Whilst you can connect multiple RealSense cameras of the same type to one PC, only one of them at a time can be accessed by the PC.
Also, the camera requires a USB 3.0 connection. So your yellow error is likely because a camera is plugged into a USB 2.0 port. If you need additional USB 3.0 ports, I recommend purchasing a mains-powered USB hub for around $15 to provide additional USB 3.0 ports. Search a store such as Amazon for 'powered hub usb 3.0'.
My fault, of course they where all connected to USB 3.0. However, I had no problem to use 3 cameras at same time with a Python wrapper of Intel RealSense SDK but as the 4th camera had problem already in the device manager I couldn't connect to it.
So, is the limit in the drivers, OS or in the hardware and is it possible to do a workaround?
The official RealSense SDK has the one-camera limit, but the open-source Librealsense SDK (which I guess you are using if you are using Python) supports multi-camera capture. So it is a software issue.
Edit: regarding your yellow error, can you tell me please if the camera that has the problem has a 'Virtual Driver' present as well as the RGB and Depth drivers for that camera? If it does, is it the Depth driver that has the yellow error?
I'am using the official RealSense SDK via a python wrapper written in C++ as I'am using python.
I'am missing the Virtual Driver device. Both RGB and Depth device is vissible but one of them got that triangle. Sometimes it's the RGB device and sometimes it's the Depth device.
All the other cameras got both Depth-, RGB- and Virtual Driver device.
If I add a R200 camera as 4th camera I can use all 4 cameras. As I can see the R200 doesn't use the Virtual Driver device, so maybe the limit is in the driver for the "Virtual Driver" - device
I've had that error myself. Here is the procedure I use to fix it.
1. In the Device Manager, right-click on each of the drivers for the affected camera - RGB and Depth - and select the Uninstall option to remove that camera's drivers from the Device Manager list. When you are finished, there should be no drivers listed for that camera.
2. Unplug that camera from its USB port.
3. Wait a couple of seconds and plug that camera back into its USB port. Windows 10 should now search the internet for the missing RGB, Depth and virtual drivers and install them one by one. Over the course of 5 to 10 seconds after insertion of the camera, each of them should appear in the Device Manager and no longer have any yellow errors.
I reinstalled the Depth Camera Manager and now it works.
Ok, great! Installing the Depth Camera Manager effectively does the same thing as my method. Installing the DCM is the preferred method when you do not have an internet connection for Windows 10 to auto-install the drivers with. Good luck, and please come back to the forum any time you need help.
I would like to use the hand tracking features from the Intel RealSense SDK with two (or more camera's) on one machine. Is this only possible using virtual machines?
From the following reference page (Intel® RealSense™ SDK 2016 R2 Documentation ) from the 2016 R2 Realsense SDK, i am citating the following line:
"An SDK session may contain multiple capture modules and each module can manage multiple devices. Use the QueryImpl function to enumerate each capture module, the CreateImpl function to create an instance of the module, and then the QueryDevice function to retrieve device specific information."
How should i interpret this?
Here's a tutorial article on device enumeration in the RealSense SDK.
The article says: "Enumerating camera devices is an important step in any application that may require selecting a specific camera device when several cameras are attached to a system. This tutorial presented a straightforward camera enumeration scheme for developers who want to build any necessary selection logic once the particular camera device and capabilities are identified. "
You can likely take that to mean that you can have different cameras attached to the computer and query their status but can only select one at a time to be the active camera out of the selection of attached cameras.