Help with an anim

[Banzai!]

I have made an animation for a pikemen in milkshape.

I used MTW2_Pike_Advance as my starting point. However my modded animation seems to have kept the first frame of this animation, so each time it repeats the guy's pike flies upwards because that first keyframe is still there.

How do I delete this keyframe? I want the pikemen to begin over at my first keyframe not the vanilla first keyframe.

Here is a link to my .cas animation (bottom of first post in this thread: http://www.twcenter.net/forums/showthread.php?t=176070

Not quite sure how to post attachments in this forum.

[Ibrahim]

I dunno, but I'll look at it. but I have a question for you: can you export .CAS files? I really need help in that department.

[KnightErrant]

@Banzai,
Are you sure what you are seeing is real?
When I animmerge your animation with a unit
I see the glitch you describe but I think you just
have to edit the first edit box in the keyframer so that
the min frame is set to 1 instead of 0. In a hex editor
the .cas looks ok, each bone has 21 position and rotation
frames. Does the anim screw up in game?

[Banzai!]

I'll try it now KnightErrant. I haven't had internet access this past week so yeah.

EDIT: no I can't get it to work. I save with the min frame set to 1 and extract anim using your v1.1 animation utilities. When I merge again I keep getting that up pike glitch.

Also is there a quick way to append one animation to another one?

My Milkshape seems to have farked up and copy/set keyframes from two open windows no longer works.

[KnightErrant]

@Banzai,

You're not the first to ask about appending an animation to another one.
This could be done, can't really commit to it right now, too many things
going on. But I'll put it on the list.

But back to your problem. This is a guess but if you go to the first keyframe
and go to the animate menu item and get it's drop down menu, can you toggle
off the "operate on selected joints only" entry? This might be what is preventing
you from removing the bad keyframe that is messing up your animation. I think
the selected joint only option leaves the keyframe for other joints so you
aren't really getting rid of it. Unfortunately, there isn't a good way to check
keyframes in Milkshape to see what's there and what's not. Max has trackview
where you can see keyframes (and the non-keyframes between them) and that
let's you know if something is wrong. Maybe you can select bones and click
through the keyframer to check if they all have 19 or 21 keyframes and then
your anim will come out ok.

[Banzai!]

I found out how to get rid of the glitch at the start. You have to delete all keyframes first to reset where the animation segues to at the start. However since you can't copy/paste keyframes from other windows that makes modding existing animations damned difficult.

[Banzai!]

Hi KnightErrant. Would it be possible to change your animations utilities so that the anim extractor sets the first keyframe as the initial basepose? That way the glitch would not appear.

[KnightErrant]

Hi Banzai,

Well try this first. I fixed the first keyframes for:
bone_Rupperarm
bone_Relbow
bone_Rhand
bone_Lclavical
bone_Lupperarm
bone_Lelbow
bone_Lhand
bone_LThigh
bone_Llowerleg
bone_Lfoot
those were the obvious frames that had large deltas in the quats
between frames 1 and 2. I uploaded it here:

http://rapidshare.com/files/13037092...ified.cas.html

(By fixing, I mean I duplicated frame 2 to frame 1 and then deleted the orginal frame 1
and then made small changes to the new frame 1 to make the motion "continuous".)

This was on the original .cas file you posted. I think you'll need to change the left arm
a little because the shield is cutting into one of the legs some.

[Banzai!]

Thank you very much.

To show you my gratitude, a link to my mod (which you may find fun): http://www.twcenter.net/forums/showthread.php?t=175571

How did you fix the first frame? There needs to an automated way to do this or get around the problem because otherwise it is nigh on impossible to edit walking animations. No copy/paste keyframe + bugged first frame = animation changes impossible.

[KnightErrant]

@Banzai,

Ah, grasshopper, I have mastered the ways of animations. Actually, I used a
method no one else seems to like using, probably because it's bugged in the
animationutilities_ver1_1. If you would like to learn this method PM me an
e-mail address and I'll send you a copy of two python files that have some new
stuff in them. Basically, there's a button to get a text representation of a .cas
file. You then can edit the keyframes using Notepad and don't have to deal
with Milkshape's keyframer at all. I can post up some screenshots of how you
do this so you can see what I did to make your .cas file not have the glitch. There's
another button to convert the text representation back to a .cas file. That's
what I did, it took maybe 25 minutes since you had 10 bones that needed some
adjustments to the frame 1 entries.

