Windows 11 KB5063878 NDI Stuttering: Fix by Changing NDI Transport

Microsoft has confirmed that the August cumulative security update for Windows 11 (KB5063878, OS Build 26100.4946) is causing severe stuttering, lag, and choppy audio/video when using NDI-based streaming flows — an issue that hits production tools such as OBS and NDI Tools and has forced broadcasters to apply a manual transport-mode workaround while Microsoft investigates. (support.microsoft.com, bleepingcomputer.com)

Background​

Microsoft shipped the August 12, 2025 cumulative update identified as KB5063878 for Windows 11 version 24H2 (OS Build 26100.4946). The patch is a routine security and reliability rollup, but within days multiple user reports and vendor advisories surfaced describing problems specifically affecting NDI (Network Device Interface) traffic and the apps that rely on it, most notably Open Broadcaster Software (OBS) and the official NDI Tools suite. Microsoft logged the symptom on its Windows release-health page and recommended an interim configuration change for impacted NDI users while it works on a permanent fix. (support.microsoft.com, neowin.net)
This is not a universal consumer streaming issue — mainstream streaming services such as YouTube, Netflix, and Disney+ are not impacted by this specific fault. The problem is transport-level and appears when NDI is used to transfer screen captures and other local PC sources between machines (for example, a capture PC sending Display Capture to an encoder PC). Microsoft’s advisory explicitly ties the symptoms to NDI’s default Reliable UDP transport (RUDP) and signals that switching NDI to use Single TCP or legacy UDP restores functional behavior in affected environments. (bleepingcomputer.com, neowin.net)

---

Overview: What’s breaking and who is affected​

• Symptoms: Severe stuttering, dropped frames, lag, and choppy audio/video in NDI-based streams even on low-bandwidth networks; problem especially noticeable when Display Capture is used on the source machine. (bleepingcomputer.com, neowin.net)
• Affected components: OBS (and other capture/production tools that use NDI), NDI Tools (NDI Access Manager), and any workflow that relies on NDI’s default RUDP transport. (bleepingcomputer.com, docs.ndi.video)
• Windows builds: Reported after installation of KB5063878 on Windows 11 24H2 (OS Build 26100.4946); parallel KBs for Windows 10 show related behavior in some reports. (support.microsoft.com, bleepingcomputer.com)
• Scope: Production/creator workflows that use multi‑PC capture and NDI over LAN are at risk; typical consumer streaming from single‑PC set‑ups (direct encoder to platform) is not implicated. (bleepingcomputer.com, neowin.net)

These are the load-bearing facts: Microsoft has acknowledged the behavior, the problem is reproducible in real-world NDI flows, and a vendor‑recommended workaround exists. Those points are corroborated by Microsoft’s support entries and independent reporting. (support.microsoft.com, bleepingcomputer.com)

---

Technical deep dive: NDI transports, RUDP, and why this matters​

How NDI moves video across a LAN​

NDI (Network Device Interface) supports multiple transport modes to balance latency, reliability, and network compatibility:

• RUDP (Reliable UDP) — NDI’s default in modern versions (NDI 5): attempts to combine UDP’s low-latency behavior with reliability mechanisms (sequencing, retransmissions, flow control) to reduce visual artifacts while keeping latency low. (docs.ndi.video)
• UDP (Legacy) — simple datagram transport with minimal overhead and low latency; no retransmission guarantees.
• Single TCP — reliable, ordered byte stream; robust in networks that mishandle UDP but susceptible to head‑of‑line blocking and higher latency under loss. (docs.ndi.video)

NDI’s Access Manager exposes these options under the Advanced tab so administrators and broadcasters can select the transport mode appropriate to their environment. When NDI applications start they read a shared configuration file (%programdata%/NDI/ndi-config.v1.json), so changes require restarting NDI clients to take effect. (docs.ndi.video)

What appears to have regressed in KB5063878​

Microsoft’s public guidance and vendor reports indicate the regression surfaces specifically when NDI uses RUDP: packet-handling anomalies — such as retransmission storms, unexpected packet drops, or altered socket timing/behavior — lead to pronounced stutter even though raw bandwidth is sufficient. The measurable effect is that streams appear jagged, audio drifts, and frame drops occur in patterns consistent with transport-layer retransmission or queuing problems. (bleepingcomputer.com, neowin.net)
At present, Microsoft and vendors have not published a line-by-line root-cause — whether the change is in the Windows network stack, a timing change in socket internals, a scheduler interaction, or an unintended side‑effect of a validation/security path is not yet confirmed publicly. That lack of a complete, public root‑cause is important: without it, any permanent mitigation must come from Microsoft and requires careful testing before broad redeployment. Treat any speculation about internals as provisional until Microsoft publishes a technical breakdown. (support.microsoft.com)

