I'm on ProPack and crosstalk may be an useful bug, but usually it's a major pain in the back. I'm working on a scene where I'd like to put at least 10 instances of same figure but with morphs set on various values. Unfortunately cross-talk makes morphs unuseable from second instance up. How to avoid it ? Would making null figure (or whatever it's called) help ? How to do it ?

Attached Link: http://ockhamsbungalow.com/Python/

Vil, get the Crosstalk Killer at ockham's.

Yes a NULL Figure should solve your problems. The procedure is:

1. Load a NULL figure, select its "Body" actor. Load your first character figure.
2. Load another NULL Figure , select its "Body" actor. Load your next character figure.
3. Repeat step #2 for subsequent figures.

---

The important thing in a NULL Figure is that it should list all the actors (body parts) that might contain a master dial. If your figure uses the standard list of actor names (most do) you should be able to use one of the existing NULL Figures such as one by Rob Whisenant (aka "rbtwhiz"):

http://www.rbtwhiz.com/rbtwhiz_ERC.html#CT

If your figure does not use standard names for the actors you may have to build your own NULL Figure.

The NULL Figure needs to list all the actors in the format:

{<br></br><br></br>version<br></br>       {<br></br>   number 4.01<br></br> }<br></br><br></br>actor BODY:1<br></br>       {<br></br>   }<br></br>actor hip:1<br></br>    { <br></br>  }<br></br>actor abdomen:1<br></br>        {  <br></br> }<br></br>[etc, etc]

Then you need a "figure" section with all the "addChild" stuff in it, you can leave out the stuff that follows the addChild section. Thats it really.

You can make a couple of improvements:

actor BODY:1<br></br>       {<br></br>   name    GetStringRes(1024,1)<br></br>        }

This will insure that you see "BODY" when you select the Body, rather than seeing "No Actor".

You can also add this just above the closing brace of the file, it will insure that the BODY actor of the NULL Figure in automaticly selected when the NULL is loaded:

doc<br></br>  {<br></br>   pickActor BODY:1<br></br>    }

I used to use the one available via DAZ. I think it was a link from their tutorial section, or maybe the tech support. Since moving to P5 and P6 I don't miss the "convenience" of crosstalk at all. Copy and paste for clothing morphs works well enough and it's at least controllable. I now wish they'd sort out the multiple figure injection problem.

Attached Link: Open letter to DAZ Developers

Quote: "I now wish they'd sort out the multiple figure injection problem." I think it is sorted, just not sure DAZ realises yet. Have you seen the above thread by Maveris?

Nope, but I'll get my ass (or online equivalent thereof) right over and take me a look. Ta for the link.

Will that Crosstalk killer of Ockham's work with poor old fashioned P4.03..or is it strictly PP and above ?

The Crosstalk killer is a Python script so it won't work on p4.

Thank you everybody for advices and suggestions so far :) The figure I want to use is Koroko, The Anime Pet by Posermatic. The critter is buildt in a bit unusual manner - hip is connected straight to head, without abdomen or neck, so I'm not sure if regular null figure would help. Also, Koroko has his eyes buildt-in as a prop into .cr2, and they morph too. Should I leave them or throw them out of the null figure .cr2 ? It looks like this: prop Korokoeyes:4 { geomCustom { What to do with this part ? Off to try out Ockham's Python thingy ;)

Attached Link: http://www.hogsoft.com

To enable or disable crosstalk in P4/PP/P5/P6 (obviously killing crosstalk is redundant in P5/6 ;-) get Crosstalker from Hogsoft. It's a freebie to the Poser community. In order to work with P4 it uses P4PyE rather than Python, so it is limited to PC only (not Mac). To extend lesbentley's input, Figure Independent Injection is now also supported as a plug-in to PBooost II also by Hogsoft. No limit to number of Mil3 figures injected per scene.

"Crosstalk killer" script doesn't work for Koroko. Before registering on yet another forum, I'd like to ask again - how to make a null figure out of something with morphing prop embedded (= geomCustom) into .cr2 ?

Quote:

"The critter is buildt in a bit unusual manner - hip is connected straight to head, without abdomen or neck, so I'm not sure if regular null figure would help"

I suspect that the parenting order is irrelivent as far as the NULL Figure is concerned, but I haven't checked this. Just so long as the actor names are correct, these are case sensitive, I think it should work.

My advice would be the try one of the standard null figures first, see if that works. If not you can try the other solutions offered in this thread, or make your own NULL Figure.

