Wardogs + new skeleton problem/warning

[Eothese]

Some of you may remember the wardog beta I put out some time ago, this took a lot of work and would not of been possible without significant help from KnightErrant and Bwian. A problem exists with the wardogs, that I thought was confined to custom battle. It seems after all that this may not be the case. Further it would seem the same problem exists with new skeleton's applied to human models. Dog statues and a CTD or half buried jesus syndrome men.

I believe this problem is caused by the way M2 uses the victim anims in conjunction with a fatality anim in descr_skeleton. The game tries to appy the anim designed for a standard human skeleton to a non-human one. In the case of the dog skeleton this is more sever than one with an added bone, for animated bowstrings for example.

The only glimmer of hope in resolving this, comes from the fact that these fatality/victim anims are not applied to cavalry or elephant riders. If a way can be found of getting the game to treat wardogs as such, this problem can be overcome. Simply using an elephant skeleton and its anims does not work.

So, does anyone know why the game doesn't apply fatality/victim anims to cavalry or elephants? Help really, really appreciated.

[Matthias the great]

is there somewhere scripts or things to look out for such wardogs?
can it be possible to "export" such files from rome to medieval?

[Taiji]

So, does anyone know why the game doesn't apply fatality/victim anims to cavalry or elephants? Help really, really appreciated.

I know this one, the impact delta (ID) is set too low for the anims to work against cav. But I suspect that if you raise it you will only see the first move, not the fatality.

Worked this and other things out by examining each melee attack animation in isolation, to know what each was. This was so I could make sets of animations that varied by troop quality (works very nicely for the 2h swords and axes). I was also testing them against cav, to see why some worked and some didn't, it turned out to be the ID. So I worked out that by raising the ID and rebuilding I could get a nice slashing move working on cav for my axes, no more prodding

[Eothese]

But the HR and CR anim sets have the no_deltas tag at the start and as their attacks do have ID positions, I assume the no_deltas tag means that impact deltas are not used for attacks against them? Either way the no_deltas tag doesn't solve the dog problems...

[Taiji]

I just answered your question, and HR anims have no fatalities... I don't really understand what the problem you're describing is. It sounds like you have working wardogs that mess up when they should be attacking or something like that... sorry, it's hard for me to work out.

edit:

But the HR and CR anim sets have the no_deltas tag at the start and as their attacks do have ID positions, I assume the no_deltas tag means that impact deltas are not used for attacks against them? Either way the no_deltas tag doesn't solve the dog problems...

I think I get you. I think it's the opposite way round to what you say there, no_deltas probably means that they don't use attack deltas not that attack deltas are not used against them. Since like I said in a previous post attack deltas are used against them. If the attack delta is incorrect then the anim will not work.

Dog statues and a CTD or half buried jesus syndrome men.

That's all you've actually mentioned about the issue so far. Are they constantly like this or only when they attack? Some more info would be cool

As a bit of insight, perhaps lowering the height of the dogs would be necessary. This way you could force a hitting downwards animation from infantry engaging them, etc. it would allow you to avoid fatalities being performed on your dogs altogether. Can it be done? I don't know anything about wardogs. I know how to adjust the height of soldiers and mounts, and how to suitably change an impact delta, but that's it. Where would the height of your wardogs be adjusted?

[Eothese]

Sorry Taiji, I thought I'd been clear enough in describling the problem, which to sum up, is when an opponent kills a dog (or any non-human skeleton) by way of a fatality, when the victim anim is applied to dog there is a spinning and distorting of the model. When this is over the default anim kicks in and you are left with a dead upright dog or a half buried jesus syndrome man.

The reason I posted as I did in my second post was due to a little confusion, are you saying with your changes you now get fatality anims applied to cav, but no victim anims?

Anyhow, I will try as you suggest and alter the height of the dogs, I can simply scale the model down. I'm not sure the hidden attribute 'height' would be much use for the dogs because of the way they are set out in EDU.

#Edit
No change with scaled down dogs

[Taiji]

Fatalities have a first hit, the hit that shows when they fail to kill. That's all I saw when I tested it. One hit, dead horse.

Maybe you read this before: http://forums.totalwar.org/vb/showthread.php?t=90707

The descr_animals thingy would give height adjustment...

