A vMix stream that keeps disconnecting can be failing in vMix, on the route to YouTube, at YouTube ingest, or only during viewer playback. First identify which part stopped, then compare vMix’s status and session log with YouTube’s timestamped stream-health messages before changing a setting.
For a 24/7 channel, a short test that starts successfully is not enough. Keep a record of the time, the visible symptoms and the exact change you test; a disconnect by itself does not prove its cause, and no single setting can guarantee an uninterrupted stream.
Classify what the disconnect looks like
Start with what you can observe at the time of the event. Is vMix’s Stream button still active, did it change colour, or did the stream stop? Does the local programme output still look and sound right? Does YouTube Live Control Room report an ingestion problem, or are viewers reporting buffering while the encoder and YouTube health panel appear normal? Those are different clues, not interchangeable descriptions of one failure.
| What you see | What it helps you distinguish | First evidence to check |
|---|---|---|
| vMix reports an error or the stream stops | The encoder session may have ended or failed to connect | vMix status and the session log around the event |
| vMix status turns amber or orange after being red | vMix is reporting a change in stream condition; its guidance says this often points to congestion | Status details, outbound conditions and the matching log time |
| YouTube Live Control Room shows a stream-health error | YouTube has detected an issue with the feed it receives | The timestamp and wording of the health message |
| Local output and ingest appear healthy, but viewers report buffering | The symptom may be in delivery or playback rather than a stopped encoder | Whether the report is widespread, and whether YouTube still shows a healthy incoming stream |
The colour or wording narrows the next check; it does not establish a root cause. A viewer’s buffering report alone does not tell you whether vMix stopped sending, the connection faltered, YouTube flagged the incoming feed, or playback was affected downstream.
Write down the time, including the time zone, and whether the local programme output remained correct. If the local output is also wrong, that is useful evidence about the production before it leaves your computer. If it remains correct while the broadcast drops, preserve that distinction rather than immediately changing an input or encoder setting.
Check vMix stream status and session log
In vMix, open the Streaming settings and use View Status to inspect technical information about the active stream. Do this during an issue if possible, or immediately afterwards if the stream has stopped. The vMix streaming help describes the available stream settings and status information. Preserve the newest session log from C:\ProgramData\vMix\streaming; vMix documents this folder for its streaming logs.
Look for the exact message and the time it appeared. An entry such as av_interleaved_write_frame(): Unknown error or Input/output error indicates that the stream disconnected, according to vMix’s FFMPEG troubleshooting guide. The guide lists possible explanations including a local internet interruption, a streaming-provider outage or closure, and third-party antivirus or security software. These are possibilities to investigate, not a diagnosis of your setup.
If the message appears as soon as you try to connect, check the destination URL and stream key against the current values in YouTube Live Control Room. A mistyped, stale or rejected credential can prevent a connection, but check the log and YouTube’s response before assuming that is what happened. If the stream was already running, a later disconnect calls for a different comparison than an immediate connection failure.
An amber or orange state after vMix was red is a distinct clue. vMix says this commonly indicates network congestion and considers inadequate computer or graphics-card performance less likely in that particular status pattern. Treat that as guidance about what to examine first, not proof that your network is at fault. Compare the event with upload conditions, other traffic on the connection, and any relevant network evidence.
The phrase “real-time buffer too full” also needs context. vMix says it can appear at startup while the provider or CDN takes longer to connect; after a stream has been red, an amber or orange change makes congestion worth investigating. The same phrase at different stages is not enough to infer the same cause.
If you cannot interpret a log, keep it intact and ask vMix support to review it. Avoid deleting or overwriting the evidence before you have recorded the event. A general live-stream quality checklist can help you review the broader setup, but it should not replace the time-matched session log for this specific dropout.
Review YouTube stream-health errors
Open YouTube Live Control Room and inspect the stream-health panel for a message at the time of the event. YouTube’s stream-health error guide describes timestamped errors, including critical red errors and moderate yellow ones. A warning can be relevant without being the cause of a disconnect, so read the timestamp and the exact wording rather than treating every warning as a repair instruction.
When the message identifies a configuration issue, check the setting it names. YouTube documents problems involving unsupported video or audio codecs, missing or multiple audio/video streams, channel counts, bitrate, frame rate and keyframe interval. Its encoder settings guidance gives recommended configurations by codec, resolution and frame rate. For H.264, for example, YouTube lists a recommended video bitrate of 5 Mbps for 1080p at 30 frames per second and 6 Mbps for 1080p at 60 frames per second. These are platform recommendations, not evidence that your actual upload path can sustain that bitrate.
YouTube recommends a keyframe every 2 seconds, and its error guidance says the interval should not exceed 4 seconds. Check what vMix is actually sending when YouTube flags a keyframe or frame-rate issue. Do not copy a bitrate or timing value from a different resolution, codec or frame rate and assume it applies to your stream.
