Documentation / Environment Setup
Table of Contents
Please review our hardware requirements page to ensure you have everything you need to set up your capture stage and that your computer has all the outlined specifications. We cannot guarantee specifications that don't match what's on that page specifically. Also, ensure your PC is setup in accordance with our installation requirements found at the top of the Capture Suite documentation here.
Capture Area Setup
Please check out this high level video for your environment setup: Environment Setup
The best environment for capturing is an open room with matte solid color walls that are a neutral infrared tone which that allows you to have about 5-10 feet between the setup and the walls. The closer you are to walls, the higher the chance that multi-path interference will affect your calibration and capture quality, especially if compounded with other issues such as poor lighting sources. The goal is to have limited invalid pixels (black pixels) in the depth view of all cameras. Your environment should also not have glass and objects that reflect infrared. In addition, bright spots that can be seen in the infrared view of the cameras should be reduced / eliminated. A matte grey rug may be able to be laid down to reduce calibration issues. Your capture area should not have visible windows which allow sunlight in. Sunlight will impact the cameras in a negative way and can cause calibration and quality issues. Curtains can solve this issue. Your capture room should be kept between 10 and 25 celsius; preferably somewhere in the middle and not too close to the bounds.
If the background of your environment has a large amount of invalid pixels (black pixels within the red squares of the image below), depth noise can become an issue. This can not only affect your calibration results, but also the quality of your content even if your calibration looks great. In the image below with the red 'x', the red squares show invalid areas in the environment. Enshrouding your environment in matte grey curtains can help reduce this issue.
Ideally, you want a strong consistent orange/yellow color in your background, as shown below in the image with a green star.
Hanging up curtains that are a matte solid color (preferably a color such as grey) around your environment will help reduce the depth noise and ensure a higher quality capture.
The Azure Kinect cameras come with a white cover that obscures the sync in and out ports. To remove the cover, use the included hex wrench to remove the two screws on the rear of each camera and slide it out.
The following is a breakdown outlining the power budget required for an 8 camera setup:
- Computer: ~1000W
- Azure Kinect: ~6W/per camera (48W total for 8 Azure Kinects)
- Monitor: ~100W
- Lights: ~40W/per light (160W total for 4 lights) Total: ~1,308W (add buffer)
The ideal flooring is matte carpet. Shiny floors, as well as flooring such as hardwood that yield a lot of invalid pixels (black) in the depth view can also cause problems. We suggest using a rug if you have an especially glossy floor or notice these issues in the depth view. Concrete can work, provided it is diffuse and not shiny - matte grip spray may do the trick for short periods of time. A matte grey rug will help reduce calibration and depth noise issues. If the rug is too dark in infrared, it will crush the dynamic range and add noise. The picture with red rectangles is an example of a bad floor. The picture with a green rectangle is an example of a good floor.
The optimal setup for a full-body, 8 camera capture is 4 "high" cameras, set at a height above the height of your subject to ensure top of head coverage, and 4 "low" cameras set approximately 3-4 feet above the ground. If you experience texture bleed (i.e. hand texture on the body) in your capture sessions, and texture bleed removal is enabled, you may have to increase the height of your cameras or utilize a top-down camera.
High cameras should be set roughly 10 feet apart from each other to form a square and tilted slightly downwards. If cameras are placed closer together, this can potentially degrade quality.
Low cameras should be placed in-between high cameras and set back about a foot to ensure ample floor coverage if a subject is moving. These cameras should not be tilted up or down, but rather straight on. If these cameras are set lower than 3 feet, they will become more susceptible to depth noise which will result in a worse quality calibration and capture.
All cameras should be placed in a vertical orientation with their color lens down. When you have the Capture Suite running, you want to head to the depth view of each camera. You want to make sure the cube is roughly in the middle to top third of each camera view and not chopped off - you want some headroom (not too much) between the calibration cube and the top of the depth view. If not adhered to, this will cause calibration to yield unfavorable results.
First, plug the power cable into your camera and the other side into the power adapter. Then, connect the USB-C cable into the camera and the other side into the active extension cable.
Each included USB-C cable can be used with an extender as-needed and plugged into a port on the StarTech USB PCIe card, avoiding the USB ports on your motherboard. Note that for every 4 cameras we recommend a StarTech USB 3.0 PCIe card with at least 4 dedicated 5Gbps channels.
If cameras do not appear in the Capture Suite, even though they are plugged into the PCIe card, try re-seating your PCIe card.
Important to note, ensure that all of your cables have consistent strain relief. If this is not done, cables are more prone to breaking and can cause corrupt JPEGs which will impact calibration, quality, and recording.
Azure Kinect Power Light Indicators:
- Not Lit: The device is not powered and not connected to the PC. Make sure that the round power connector cable is connected to the device and to the USB power adapter. Also, ensure that the USB-C cable is connected to the device and to a USB 3.0 port on your PC.
- Solid White: The device is powered on and working correctly.
- Flashing White: The device is powered on but doesn't have a USB 3.0 data connection. Make sure that the round power connector cable is connected to the device and to the USB power adapter. Make sure that the USB-C cable is connected to the device and to a USB 3.0 port on your PC. Connect the device to a different USB 3.0 port on the PC. On your PC, open Device Manager (Start > Control Panel > Device Manager), and verify that your PC has a supported USB 3.0 host controller.
