Modding extras by CA (unsupported)

This needs a new thread with a clearer title:

Download

Alternative download location

This is a zip of files that are stored in binary format, within the animation files in data/animations folder. Vercingetorix's program cannot extract them, and they have been invisible and unavailable for all this time. Until now. Jerome Grasdyke has very generously provided us with these files, which now completely unties our hands and lets us do everything with skeletons that we could ever dream of.

IMPORTANT NOTICE DISCLAIMER:

These files are unofficial, and completely unsupported by CA. Use them at your own risk. By downloading this file you accept this disclaimer.

(Signifier, I took the freedom of using your description)

Many thanks to Jerome:bow:

For more information and posts by the man himself Click Here

I've uploaded the file to the LM's webspace if you want to get them from there:

http://www.totalwar.org/patrons/lm/M..._evt_files.zip

Myrddraal, it's probably not a good idea to hotlink the TWC download mirror. I'd suggest linking the download page itself.

OK, why not?

I'm not affiliated with TWC in any way, but I know that the general rule is that hotlinking is not good. It takes advantage of the the bandwidth of the content provider, without giving them any recognition. If linked to the download page, then the person knows that this is a TWC forum, they are the ones hosting the file, etc.

It's just like, why is it wrong to hotlink to pictures hosted by other websites? The same principle.

Or, put more simply: they don't get money from ads if people don't go to their pages, and they're the ones paying for us to be able to download.

-Simetrical

Hey Myrddraal:

Do you think the release of those files make the animations for a volley fire system possible?

Do you mean just making a firearm animation?

This won't change the current situation. You could make a new animation without this, but you could only base it on the existing skeletons.

If you wanted to make a new shooting animation, you can. As far as firing in a volley, the current archers do just that.

What the new files were injtended ( had they actually worked ) was to allow you to run new ani mations outside of the pack files, and to use completely new skeletons. This would have given you far greater control over the process, and allowed you to do things that just can't be done now.

No it's more than just firing in a volley like the archers do. If you all have ever seen a movie with muzzle loading firearms in it, you will notice that the front row fires in unison then goes to the back row. At that point the second row steps forward and fires and the process is repeated. It is a very complicated animation but it was present in all the previous Total War games where they had firearms.

Hrm...

Couldn;t that just be a continuous Looping Change in Formation?

I mean a formation could be defined in having the first or two first rows usetheir weapon and then a call to change the formation comes and you have the first row exchange position with the second and then the first row fires.

The firing delay is then not a question of every how many seconds the volley is shot, like the archers, but rather how long it takes for them to re-form, once re-formed then the firing delay comes in to effect.

The complete loop would have to be probably associated with a sorts of "Mode".

Not unlike the Phalanx, which I see more as a mode of fighting rather than a simple formation...

Its a way of looking at it I guess :charge:

with wich program ca you open the .evt files?

Notepad

has anyone actualy managed to get this to work ? deleting the paks and dats in data\animations so that the descr_skeleton.txt file is used ?

I just gave it a try (deleted the files, placed the extracted animation files into the animation folder, along with the evt files). But when I launch the game it crashes back to desktop on loading (no error message). After some experimenting I noticed it crashes after reading the first animation entry from descr_skeleton.txt. If the animation file doesn't exist I get an error message, if it does exist I dont get any. I then added a second entry with an animation file that didn't exist but got no error message. So it seems it crashes right after loading in the first animation file.

Could there be a problem with Vercingetorix's extractor so that it creates (slightly) corrupted cas files ? I noticed that some units did weird things ingame after changing animation speeds with Vercingetorix's tool (especially horses). So my best - and only - guess is that there is a problem with the cas animation files.

Nope it seems there's a big part within the packs, which is never stored in the text files. So if we delete the packs, that part becomes missing and the game crashes. I think it might be an oversight on CA's part, one which they will hopefully fix in the expansion and let the game parse this descr_ file just as it parses many of the others without so much as a hitch.

hm not sure. The descr_engine_skeleton.txt file has exactly the same format as descr_skeleton.txt. And those files are NOT located in the pak files, but seperate files (including evt ones). If it works for those then why wouldn't it work for the unit animations ?

EDIT: if I use a siege engine cas animation file (the ones that are not located in the packs) for the first entry descr_skeleton.txt, and an inexisting file for the second entry, I get an error message ("file does not exist"). That means it accepts the first animation file. So I'm still guessing there is a problem with the extracted cas files.

