pylibfreenect2
pylibfreenect2 copied to clipboard
r9y9
Packet buffer is NULL
I am getting the following error on Ubuntu 14.04 while running the examples:
[Error] [DepthPacketStreamParser] Packet buffer is NULL [Error] [DepthPacketStreamParser] Packet buffer is NULL [Error] [DepthPacketStreamParser] Packet buffer is NULL [Error] [DepthPacketStreamParser] Packet buffer is NULL [Error] [DepthPacketStreamParser] Packet buffer is NULL [Error] [DepthPacketStreamParser] Packet buffer is NULL [Error] [DepthPacketStreamParser] Packet buffer is NULL ..... it continuously prints the above and eventually crashes. Am I missing something in the installation? since they worked fine on OSX.
--Edit-- Just to add, the Protonect binary seems to work fine in libfreenect2
Nope. Protonect works fine. I get that error as soon as I run the examples and any other scripts I wrote which use pylibfreenect2 which were working fine on OSX. I am starting to wonder if it might be a graphics card issue? (I am running ubuntu 14.04 on a macbook pro with amd Radeon HD 8870M)
Um, it might be the issue related graphics, but I'm afraid that I don't have a workaround right now. It's worth trying to see which packet processor triggers this; OpenGLPacketPipeline, OpenCLPacketPipeline or CpuPacketPipeline to dig into the problem.
I am not 100% sure if it's a graphics card issue. It seems very odd that Protonect works but not this. I might be missing some configuration settings
How did you resolve this issue? I am also getting recursively same error.
./Protonect doesn't present any isssues ./Protonect cl DOES present the same error
Changing from OpenCL to OpenCV seem to have solved the issue
EDIT: As mentioned below by @r9y9 using
from pylibfreenect2 import OpenGLPacketPipeline
pipeline = OpenGLPacketPipeline()
instead of
from pylibfreenect2 import OpenCLPacketPipeline
pipeline = OpenCLPacketPipeline()
Seems to have solved the issue. (CpuPacketPipeline() also works just fine)
@LucasSubli You can switch to CPU or OpenGL packet processor. See:
- https://github.com/r9y9/pylibfreenect2/blob/master/examples/multiframe_listener.py#L11-L21
- http://r9y9.github.io/pylibfreenect2/latest/api.html#packet-pipelines
I am facing the same problem. Protonect works normally. Do you think the problem lies with the GPU? My PC uses only CPU. Can not I use Kinect V 2 with pylibfreenect with CPU alone? However, from pylibfreenect 2 import CpuPacketPipeline pipeline = CpuPacketPipeline () I guess that this is programmed to work with only the CPU
I do not know what to do
Well, since I don't have Kinect v2 anymore, it's hard to say what is the real cause of this. Could anybody confirm whether the issue comes from pylibfreenect2 or libfreenect2? According to the https://github.com/r9y9/pylibfreenect2/issues/34#issuecomment-331284252, there's something wrong with OpenCL. Does Protonect work even with OpenCL backend?
I tried $ ./Protonect gl It works normally $ ./Protonect cl It does not work properly, Packet buffer is NULL $ ./Protonect cpu It works normally
OpenCL doesn't work, but the problem was solved. I modified the example code(multiframe_listener.py) and tested it. I made the order of OpenCL the last.
【before】 try: from pylibfreenect2 import OpenCLPacketPipeline pipeline = OpenCLPacketPipeline() except: try: from pylibfreenect2 import OpenGLPacketPipeline pipeline = OpenGLPacketPipeline() except: from pylibfreenect2 import CpuPacketPipeline pipeline = CpuPacketPipeline()
【after】
try:
from pylibfreenect2 import OpenGLPacketPipeline
pipeline = OpenGLPacketPipeline()
except:
try:
from pylibfreenect2 import CpuPacketPipeline
pipeline = CpuPacketPipeline()
except:
from pylibfreenect2 import OpenCLPacketPipeline
pipeline = OpenCLPacketPipeline()
This code worked without problems! Thank you for your help!
@tokyokuma Thank you very much! So there should be something wrong with OpenCL setup. I used OpenCL backend because it was the fastest backend for me, but I changed example code to use the OpenGL backend instead by default.
I had the same issue, so I edited the parameters to:
depth_method to opengl reg_method to cpu
After that, it is working well.
I had the same issue, so I edited the parameters to:
depth_method to opengl reg_method to cpu
After that, it is working well.
hi, but I still want to use opencl for registration, any idea?