---

The immediate, tested workarounds (step-by-step)​

For streamers and broadcast engineers who are seeing NDI stutter after KB5063878, vendors and Microsoft recommend switching NDI’s Receive Mode away from RUDP. The change is reversible and does not require uninstalling security updates.

• Download and install the free NDI Tools pack (if not already installed).
• Run NDI Access Manager on the receiving machine.
• Select the Advanced tab.
• Change Receive Mode from RUDP (default) to Single TCP or UDP (Legacy).
• Click OK and then Restart any NDI‑receiving applications (OBS, vMix, Studio Monitor, etc.) so they re-read configuration. (neowin.net, docs.ndi.video)

Practical notes and trade-offs:

• Single TCP is the safest universal fallback across complicated network stacks and equipment that mishandle UDP variants, but it can add latency and is vulnerable to head‑of‑line blocking if the network experiences packet loss. Use it for reliability when latency budget permits. (docs.ndi.video)
• UDP (Legacy) preserves better latency behavior on stable LANs but removes RUDP's reliability mechanisms — it’s a good option on a dedicated, low‑loss LAN. (docs.ndi.video)
• After changing the setting, restart all NDI apps — they read the config at startup; running apps will not pick up the change automatically. (docs.ndi.video)

If changing transport is not immediate or feasible (for example, remote contributors you cannot reconfigure), some organizations may consider rolling back the KB in test or maintenance windows. Rolling back security updates carries risk — document the decision, apply compensating controls, and coordinate with security teams. Microsoft is investigating and will publish a permanent remediation; prefer that route for long-term stability. (bleepingcomputer.com, support.microsoft.com)

---

How to detect if you’re impacted​

• Check whether KB5063878 is installed and which OS build you’re on:
• Run winver or go to Settings > System > About to confirm OS Build 26100.4946 if KB5063878 is present. (support.microsoft.com)
• For scriptable checks, use PowerShell: Get-HotFix -Id KB5063878 (note: Get‑HotFix uses Win32_QuickFixEngineering and may not list every update in every configuration; consult Windows Update history if needed). (learn.microsoft.com)
• Test your production flow: perform a 10–15 minute sustained NDI capture test (the same capture modes you run live) after changing transport to validate that frame rate, sync, and audio continuity are stable.

---

Operational guidance for streamers, studios, and IT admins​

• For live shows, treat KB5063878 as a known risk if your stack uses NDI/RUDP — test early and decide whether to:
• Apply the NDI transport change on all receiving machines, or
• Delay or roll back the KB in controlled maintenance windows for mission‑critical systems while balancing security risk.
• For studios managing many devices:
• Pause automatic approvals for KB5063878 in WSUS/Windows Update for Business test rings until Microsoft issues a KIR (Known Issue Rollback) or patch.
• Pilot the NDI Receive Mode change on a small device subset and collect quality metrics.
• If rollback is necessary, document compensating controls and schedule reinstall of security updates after remediation is available.
• Communicate changes to remote contributors and show partners: include a pre-show checklist that confirms NDI transport mode and OS build to prevent last‑minute surprises during live events.

---

Risk assessment and trade-offs​

• Security vs. reliability: Uninstalling a security update is never an ideal long-term strategy; it reduces exposure to the vulnerabilities the monthly rollup mitigates. Wherever possible, prefer configuration workarounds (NDI transport change) and wait for Microsoft’s patch. (support.microsoft.com)
• Performance vs. latency: Switching to Single TCP trades some latency for reliability — that may be acceptable for interview streams and slide decks but less so for ultra-low-latency competitive game capture. UDP (Legacy) maintains low latency but loses the retransmission and flow-control benefits of RUDP. Choose based on the real-time requirements of your show. (docs.ndi.video)
• Operational complexity: Updating NDI settings across dozens of distributed machines adds coordination overhead; consider using scripted configuration management or an image update if you manage many endpoints. The NDI Access Manager settings are stored centrally in %programdata%/NDI/ndi-config.v1.json, which IT teams can deploy or modify as part of configuration automation. (docs.ndi.video)

