Forum Replies Created
-
Posts
-
Ok, im not really sure how to explain this, but.
We wrote Syphon Recorder. Its not adding a thing to the image – really.
Do yourself a favor and see if you see it in any other Syphon enabled app, like Simple Client, or VDMX, etc.
I even double checked and looked at the code and we don’t add anything anywhere. I’m not really sure why you are so confident, but take a moment and think about what you’re saying and why I feel confident in taking the time to give the advice I am giving to you.
Just because what you see in Resolume’s preview does not mean Resolume isnt outputting something additional to Syphon. Syphon isn’t a screen grab. You can shit out whatever you want however you want to Syphon. Notice how the FPS panel isnt in Syphon?
Go talk to Resolume, its likely you have a video capture device or audio input set up from a preset and the device (audio or video) is no longer present.
That sounds like a Resolume question then, no? That has nothing to do with Syphon.
To be clear – Syphon Recorder / Syphon at all does nothing to change the image. Its either there from the serving app or Syphon doesnt see it.
Check the Resolume forums.
Interesting:
“com.apple.DesktopVideoOut”
Are you attempting to record to “Desktop Video Out” from an older FCP 7 Quicktime Codec / Component install?
Unless I missed it, you didn’t answer what source software you are sending from.
Try switching the codec (try Pro Res or Intermediate) and audio input you from in Recorder, and see if you can isolate.
Hi
Interesting. That error is:
“The specified audio hardware device could not be located.”
Can you check your input device preferences and see if changing it solves the issue for you?
For what its worth, I recorded some things this weekend with Recorder & 10.8.3 – please post a crashlog and the app you were sending via.
Ideally you’d have support native in the browser. At one point, a friend at Google mentioned they might put Syphon support right in to Chrome. I can ping them and see if its feasible at all, and point them to this thread.
I suspect it would never get to be in an official version, but perhaps an alternative build of Chrome could have Syphon support.
This is *not* an indication that anyone is working, will work, or ever will make such a thing, just that it was mentioned to me as a possibility. In other words, no promises now or ever 🙂
Try to use GL Profiler and break on error, and introspect your GL State, or see if you can reproduce on another GPU, you might be hitting an edge case.
Additionally, OpenFrameworks uses OpenGL, not Direct X, so “Wyphon” is out of the question until they support OpenGL (and gain a ofxWyphon add on).
Secondly, Movie playback on OpenFrameworks has some subtle issues as well.
I saw this too. The issue is they dont up-rez for retina resolution, so they send a screen size frame in points, not retina “scalefactor 2” pixels.
Is the GL context active, and the same as the one you’ve initted Syphon with? Just check if the CGLCOntextObj returned from CGLGetCurrentContext() matches that you’ve initted Syphon with.
Well, a few things.
1) ensure that the CGLContext you init the SyphonServer on is actually the active context used during rendering. Sometimes contexts get destroyed and re-created. Maybe lazy init in the render loop?
2) ensure that the texture you are publishing is not being deleted before you publish.
3) OpenGL Profiler, enable break on error, and ensure there are no GL errors. This is also a great place to go to Resources (when you are stopped at a break point) and verify that your textureID is valid, and matches your expectations.
Is the deform a nurbs deform,or a vertex deform? Can you bake the deform in, or, if its an animation, ensure its a skinned animation and not a per vertex / frame animation (ie, the model has bones, joints, joint weights and animates by moving the joints, rather than swapping out sets of geometry?)
Also, try removing the light and the camera? I dont know blender at all, so im unsure how I can help, but post the file too, maybe?
If you are viewing an image larger or smaller than its original size, filtering has to be applied for scaling regardless of Syphon or not.
Default filtering is GL_LINEAR, or a linear filtering model, not a pixel art esque GL_NEAREST, for nearest neighbor filtering.
Thus if you are viewing a 10×10 image at anything larger or smaller than 10×10, it will be filtered with GL_LINEAR filtering. Understand, the actual Syphon stream has *no* filtering applied as it is not being scaled at all – it is only on the onscreen viewing where filtering is applied.
Simple Client is not meant to be anything other than a basic QA test to ensure Syphon streams are being properly published and received and driven by the renderer appropriately.
If you want to force nearest neighbor filtering on your output, you have to do that via a feature in the application your are actually intending to use as a Syphon Client.
Simple Client draws with a different blend mode. Alpha is in there, its just a difference between premultiplied and non-premultiplied alpha, as far as I can tell :
Also, native network support in Syphon is something in the works. We have one solution working internally, and are looking for some other possibilities. Time, not money, is the issue here.