Re the customGeom. Leave "geomCustom" out, just use:

prop Korokoeyes:1<br></br>        { <br></br>  }

Message edited on: 03/08/2006 13:12

None of regular null figures available works, I have to make my own. Thanks for the advice on Koroko's eyes, I'll try it too. Thank you ! Off to work :)

I'm depressed. Nothing works as it should. I made .cr2 according to all the advices, I compared it to existing null figures and it looks similar, I loaded it into Poser according to the route described in post #3, and... body morphs in any figure other than the first one do not work at all (eye morphs seem to be fine though). Waaaaah ;( Where's the mistake ? Somebody help ? { version { number 4.2 } actor BODY:2 { name GetStringRes(1024,1) } actor hip:2 { } actor head:2 { } actor lShldr:2 { } actor lForearm:2 { } actor lHand:2 { } actor rShldr:2 { } actor rForearm:2 { } actor rHand:2 { } actor lButtock:2 { } actor rButtock:2 { } prop Korokoeyes:2 { } prop ThunderTail_1:2 { } figure { name Figure 1 root BODY:2 addChild hip:2 BODY:2 addChild head:2 hip:2 addChild lShldr:2 head:2 addChild lForearm:2 lShldr:2 addChild lHand:2 lShldr:2 addChild rShldr:2 head:2 addChild rForearm:2 rShldr:2 addChild rHand:2 rShldr:2 addChild lButtock:2 hip:2 addChild rButtock:2 hip:2 addChild Korokoeyes:2 head:2 addChild ThunderTail_1:2 hip:2 inkyChain lLeg { off name lLeg } defaultPick hip:2 displayOn 1 } displayMode USEPARENT locked 0 } }

Try the file attached above, see it that works. Copy the text and paste it into a text editor, save it as "TEST_NULL_01.cr2".

I regards to props in NULL Figures, I think there is a problem. The NULL will refer to a specific prop/number, eg box_1, but if you load another instance of the box it will be named box_2, so the NULL will not refer to it. If you wanted to have 3 instances if the box in the scene, you would need to include box_1, box_2, and box_3, in the NULL.

If the prop is contained in the same cr2 as the figure, there is a workround. Just convert the prop to an actor. Open the cr2 in a text editor, replace all instances of the srtring "prop box_1" with "actor box_1".

Of course all this is only necessary if the prop contains a master channel. If the prop does not contain a master it can be left out of the NULL Figure. Things that are slaved don't need to be in the NULL, only things that contain master channels.

In the file attached to my last post I included 4 instances of each prop, this should be enough to test, but re-reading your first post "I'd like to put at least 10 instances of same figure ...", not enough for that. I choped some stuff of the end of your file, I think there may have been a missmatched brace there, but not sure. P.S. I never tried putting a prop in a null figure, I'm just going on theory here, but I think it should work.

Message edited on: 03/08/2006 16:47

Can't believe it... This doesn't work either O.o Again followed the instructions, used three instances of your new NULL and three Korokos in total, in proper order, with BODY selected. No effect other than morphs not working (eyemorphs still seem to be fine, but not bodyshaping morphs) :( Thanks for the help anyway Les. You've tried hard and I appreciate it. I've learned a lot and while this knowledge doesn't seem to be useable in this case, maybe someday later it'll help with another project. Now off to do a whole lot of renders and to compose them into one :(

Attached Link: TestNull2.zip

Hi Vilian,

I was concerned surprised to hear that the NULL is not working for you. I thought that perhaps there was some defect in my method, or in my understanding of Null Figures. So I decided to run a test.

I constructed a figure out of primitives "Z-Pet.cr2", using the same actor names and hierachy as your NULL in post #14. I made a morph in each actor and prop, then made an FBM in Poser. I then edited the cr2 and slaved all the morphs to another master dial in the hip named FBM2.

Loaded TEST_NULL > Z-Pet > TEST_NULL > Z-Pet > TEST_NULL > Z-Pet > TEST_NULL > Z-Pet. Everything worked fine, no crosstalk in the actors or props.

So I'm wondering why the null seems to work for me in P4, but not for you in PP. Is it the way you are using it? I don't think so, I sounds like you are using it correctly. You are selecting the BODY of the Nnull after loading it aren't you?

Is it something about the Anime Pet? But I can't see that either, as the actor names and hierachy are the same in the Anime Pet and the Z-Pet I made to test the null, I can't think what else would make a diffrence.