Ok, perhaps someone could assist me with an experiment... as some of you may remember from a while back, the attempt to get the game to load in the animations database from text file direct proved a little problematical. Recently a bunch of questions we were answering for the forums jogged my memory into an alternative approach to the problem.

The idea was to extern the ability to build the anim db files instead of running direct from the text files - these can be generated by the main exe in dev mode from the data you already have (in the zip file above) and do more or less the same thing as reading direct from the text file, except that this also does some essential post-processing before spitting out the various packed data files. But when I checked the code, I was surprised to find that this was already available in the release build.

I don't have a machine to hand with a retail build on it atm, so could someone test for me wether putting -util:animdb on the command line correctly rebuilds the animation packs for you... you will need the pack with the .evt files for this.

will give this a try as soon as I am able...hopefully tonight!

Is there any chance we could get an list of what the engine is capable of "compling/converting"? It took me a few weeks to find out that vegetation can be edited/added by removing some files and starting the game. The same counts for battlefield buildings.

Perhaps we will be able to discover everything, but seeing that PC Gamer has already placed quite a few reviews of R:TW mods I think that both CA and us can profit if we are given more insight of how the game works. I noticed with Middle Earth: Total War that people bought M:TW and VI just to play the mod and I start to notice the same with N:TW2 and Sengoku Jidai and I am sure that Europa Barbarum will also increase sales.

CA made a great engine and I think it would be a loss if CA does not fully understand what hundreds of modders and artists are capable of and what they can mean to CA.

If this came off like a rant, then I blame Brad Pitt for telling on Live 8: "Let's be bold." :wink:

As to your own request, I am not able to at the moment, but I do appreciate your contribution alot!

Trying to test now!

EDIT: Jerome, you'll have to give us more tips on what to do to test this out: should the four animation pack files be deleted? Should descr_skeleton.txt be deleted? Am I supposed to see anything happen when running the game with -util:animdb enabled? Etc.

my guess is that the four anim files need to be removed, and that those are regenerated by using that parameter. Have you tried that dsyrow1 ?

Quote:

---

Originally Posted by [cF]Adherbal

my guess is that the four anim files need to be removed, and that those are regenerated by using that parameter. Have you tried that dsyrow1 ?

---

That's pretty much correct. What should happen is that the game will parse descr_skeletons.txt, load the .evt and .cas files and generate a new batch of animation pack files. Your best bet for checking wether the process has run correctly is probably just to sort your directory view by date, any files that were just created should be easy to recognise.

Okay, here's what I did: I unpacked the animation packs using Verc's XIDX program, then I extracted the EVT files into the same folder, and then removed the four animation files. When I ran the game, it complained that it couldn't find the very first file in descr_skeleton.txt. Then I realized that I extracted the animation files in their native directory, i.e. into data/animations/data/animations, so then I simply moved the entire folder structure into the same folder as where the four anim pack files used to be. Then when I ran the game, it crashed without a comment, like it used to in our previous tests. I did add the -util:animdb to the command line, but have not noticed any different behavior because of it. The game still crashed without saying anything, as before. That's why I asked if there was something else I should have done, to bring out the new -util behavior. Maybee someone else can confirm this while Jerome is still in office.

Ok here's some new interesting behavior. I decided to play the game, so I restored all folders to their original state, and when I ran the game it again said that I was missing the first animation referenced in descr_skeleton.txt. But when I removed the -util:animdb from the shortcut, the game ran fine. This tells me that the game IS reading descr_skeleton.txt and is trying to parse it, even though the four animation pack files are still present.

What the code does is this:

* if you run with -util:animdb on the command line, it parses descr_skeleton.txt and tries to build the animdb files from it, regardless of other factors;
* if you run without the command line option, it looks for the packed files, and if it finds them it does not use descr_skeletons.txt as all the information is inside the packs;
* if it fails to load the packs, it falls back on loading descr_skeletons.txt, but does not regenerate the packs -- renaming pack.dat, pack.idx, skeletons.dat and skeletons.idx inside data/animations/ should cause it to take this code path;

In order to load from descr_skeletons.txt you need to have a complete set of all the files it references in the correct places - that includes all the .evt files in the zip file that I supplied some time ago, plus the .cas files generated by Vercingetorix's unpacking tool. If it can't find any of the files, the code should warn you, providing that you have -show_err on the command line. In general there should be matched pairs of .evt and .cas files.

This process works correctly when I try it with a dev build in release mode, so in theory it should work for you, but to the best of my knowledge no modder has yet managed to get it to complete... Exactly why is not really clear at this point.

