• Thread Author
The August cumulative update for Windows — most notably KB5063878 (Windows 11 24H2, OS Build 26100.4946) and companion rollups for Windows 10 — has been identified as the root cause of a wave of streaming and system problems that emerged in mid‑August and persisted into September, producing severe stuttering, dropped frames, and audio drift in multi‑PC NDI workflows used by Twitch and YouTube broadcasters. Microsoft has acknowledged the regression and vendors including the NDI tools and Streamlabs have published immediate mitigation steps while engineering works on a permanent fix. (support.microsoft.com)

Futuristic multi-monitor workstation with RGB-lit towers and glowing data streams.Background / Overview​

NDI (Network Device Interface) is a widely used video‑over‑IP protocol that enables low‑latency, high‑quality audio and video transmission across local networks. It powers multi‑PC streaming rigs where one machine captures Display Capture and sends high‑frame‑rate, high‑bitrate content to an encoder or switcher over Ethernet. In the August 12, 2025 Patch Tuesday rollup, Microsoft shipped KB5063878 for Windows 11 (and parallel updates for Windows 10), a security and servicing package that fixed multiple vulnerabilities but also introduced unexpected changes in network and installer behavior. Within days, broadcasters and AV professionals reported reproducible regressions specifically when NDI used its default Reliable UDP (RUDP) transport: severe jitter, rhythmic micro‑pauses, and audio‑video desynchronization. Microsoft recorded the behavior on its Release Health pages and pointed affected users to vendor guidance while investigations continue. (support.microsoft.com)
Why this matters: the problem is not general playback on Twitch or YouTube; it targets the production chain used by creators who use NDI to move local capture sources between machines. For those creators, the observable impact is immediate and severe — live shows rely on consistent timing and low latency, and the regression undermines both. Independent coverage from the specialist press and vendor advisories corroborate Microsoft’s description of the symptoms and the temporary mitigations. (pcgamer.com) (tomsguide.com)

What exactly broke: symptoms and scope​

  • Primary symptom: Severe stuttering, dropped frames, and choppy audio/video in NDI-based streams even under low LAN load, often worse when the sending machine uses Display Capture rather than a camera feed. (support.streamlabs.com)
  • Transport fingerprint: The failures consistently surface when NDI uses RUDP (Reliable UDP) — NDI’s modern default transport — while UDP (Legacy) and Single TCP transports appear unaffected in tests. This observation led vendors to recommend switching receive mode away from RUDP as a practical workaround. (support.streamlabs.com) (pcgamer.com)
  • Affected builds: KB5063878 (Windows 11 24H2, OS Build 26100.4946) is central to the reports; companion Windows 10 rollups such as KB5063709 have been implicated in parallel reports. Microsoft’s Release Health messaging lists the August rollups as associated with the regression. (support.microsoft.com)
  • Collateral issues reported: Separately, users reported Windows Update install failures (error code 0x80240069), problems with Windows Installer elevation behavior (unexpected UAC prompts or MSI repair failures), and community‑reported storage glitches on certain SSDs. Microsoft acknowledged and is investigating several side effects from the August updates; some are fixed or have targeted mitigations in follow‑up releases. Community testing suggests the storage incidents are more sporadic but serious where they occur. These broader impacts underscore the wide scope of the August cumulatives. (learn.microsoft.com) (windowscentral.com)

Technical analysis: why RUDP trips on the update​

NDI supports multiple transport modes to balance latency and reliability:
  • RUDP (Reliable UDP): The modern default for NDI 5. It overlays sequencing, retransmits, and pacing atop UDP to achieve low latency with reliability.
  • UDP (Legacy): Simple datagram transport that offers low overhead but no retransmission guarantees.
  • Single TCP: Ordered, reliable byte stream; robust but can suffer head‑of‑line blocking under loss.
RUDP depends on precise timing, socket buffering, scheduler behavior, and subtle interactions with the underlying OS networking stack. The August rollup appears to have altered how certain RUDP flows are processed (timing/acknowledgement/queue handling), which for NDI’s workload produced retransmission storms, aggressive pacing, and rhythmic micro‑pauses instead of smooth flow. That behavior matches independent reproductions by vendors and AV engineers: flows using RUDP show severe stutter while Single TCP or UDP (Legacy) flows remain stable. This narrows the likely failure surface to transport‑level timing and socket semantics rather than an application bug in OBS or NDI itself.
This is a classic example of how OS‑level networking adjustments — often introduced for security, reliability, or performance reasons — can produce regressions in higher‑level protocols that rely on implicit timing assumptions.

Who is affected (and who likely isn't)​

  • Affected: Creators and production teams that use multi‑PC setups with NDI to move Display Capture, gameplay, and multi‑camera feeds between machines; broadcast facilities and lecture capture systems that depend on NDI RUDP for low‑latency, synchronized streams. OBS Studio users in multi‑PC configurations, vMix and other switchers that rely on NDI have been prominently affected.
  • Not broadly affected: Single‑PC streaming setups that encode and send directly to Twitch/YouTube without NDI (i.e., OBS on one machine streaming directly) are generally not implicated by this specific regression, unless they rely on local NDI sources. Consumer playback of Twitch/YouTube streams by viewers is also unaffected by the underlying NDI transport regression.