I'm linking a zip file with the "Z-Pet.cr2" (I didn't bother making it properly posable, as I only wanted to check for crosstalk), and a slightly modified version of the null from post #15 "TEST-NULL-2.cr2", you should not need to select the BODY when useing this null, as it should be selected automaticly when the null is loaded. If you could test the combination of the "Z-Pet.cr2" and the "TEST-NULL-2.cr2" and report back, it would rule out some possibilities and perhaps help us in tracking down the problem. There are two master dials in the character "FBM" in the body, and "FBM2" in the hip, both control the same set of morphs.

Can you give any more details on the crosstalk you are getting with the Anime Pet, is this master dials in the BODY controling morphs in the actors and props, or something else?

P.S. It turned out that it was not necessary to use multiple instances of the props in the null, one instance works ok.

P.P.S. This is bugging me, I want to know why it's not working!

I've finished the pic as a multilayered composition - every instance of Koroko as a separate render. I'd like to know too if this always gonna be like that :( The problems occur with FBMs controlled from BODY. Morphs on separate bodyparts are working fine, but it takes too much time with setting values "by hand", especially when using combination of morphs. Eyes morphs work just fine all the time too. I did an experiment with your Z-Pet.cr2 and new NULL. With three instances of both (loaded according to instructions), everything worked perfectly fine. With three Korokos and three nulls, everything blows up - I can set morphs on the Koroko #1, but when trying to do that on any other instance, it automatically take shape of the #1, and nothing else works. So far, this is the weirdest thing I've seen while working with ProPack, and I have no idea how to solve it :(

Sorry to take so long getting back to you, but I have been having connection problems. This does sound rather strange. Things to check. There is just a slight chance that the root actor in Koroko is not using the internal name "BODY", can you check that it is? Do a search for the string "linkParms", and report back if you find any instance of it in the file. The morphs are loaded with the figure, right? Not injected into it? Finally, can you post the block of slaving code from one of the targetGeom channels that is crosstalking? It will probably look something loke this:

                   valueOpDeltaAdd
                                Figure 1
                                SomeActor:1
                                SomeChannel
                        deltaAddDelta 1.000000

Sorry for not responding for so long - usually I'm not home around the weekends. 1. CR2 starts with: { version { number 4.2 } figureResFile :Runtime:Geometries:Koroko:Koroko.obj actor BODY:2 { I guess that means that the "root actor" is the BODY ? 2. No "linkParms" in the file. 3. Yes, the morphs are build-in, no injection mess. 4. Is this the stuff you asked for ? channels { targetGeom Ball { name Ball initValue 0 hidden 0 forceLimits 4 min -100000 max 100000 trackingScale 0.000790989 keys { static 0 k 0 0 } interpStyleLocked 0 valueOpDeltaAdd Figure 2 BODY:2 PBM-Ball deltaAddDelta 1.000000 indexes 155 numbDeltas 169 deltas Hope this helps in solving the mystery.

Well, everything from your last post seems to check out ok. I'm sorry, but those were my last desperate attemps to find some reason why the null is not working. As far as I can see it should work, but it doesn't. I'm out of ideas! If I think of anything else I'll let you know. Right now I'm stumped :(

I think I see the problem. Sometimes you need a slightly diffrent Shield figure cunstruction. You have to keep all the body parts and delete all the channels from them. (You'll probably wear out your delete key!)

{
version
{
number 4.01
}
actor hip:1
{
*A totally empty geometry designation
}

...

• This is where the second geometry line usually is

actor hip:1
{
name Fog Control
on
bend 0
dynamicsLock 1
hidden 0
addToMenu 1
castsShadow 1
includeInDepthCue 1
parent BODY:0
endPoint 0 0.021 0
origin 0 -0.079 0
orientation 0 0 0
displayOrigin 0
displayMode USEPARENT
customMaterial 0
locked 0

• Look Ma No Channels!

}
...

figure
{
name EMC Shield
root BODY:0
addChild hip:1
BODY:1

...
}
}

Something like this. See in the middle where the Actor is included but all the channels have been removed.

Nerd3D

Nerd,

Thanks for dropping by.