[Banzai!]

Yeah KnightErrant I'm up for that.

I've PMed you.

[KnightErrant]

Ok, here's how to do keyframing work without a keyframer. You run the
GameObjectApplicationToolbox.py script by double-clicking on it in explorer.
Choose the "Convertcastotxt" button (third button down on the right
side). In the File Chooser select the .cas file to convert, in this
case
MTW2_Pike_advance_level.cas

This will produce a .txt file of the same name in the same directory

MTW2_Pike_advance_level.txt

This can be opened in any text editor like Notepad. Here's the top part:

3.210000 38 9 0 1.04999995232 1 0 0 0 0 1 0 102 102 102
 13060   0 
 21 0 0 1 2 3 1 5 6 7 7 6 10 11 12 6 14 15 16 1 18 19
 21  0.00000  0.05000  0.10000  0.15000  0.20000  0.25000  0.30000  0.35000  0.40000  0.45000  0.50000  0.55000  0.60000  0.65000  0.70000  0.75000  0.80000  0.85000  0.90000  0.95000  1.00000 
          Scene_Root   0   0      0   6720   0   1   0 
         bone_pelvis  21  21      0   6720   0   1   0 
         bone_RThigh  21  21    336   6972   0   1   0 
      bone_Rlowerleg  21  21    672   7224   0   1   0 
          bone_Rfoot  21  21   1008   7476   0   1   0 
            bone_abs  21  21   1344   7728   0   1   0 
          bone_torso  21  21   1680   7980   0   1   0 
           bone_head  21  21   2016   8232   0   1   0 
            bone_jaw  21  21   2352   8484   0   1   0 
        bone_eyebrow  21  21   2688   8736   0   1   0 
      bone_Rclavical  21  21   3024   8988   0   1   0 
      bone_Rupperarm  21  21   3360   9240   0   1   0 
         bone_Relbow  21  21   3696   9492   0   1   0 
          bone_Rhand  21  21   4032   9744   0   1   0 
      bone_Lclavical  21  21   4368   9996   0   1   0 
      bone_Lupperarm  21  21   4704  10248   0   1   0 
         bone_Lelbow  21  21   5040  10500   0   1   0 
          bone_Lhand  21  21   5376  10752   0   1   0 
         bone_LThigh  21  21   5712  11004   0   1   0 
      bone_Llowerleg  21  21   6048  11256   0   1   0 
          bone_Lfoot  21  21   6384  11508   0   1   0

I won't spend too much time here because we don't want to modify any of this.
The top line is the header of a .cas file. The first float, 3.21000, is the version
number. The 1.04999 is really 1.05, the length of the animation in seconds.
(21 frames time 0.05 seconds per frame is 1.05.) The second line with the 13060
is the chunk size in bytes of the file, basically the file size minus the size
of the header and footer. The third line is the bone hierarchy for 21 bones, the
20 usual bones plus 1 for Scene_Root, the root bone found in .cas files.
The fourth line are the time ticks of the animation going from 0.0 to 1.0.
Next come the bone names and several columns of data. First column are names,
second column the number of rotation frames, third is number of position frames,
fourth and fifth are offsets in bytes to where the data is, and then some more columns
which we'll skip.

On to the rotation keyframes. Below the bit I showed above is the actual data
for the rotation keyframes. .cas files store this as quaternions but that's a little
hard for most of us to visualize so I convert the quats to Milkshape Euler angles
in degrees and write out both. Here's the keyframes for bone_pelvis:

0 bone_pelvis  quaternion data and Milkshape euler angles
+0.0359911174 +0.1861125529 +0.0129105933 +0.9817841053      +4.6491146807 -21.3777595644  -2.3847970475
+0.0452478752 +0.1842074990 +0.0366826244 +0.9811598063      +6.2907651619 -20.9868917438  -5.4485520377
+0.0422854945 +0.1819241792 +0.0395931974 +0.9816047549      +5.9786758895 -20.7203367813  -5.7135223964
+0.0189136527 +0.1653417349 +0.0077497289 +0.9860244989      +2.4163796288 -19.0121628189  -1.3053071003
+0.0129362186 +0.1500629038 -0.0063095223 +0.9885717034      +1.4211447607 -17.2690524389  +0.5155516965
+0.0159254130 +0.1302146614 -0.0177047625 +0.9911997914      +1.5993224115 -14.9930216716  +1.8361420388
+0.0239474047 +0.1168485656 -0.0183651950 +0.9926910996      +2.5491034379 -13.4660210383  +1.8187604880
+0.0341287106 +0.1089670956 -0.0024901116 +0.9934561849      +3.9510992297 -12.5139925595  -0.1461489594
+0.0421253070 +0.0961932912 +0.0235358328 +0.9941923022      +5.1587044226 -10.9111392604  -3.2052756560
+0.0453248061 +0.0934590176 +0.0289297961 +0.9941700697      +5.5764212898 -10.5565465746  -3.8491898190
+0.0441512577 +0.0962586179 +0.0138846142 +0.9942797422      +5.2874487927 -10.9638691746  -2.1079071550
+0.0410093963 +0.0985871479 -0.0067886035 +0.9942598939      +4.6923817947 -11.3381643949  +0.3163328743
+0.0448558517 +0.0958925188 -0.0055649662 +0.9943649173      +5.1518473395 -11.0230305656  +0.1438616865
+0.0524983667 +0.1139110848 +0.0320142135 +0.9915862679      +6.5616666861 -12.8584953484  -4.4386018472
+0.0526245795 +0.1174385622 +0.0349323936 +0.9910693765      +6.6375663095 -13.2442627252  -4.8087989540
+0.0510141067 +0.1293480545 +0.0355125479 +0.9896491766      +6.5370054343 -14.6192873806  -4.9496609335
+0.0461298786 +0.1382496357 +0.0264909398 +0.9889677763      +5.8773103327 -15.7240687713  -3.8810415524
+0.0355201550 +0.1435534060 +0.0095693376 +0.9889586568      +4.3656323750 -16.4546872767  -1.7402967689
+0.0212689433 +0.1711403877 -0.0070993681 +0.9849914908      +2.4030483802 -19.7213242850  +0.4081507071
+0.0237362534 +0.1794726700 -0.0048454227 +0.9834645987      +2.7538737899 -20.6856206373  +0.0618923461
+0.0359911174 +0.1861125529 +0.0129105933 +0.9817841053      +4.6491146807 -21.3777595644  -2.3847970475

The first four columns are the quaternions, last three are Euler angles in x, y, and z order, in degrees.
The way it works is you modify the angles and when back converting the script ignores the
quaternions and converts the Milkshape angles back to new quaternions and writes those
into the new .cas file. This is how you can modify the keyframes.

Now bone_pelvis looks fine, no glitches here. In fact, page down through bone_RThigh,
bone_Rlowerleg, bone_Rfoot, bone_abs, bone_torso, bone_head, bone_jaw, bone_eyebrow,
bone_Rclavical, and bone_Rupperarm. They all look fine so I won't show them here.
Now look at bone_Relbow, just the first few keyframes:

11 bone_Relbow  quaternion data and Milkshape euler angles
+0.0027312329 -0.2731546164 +0.0214967914 +0.9617260695      -0.4371054265 +31.7030709383  -2.6850724406
-0.0711576417 -0.7427593470 +0.0818803832 +0.6607123017     -117.7589909519 +75.8942685098 -118.6344709393
-0.0711576343 -0.7427592874 +0.0818803757 +0.6607123613     -117.7589636311 +75.8942753400 -118.6344436185
-0.0711576343 -0.7427592874 +0.0818803757 +0.6607123613     -117.7589636311 +75.8942753400 -118.6344436185
-0.0711576343 -0.7427592874 +0.0818803757 +0.6607123613     -117.7589636311 +75.8942753400 -118.6344436185

The red first frame is definitely part of the "glitch" you're seeing. So copy the second
key frame and delete the first key frame so you have

11 bone_Relbow  quaternion data and Milkshape euler angles
-0.0711576417 -0.7427593470 +0.0818803832 +0.6607123017     -117.7589909519 +75.8942685098 -118.6344709393
-0.0711576417 -0.7427593470 +0.0818803832 +0.6607123017     -117.7589909519 +75.8942685098 -118.6344709393
-0.0711576343 -0.7427592874 +0.0818803757 +0.6607123613     -117.7589636311 +75.8942753400 -118.6344436185
-0.0711576343 -0.7427592874 +0.0818803757 +0.6607123613     -117.7589636311 +75.8942753400 -118.6344436185
-0.0711576343 -0.7427592874 +0.0818803757 +0.6607123613     -117.7589636311 +75.8942753400 -118.6344436185

The new red first frame replaces the bad keyframe but you still have 21 frames for this bone.

