What is also weird is when i open the solution in visual studio it shows Win32 as default and i have to enter x64 manually, not sure there that is the right way.
I get these errors when i build in visual Studio:
The type 'Rectangle' is defined in an assembly that is not referenced. You must add a reference to assembly 'System.Drawing, Version=4.0.0.0, Culture=neutral, PublicKeyToken=b03f5f7f11d50a3a'.
and
Could not copy the file "C:\Users\User\Downloads\_latest SDK\Fbuild\bin\x64\Release\ZED_Body_Fusion.exe" because it was not found.```
There is no csharp sample to send fused data to Unity, can you link me to what you were using?
Then, once the c++ sender is compiled, you need to specify the calibration file as argument when running the executable.
For that, open a terminal in the folder containing the executable and run
I know how to start the fusion in cmd, i use it with non blackwell GPU in other setups.
But im not sure how to reproduce the Fusion, maybe i use the wrong source in the SDK?
The only guidance i have is the tutorial with the SDK examples from your website.
i also compiled it and the fusion is now kinda working.
However i get this error very often:
C:\Users\User>C:\Users\User\Downloads\_latest_SDK\liveLinkBuild\Release\ZED_Unity_Livelink_Fusion.exe C:\Users\User\VHS-dev-calibration.json
Try to open ZED 31993813..[2025-05-27 11:42:40 UTC][ZED][INFO] Logging level INFO
Error: CAMERA NOT DETECTED
Try to open ZED 38295136..[2025-05-27 11:42:41 UTC][ZED][WARNING] CAMERA NOT DETECTED in enum sl::ERROR_CODE __cdecl sl::Camera::open(struct sl::InitParameters)
[2025-05-27 11:42:41 UTC][ZED][INFO] Logging level INFO
Error: CAMERA NOT DETECTED
Unable to subscribe to 31993813 . MEMORY ALREADY USED
Unable to subscribe to 38295136 . MEMORY ALREADY USED
no connections
that is something i nevery see on the other pc’s with fusion active.
Any clues what that could be?
I also answer this question by myself, for all after me with not so much experiences in this stuff.
How i got Fusion working on windows with SDK 5.0.1
Use this source in Cmake: from github - stereolabs/zed-unity-livelink
follow the Cmake tutorial but use paths to the zed-unity-livelink. only the how to make the visual studio solution → stereolabs website /docs/app-development/csharp/use
use latest visual studio community (with unity)
once you compiled the first (release) build try optimizing, the zed-unity-livelink has old settings within: “senderRunner.cpp” replace them with the new settings: stereolabs website /docs/depth-sensing/depth-settings
recalibrate your ZED cam with neural as in the senderRunner.cpp
if not yet train all AI models
if you get the Unable to subscribe to [cam ID]. MEMORY ALREADY USED error your cable is almost certain the problem. If you have an original zed cable with this error it is broken. And if you use fiber cable like me, be sure the Cable plus ist perfectly stable and secure in the cam or it will fail!
@BenjaminV Hi Benjamin
I made now the first build from the Unity Project and for a few Minutes everything works fine.
After about 3 minutes suddently the fusion stops working and cmd has errors.
The Unity Experience still runs.
[2025-05-27 14:59:41 UTC][ZED][WARNING] IExecutionContext::enqueueV3: Error Code 1: Cuda Runtime (an illegal memory access was encountered)
[2025-05-27 14:59:41 UTC][ZED][WARNING] [Grab] CUDA error returned: an illegal memory access was encountered ( 700 )
[2025-05-27 14:59:41 UTC][ZED][ERROR] Skeleton detection inference failed!
[2025-05-27 14:59:41 UTC][ZED][WARNING] CUDA ERROR in enum sl::ERROR_CODE __cdecl sl::Camera::grab(struct sl::RuntimeParameters)
in enum sl::ERROR_CODE __cdecl sl::Mat::updateGPUfromCPU(struct CUstream_st *,int) : Err [700]: an illegal memory access was encountered.
in enum sl::ERROR_CODE __cdecl sl::Mat::copyTo(class sl::Mat &,enum sl::COPY_TYPE,struct CUstream_st *) const : Err [700]: an illegal memory access was encountered.
