Love Houdini, but coming from ICE, VOPs are an absolute horror to use IMO. They need to drag it out of the 90's and introduce some more modern interaction workflows...
DAN On Thu, Mar 3, 2016 at 4:45 PM, Christopher Crouzet < christopher.crou...@gmail.com> wrote: > VOP graphs are being converted to VEX code, which means that for each VOP > node you'll find a corresponding VEX function, and most likely vice versa > too. If you want to see the generated code, just right click on the VOP > node, and select “View VEX Code...”. It can be informative to check it out > sometimes to help with debugging, but it's overly cluttered in comparison > to hand-written VEX code and thus can be hard to read. > > Regarding the line `int pt1 = min(neighbours(input, pt0));`, it's picking > the neighbour with the smallest point number. The reason is simple—the > `Sort SOP` node reorder the points so for example the lowest point numbers > for a sort by X become the ones having the lowest X value. And hence, you > want the neighbour of `pt0` with the lowest point number to have `pt0` and > `pt1` representing the edge that matches the best to what's expected from > the sorting. If instead you'd simply do `int pt1 = neighbours(input, > pt0)[0];`, then you'd sometimes get what you'd hope, sometimes not. > > Also yeah, there's never a single way to do things. If you compute a sort > of normal as you say for each point, that you can make in a predictable way > accross the entire mesh, then you'd probably avoid the problem that I > mentionned in my previous email. Maybe the `PolyFrame SOP` could be used > for that, but I don't know how it internally works and thus I don't know > how predictable it is. Worth a try! I only provided the scene this way to > give the idea that the essential of it can be done easily enough in > VOP/VEX. But it's always the remaining 10% that takes 90% of the time! :) > > > On 3 March 2016 at 21:28, Olivier Jeannel <facialdel...@gmail.com> wrote: > >> You seem very dispointed by Houdini Anto, are you ? >> >> On Thu, Mar 3, 2016 at 2:16 PM, Anto Matkovic <a...@matkovic.com> wrote: >> >>> Thank you kindly Christopher, >>> Yes as far as I know, FE presents the whole points data as an array. >>> Beside that, at this moment I just feel FE as much better solution for me >>> (and for Olivier of course :) ), instead of H. Can't really say more for >>> now. Also really need to say, while I'm not programmer, unfortunately or >>> not I have to communicate with them every day, even in private life, so >>> maybe I become a bit resistant. I think we all know how is going with this >>> kind of people - one has this or that opinion, another has completely >>> opposite, all long stories full of smart words. No need to waste your time >>> on me, let's say in short :) Of course I've read everything you said here. >>> thanks again... >>> >>> ------------------------------ >>> *From:* Christopher Crouzet <christopher.crou...@gmail.com> >>> *To:* Softimage Mailing List <softimage@listproc.autodesk.com> >>> *Sent:* Thursday, March 3, 2016 4:40 AM >>> *Subject:* Re: OT Houdini build Array VOP question (and a bit of rant) >>> >>> You can think of each geometry attribute as an array which length equals >>> the number of points (if dealing with points). VOPs simply iterate over >>> these points and process the graph for each of them. Since each attribute >>> is an array, at any time you can pick the value of any element with `Get >>> Attribute` provided you know the point number. This is so straightforward >>> that you can even do this outside of VOP/VEX, with the `point` expression >>> for example. >>> >>> Sometimes it can be useful to define attributes that can themselves hold >>> arrays, for example a list of neighbours, to then pass these down to other >>> nodes for further processing. That's why they added that feature in H14, >>> but otherwise there's no real need to directly create arrays yourself. >>> Shamefully, I haven't used Fabric Engine yet but from what you're saying >>> maybe they're just presenting the data differently by providing you with >>> the whole points data as an array instead of letting you directly work on a >>> per-point context? In any case, there's plently of good reasons to use >>> Fabric Engine but this definitely isn't one of them—just wrap your head >>> around how Houdini works instead, everything will eventually start to make >>> sense :) >>> >>> On 3 March 2016 at 05:43, Anto Matkovic <a...@matkovic.com> wrote: >>> >>> Nope. I think the only way to create an array directly in VOP, is pcfind >>> (pcopen that returns array), or array version of point neighbors. Or, to >>> stack the 'append' node several times :). There are examples how to loop >>> over arrays of indices, later in network, here: >>> >>> https://www.sidefx.com/index.php?option=com_content&task=view&id=3148&Itemid=412 >>> By the way, had to do some pretty interesting networks :) for Kristinka >>> Hair for H - while everything worked at the end of day, anyway. >>> >>> If you're around arrays and nodes, Fabric is waiting for You.... >>> >>> >>> ------------------------------ >>> *From:* Olivier Jeannel <facialdel...@gmail.com> >>> *To:* "softimage@listproc.autodesk.com" <softimage@listproc.autodesk.com> >>> >>> *Sent:* Wednesday, March 2, 2016 1:29 PM >>> *Subject:* OT Houdini build Array VOP question (and a bit of rant) >>> >>> Hello serious list :) >>> >>> I'm a bit confused with houdini vop array. >>> While I managed to do it in vex, I would like to make a build array >>> (like build array from set) of the pointposition (P) in VOP. >>> >>> I understand you need to for-loop on each Ptnum and probably append the >>> P values and this will buid an array of P. >>> But you know what ? Well I can't manage to make it work. >>> >>> I found no example on the net (sideFX, odforce). >>> The doc is just words, no schemes, no graphics. >>> The examples hips are bizarre, not so simple, and use the old loop node. >>> >>> So I'm wondering if someone from here could provide a screen shot of how >>> that should be connected ? >>> >>> Thank you :) >>> >>> >>> ------ >>> Softimage Mailing List. >>> To unsubscribe, send a mail to softimage-requ...@listproc.autodesk.com >>> with "unsubscribe" in the subject, and reply to confirm. >>> >>> >>> ------ >>> Softimage Mailing List. >>> To unsubscribe, send a mail to softimage-requ...@listproc.autodesk.com >>> with "unsubscribe" in the subject, and reply to confirm. >>> >>> >>> >>> >>> -- >>> Christopher Crouzet >>> *http://christophercrouzet.com* <http://christophercrouzet.com/> >>> >>> >>> ------ >>> Softimage Mailing List. >>> To unsubscribe, send a mail to softimage-requ...@listproc.autodesk.com >>> with "unsubscribe" in the subject, and reply to confirm. >>> >>> >>> >>> ------ >>> Softimage Mailing List. >>> To unsubscribe, send a mail to softimage-requ...@listproc.autodesk.com >>> with "unsubscribe" in the subject, and reply to confirm. >>> >> >> >> ------ >> Softimage Mailing List. >> To unsubscribe, send a mail to softimage-requ...@listproc.autodesk.com >> with "unsubscribe" in the subject, and reply to confirm. >> > > > > -- > Christopher Crouzet > *http://christophercrouzet.com* <http://christophercrouzet.com> > > > ------ > Softimage Mailing List. > To unsubscribe, send a mail to softimage-requ...@listproc.autodesk.com > with "unsubscribe" in the subject, and reply to confirm. >
------ Softimage Mailing List. To unsubscribe, send a mail to softimage-requ...@listproc.autodesk.com with "unsubscribe" in the subject, and reply to confirm.