Quote:
"You have to keep all the body parts and delete all the channels from them"

We have tried a Shield figures with no channels, vilian's one in post #14, and my one linked in post #18 in "TestNull2.zip". The format is:

{<br></br><br></br>//"TEST-NULL-2.cr2"<br></br>version<br></br>  {<br></br>   number 3<br></br>    }<br></br><br></br>actor BODY:1<br></br>       {<br></br>name GetStringRes(1024,1)<br></br>      }<br></br>actor hip:1<br></br>    {<br></br>   }<br></br>actor head:1<br></br>   {<br></br>   }<br></br>[stuff left out here for reasons of space]<br></br>actor rButtock:1<br></br> {<br></br>   }<br></br>prop Korokoeyes:1<br></br>      {<br></br>   }<br></br>prop ThunderTail_1:1<br></br>   {<br></br>   }<br></br>figure<br></br> {<br></br>name TEST_NULL 1<br></br>       root BODY:1<br></br>addChild hip:1 <br></br>      BODY:1<br></br>addChild head:1 <br></br>  hip:1<br></br>[stuff left out here for reasons of space]<br></br>addChild rButtock:1 <br></br> hip:1<br></br>addChild Korokoeyes:1 <br></br>     head:1<br></br>addChild ThunderTail_1:1 <br></br> hip:1<br></br>defaultPick BODY:1<br></br> }<br></br>doc<br></br>    {<br></br>   pickActor BODY:1<br></br>    }<br></br>}

This works fine with the "Z-Pet.cr2" test figure (also included in the zip) which uses the hierachy from vilian's post #14. Unfortunatly, although it works on my test figure, vilian reports that it does not work on the Koroko figure.

P.S. I don't understand the use of "root BODY:0" in your post, why ":0" instead of ":1"?

Root BODY:0 == typo You need both sets of Actor lines. The geometry declarations, blank inside, AND the actor setup, with no channels. You need to include all (I think "all" any way) this junk ... actor hip:1 { name Hip on bend 0 dynamicsLock 1 hidden 0 addToMenu 1 castsShadow 1 includeInDepthCue 1 parent BODY:0 endPoint 0 0.021 0 origin 0 -0.079 0 orientation 0 0 0 displayOrigin 0 displayMode USEPARENT customMaterial 0 locked 0 } (Obviously you'll need to repeat for each actor) And, no, I've never figured out why it needs to be included. This is the way I build my commercially distributed shields. The tutorial on my site shows the easy way. That way works about 99% of the time. It seems like P4PP is most likely to need the longer shield format to work properly. This setup is really a "Shield" figure, not a Null. I need to go back and re-do my tutorial covering this to reflect this problem. Unfortunately the quick and easy "Null" figure isn't 100% reliable. Nerd3D

Well I must say that it seems strange to me that one should need to include a definition section in addition to the declaration section, and if it is needed it seems strange to me that all the "junk" is necessary, as the 'name' and 'parent' lines are the only ones that could by any strech of my imagination rlelate to EMC/ERC. Rob's Millenium Null has none of this, and and has always seemed to work ok. Never the less, as my custom NULL Figure does not seem to work on Koroko, I have made a Shield Figure in the format sugested by Nerd. The text of this Shield is attached above. Vilian, I would be grateful if you could test this Shield on Koroko and report back with the results. Nerd, I hope my above comments don't sound too ungrateful, I do appreciate you taking the time to give advice on this matter, I guess it's just that a long association with Poser has imbued me with a sceptical outlook, and I have always been one who has to see to believe.

It's great seeing the thread alive. Thank for all the tries guys ! hugs But... (just don't scream...) That "Shield" figure doesn't work either O.o ! As usual I used three NULL/shield/test (whatever you call it now ;) ) figures and three Korokos, all according to usual instructions. As usual morphs in Figure 2 and Figure 3 are blocked (spinning the dials doesn't give any results). hits her head against the wall

Big SCREAM :( Ok, new idea :) Load Koroko, save Koroko back to a pallet with a new name, say "Koroko2" for example. Try the TEST-NULL-2.cr2 from post #18 with Koroko2. I am feeling optimistic about this, of course it is not a good solution if you need to redistribute your work, but I get the impression this is just for your own use. Any rate, if this does work it will narrow down the possibilities and may help us arrive at a better solution. Cheers, Les.

P.S. The idea is that when the figure is the only one in the scene Poser will reparse the figure at save time seting everything to ":1" and "Figure 1".