The same problem occurs for bone_Rhand:

12 bone_Rhand  quaternion data and Milkshape euler angles
-0.0711576343 -0.7427592874 +0.0818803757 +0.6607123613     -117.7589636311 +75.8942753400 -118.6344436185
+0.0101752775 +0.1596707702 -0.2204553485 +0.9621856213      -3.0659306189 -18.1648590693 +26.2998995312
+0.0101752775 +0.1596707702 -0.2204553485 +0.9621856213      -3.0659306189 -18.1648590693 +26.2998995312
+0.0101752775 +0.1596707702 -0.2204553485 +0.9621856213      -3.0659306189 -18.1648590693 +26.2998995312
+0.0101752775 +0.1596707702 -0.2204553485 +0.9621856213      -3.0659306189 -18.1648590693 +26.2998995312

Same solution, copy the good second keyframe values and delete the bad first frame so that
you end up with 21 good frames. Repeat for bone_Lclavical, bone_Lupperarm, bone_Lelbow,
and bone_Lhand.


Now we come to bone_LThigh. This gets interesting so I'll show the whole entry:

17 bone_LThigh  quaternion data and Milkshape euler angles
-0.0081541529 -0.0539951548 -0.0931290835 +0.9941554070      -0.3548216840  +6.2506369468 +10.6839364328
-0.3135873079 -0.2903384566 -0.0319196284 +0.9035196304     -40.8125488759 +33.0022651260  -8.5312708347
-0.2785103023 -0.2977492511 -0.0440053456 +0.9120531678     -35.8209749135 +34.5856908058  -5.9668340490
-0.2299130857 -0.3079921305 -0.0393587835 +0.9223511815     -29.5772531063 +35.8914897864  -4.8871446389
-0.1462429315 -0.3401504755 -0.0080502890 +0.9288949370     -20.1416985227 +39.3668027343  -6.2774629416
-0.1036637649 -0.3524932563 +0.0015064689 +0.9300537705     -14.8742378518 +40.9471821497  -5.7660337445
-0.0637732372 -0.3406051695 +0.0086891297 +0.9380008578      -9.3828818723 +39.6329956968  -4.4491702843
-0.0332119726 -0.3033235073 +0.0076715224 +0.9522777796      -4.7701200196 +35.2529251557  -2.4395102779
+0.0007814851 -0.2596846521 -0.0071375719 +0.9656667709      +0.3454649384 +30.1010466534  +0.9398636984
+0.0743451640 -0.2032247186 -0.0318042003 +0.9757874012      +9.8898415644 +23.0714825540  +5.7569564231
+0.1082658917 -0.1750962585 -0.0469822362 +0.9774520397     +13.9945419301 +19.3976979692  +7.9071768134
+0.0830243677 -0.1645217985 -0.0565966964 +0.9812422395     +11.0224474382 +18.2686950202  +8.3797994370
-0.0152230021 -0.1544868201 -0.0668562204 +0.9856126904      -0.5628938187 +17.8522737544  +7.6726921504
-0.1626888365 -0.1362153441 -0.0832804590 +0.9736745358     -17.9123264852 +16.9990840894  +7.0790761407
-0.3695906103 -0.1503919512 -0.0903780982 +0.9124783874     -43.5228499008 +19.9539400375  +3.2792969774
-0.3874243796 -0.1564243883 -0.0880874693 +0.9042534828     -45.9621085289 +20.5576126469  +2.3321753492
-0.3960159719 -0.1793194711 -0.0696822107 +0.8978642821     -47.8058600388 +22.1603284885  -1.0453660020
-0.3553634286 -0.1639790386 -0.0590253100 +0.9183374643     -42.3970639709 +20.0675073004  -0.4961728490
-0.3606556952 -0.1648182273 -0.0403692648 +0.9171329141     -43.3983492126 +19.3561550512  -2.7239214894
-0.3649627268 -0.2295900285 -0.0247732513 +0.9019295573     -45.8439872374 +25.6092086042  -7.8328916760
-0.3515519798 -0.2539477348 -0.0253855959 +0.9007093310     -44.8413871842 +28.3798560580  -8.6828660295