[Eothese]

Thanks again Taiji, I hadn't seen that thread. So I've tried a variety of approaches now, all testing done with dogs v's billmen (MTW2_2H_Axe)

Scaled down dogs;- No change

Reduced height in descr_animal by small increments;- Dogs never get killed in melee, or no change

Given the fatality attack an ID higher than the dogs height (it had none at all);- No change

Swapped the victim attack for a dog death anim;- No spinning but still a dog statue. (Causes a reversal of the problem, ie when a MTW2_2H_Axe fatality kills a man, you get jesus dying, but it proves my theory in a additional way)

Gave the dogs a fatality and victim anim entry from knifeman, with the fatality a dog attack, the victim as was;- No more CTD's!!! But still a spin and statue effect on any fatality against the dogs.

Made a dog death anim the same length as the victim anim;- No change

In conclusion, I'm still not sure why the fatality/victim anims are not applied to cav and elephants, other than perhaps their EDU tag as cavalry...

[Taiji]

When I was examining them this animation stood out as being for a reduced height target:

;Axemen do not attack - maybe test on a severe slope?
anim		eager_attack_centre_lo_c_stab_fail						data/animations/MTW2_2HSwordsman/MTW2_2HSwordsman_at_lo_c_stab_fail.cas						-fr
anim		eager_attack_centre_lo_c_stab_success					data/animations/MTW2_2HSwordsman/MTW2_2HSwordsman_at_lo_c_stab_success.cas					-fr	-id:0.158,-0.150,1.774	-if:18	-evt:data/animations/MTW2_2HSwordsman/MTW2_2HSwordsman_at_lo_c_stab_success.evt

This one is clearly anticav with it's high ID and name including 'overhead':

; Anti cav stab + l-r slash vs inf
;anim		eager_attack_centre_overhead_c_stab_fail				data/animations/MTW2_2H_Axe/MTW2_2H_Axe_attack_centre_Hi_Stab_Fail.cas						-fr
;anim		eager_attack_centre_overhead_c_stab_success				data/animations/MTW2_2H_Axe/MTW2_2H_Axe_attack_centre_Hi_Stab_Success.cas					-fr	-id:0.202,0.839,1.552	-if:17

This is a fatality, note the ID for success.

; Flipping opponent. starts with a slow L-R downward slash.
;anim		eager_attack_centre_mid_b_slashlr_fail					data/animations/MTW2_2HSwordsman/MTW2_2HSwordsman_at_mid_c_slashlr_v0_fail.cas				-fr
;anim		eager_attack_centre_mid_b_slashlr_fatality				data/animations/MTW2_2H_Axe/MTW2_2H_Axe_attack_centre_mid_c_slashlr_v0_fatality.cas			-fr									-evt:data/animations/MTW2_2H_Axe/MTW2_2H_Axe_attack_centre_mid_c_slashlr_v0_fatality.evt
;anim		eager_attack_centre_mid_b_slashlr_success				data/animations/MTW2_2H_Axe/MTW2_2H_Axe_attack_centre_mid_c_slashlr_v0_success.cas			-fr	-id:0.368,-0.217,1.700	-if:38	-evt:data/animations/MTW2_2H_Axe/MTW2_2H_Axe_attack_centre_mid_c_slashlr_v0_success.evt
;anim		eager_attack_centre_mid_b_slashlr_victim				data/animations/MTW2_2H_Axe/MTW2_2H_Axe_attack_centre_mid_c_slashlr_v0_victim.cas			-fr									-evt:data/animations/MTW2_2H_Axe/MTW2_2H_Axe_attack_centre_mid_c_slashlr_v0_victim.evt

By reducing their height so much that no animations are working against them, we can then work out which animation works first (hits lowest) by gradually raising their height...

A better idea!: If your 'eager_attack_centre_lo_c_stab_success' animation is missing for axe, which it should be by default, then take it from 2h sword. It may not be the ID at all but the low hitting move may require a '...lo_c_stab'. Assuming this get's it working you'd have to set up every animation set with this attack and a replacement animation (anything should work).

I hope that last helps!

[Eothese]

This is a fatality, note the ID for success.

; Flipping opponent. starts with a slow L-R downward slash.
;anim        eager_attack_centre_mid_b_slashlr_fail                    data/animations/MTW2_2HSwordsman/MTW2_2HSwordsman_at_mid_c_slashlr_v0_fail.cas                -fr
;anim        eager_attack_centre_mid_b_slashlr_fatality                data/animations/MTW2_2H_Axe/MTW2_2H_Axe_attack_centre_mid_c_slashlr_v0_fatality.cas            -fr                                    -evt:data/animations/MTW2_2H_Axe/MTW2_2H_Axe_attack_centre_mid_c_slashlr_v0_fatality.evt
;anim        eager_attack_centre_mid_b_slashlr_success                data/animations/MTW2_2H_Axe/MTW2_2H_Axe_attack_centre_mid_c_slashlr_v0_success.cas            -fr    -id:0.368,-0.217,1.700    -if:38    -evt:data/animations/MTW2_2H_Axe/MTW2_2H_Axe_attack_centre_mid_c_slashlr_v0_success.evt
;anim        eager_attack_centre_mid_b_slashlr_victim                data/animations/MTW2_2H_Axe/MTW2_2H_Axe_attack_centre_mid_c_slashlr_v0_victim.cas            -fr                                    -evt:data/animations/MTW2_2H_Axe/MTW2_2H_Axe_attack_centre_mid_c_slashlr_v0_victim.evt