So I have to do everything I do with normal NULL, but with Koroko resaved to the library ? Off to try :) Edit: runs in circles around the desk jumping and screaming It works, it works !!!! Now I have three differently morphed Korokos in one scene !!! Wooohooo ! hugs Les How and why ? I got lost in those technical issues way earlier, and now a miraculous solution - it's so weird ^^;

Message edited on: 03/26/2006 16:47

See ? First one has "Motoko", second "Ball" and third "Forehead" morph set to 1. Actually it's Sparky - Koroko with built-in tail prop and loading with different texture, but they both crosstalk exactly the same (same bodyparts, same morphs, same JP's and such). Just wow ! I think I'm satisfied with this solution - resaving figure to the library takes only seconds - and I thank you for that, Les. I already lost hope for pure Poser renders with multiple Korokos - and what a surprise ! *hugs again* I have to go catch some sleep now, and I'm going with smile on my face. Thank you very much ! Bookmark again - I'd like to know why and how, really :-)

Your welcome :) I think this issue was bugging me almost as much as it was bugging you! My bedroom has cloths and computer parts laying all over the floor, but I hate to see an unsolved problem laying around. As to the how and why, I'm not sure of the details yet, but it seems to be something to do with the with the figure and actor numbers. Saving the figure from an otherwise empty scene reset those to "1". Some quick experiments sugest that this is is a general problem with figuires in P4, not just one that relates to Koroko. I find it strange that I have not heard of this problem before, or come acrss it in my own work. Well now I'm off on a quest to find a solution that does not involve resaving the figure, and to save more damsels from the clutches of the crosstalk dragons. Farewell for now sweet maden! P.S. If I can figure out the why and how, I'll let you know.

That's really weird. Have you compared the original file to the resaved version to try and see what changed. Nerd3D

Nerd,

I don't have the original Koroko figure, but I am quite confident that I know what has changed. We can see from Vilian's post #21 that in the slaving code is:

valueOpDeltaAdd
Figure 2
BODY:2
PBM-Ball
deltaAddDelta 1.000000

Note the "Figure 2" and "BODY:2". It's a good bet that the actor number throughout the whole file is also ":2". Resaving Koroko to a pallet when it was the only figure in the scene reset all these numbers to "1".

It seems (admitidly at first sight, and without a thourough investigation) that you will get this crosstalk problem either with a Rob The Whiz type NULL Figure, or a Nerd type Shield Figure, when the actor number (:#) in the character cr2 you are loading is not ":1"!

I did the following experiment. Created a figure with an actor number of ":4". Made an FBM where the slaving code was:

                   valueOpDeltaAdd<br></br>                             Figure 4<br></br>                            BODY:4<br></br>                              FBM-01<br></br>                      deltaAddDelta 1.000000

Loaded NULL > character > NULL > character > NULL > character. Saved a pz3.

On examining the pz3:
In the first character (actor number :2), the slaving code was correct. In the second character (actor number :4) the slaving code pointed to ":2". In the third character (actor number :6) the slaving pointed to ":4". Make of this what you will. The only thing I can say for certain at the moment is that the second and third characters are crosstalking, and that the NULL did not do it's job. When the character cr2 is edited to use ":1" and "Figure 1" throuhgout' the NULL works fine and I can load as many instances of the character as I like without crosstalk.

From the above experiment, it looks to me like there is a serious limitation in the NULL/Shield approach to preventing crosstalk. I am looking for a solution, but fear I may be on a wilde goose chase, or that the solution may be so complex as to impractical.

Add to all this the complication that if you load three instances of posette and resave the third instance to a pallet the actor number will be ":3", but the slaving code will still point to "Figure 1" and "BODY:1".

There does seem to one almost bullit proofe solution to crosstalk. Inject the slaving code after the character is loaded, in the format:

                     valueOpDeltaAdd<br></br>                             _NO_FIG_<br></br>                            BODY<br></br>                                Morph<br></br>                       deltaAddDelta 1.000000

However there are some drawbacks to this method, loading the character becomes a two step process, load the cr2 then the pz2, even if you only want one instance of the character in the scene, another drawback is that the character must be created (or edited) with this in mind, yet another is the extra work in implimenting this, and a final one is that if you save this character back to a pallet from the Poser interface you are back to square one.

Of course all this only relates to P4/PP, in P5 and up the crosstalk problem seems to be solved.

As far as I can tell by just looking at both versions of Koroko in EditPad, the only thing that had changed is what Les foreseen - that all those 2's changed into 1's, from actor BODY:2 { } actor hip:2 (and so on) to actor BODY:1 { } actor hip:1 (and so on), including all the referrals deep in the file. It's late at night here, so tomorrow I'll try to compare .cr2's in some kind of file-comparing software. And I'll try to understand last post by Les, as it's very interesting and yet too complicated for my poor tired brain ^^ Thanks for help once again :)

Thanks Les:):):) I just tried your proceedure from post #14 on Eldrich's Companion Wyrm and it works. Yeah!!!!