» cross-process texture sharing, depth rendering
This site relies heavily on Javascript. You should enable it if you want the full experience. Learn more.

cross-process texture sharing, depth rendering

helo nerds,

this is to inform you that latest alphas come with a new commandline: argument


starting vvvv with this option (on windows >= vista) brings you:

  • a VideoTexture that works on two heads of one graphiccard (as you'd have always expected)
  • texture sharing
  • broken Flash (EX9) as long as original movie's resolution is non-power of two
  • broken Text (EX9.Geometry)

texture-sharing can be used to share textures e.g. between two instances of vvvv (without noticable performance penalty) and even with any other software that also supports DX9EX. check the helppatch of SharedTexture (EX9.Texture) for instructions.

apparently this can also work as a bridge to opengl if someone wants to give this a try...and for those wondering, yes this enables stuff on windows your macfriends have been bragging about for a while now using syphon.

depth rendering

another new thing (not related to dx9ex) is the possibility to render the deph of scenes directly without using an extra pixelshader pass. just set the Texture Format of a DX9Texture (EX9.Texture) to INTZ and you should get the depth of your scene rather than a colorbuffer. disclaimer: may not work on certain graphiccards, see here

available now in latest alphas, enjoy.

joreg, Friday, Jul 6th 2012 Digg | Tweet | Delicious 18 comments  
elektromeier 06/07/2012 - 18:17

dx9ex also helps for VideoTexture span across multiple graphiccards?

joreg 06/07/2012 - 18:46

no, not across multiple cards, only with non-spanned dualview setups.

sebl 06/07/2012 - 21:51

s u p e r !

and thanks

manuel 07/07/2012 - 05:48

hey joreg, why this cross process texture is activated in commandline and not directly added to vvvv default functionality ?

joreg 07/07/2012 - 14:30

since the feature is rather of special interest and we have the feeling it is not too much tested in the wild with bigger projects we thought in a first step we make it optional and if it turns out to be _the thing_ we can still make it default in a later release.

catweasel 08/07/2012 - 15:07

Would this be possible from opengl->dx?
This would be good too for ofx processing or Avenue to v4...
But one way is better than no way!
Sounds great, I'll try and test soon, thanks guys :)

sven 08/07/2012 - 18:19

Is the depth evaluation radiometric from the current point of view or is it aligned to the z-axis?

joreg 08/07/2012 - 18:49

depth is using the same camera as color.

elliotwoods 09/07/2012 - 21:33


@sven - aligned to Z of the camera view, and often not a linear response with respect to euclidean xyz world space

robi_h 11/07/2012 - 11:14

hell yeah thanks

mrboni 11/07/2012 - 11:56

@elliot - if it aint linear, what is it and why?

circuitb 31/07/2012 - 14:13

launching last alpha with the /dx9ex switch seems to break the kinect texture nodes! white renderer!!
tested with both microsoft & openni nodes

circuitb 31/07/2012 - 14:22

also changing the depth mode of the kinect (Microsoft) node throw some errors

joreg 31/07/2012 - 23:33

@circuit: thanks. fixed for openni-nodes in latest alpha.

kyy921 30/11/2012 - 18:59

will it support for unity3d?

joreg 30/11/2012 - 19:52

@kyy921: you have to check with unity3d if it supports dx9ex texture sharing. vvvv does.

vj_raul 13/09/2013 - 14:48

@joreg/electromeier: /dx9ex texture sharing works also on 2 graphic cards here ...

But I have to right-click on the renderer node (close and open) once. It's the same when both screens connected to one graphic card (GeForce 770).

I don't use SharedTexture(EX9.Texture), but VideoOut(SharedMemory) and SharedMemory(EX9.Texture)

edit: Well, it seems I don't even need the /dx9ex for this nodes. Sorry!

joreg 13/09/2013 - 15:04

vj_raul: yes, in your case it works on multiple cards because those nodes you mentioned don't make use use of gpu-texture sharing. they share via cpu-memory.

  • 1

anonymous user login


~27d ago

manuel: very cool to generate matcap / cubemap for reflections https://cables.gl/p/pDCOCw

~28d ago

joreg: Join us for the 19th #vvvv meetup on November 11th: https://thenodeinstitute.org/event/19-worldwide-vvvv-meetup/

~1mth ago

~2mth ago

joreg: @zaid please ask this in the forum: https://discourse.vvvv.org/

~2mth ago

zaid: how can i detect a changing spread in gamma? the old workflow with a change and select node isnt possible. cheers

~2mth ago

joreg: @dawoof such a question is probably better asked in the forum or chat

~2mth ago

dawoof: hey vvvvers, What are yall's fav depth-camera for body tracking? (dancing in this case) eyeing the ZED but the delivery time is long