Vendor and Microsoft response​

  • Microsoft: Acknowledged the behavior via its Release Health / support channels for the August cumulative and posted guidance that points to vendor‑provided mitigations while engineering investigates. The company also documented related install and Windows Installer side effects and offered targeted workarounds or temporary fixes for some issues. (support.microsoft.com)
  • NDI vendors & Streamlabs: Published explicit step‑by‑step workarounds: change NDI Receive Mode from Auto / RUDP to Single TCP or UDP (Legacy) using the NDI Tools Access Manager (or Access Manager in the NDI Tools toolbox). Streamlabs and the NDI documentation emphasize that this must be done on both sending and receiving machines for the setting to take effect. (support.streamlabs.com) (pcgamer.com)
  • Press & independent labs: Multiple outlets reproduced the problem and documented the mitigation, increasing pressure for a fix. Coverage by specialist outlets also raised awareness of parallel reports around storage and installer regressions. (tomsguide.com) (windowscentral.com)

Step‑by‑step: practical mitigations for streamers (short term)​

If your stream quality dropped suddenly after mid‑August updates and your setup uses NDI, follow this prioritized list. These steps minimize disruption while preserving security posture when possible.
  • Check whether your machines installed KB5063878 (Windows 11 24H2) or companion rollups for Windows 10. If the affected update is present, proceed with the mitigations below. (support.microsoft.com)
  • Install or open NDI Tools on both the sending and receiving PC (you may need to download the NDI Tools toolbox). Open Access Manager. (support.streamlabs.com)
  • In Access Manager, go to Advanced, and change Receive Mode from Auto (Recommended) / RUDP to Single TCP or UDP (Legacy), then click OK. Repeat on all affected machines (sending and receiving). This is the single most effective temporary fix documented by vendors and Microsoft. (support.streamlabs.com)
  • If changing transport is impractical (e.g., tight latency budget or network design incompatible with TCP), consider:
  • Using a single‑PC streaming workflow (encode locally), if feasible.
  • Temporarily backing out the KB update (control‑panel uninstall or use Windows Update history) — only as a last resort because of security implications. Microsoft’s guidance and advisories recommend mitigations over rollback where possible. (support.microsoft.com)
  • Test thoroughly before going live: perform a local recording and a closed stream to a private channel to validate lip‑sync and frame‑rate stability after the change. If problems persist, collect logs (OBS log files, NDI logs) and vendor support may assist.
A short checklist for live shows:
  • Switch NDI Receive Mode on all machines to Single TCP or UDP (Legacy). (support.streamlabs.com)
  • Confirm Network settings: use wired Gigabit Ethernet, disable Wi‑Fi on production machines, and avoid network switches with packet‑inspection features that could exacerbate timing issues.
  • Run a 10‑minute dress rehearsal to verify sustained stability.

How to detect the issue during a stream​

  • Repeating rhythmic micro‑pauses where video keeps playing for short bursts then halts briefly.
  • Audio gradually drifting out of sync with video.
  • Symptoms show up even when LAN utilization is low, indicating transport anomalies rather than bandwidth saturation.
  • Problem predominantly occurs when the Display Capture source is used on the sending machine (more than with camera sources).
If you observe these symptoms immediately after an update, apply the NDI Receive Mode change and retest.

Risks and trade‑offs: security vs. uptime​

  • Rolling back KB5063878 will likely restore the affected behavior but reintroduce the security fixes the update provided. For production environments where uptime is critical, IT teams must weigh the immediate operational need against the security posture. Microsoft generally recommends using mitigations where possible to retain security patches while limiting functionality regressions. (support.microsoft.com)
  • Switching NDI Receive Mode to Single TCP increases robustness but may add small amounts of latency compared with RUDP; streaming professionals should measure and confirm whether the latency remains acceptable for their workflow. UDP (Legacy) may preserve latency characteristics but lacks RUDP’s retransmit semantics. Choose the transport that best fits your latency and reliability needs. (pcgamer.com)

Broader system reports: installer and storage issues (what to watch)​

In addition to the NDI regression, the August rollups produced several community reports and Microsoft troubleshooting notes:
  • Windows Update install error 0x80240069: Some WSUS and Update Catalog deployments failed with that code; Microsoft documented workarounds and follow‑ups. (support.microsoft.com)
  • Windows Installer / UAC changes: The security hardening introduced stricter elevation checks that caused some MSI repair operations to prompt or fail under standard user contexts (Error 1730 in some reports). Microsoft provided Known Issue Rollback and guidance for enterprise admins.
  • Storage / SSD anomalies: Community testing reported severe issues on some SSDs when handling very large file transfers, with brand and controller patterns (Phison controllers cited in community tests). These reports are under investigation by Microsoft and hardware vendors; users with critical storage workloads should exercise caution and ensure backups. The storage claims are serious but are still being validated across broad datasets — treat them as active investigations rather than confirmed, universal failures. (windowscentral.com)
Flag: the most dramatic storage failure claims (for example, “bricking” across broad sets of SSDs) were reported by community testing and specialist outlets; while they are credible and under active investigation, they are not universally reproduced across all hardware and Windows configurations. Consider them high‑priority but not universally confirmed until vendors provide definitive guidance or fixes. (windowscentral.com)

