Reliable spatial streaming depends on more than a fast internet connection. This checklist helps creators and event teams estimate bandwidth, measure end-to-end latency, test each part of the delivery chain, and choose a practical fallback before a live holographic or 3D event begins.
Overview
Spatial streaming can describe several workflows: a conventional camera feed shown inside an immersive environment, a real-time 3D scene, volumetric video, a digital avatar, or a feed sent to a venue display that creates a holographic effect. Each workflow has different network demands. A flat 1080p program delivered to a venue is not equivalent to a multi-view volumetric stream rendered for interactive viewers.
The useful planning question is not simply, “How much bandwidth does a hologram need?” Instead, map the complete path:
- Capture: cameras, tracking systems, microphones, and sensors produce source data.
- Production: a workstation or switcher composites, renders, encodes, and monitors the program.
- Contribution: the encoded feed travels from the production site to a platform, cloud service, or remote control room.
- Distribution: the platform creates one or more audience streams and delivers them to viewers or venue endpoints.
- Playback: a headset, mobile device, browser, LED wall, projection system, or holographic display decodes and renders the experience.
Plan each leg separately. A venue may have enough download capacity for playback but not enough upload capacity for contribution. A cloud platform may accept the source feed while the audience experiences buffering because the delivery profile, device support, or regional network path is unsuitable.
Use the figures below as planning starting points rather than universal specifications. Final requirements depend on the codec, frame rate, color depth, scene complexity, number of views, platform limits, and whether the stream is compressed video or a real-time 3D data feed.
What to track
1. Bitrate and usable bandwidth
Record the actual encoded bitrate, not only the advertised quality label. A stream that averages 8 Mbps may produce higher short-term peaks, so the connection needs headroom for those bursts and for ordinary venue traffic. As a working rule, reserve capacity above the measured stream rate rather than operating the link at its apparent limit. A separate production network or dedicated wired connection is preferable to sharing the contribution path with guest Wi-Fi, cloud backups, or large software updates.
The following table is a deliberately broad planning matrix. It describes the program feed or delivery stream, not the total bandwidth of a volumetric capture system.
| Workflow | Starting bitrate to test | Practical connection target | Notes |
|---|---|---|---|
| HD presenter, avatar, or event feed | About 4–10 Mbps | At least 2x the sustained upload rate where possible | Test motion, graphics, and scene changes; avoid relying on the average alone. |
| 4K single-view program | About 12–35 Mbps | At least 2x the sustained upload rate where possible | Codec, frame rate, and platform ingest limits can change this substantially. |
| Multi-view or stereoscopic delivery | About 20–60 Mbps combined | Prefer dedicated capacity with measured headroom | Count every view or eye delivered, unless the platform packages them together efficiently. |
| Volumetric or real-time 3D feed | Measure from the chosen encoder and scene | Use a sustained test above the measured peak | Point clouds, meshes, textures, and metadata can vary too widely for one fixed number. |
For audience scale, estimate platform egress rather than multiplying the venue upload by the viewer count when a managed delivery service is used. If you are distributing directly, calculate the encoded bitrate multiplied by concurrent viewers, then add operational margin. Confirm whether the platform supports the intended resolution, frame rate, immersive format, simultaneous viewers, and geographic audience before committing to the workflow.
2. Latency across the whole chain
Measure glass-to-glass latency: the time between an action at capture and the corresponding image or sound at playback. Break it into capture and processing delay, encoder and buffer delay, network travel, platform processing, decoder delay, and display rendering. A low-latency contribution feed can still feel delayed if the player adds a large buffer.
Choose a target based on the experience. A prerecorded-looking presentation can tolerate more delay than a performer responding to a remote audience. Interactive spatial live events, synchronized remote performers, and audience participation generally need tighter control than a one-way showcase. Set an acceptable range, a warning threshold, and a failure threshold before rehearsal; do not define success only as “it worked.”