Arrrgh!! I think we have different and contrasting view's as to how combat anims work. IMO the game plays one of the first three for the attacker depending on whether (in the order they appear) a miss, a fatality or a normal hit (success) has been generated. It does not play them one after the other, the success has no relation to the fatality, other than it looks similar. Fatality's are generated by chance, dependant on the last number in EDU line stat_pri. Victim anims are applied to the victim of a fatal attack and both are played at the same time. Fatality's have no ID or IF because they are not needed due to the victim anim being played in conjunction.

I see it this way because I have anim merged these .cas files with the models and played them in MS3D. Commenting out the victim anims allows the normal death anim, for whatever descr_skeleton entry type, to play. The fatality anim still plays but generally goes on long after the killed man has died.

Indeed commenting out the fatality/victim anims solves the dog statue problem altogether, but the game loses something of its self. Hence my quest for an alternate solution.

Introducing more attack anim types won't solve the problem, it will just reduce the chance of it happening. Changing ID is similar to changing strike distances in that the attacker has to be in a certain position to be able to use certain attack anims.

[Taiji]

Are you not going to test the suggestion I've just given you?

To be honest it's pretty annoying that you seem to have ignored it but it's at least partly my fault for including anything else in the post... so here it is again:

If your 'eager_attack_centre_lo_c_stab_success' animation is missing for axe, which it should be by default, then take it from 2h sword. It may not be the ID at all but the low hitting move may require a '...lo_c_stab'.

There I am giving you a potential method to avoid the possibility of a fatality or any other normal move from occuring against your dogs.

Either you already have it working fine or you should try the method I outline there. If it's not clear what you should do, let me know and I'll try to explain further. If you can see a problem with it then tell me what it is, please.

Introducing more attack anim types won't solve the problem, it will just reduce the chance of it happening. Changing ID is similar to changing strike distances in that the attacker has to be in a certain position to be able to use certain attack anims.

Actually it looks to me that this is exactly what you need to do. You are using an animation set which cannot hit a short target. Thus you get no attacks when you make the target small. The exact same thing happens when trying to use an animation set which has no relevant animation for attacking high, no attacking takes place versus cavalry.

edit:

In this post below I was mistaken. I'd forgotten the way I had actually fixed the issue:

I know this one, the impact delta (ID) is set too low for the anims to work against cav. But I suspect that if you raise it you will only see the first move, not the fatality.

Worked this and other things out by examining each melee attack animation in isolation, to know what each was. This was so I could make sets of animations that varied by troop quality (works very nicely for the 2h swords and axes). I was also testing them against cav, to see why some worked and some didn't, it turned out to be the ID. So I worked out that by raising the ID and rebuilding I could get a nice slashing move working on cav for my axes, no more prodding

The way I fixed it was by using the correct animation type. Nothing to do with IDs or actual animations (which are probably just aesthetic). It's been a while, sorry about that

edit: Oh and you don't even need to rebuild if your 2h swords have the anim type set up already. Just make your dogs small enough to not be attacked by your billmen and then test with a 2h sword unit.

[Eothese]

Taiji, I did not mean to cause you any offence. I have tried your method, like I said it only reduces the chance of a fatality happening, other normal attacks are fine. TBH I've spent more than 50 hours trying various methods of fixing this, and I've become something of an obssessive. I realise what seems crystal clear to me may seem doubtfull to others. The statue problem is due to the victim anim being played on a non-human skeleton, the only soloution, other than commenting all these out is to get the game to treat dogs as cavalry.

If you want to have a crack at this yourself, I'd be happy to upload the necessary components...

#Edit

No need, you gave me an idea, all it needed was an increase in height in descr_animals, to match that of a mount. It makes them much weaker but that can be balanced, the point is not one dog statue. I spent far too much time playing with descr_skeleton...

[Taiji]

Yep, that should work too. ...but would make them more susceptible to missiles, I'd imagine.

