I'm not comfident enough to tell, and I can only speak for vop (not vex). Having to build a for each loop set of nodes, well it took me time to figure (in fact Mikael tut was the answer) Having to work with per prim vertex array seems to be a limitation when coming to sort those vertex. I have overal sorting issues : it's ok to sort ptnum but an arbitrary integer attribute I'm not sure I can. I wish there was more example (in vop) of array building and sorting. I miss the "sort array with key"
I'm just starting with arrays in H but I find it over complicated. On Wednesday, March 29, 2017, Andy Nicholas <a...@andynicholas.com> wrote: > Heh! Flattery will get you everywhere ;) > > Yes, they make sense in both packages to me, and you have the added > benefit of being able to see the arrays directly in Houdini using the > Geometry Spreadsheet, rather than turning on the visualisation in Softimage > and have them disappearing off the top of the screen. You can do Python > style array slicing in VEX too which is awesomely useful. > > Anyway, that's why I wanted to ask Olivier an honest question to try and > understand where Houdini is lacking. Maybe I'm too close to it to see the > issues. > > A > > > On 29/03/2017 13:06, Jordi Bares wrote: > > I was wondering if using arrays in Houdini makes as much sense as in > Softimage… Andy??? You are the expert here. > > jb > > On 29 Mar 2017, at 10:57, Andy Nicholas <a...@andynicholas.com > <javascript:_e(%7B%7D,'cvml','a...@andynicholas.com');>> wrote: > > Hi Olivier, where do you see the biggest difference with ICE arrays and > Houdini arrays? In both you have ways of adding, removing, sorting, etc. > elements in the array in whatever context you're in, no? > > A > > On 29/03/2017 09:48, Olivier Jeannel wrote: > > An example of something we own as ice user : > One of the first thing I replicated was the Modulate by Volume. > When I arrived on Houdini and saw all those tutorial with people using the > Attribute transfer, I tried to use it myself and was horrified : I found it > slow and not precize. > Nobody was using a UV location + dotproduct method, and imho that's by far > the most efficient method. > The hardest thinking was "how should I wire it in Houdini way of working > ?". I came up with one Vop HDA, and one SOP HDA (which is simply the vop > compounded). > The great "plus" with Houdini, is that it's able to transfer values of any > context (point, prim, you name it, ..) > I'll try to record something. > > I think there's a large place for improvement in H for everything that > concerns arrays. In ice, there was a lot of things to do with arrays , > hence the speed. And the ice tools were super efficient for that. In > Houdini, there's a kind of "thinking" that as VOP by nature is looping > through points it is an "enough" solution. > Well, "maybe" but it's so criptic that unless you're a vex/C/python > programmer I find it very time consumming to understand. Plus there are no > doc samples or tut, a part from the one from Mikael Perterssen http:// > shortandsweet3d.blogspot.fr/. That makes me wonder if other people really > consider or understand this. > Also, if you compare Peter Quint and Mikael way to deal with States, hell, > I'm 200% on the ice method. > > > 2017-03-28 20:39 GMT+02:00 Olivier Jeannel <facialdel...@gmail.com > <javascript:_e(%7B%7D,'cvml','facialdel...@gmail.com');>>: > >> I'm a morron, but I'd love to have exclusive pure Ice minded HDA library. >> >> 2017-03-28 20:29 GMT+02:00 Rob Chapman <tekano....@gmail.com >> <javascript:_e(%7B%7D,'cvml','tekano....@gmail.com');>>: >> >>> Thought I'd pipe in since tekano got invoked, also slowly attemting to >>> transition and agree with most already said and thanks, is already a huge >>> pointer to as yet unknown aspects and features of how complex houdini is. >>> also would be interested in a more 'compounded' way of learning Houdini >>> like ice was introduced. Everything a compound node of nested compound >>> logic with exact same UI logic and Core nodes and complexity under the hood >>> but still accessable in a single click and an 'easy for artists' ability to >>> follow the logic flow into further nested compounds and see how it was >>> made. Not so with houdini yet 😎 open one compound and is equivalent to >>> inside of the neighborhood telephone junction box. Part of the enjoyment, >>> for me, was building own logic and then seeing the contrast of the >>> 'Softimage' way, and for sure, if you are building something fairly complex >>> requiring macro detailed interactions with something of a much larger >>> scale, eg characters running through a several fields of flowers, then >>> somethings can be improved or optimised from the off the shelf examples. >>> Otherwise prepare for big data and long iteration times. It seems covering >>> all bases like the 'houdini' way is fine for examples and base setup but >>> not so in more complicated tasks is better to be good at understanding >>> which bits to leave out. 😀 or be able rapidly prototype your own. I think >>> like Mr Bolland has done and Pooby is asking for is these intermediate >>> compounds between that Softimage bought with it to help us poor artists out >>> 😂 >>> >>> ------ >>> Softimage Mailing List. >>> To unsubscribe, send a mail to softimage-requ...@listproc.autodesk.com >>> <javascript:_e(%7B%7D,'cvml','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 > <javascript:_e(%7B%7D,'cvml','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 > <javascript:_e(%7B%7D,'cvml','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 > <javascript:_e(%7B%7D,'cvml','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.