I gave it a try, and like dsyrow it crashed with no error (-show_err parameter is there).

I will try it with an unmodded RTW version later today, just to be sure the new skeletons I'm using in descr_models_battle.txt (which are not defined in descr_skeleton) dont cause the problem - although I would expect it to return some kind of error message then.

Well, if Vercingetorix' tool for modifying the animation pack files works reliably for you, there's not a huge amount of pressure on sorting this out ;) If you guys want to pursue it, I'll help of course, although we're still in the runup to the expansion pack beta over here...

well there is only one thing that caught my interest from the descr_skeleton file:

Quote:

---

-id:x,y,z impact delta - specifies the point of weapon impact relative to the starting position of the root node

---

my question is: would moving this point forward by a "meter" or so allow soldiers from the second line in a unit to attack "throught" the units of the first line ? That would be very usualy for some spear/halberd/pike formations which dont fit with the phalanx formation.

if this doesn't work then IMO the descr_skeleton file doesn't give us much more options then what we can achieve with Verci's tools.

The ability to parse descr_skeleton.txt is simply invaluable, because not only will we be able to give units the ability to hit from further back, as you said, but we can also define which frame in the attack anim actually is the attack frame, so that the defender's animation can run synchroniously, not wildly irrelevantly like it does now (you attack, the guy ducks like 10 seconds later, etc).

Jerome, the biggest problem right now is the crash without the error message. That may be the best clue as to what is going wrong here, and you knowing that we have such an error might be a hint as to where to look for a solution. Would NOT having a matching .evt for a .cas cause such a crash, or would an error pop up? I really hope you can get your hands on the same build as us, so that you can at least get this parsing ability by the time the x-pak comes out, because you devs will see some amazing wonders from us in that case. Please try to find the same build as what we're using, so you can run through your steps and verify that you get the same nameless error as we do, so that you know that the error does not lie in some wrong step of ours, and are more able to pin down the issue on your side in time for the release. Or, if it is a mistake on our side after all, this would be the very BEST case scenario, and we can enjoy the wonderful ability to parse descr_skeleton.txt right away :charge:

Jerome, here's the deal: I just tried a vanilla version of everything, unpacked the CAS files and the EVT files. Here are the entire contents of my descr_skeleton.txt:

Quote:

---

type strat_named_with_army
scale 0.7
anim idle_1 data/animations/SG 01 idle 01.cas
-evt:data\animations\strat_named_with_army__SG 01 idle 01.evt

---

And the game crashes within seconds of showing the loading screen.


EDIT: Another test. Here are the new ENTIRE contents of my descr_skeleton.txt:

Quote:

---

type strat_named_with_army
scale 0.7
anim idle_1 data/animations/SG 01 idle 01.cas

---

I removed all references even to the EVT files, just the CAS itself. Again, same results, crash immediately upon showing the loading screen; in other words, the game doesn't get far enough to tell me that there are other skeletons lacking, it crashes immediately upon finding that one valid animation entry.

Ok, I've finally found a moment to get back to this. My sincere apologies that it's taken so long, but getting new projects off the ground is a very time-consuming business as I'm called upon to contribute to the business, technical and design parts of it all at the same time.

So, I've installed a vanilla copy of Rome, then applied BI to bring it up to version 1.3 / 1.4, deleted the animation pack files, and added in all the .cas and .evt files referenced by the game. It took a little while, but it was worth it... In short, it works as it should -- it loads and runs correctly, without crashing or other strange behaviour. And it even gives error messages for missing .cas files... now, that leaves a couple of possible explanations for the behaviour you guys have seen:

1) some small change in 1.3/1.4 has caused this to suddenly work (unlikely but possible)
2) the unpacking tool generates faulty .cas files which cause the game to crash before an error message is generated (probable)

I've even double-checked by archiving the .cas and .evt files, then deleting the contents of the directory, and then repopulating it from the archives -- there is nothing funny going on, and no further information is needed.

I've also uploaded a new .evt archive which is up to date with BI, in the same place as before. I would upload the .cas archive as well, but it would have to be cleared with legal, and at 10 mb it is some ways beyond the size limit.

The archive is called evt_files_bi.zip, if you are having trouble finding it...

It's in /patrons/lm/Modding? I only see the original one there...

Quote:

---

Originally Posted by The samnite

It's in /patrons/lm/Modding? I only see the original one there...

---

It in http://www.totalwar.org/Downloads/Rt...ads/RTWupload/ directory.

evt_files_bi.zip

Thanks!