Sure, upload the stuff, I'll have a crack if I get time

Just to make sure I wouldn't be wasting my time, did you test this exactly?:

...if your 2h swords have the anim type set up already. Just make your dogs small enough to not be attacked by your billmen and then test with a 2h sword unit.

That's the solution I see that should work as well as (or better than) any 'making them cav' based solution.

edit:

You make a very interesting comment here:

Fatality's are generated by chance, dependant on the last number in EDU line stat_pri.

In what way dependant? Can you tell me the results expected from using different vaules? I'd assumed that this was just not used at all but if you have a way to show it affects fatalities and how, I am very very interested!!

[Eothese]

Hmmm, its not flawless, 20 units of dogs v's 20 units of billmen results in 2 staues and a CTD, which I thought had been eliminated. But I think with a little more tweaking I should be able to eliminate them altogether.

[Taiji]

I just checked vanilla's descr_skel, Zweihanders and other 2h swordsmen should be able to attack them where billmen never could when you lower the dogs height. Have you tested this? (+rep for you if you do... Now I'm interested I just have to know!!!)

The reason that it interests me is the suspiciousness of it all. Why have this anim type if it can hit nothing? Why is it missing from every unit bar one? Why have a lo if a hi can achieve the same thing in all respects? Is it left over from RTW? Does RTW have a lo for every animation setup to be able to attack wardogs and pigs? (I can't be bothered to dig out my disks and install it)

You see my mind is bubbling over with all this, I am happy to test it out if you would rather focus on another approach

[Eothese]

Tested, with the height reduced to 0.4 where billmen can only very, very rarley get a kill (2-3), the Zwiehanders fair only a little better, (10-12). They just seem to miss, even though their weapon often penetrates the dog.

As for the why, my guess is its all just down to asthetics and variety. Although fighting on a hill could come into play. In the case of RTW, dogs were very hard to kill before they released a patch, can't remeber which one. But descr_skeleton contains this line in the riders section.

;Bob changed the impact points for low attacks to solve a problem of not being able to hit dogs

Further, and just to add to the confusion, a lot of the infantry sets have their low attack anims commented out...

[Taiji]

Tested, with the height reduced to 0.4 where billmen can only very, very rarley get a kill (2-3), the Zwiehanders fair only a little better, (10-12). They just seem to miss, even though their weapon often penetrates the dog.

That nearly answers my question. If you lower the dogs height to the point where billmen cannot attack(0.39, 0.38, etc.), do the zweihanders still attack?

As for the why, my guess is its all just down to asthetics and variety. Although fighting on a hill could come into play. In the case of RTW, dogs were very hard to kill before they released a patch, can't remeber which one. But descr_skeleton contains this line in the riders section.

;Bob changed the impact points for low attacks to solve a problem of not being able to hit dogs

Interesting. Seems to back up what I have been saying. Also suggests maybe IDs have a more than aesthetic role.

Further, and just to add to the confusion, a lot of the infantry sets have their low attack anims commented out...

I see, in RTW lo anims were the exception. In m2tw they may be required, if dogs are not to be subject to fatalities.

[Eothese]

Interesting. Seems to back up what I have been saying. Also suggests maybe IDs have a more than aesthetic role.

I would say the ID are critical in lining the attacker up with an opponet before an attack takes place. The different attack types (5 in the case of 2 handed axe) are more what I would say are for asthetics.

When I was playing with the dogs attack anim ID's the dogs would enter a kind of 'phasing' or 'teleporting' back to the start point of their attack, when I set the values incorrectly.

But still, problem solved. Now to see how this affects other thing, ie setting a mount's height to a lower value, will this mean fatality anims are applied to cav, or that they will be allowed on walls??

[Eothese]

Both can attack at this height, its just the Billmen miss 99% of the time where the Zwiehanders miss 90%. This suggests to me that the game does not choose which anim to use depending on anything situational, it just randomises between its options.

[Taiji]

If both can attack then you are not testing it. If you lower the dogs height to the point where billmen cannot attack(0.39, 0.38, etc.), do the zweihanders still attack?

What I am asking you to test is whether a lo attack anim type can hit a smaller creature. If billmen are still able to attack then you are not testing it at all.



Anim types are not just aesthetic that's why I am suggesting this test. To reword what I said earlier: If you remove the hi and overhead anim types the soldier will no longer attack cavalry.

I suspect that because zweihanders have a lo attack anim they may be able to hit the dogs when they are made smaller. Billmen use 2haxe and so will be unable to hit. Easy to test... for you.