---

What Microsoft and vendors have said — and what remains unknown​

Microsoft has acknowledged the NDI streaming symptom set and posted guidance on the Windows Release Health dashboard; the company said it is investigating and advised the temporary workaround of switching NDI transport away from RUDP. Independent reporting by industry outlets (and vendor statements from the NDI team) corroborate Microsoft’s advisory and the efficacy of switching to TCP/UDP as a mitigation. (support.microsoft.com, bleepingcomputer.com, neowin.net)
What is not yet public is the definitive technical root cause. Multiple plausible mechanisms exist — a timing change in the network stack, altered socket semantics, unexpected interaction with retransmission logic, or a security validation pathway interfering with RUDP — but none are confirmed by Microsoft at this time. Until Microsoft publishes a formal technical breakdown, treat detailed root-cause narratives as speculative. (support.microsoft.com)

---

Checklist: Quick actions for creators and sysadmins​

• Immediate (first 1–2 hours)
• Verify whether KB5063878 is installed (winver / Settings or Get-HotFix). (support.microsoft.com, learn.microsoft.com)
• If NDI-based workflows show stutter, change NDI Receive Mode to Single TCP or UDP (Legacy) and restart NDI apps. (neowin.net, docs.ndi.video)
• Run sustained capture tests in the exact configuration used for live shows for at least 10–15 minutes.
• Short-term (same day)
• Communicate the change to remote guests and partners.
• If you manage many endpoints, pilot the change on a test ring before mass deployment.
• Medium-term (days to weeks)
• Watch Microsoft’s Windows Release Health and vendor advisories for a KIR or replacement LCU that addresses the regression.
• Plan to revert temporary transport changes only after the fix is validated end‑to‑end on a representative set of hardware. (support.microsoft.com)

---

Wider context: Why this incident matters for update management​

This episode is another reminder that cumulative rollups carry systemic risk for specialized, real‑time workflows. NDI is widely used in multi‑PC production environments because it offers flexibility and high quality with commodity hardware. A change in how Windows handles RUDP semantics — whether intentional or accidental — can cascade into real, revenue‑impacting outages for live events, e‑sports broadcasts, corporate webinars, and creators. The takeaways for IT and production teams are clear:

• Maintain a staged update policy for production systems (test rings and pilot deployments).
• Keep a playbook for emergency mitigations (configuration changes, rollback procedures, communications templates).
• Test the exact capture modes and network conditions you operate under; synthetic tests often fail to stress the same code paths as real Display Capture + NDI flows.

---

Final analysis and verdict​

The facts are plain: Microsoft’s August rollup (KB5063878) correlates with reproducible NDI stuttering when RUDP is used; Microsoft has acknowledged the issue and recommended a practical, reversible workaround — change the NDI transport to TCP or UDP while a permanent fix is developed. Independent reporting and vendor documentation back that guidance, and NDI’s own docs explain why conservative transport choices are viable when the default RUDP path behaves unexpectedly. (support.microsoft.com, bleepingcomputer.com, docs.ndi.video)
Strengths of the response so far:

• Microsoft and vendors published an interim mitigation promptly, which enables show continuity for many affected teams. (neowin.net)
• The workaround is configurable and reversible; it does not require uninstalling security updates for many environments. (docs.ndi.video)

Risks and remaining concerns:

• The permanent root cause is not yet public; until Microsoft provides a technical remediation, the issue could reappear in related updates or in complex network environments. Treat claims about internals as provisional. (support.microsoft.com)
• Some production teams may find the transport trade-offs unacceptable (added latency under TCP, reduced reliability under UDP), forcing more complex operational workarounds or last‑minute rollbacks with security implications.

Immediate priority for production teams is stability: test the NDI Receive Mode change now, stage updates conservatively, and follow Microsoft’s updates for an official patch. For administrators, the long-term lesson is to align update cadence with the needs of real‑time workflows and to keep an incident playbook ready for rapid configuration rollouts.

---

This is a live issue under active investigation; keep an eye on Microsoft’s Windows Release Health notices and vendor advisories for the formal fix and KIR announcements. (support.microsoft.com, bleepingcomputer.com)

---

Source: PCWorld Latest Windows 11 update kills streaming app performance