CUDA error at C:\builds\sl\ZEDKit\lib\src\sl_zed\DetectionHandler.cpp:2148 code=700(cudaErrorIllegalAddress) "cudaEventRecord(ev_image, sdk_stream)"
CUDA error at C:\builds\sl\ZEDKit\lib\src\sl_zed\DetectionHandler.cpp:2236 code=700(cudaErrorIllegalAddress) "cudaStreamWaitEvent(cu_strm, ev_image, 0)"
CUDA error at C:\builds\sl\ZEDKit\lib\src\sl_zed\DetectionHandler.cpp:2178 code=700(cudaErrorIllegalAddress) "cudaEventRecord(ev_depth, sdk_stream)"
in void __cdecl ai::BatchedBuffer::memcpy_d2h(unsigned __int64,struct CUstream_st *) : Err [700]: an illegal memory access was encountered.
in void __cdecl ai::BatchedBuffer::memcpy_d2h(unsigned __int64,struct CUstream_st *) : Err [700]: an illegal memory access was encountered.
CUDA error at C:\builds\sl\ZEDKit\lib\src\sl_zed\DetectionHandler.cpp:2252 code=700(cudaErrorIllegalAddress) "cudaEventRecord(ev_bt, cu_strm)"
CUDA error at C:\builds\sl\ZEDKit\lib\src\sl_core\utils\util.cu:482 code=700(cudaErrorIllegalAddress) "cudaCreateTextureObject(tex, &resDesc, &texDesc, NULL)"
in enum sl::ERROR_CODE __cdecl sl::Mat::updateCPUfromGPU(struct CUstream_st *) : Err [700]: an illegal memory access was encountered.
in enum sl::ERROR_CODE __cdecl sl::Mat::updateCPUfromGPU(struct CUstream_st *) : Err [700]: an illegal memory access was encountered.
CUDA error at C:\builds\sl\ZEDKit\lib\src\sl_zed\DetectionHandler.cpp:2274 code=700(cudaErrorIllegalAddress) "cudaStreamWaitEvent(cu_strm, ev_depth, 0)"
in enum sl::ERROR_CODE __cdecl sl::Mat::updateCPUfromGPU(struct CUstream_st *) : Err [700]: an illegal memory access was encountered.
CUDA error at C:\builds\sl\ZEDKit\lib\src\sl_zed\DetectionHandler.cpp:2306 code=700(cudaErrorIllegalAddress) "cudaStreamWaitEvent(cu_strm, ev_bt, 0)"
[2025-05-27 14:59:41 UTC][ZED][WARNING] IExecutionContext::enqueueV3: Error Code 1: Cuda Runtime (an illegal memory access was encountered)
[2025-05-27 14:59:41 UTC][ZED][WARNING] [Grab] CUDA error returned: an illegal memory access was encountered ( 700 )
[2025-05-27 14:59:41 UTC][ZED][WARNING] CUDA ERROR in enum sl::ERROR_CODE __cdecl sl::Camera::grab(struct sl::RuntimeParameters)
[2025-05-27 14:59:41 UTC][ZED][ERROR] Skeleton detection inference failed!
Any Idea what this can cause?
We are using:
Nvidia GeForce RTX 5080
32gig ram
I9 CPU 14900KF
Win 11 Pro
Cuda compilation tools, release 12.9, V12.9.41
Build cuda_12.9.r12.9/compiler.35813241_0
Thanks for the report,
Could you please specify which depth mode you are using? I’m mostly interested in knowing if you were using one of the “NEURAL” modes?
I do not know why yet, but the crash most likely comes from that, and not the body tracking AI model.
As a workaround, you can try to use any non-neural depth mode.
The “MEMORY ALREADY USED” error code is not very clear and self-explanatory, but when you see this error, it means the camera can not be opened.
@BenjaminV
Thanks for this Update, i will use the other modes then.
I have a general quest, if i compile my Unity Exe on a Directx 11 PC setup, does this have a negative impact when the exe runns on the DirectX 12 System setup?
We use older Pc’s to work on the project, and deploy it then to the new *blackwell Hardware.
It should be working fine, as DirectX 12 has backwards compatibility as far as I know.
I’ve not tested it but you might be able to configure your graphic settings so it uses DX11 on your old pc and DX12 on the new one.
In Edit > Project Settings > Player > Other Settings > Graphics APIs for Windows, set DX12, then DX11. That way, it will use DX12 if possible and fall back to DX11 if not.
The only downside is that you are not testing the same version you are deploying into production, which can generate unexpected behaviours.