At that time, I prepare a vj gig, I like to control all my vvvv patch with one vvvv interface.
There’s one problem that I can’t resolve, how can I load a patch during my gig without freezing my EX9 renderer.
thanks for your help.
Loading things, like patches, images or whatever, will always freeze your system for a short period of time. AFAIK this problem is not solved yet as vvvv probably runs everything in the same thread (except for the DirectShow stuff).
I’m sure Joreg can explain those things a bit more detailed.
What you might try is to put all your patches into seperate subpatches and add an “Enable” flag connected to basic nodes to reduce their computing-power consumption.
Add an EX9 Layer Output to every subpatch and feed them into a main input switch (node) which you connect to one basic renderer.
All your patches are preloaded this way and all you have to do is to use the node switch to select a certain render output.
sven is right. the vvvv user interface and the render loop is running in the same thread. therefore you will get one very long render frame while a patch is loading. this will obviously make your rendereres freeze.
note that this behaviour makes patching much easier, as you will always have a consistent state where everything is fully ready and evaluated - otherwise you´d need to deal with IsLoaded pins all over the place. so you can be sure that only fully loaded patches go to the screen.
i agree that there are things to improve, and asynchronuous texture loading is one of them. increasing the speed of patches which are built after sven´s proposal is another…
for now having a second machine and a video switcher seems to be good practice at a live gig anyway…
Thanks guys, with your answers I can see now that I’m not in the wrong way. I used the subpatches to collect all my patches and I used 2 laptops et 1 videomixer.
But I didn’t know that I must use an Enable node to optimize my patches
I m waiting for the asynchronuous texture loading.
what means asynchronous texture loading exactly? and will the result be no more freezing when loading new textures in the graphiccard?
“increasing the speed of patches which are built after sven´s proposal is another…”
would be cool. at the moment it seems that even pathes feeded with empty spreads consumes computing power. would be a real advantage to have them ready in ram but without making the rest any slower… dont know if such a thing can be easily achieved but would be certainly usefull for live gigs.
i agree…not being able to really kill parts of a complex patch makes things complicated. i can imagine having some sort of ‘kill node’ which disables all following nodes in the thread.