We can't just copy the second keyframe here because the frames are changing. This is a
looping animation so we want to "linearly interpolate" between the last keyframe and the
second keyframe. So for the x rotation I'll pick -42.8 as being halfway between -44.84
(the last keyframe x) and -40.81 (the second keyframe). Similarly for the y and the z.
So my new first keyframe would be something like:

17 bone_LThigh  quaternion data and Milkshape euler angles
-0.0081541529 -0.0539951548 -0.0931290835 +0.9941554070     -42.8548216840 +31.2506369468  -8.6139364328
-0.3135873079 -0.2903384566 -0.0319196284 +0.9035196304     -40.8125488759 +33.0022651260  -8.5312708347
-0.2785103023 -0.2977492511 -0.0440053456 +0.9120531678     -35.8209749135 +34.5856908058  -5.9668340490

That's just roughing it out for the first few significant digits, little half degree differences won't
be noticeable.


bone_Llowerleg requires the same treatment:

18 bone_Llowerleg  quaternion data and Milkshape euler angles
-0.3135873377 -0.2903382778 -0.0319198929 +0.9035196304     -40.8125352155 +33.0022514656  -8.5312298536
+0.2878805399 +0.0901421756 +0.0485884547 +0.9521755576     +34.2514418384  -8.2612546738  -8.3919640039
+0.2880963683 +0.1072372198 +0.0556050837 +0.9499519467     +34.5908544288  -9.8867432198  -9.7853567456
+0.2766011953 +0.1221724972 +0.0595879853 +0.9513227344     +33.4993458032 -11.5069460532 -10.6420537128

should go to something like:

18 bone_Llowerleg  quaternion data and Milkshape euler angles
-0.3135873377 -0.2903382778 -0.0319198929 +0.9035196304     +35.8125352155  -6.0022514656  -7.0312298536
+0.2878805399 +0.0901421756 +0.0485884547 +0.9521755576     +34.2514418384  -8.2612546738  -8.3919640039
+0.2880963683 +0.1072372198 +0.0556050837 +0.9499519467     +34.5908544288  -9.8867432198  -9.7853567456
+0.2766011953 +0.1221724972 +0.0595879853 +0.9513227344     +33.4993458032 -11.5069460532 -10.6420537128

And bone_Lfoot:

19 bone_Lfoot  quaternion data and Milkshape euler angles
+0.2878805101 +0.0901420563 +0.0485892780 +0.9521755576     +34.2514418384  -8.2612136926  -8.3920502351
-0.0289541371 +0.0109546594 -0.0896497965 +0.9954921007      -3.4179801265  -0.9522474717 +10.3202792321
-0.0662014484 +0.0151521685 -0.1078526527 +0.9918445349      -7.7359670228  -0.9040037112 +12.4729815424

should go to something like:

19 bone_Lfoot  quaternion data and Milkshape euler angles
+0.2878805101 +0.0901420563 +0.0485892780 +0.9521755576      -3.0514418384  -0.7812136926  +8.3920502351
-0.0289541371 +0.0109546594 -0.0896497965 +0.9954921007      -3.4179801265  -0.9522474717 +10.3202792321
-0.0662014484 +0.0151521685 -0.1078526527 +0.9918445349      -7.7359670228  -0.9040037112 +12.4729815424

If you save your changes and back convert using the converttxttocas button (fourth down on right
hand side) you should have a .cas file with no glitches. You can animmerge to check it out.