I'm a bit confused on what the .evt files actually do. So what is their function, game wise, and can we add/remove/swap them around in desc_skeleton.txt


For a simple example: horse skeleton has no .evt for attacking, thus they cannot be used for replacing a wardog or elephant. But if the wardog's attack .evt can be added to the skeleton, then they can attack and use an assigned attack animation?

Will the engine acknowledge a new .evt entry in a skeleton in desc_skeleton.txt?

Quote:

---

Originally Posted by JeromeGrasdyke

I've also uploaded a new .evt archive which is up to date with BI, in the same place as before. I would upload the .cas archive as well, but it would have to be cleared with legal, and at 10 mb it is some ways beyond the size limit.

---

If at all possible this would be much appreciated! One problem that may be causing the CTD is that there are several files in pack.idx/dat that have the same exact filename/path but are not the same file. The only difference is that they have different scale values. So on extraction they get de-scaled and on packing the get re-scaled, so you have the original animations when packed but not unpacked. On the other hand that would just cause some graphical oddities and shouldn't be reason for a CTD. If we could get those .cas files that would be most excellent.

The .evt files contain sound event markers linked to various animation frames.

There are some fairly subtle game balancing facts involved with the animations, which i should probably mention. For example there is only one attack per animation, so the length of the attack animation determines attack frequency for that troop type. The attack delay values in export_descr_unit.txt are used to balance this, so that it doesn't become too restrictive for artists.

About the scaling... generally things with the same name are referring to *one* file on disk which is getting rescaled at load time by the amount stated. The requested scaling factor is stated in the animations text file, and is preserved so that the memory-saving device which maps loaded animations into skeletons can distinguish between identical animations at different scales.

The other thing to beware of is named nodes. The animation system looks for a number of named nodes in animations, namely these ... bone_H_Saddle, bone_camel_root, bone_E_platform, bone_wheel, bone_Lhand, bone_Rhand ... for marking a number of things in animations. Saddle and camel root should be self-explanatory - this is where the riders are attached, wheel and platform are used for siege engine animations, and the hand bones are used for various display things such as the pikes. These things not being present might well be at the root of the problem, and I imagine it wont be easy to reconstruct all the data in .cas form from the animations database.

I'll look into making the .cas archive available... it would almost certainly need some sort of EULA though, so it may take some time and the powers-that-be may decide it's not worth the effort.

Crashes while starting up with no error message.
I guess I'm stuck with it just like everyone else.

Oh Jerome, if you could do that, that would be SO WONDERFUL! :) :)

Here's the thing, while like you said the CAS files extracted with Verc's program may indeed be somehow incompatible with the game, thus causing the crash, when they are packed into the archive -- again using Verc's program -- they work properly! That is how we've been able to modify animations using Verc's programs -- the stuff works! It's just that it doesn't work when not packed...

So, a couple of issues: if you could make the CAS files available, that would of course be really wonderful, because one of the things it could possibly do would be to help Verc isolate the differences between the CAS files his script produces and the originals.

However, here's a better solution: would it be at all possible to release the native CA CAS tool? You could release that entirely without support, and thus not bother the powers-that-be too much, and yet really help out the modding community! :)

Jerome, even if you can't help with he CA script, or the entirety of CAS animation files, maybe you could provide just a couple, so that Verc would be able to figure out the differences between them and his own files? That would still help a lot, if for some reason you cannot provide the other things mentioned.

That would be nice. We desperately need a proper CAS import/exporter. The flippering effect is frustrating me :(

Jerome, I'm bumping this thread because we're still hoping for an answer from you, and I don't want it to be too hard for you to find the thread when you are going to do so.

Hey guys, i'm new here. Thanks for the files man.

I guess this is the thread to watch if you want to change the impact frame for animations. I guess everyone is still waiting to see if V's .cas files are somehow different from the ones the game wants when it reads descr_skeleton.txt.

I did the same tests as everyone else with v1.5 and got the same results, so I think that rules out the possibility of a change from 1.2 to 1.3 making it work. If it finds a .cas file, it dies.

PLEASE make those available (or at least just one so V can figure out what the diff is). Making mumakil behave correctly means changing the impact frame of the elephant animations! and we want well behaved mumakil dangit!

shameless bump.

I don't want you to think we stopped caring about getting at least one original .CAS file so that we can do things with animations. Just give us one original and I am sure Vercingetorix (or somebody else) will figure out how to parse it out of the pack file such that reading skeletons.txt won't crash the game.