Testing and verification: how to confidently confirm the issue on your rig​

  • Verify update presence: confirm KB5063878 (or the equivalent Windows 10 package) is installed via Settings → Windows Update → Update history. (support.microsoft.com)
  • Reproduce the symptom locally: on a simple two‑PC NDI test (one sender, one receiver), send a Display Capture source and watch for rhythmic stutter, audio drift, or retransmission patterns in NDI graphs. If the issue appears with RUDP and disappears with Single TCP or UDP (Legacy), it matches the known regression fingerprint.
  • Collect diagnostic artifacts: OBS logs, NDI Access Manager screenshots, and Windows Event Viewer network‑related entries can help vendors triage root causes. If reporting to Microsoft or vendor support, include OS build, KB IDs, and exact NDI version.

What to expect next: timeline and outlook​

Microsoft has acknowledged the regression and vendor engineering teams (NDI, Streamlabs, OBS maintainers) are coordinating mitigations and tests. Historically, platform regressions tied to cumulative updates receive targeted follow‑ups within a few Patch Tuesday cycles or via out‑of‑band servicing updates if the user impact is broad and severe. Expect one of these outcomes in the near term:
  • A targeted Microsoft fix that restores previous RUDP behavior while preserving security mitigations.
  • A vendor‑side adjustment in NDI implementation that tolerates the timing change (less likely because the regression appears at OS transport semantics).
  • Continued guidance for using Single TCP or UDP (Legacy) until a fix is widely available. (support.microsoft.com) (support.streamlabs.com)
Creators should watch official Microsoft Release Health updates and NDI vendor pages for firm timelines. In the interim, the transport change is a validated, low‑risk mitigation that preserves security patches while restoring usability for most multi‑PC streaming scenarios. (support.streamlabs.com)

Recommendations for creators and IT teams (actionable summary)​

  • If you run multi‑PC NDI workflows, immediately change NDI Receive Mode on all affected machines to Single TCP or UDP (Legacy) using NDI Tools Access Manager. This is the most reliable short‑term fix. (support.streamlabs.com)
  • Maintain up‑to‑date backups and avoid large, critical transfers during pending investigations if your storage hardware has any of the risk indicators reported by the community. Flag drives with symptoms for vendor diagnostics. (windowscentral.com)
  • Avoid uninstalling security updates on production machines unless absolutely necessary; instead prefer configuration mitigations. If rollback is unavoidable, plan for compensating security measures. (support.microsoft.com)
  • Test changes in a private or rehearsal stream before going live. Validate audio/video sync and frame stability for at least 10 minutes under real workload conditions.
  • Monitor official Microsoft Release Health advisories and NDI vendor pages for a permanent patch and follow their guidance for re‑enabling RUDP when safe. (support.microsoft.com)

Final analysis: strengths, risks, and what this episode reveals​

Strengths:
  • The issue was identified quickly by the creator community and reproduced by vendors, enabling a focused mitigation (transport mode change) that restores usable behavior. This quick feedback loop between users, vendors, and Microsoft limited the window of disruption for many creators.
  • Microsoft’s Release Health process and vendor advisories provided clear actionable guidance rather than leaving users to trial‑and‑error. (support.microsoft.com)
Risks and blind spots:
  • OS‑level changes meant to improve security or reliability can have outsized side effects on timing‑sensitive protocols like RUDP. This episode highlights a testing gap: real‑world production workflows (multi‑PC live production) are hard to fully emulate in platform QA, and regressions that affect niche but mission‑critical workloads can be missed until wide deployment.
  • Community reports of storage anomalies merit attention: while the NDI regression has a clear temporary workaround, storage failures and installer elevation regressions pose data‑loss and deployment risks that require separate tracking and more cautious mitigation strategies. Flagging unverified claims: the most severe storage “bricking” narratives are community‑sourced and under active investigation — treat them as high‑priority leads but not universal facts until vendors confirm root cause and remediation. (windowscentral.com)
Bottom line: For Twitch and YouTube creators who rely on NDI, the August rollup produced a production‑blocking regression — but a vendor‑endorsed, practical workaround exists and should be implemented immediately. IT teams should balance the urgency of live production against the security benefits of the update when considering rollbacks. Watch official vendor and Microsoft advisories for the final fix and re‑enable the default transport only after it’s confirmed safe.

Creators and system administrators who follow the mitigations above should be able to restore multi‑PC streaming functionality quickly while preserving security updates. Continue to document any anomalous behavior, remain cautious about large storage operations on systems that installed the August cumulative, and monitor Microsoft and vendor channels for the permanent remedy. (support.streamlabs.com)

Source: islaBit Windows update is responsible for issues with Twitch and YouTube live broadcasts - islaBit
 

Back
Top