import the obj in Blender.
After, i separate first the lods in another layer, using the outliner.
after that, still with the outliner now in the layer with all the lods, i put doors, hood, trunk, wipers, suspension, wheel, dash, etc, each part that need to be adjust or rescaled in its layer.

Next, i just have to rescale the dash, steering wheel and wipers, an dmove the trunk and hood.
For the doors, there is always one which is good, and the other just need to be adjusted.

Now there is always some bad surprise, as having to look for some part in the lod1 (rims, pars of the dash, some hoods, etc. ) or dash so distorted that i take half an hour to adjust. And these fu...g wipers.

Wipers are worst jaja

[shakotay2]

-jm6, fixed AMC Rebel, hood_a, reported by Doc0, feedback from SoniKalien, thanks!

edit: AMC wipers use uv channels 1 to 4 but mainly 3 and 4:

AMC_wipers_uvchannel4.JPG

[Doc0]

-jm6, fixed AMC Rebel, hood_a, reported by Doc0, feedback from SoniKalien, thanks!

thanks for that, i don't have the time to make a list of all the problem i encountered, but thanks to Sonik i tried to find the issue for some of them:

Henessey Venom : error message : more than one V block! the engine is the problem, like the AMC by converting the engine line alone, and the rest without it, it works. I'm disapointed to see there is no real engin inside.

Camaro 2016 : error message : more than one V block problem are the Seats and the line below, work without them, not a big deal as the others camaro of the game are good, so i guess we can find seats, calipers wheels and rotor here.

Fiat Abarth 595 68 : problem is the body, this time always give message : a more than 15 submeshes - break. Without the body, no error message but the mesh is still a mess.

[shakotay2]

Henessey Venom : error message : more than one V block!

renewed the -jm6 link. Hopefully that "v block" error should have vanished now. (edit: forget about -jm6, use -jm7 instead)

Fiat Abarth 595 68 : problem is the body, this time always give message : a more than 15 submeshes - break.

yeah, too many false finds; trying to exclude them -> edit: fixed

ahh yeah I see what you mean, thank you for explaining , I'll try to think of something to help with UV's

You might start with the AMC_Rebel_70 because for Lod01 and LS0 there's a high percentage of messy uvs (when using 4 4 4 4 4 4 4 as uv offsets with -jm6). Well, -jm7 looks much better now...

-jm7/8 improves AMC; engine_a uv maps better (but still assumed to be uncomplete):

AMC_rebel-engine.JPG

[Doc0]

-jm7/8 improves AMC; engine_a uv maps better (but still assumed to be uncomplete):

AMC_rebel-engine.JPG

Many thanks for the jm8, i don't know what you did but before there was some obj generated that blender was unable to open.

With your jm8, i can open them without problem (for what i remember that was the plymouth prowler, the ferrari f50, the vauxhall astra, the two ultima cars, etc.)

And fiat 595 is ok!

Thanks again!!!!

[TomWin]

Where can I find gr2 skeleton importer or gr2 to obj? Somebody posted here something but it didnt work. I use granny viewe but it's only viewer

[Andrakann]

Where can I find gr2 skeleton importer or gr2 to obj? Somebody posted here something but it didnt work. I use granny viewe but it's only viewer

Only this tool can convert skeleton to DAE (it's also have text/xml structure and can be edited in text editor like OBJ).
And you need to put fresh granny2.dll to tool root folder to work.
Archive with Granny Viewer + GR2-to-DAE + granny2.dll is uploaded here.

[xInno3Dx]

@shakotay2 I would like to report a crash which occours when I'm converting the agera one from Horizon 3 with the tool. But when i extract the one from Apex it works fine.

FH3: https://drive.google.com/file/d/0B19AdD ... sp=sharing

Apex: https://drive.google.com/file/d/0B19AdD ... sp=sharing