- Flashing Amber: The device doesn't have enough power to operate. Make sure that the round power connector cable is connected to the device and to the USB power adapter. Make sure that the USB-C cable is connected to the device and to a USB 3.0 port on your PC.
- Amber, then Flashing White: The device is powered on and is receiving a firmware update, or the device is restoring the factory settings. Wait for the power indicator light to become solid white.
More Azure Kinect troubleshooting instructions can be found on Microsoft's website here.
To help with synchronizing the cameras, daisy-chain the hardware sync cables, also known as 3.5mm audio (AUX) cables, starting from the front camera’s rear port marked “Out” and going into an adjacent camera’s port marked “In”. The front camera is the camera which faces the "p" marker on the calibration cube (2 squares over 3 squares). With an 8 camera setup, you will be using 7 cables total, as the cameras don’t need to form a completely connected chain.
Both warm and cold temperatures, can negatively impact the quality of your calibration and captures. For this reason, use a temperature controlled room, with quality airflow, with temperatures set between 10 and 25 celsius. It is ideal to be closer to the middle of that temperature range. The camera tabs within the Capture Suite do have a temperature indicator to alert you when the camera is outside of the operating range. Going to that camera tab and opening the device statistics section will show the current camera temperature.
We strongly suggest a room with minimal fluorescent, neon, halogen, incandescent, and sunlight, as all of these light sources can interfere with the depth sensors and generate more depth noise/yield worse calibrations. This can cause more texture bleed issues, such as fingers bleeding onto the torso.
We recommend using freestanding LEDs to light your subject, set at 5600K. These lights should be placed behind the camera setup at a height lower than the high cameras - this placement is important as you want to limit lens flares; you may need to modify the placement depending on your environment. Utilizing 4 lights should be fine for most use cases. Make sure your light source has a way to dim the lighting intensity so that you have more control. Lights that are too bright could not only blow out the color of your capture quality, but also yield calibration issues.
Make sure that when you setup your lighting, no lens flares are seen on any color views for the cameras. This can impact calibration and quality of results.
Ideal lighting placement is set at a height of 4 to 5 feet and set behind the low corner cameras.
If you elect to record audio and choose to use a microphone on a stand in lieu of a wireless microphone, position the microphone so that it is near the subject being captured but does not fall on the mesh when viewing the preview in the Capture Suite.
The audio interface can be plugged into a USB 2.0 port.
Calibration Cube Placement
The spot at which you place the calibration cube is where the subject should stand when being captured. This should be roughly in the center of your capture area. It’s a good idea to place a piece of tape as a marker wherever the stand is so that you know the spot of calibration. The marker face which has 2 squares on top of 3 squares ('p') should face your front high camera, which is how your subject should orient themselves.
The cube should sit on a stand at a height of about 4 to 5 feet.
To make a capture, you will need to calibrate your cameras using our calibration cube. Calibration cubes ship unassembled. Follow the instructions below to assemble your cube, or check out the instructional video here.
Lay out and identify all of the pieces of your calibration cube (from perspective of looking at side "P", the front of the cube):
- “P” - front facing panel (this will face your front high camera)
- “T” - left facing panel
- “S” - right facing panel
- “L” - back facing panel
- Square “L” - top facing panel
- White square - cube base
Seat the bottom of the “P” panel to the pegs on the white base of the cube. The panel may be a little less sturdy until you start to get your other panels attached.
Next, you will want to seat the bottom of your “T” panel into the base, to the left of the “P” panel. From there, you will want to gingerly snap the sides of the panels “T” and “P” together. You will hear some little clicks when you know the panel is fully seated. Sometimes, the panels become unseated when attempting to place the following panel. This is ok! Just re-seat/attach your disconnected panels and continue forward!
Perform the same procedure described in step #3 for panel “S,” to the right of panel “P.”
Perform the same procedure described in step #3 for panel “L,” directly behind panel “P.” This is the back panel of your cube.
- Place the smaller, square shaped “L” panel on top of the cube. You want to make sure this panel’s marker tail is facing down towards the front facing “P” panel.
- Once you’ve finished all of these steps, you want to go around the entire cube and make sure all of the tabs are securely snapped together.
- Locate the small screw hole on the bottom of your cube’s base. Screw your tripod’s ball head into this screw hole.
- Place the cube in the center of your camera array with the "P" panel facing the front high camera - this is how the subject should orient themselves.
- You’re ready to calibrate!
Wardrobe, Hair, and Object Capture
When selecting a wardrobe for your subject, avoid materials that absorb infrared light such as shiny, plastic, or reflective items as well as black cloth. These materials will not appear properly in the capture.
Some hair types and objects will not work great with the Azure Kinect cameras due to lack of infrared reflectivity. Using Vitamin E or Keratin hairsprays can bring back some volume for hair. These sprays, in addition to body powder sprays such as Gold Bond Talc-Free Body Powder Spray, can also make some objects appear as well. You can note this issue by going to a depth view of a camera and seeing invalid pixels (black pixels) on the object or hair of the subject.
Long sleeve shirts and pants can help reduce any potential background bleed on arms and legs.