While on the subject, after the rotation data for bone_Lfoot comes the position animation frames
for bone_pelvis, shown below:

0 bone_pelvis  animation data and deltas
+0.0000089781 +0.9089912176 -0.0000567846
-0.0119566014 +0.9362711310 +0.0770919174
-0.0188270491 +0.9536758661 +0.1445211619
-0.0227415580 +0.9556493163 +0.2371863872
-0.0248528998 +0.9503000379 +0.2990277708
-0.0260876603 +0.9372131228 +0.3798140585
-0.0218601450 +0.9144630432 +0.4702308476
-0.0109665627 +0.8891604543 +0.5713204145
+0.0079560671 +0.8697217107 +0.7156654000
+0.0203840956 +0.8796760440 +0.8177018166
+0.0290276278 +0.9047831297 +0.8947846889
+0.0362256058 +0.9322296977 +0.9549288154
+0.0413517728 +0.9499911070 +1.0092128515
+0.0424700268 +0.9463566542 +1.1409735680
+0.0423726998 +0.9442256689 +1.1588103771
+0.0409461260 +0.9320024848 +1.2297110558
+0.0356029756 +0.9121266007 +1.3017691374
+0.0285389069 +0.8945383430 +1.3617962599
+0.0109724132 +0.8757258058 +1.5074499846
+0.0064266389 +0.8848003149 +1.5499432087
+0.0002717352 +0.9089912176 +1.6206053495

These are in x, y, z order. Note the third column, the z's. The z-axis points out of the computer
screen so these values are what makes the unit move forward, in fact the unit moves 1.62 meters
in 1.05 seconds. You can adjust these slightly higher to make the unit move faster or lower
to slow it down. (You can't adjust too much or the unit will look like it is speed skating.)

Likewise, if you add 10.0 to each entry in the second column, the y-axis or vertical axis, you can
make your unit fly 10 meters above the ground.

Note that the position entries for all the other bones are zero or nearly zero. This is the
usual way CA does animations, only bone_pelvis ever moves, all other bones simply rotate relative
to bone_pelvis.

Finally, the last bit of data in a .cas file is the basepose data:

0    skeleton pose data, all bones including Scene_Root
+0.0000000000 +0.0000000000 +0.0000000000
+0.0000000000 +0.0000000000 +0.0000000000
+0.0952388123 +0.0007522846 -0.0000000077
+0.0225613099 -0.4644489586 +0.0143959792
+0.0241626594 -0.3995067179 -0.0316339023
-0.0000000069 +0.2124620527 +0.0000000008
-0.0002945330 +0.2115577906 +0.0000000299
-0.0000617082 +0.2349731028 +0.0000000668
+0.0003562596 +0.0108105801 -0.0034470388
+0.0016836961 +0.1178482771 -0.0744605809
+0.0132546443 +0.1300113499 -0.0273839142
+0.1653589010 -0.0517836586 +0.0034832826
+0.3022063971 +0.0111386608 -0.0137691945
+0.2838369906 -0.0030556759 +0.0263375100
-0.0102220895 +0.1300113499 -0.0273839142
-0.1678023189 -0.0517838039 +0.0034833583
-0.3021733165 +0.0111954408 -0.0144314393
-0.2838012278 -0.0032003860 +0.0267044436
-0.0952387825 +0.0007523639 +0.0000000220
-0.0216093510 -0.4641438425 +0.0230784956
-0.0250775032 -0.3986377716 -0.0406115167

This is where the skeleton for human units exists, it isn't in the .mesh file. All .cas files for
humans have this exact skeleton.


And since we've gone this far, the very last thing in a .cas file is the footer, a collection
of ints, floats, and a node string:

104 1 1 CaozSceneCustomAttribNode 1 0 1 0 0 0.00000 0.00000 0.00000 1.00000 0.00000 0.00000 0.00000 0 0 -1 0 0 1 16 2 0 0 16 3 0 0 16 8 0 0 16 10 0 0 12 5 0 12 12 0

And that is the basics for keyframing without a keyframer.

[Banzai!]

Btw KnightErrant why does your animationutilities come up with an error when trying to extract/merge horse rider animations?

[KnightErrant]

Hmmm, wasn't aware that it did. Tried the old one and the new one I sent you
on HR_lance_Attack_stab_high_fail.cas and it worked fine. Let me know which
.cas file produced an error and I'll take a look at it.

[Banzai!]

Horse archer anims.

[Banzai!]

HR_Bow_attack_missile_ready:

animationsutilities:

"in getcasfooter
int_zero2 = int(tokens[16])
ValueError: invalid literal for int() with base 10: '+0.00000'
"
Also GOAT totally borks up any animations it touches. Ppl go floppidy flop with GOAT.