3. Stability, jitter, and loss
Track packet loss, jitter, round-trip time, disconnects, encoder queue growth, dropped frames, decoder errors, and playback stalls. A connection with sufficient headline speed can still fail when latency varies sharply or packets arrive out of order. Log both the median and the worst observed result during a sustained test.
4. Device and venue constraints
Record the playback device, operating system, browser or application version, display output, refresh rate, and rendering load. For a venue, also record the playback computer, switcher, media server, display processor, cable path, and network port. A 3D live streaming network requirement is incomplete if it ignores local rendering capacity or a display pipeline that adds several frames of delay.
Cadence and checkpoints
Run a short health check before every rehearsal, a full path test before every venue move, and a documented readiness test before the broadcast. Recheck the baseline monthly or quarterly for recurring programs, and whenever the platform, codec, venue network, capture resolution, audience geography, or playback device changes.
Preflight checklist
- Confirm the chosen ingest protocol, codec, resolution, frame rate, audio format, and platform limits.
- Measure wired upload and download performance at the actual production position.
- Run a sustained contribution test for at least the length of the longest expected segment, including scene complexity and peak motion.
- Record bitrate, peak bitrate, packet loss, jitter, round-trip time, dropped frames, encoder queue, and reconnect behavior.
- Test the complete route from capture through cloud delivery to the real venue display or audience device.
- Test at the same time of day when possible, particularly in shared venues where network use may vary.
- Verify a backup connection, alternate encoder profile, local recording, and a static or prerecorded fallback.
- Assign one person to watch technical metrics while another monitors the audience or venue experience.
Use this compact test log for every run:
| Field | Record |
|---|---|
| Date, venue, and test type | Rehearsal, move, platform change, or event readiness |
| Source and delivery profile | Resolution, frame rate, codec, bitrate, views, and audio format |
| Connection | Provider or network type, wired or wireless, measured upload and download |
| Quality metrics | Peak bitrate, packet loss, jitter, round-trip time, dropped frames, and stalls |
| Latency | Capture-to-display result and method used to measure it |
| Playback endpoint | Device, application, display, render load, and output path |
| Decision | Pass, pass with mitigation, or fail; include the owner and next action |
How to interpret changes
When quality declines, identify where the symptom begins instead of immediately increasing the bitrate. If the encoder queue grows before the network test changes, investigate GPU or CPU load, capture settings, and storage. If contribution is clean but venue playback stalls, inspect the platform profile, local download path, decoder, and display system. If only remote viewers report problems, compare device, region, and player data rather than changing the on-site connection blindly.
Use this fallback decision tree:
- Is the source unstable? Reduce capture complexity, resolution, frame rate, or active views; keep a local recording running.
- Is upload capacity insufficient? Move to the dedicated wired path or backup connection, then switch to a lower tested profile.
- Is latency too high but the feed stable? Review encoder and player buffers, platform mode, and unnecessary processing stages.
- Is only the 3D layer failing? Deliver a conventional camera or rendered 2D program while preserving audio and essential event information.
- Is the venue endpoint failing? Use a local playback file or pre-rendered scene, and keep the live feed available as a secondary source.
Do not treat a higher bitrate as a universal fix. More data can increase congestion, encoding load, and buffering. A lower, stable profile with known device support is often the safer live production choice.
When to revisit
Review this plan on a monthly or quarterly cadence for recurring shows, and immediately after a failed rehearsal, venue move, platform update, codec change, audience expansion, or new immersive device rollout. Revisit the assumptions whenever you change from a single-view program to multi-view, add remote performers, introduce interactive controls, or move from a controlled venue network to public internet delivery.
At each review, compare the latest test log with the previous baseline. Look for rising latency, smaller bandwidth margins, more frequent reconnects, increased render load, or a growing gap between the production feed and viewer experience. Update the requirements table, fallback profiles, supported devices, and contact list. Keep the last known-good configuration available so the team can restore it quickly.
Before the next live holographic event, schedule one complete path test, confirm the pass criteria in writing, and rehearse the fallback rather than merely documenting it. That routine turns bandwidth and latency from last-minute risks into measurable production variables.