> For the complete documentation index, see [llms.txt](https://help.movin3d.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://help.movin3d.com/movin-tracin-stage/setup/live-settings.md).

# Stage 2 · Settings

Set the capture zones, learn the empty stage, and configure OSC output. The pipeline runs in preview here — nothing is broadcast until you go live.

Entering Settings starts the pipeline — the sensor-data processing system — in preview mode. The sensors stream, but nothing is broadcast yet.

Give it 5–15 seconds to warm up — the view may look empty for a moment.

For the first setup, complete **Set capture zones**, learn the empty stage in **Post processing**, then configure **OSC output**. Leave optional tuning at its current values until your first output is confirmed.

## Set capture zones

<figure><img src="https://1990205953-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FsiYoXLAF78e1PzseNXtd%2Fuploads%2Fgit-blob-bdf0b88549740f5858e70e8864e038565679e595%2Flive-01-zones.png?alt=media" alt="Capture zones over the live point cloud"><figcaption><p>The two zones over the live scene: cyan (outer) and amber (inner).</p></figcaption></figure>

* **Point cloud capture zone** (cyan, outer) — the volume sent when **Source** is **zone**. Enter min and max for each axis, or drag the box in the viewport.
* **Motion capture zone** (amber, inner) — the volume used by the **human** source and motion tracking. It must stay inside the cyan zone.
* **Zone presets** — save the layout under a name and reload it for the next show.

{% hint style="info" %}
If the two zones are exactly the same size, only one box is visible. A note appears with a **Pad capture zone** button that separates them so you can edit both.
{% endhint %}

### Check the capture area

Values are in meters. Use the viewport's axis labels and Top, Front, and Side views when editing; do not assume screen-left is stage-left — check the axis labels. The amber zone should cover the performer's full height and intended movement area, and stay inside the cyan zone. Changes save automatically.

After learning the background, have one person walk the center and edges of the area. Check that their points remain inside the intended zone and that **People** responds when they enter. If they disappear at an edge, check the zone bounds and sensor coverage before going live.

## Post processing

<figure><img src="https://1990205953-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FsiYoXLAF78e1PzseNXtd%2Fuploads%2Fgit-blob-928a043d1feebd223b223de08a9b6bf52672c8fd%2Flive-02-postproc.png?alt=media" alt="Post processing — orientation, background removal, voxelization, and smoothing"><figcaption><p>World orientation, background removal, voxelization, and motion smoothing.</p></figcaption></figure>

Start with **Learn the background** below; it is required before GO LIVE. If the stage direction needs changing, apply **World orientation** first, then learn the background in that orientation.

### Learn the background

Background removal uses a point-cloud scan of the empty stage. It affects the **human** source and motion capture, not **all** or **zone**.

1. Clear the stage. Fixed set pieces that belong to the background can stay.
2. Turn on **Remove background**.
3. Confirm the dialog that appears. On a first setup its only button is **Learn now**. If a background was saved earlier, the dialog offers **Use the saved one** — take that if the stage has not changed, or **Learn a new one** if it has.

Learning takes about 10 seconds. Progress appears in the section.

{% hint style="warning" %}
Keep everyone off the stage while it learns. Anyone standing there becomes part of the background — and people in that spot later get cut out of the cloud.
{% endhint %}

Learn it again whenever the set changes. Anything added afterward stays in the cloud, and a person standing where something was removed loses part of their points.

{% hint style="info" %}
A **Learn now** or **Re-learn** button appears beside the section title only when the background needs attention. To replace a background that is still good, turn **Remove background** off and on again, then choose **Learn a new one**.
{% endhint %}

## OSC output

<figure><img src="https://1990205953-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FsiYoXLAF78e1PzseNXtd%2Fuploads%2Fgit-blob-d089927ef9f399eacd024f63a507fbffd22f378d%2Flive-03-osc.png?alt=media" alt="OSC output — target, source, and mocap settings"><figcaption><p>The streaming destination and source, configured before you go live.</p></figcaption></figure>

Connect the receiving computer to the server's **OSC** port with a network cable, directly or through your show-network switch. The **LIDAR** port is for the sensors only.

1. Create an OSC UDP input in your receiving application. Enter that computer's wired IPv4 address and the input's port as **Target IP** and **Target port**, then select **Apply**. A Wi-Fi address or `localhost` will not work, and the computer's firewall must allow the application's inbound UDP.
2. Select **Test connection** and look for `/movin/test` in the receiving application. The test is sent even in preview; show data flows only after GO LIVE.
   * **The message appears** — the connection works.
   * **The computer answered but no message** — check the receiving port, UDP input, address filter, and firewall.
   * **The computer did not answer** — check the IP and cable, but look at the receiver too; some computers block ping while still receiving UDP.
3. Choose the point-cloud **Source**:
   * **all** — the full fused cloud.
   * **zone** — only the points inside the cyan zone.
   * **human** — the background-removed cloud inside the amber zone.
4. Under **Mocap · on start**, set **Max tracked people (slots)**. The read-only **Slot limit (this host)** field below it shows the cap for your server. Changes apply on the next pipeline start or restart. Restart briefly interrupts processing and output.

Skeleton data is a separate mocap output — it streams regardless of the source. See [OSC message format](/movin-tracin-stage/reference/osc-format.md) for the message layout.

If a **Mocap backend** menu appears here, leave it alone unless your MOVIN support team asks you to switch. A change applies on the next start or restart.

{% hint style="info" %}
Individual prop tracking is not supported. Props can appear in the cloud and, depending on their shape and the sensor view, can become tracking candidates. Check your actual set and props before the show. A distant or mostly hidden person may also be missed; check coverage and contact MOVIN support if needed.
{% endhint %}

**If you changed the person limit or backend**, select **Restart** in the top bar while still in preview and confirm. Wait for the preview to return, then check **People** and background readiness.

## Optional tuning

Come back to these once the basic setup works. Each applies immediately unless noted.

* **World orientation** — rotate the stage (**Yaw** 0, 90, 180, or 270 degrees) or select **Mirror X** so forward matches your content, then select **Apply**. The capture zones turn with the stage. The background does not — learn it again afterward.
* **Point cloud voxelization** — merges nearby points to reduce the data sent. Turn on **Voxelize** and set the **Voxel size**. In the viewport it shows on **Zone** and **Human** only.
* **Motion smoothing** — steadies the skeletons your show software receives, with a tradeoff between steadiness and responsiveness. Drag **Strength** (0–100%). The viewport does not reflect it, so check both slow and fast movements in the receiving application.

Continue to [Stage 3 · Live](/movin-tracin-stage/running-a-show/running-a-show.md).


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://help.movin3d.com/movin-tracin-stage/setup/live-settings.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.