For a URL or key error, retrieve the current stream details from Live Control Room and compare them with vMix’s destination settings. YouTube’s encoder troubleshooting steps direct operators to update the encoder with the current stream key when needed. If the connection uses RTMPS, verify that the server URL supplied by YouTube is the one configured, and check the official guidance for any required connection settings. A key or URL check is especially relevant when failure happens at connection time; it does not explain a mid-session failure unless the evidence points there.
If YouTube indicates that the feed is arriving but has a quality warning, distinguish that from a stopped connection. A bitrate or format warning may affect the viewing experience even if vMix continues to send. Conversely, a healthy-looking local preview cannot show whether YouTube is receiving a valid stream, so compare both sides.
Compare timestamps across systems
A useful diagnosis depends on lining up events. Record the local time and time zone when vMix changed state, when the log reported an error, when YouTube added a stream-health message, and when viewers first reported a problem. If systems display different time zones, convert them consistently before deciding which event came first.
A simple incident note can contain the start time, end time, vMix status, exact log excerpt, YouTube message, local programme condition, and any network or security alert visible at that time. Note whether the stream recovered by itself or needed a manual restart. Avoid paraphrasing the error if you can preserve its exact text; later, the wording may distinguish a connection rejection from a session that ended after running.
If vMix reports a disconnect and YouTube records an ingest error at roughly the same moment, that alignment supports investigating the outbound feed and its response. It still does not prove whether the interruption was local, on the network route, at a provider, or caused by configuration. If YouTube shows no matching ingest error while viewers report buffering, keep the playback symptom separate and check whether the incoming stream remained healthy.
Look for repeatable patterns without turning coincidence into certainty. For example, note whether the event coincided with a network change, a security-software alert, a high-traffic period, or a vMix configuration edit. The note should say “coincided with”, not “caused by”, until further evidence supports that conclusion.
For a long-running channel, keep a small incident history rather than relying on memory. It helps distinguish an isolated interruption from a recurring time or condition, and gives vMix, YouTube or your internet provider something concrete to examine. If the content itself is stored on a network drive, test whether file access is implicated separately; the guide to buffering when FFmpeg reads from a network drive addresses that different part of the chain.
Test one change at a time
Once you have preserved the evidence, choose the next test based on the message, not on a generic list of tweaks. Change one variable, record exactly what changed and when, then run a representative test. If several settings change together, a better result will not show which change mattered, and a worse result may be harder to reverse.
For a suspected connection-capacity issue, compare the outgoing bitrate with upload capacity that is actually available during the stream, not simply a best-case result from an unrelated time. YouTube recommends selecting a quality that the connection can sustain, testing with representative audio and movement, and monitoring stream health. A rehearsal should use the intended resolution, frame rate, bitrate, content, audio routing and expected network use. A short, low-motion test may not represent a channel that carries continuous music visuals or a busy news loop.
If YouTube flags a bitrate or format mismatch, test the relevant setting against the current YouTube table for the chosen codec and mode. If the warning concerns keyframes, inspect that setting rather than lowering resolution as a first reflex. If the key or server URL is rejected, verify the Live Control Room values. Each test should answer one question and leave the rest of the configuration unchanged.
vMix provides a Network Buffer control. Its documentation says increasing the buffer may help when the maximum available network speed is unreliable, while adding latency from vMix. Consider it only when evidence points to variable speed rather than a complete outage, rejected key or provider closure. It is a tolerance setting, not a cure for a failed connection; test the delay and the effect on your programme before relying on it.
Use the FFMPEG application recommended in current vMix guidance unless a compatibility requirement gives you a reason to do otherwise. vMix describes FMLE as legacy. Do not switch application, protocol, codec and bitrate at the same time: if you need to test compatibility, make that the sole change and verify that your provider supports the selected option.
A speed test can help investigate outbound capacity, but it does not reproduce a continuous broadcast or establish what happened during an earlier incident. YouTube’s troubleshooting flow recommends checking the outbound connection when encoder output appears healthy but the stream fails, and contacting your internet provider if a connection test points to a problem. If local encoder output itself is poor, inspect source routing and system load as a separate test. Trying another encoder can be a comparison, but it is not proof that vMix was the cause.
For a channel that must play prerecorded material around the clock, decide whether operating a live production through vMix is the right workflow before tuning it indefinitely. Continuous prerecorded-video playout is a different use case from an always-live production; YouTube’s encoder setup information also discusses cloud services for prerecorded 24/7 streaming. If your workflow includes a playlist, the guide to looping a YouTube lofi playlist may help clarify the distinction. Keep the choice tied to your needs and current official guidance rather than assuming one method guarantees continuity.
Document the result and next step
After each test, write down the single change, when it was made, the test conditions and the observed result. Include whether vMix stayed red, turned amber or stopped; whether the local output remained correct; and whether YouTube reported a new message. If there was no recurrence during the test, record that accurately without treating it as proof of a permanent fix.
A practical record might read: “Changed only the YouTube output bitrate; test used the usual resolution and audio; vMix remained red; no matching health error appeared during the observation.” That says what was tested and seen without claiming the cause has been solved. If a failure returns, the same notes let you compare conditions rather than repeat random changes.
Escalate with the relevant evidence. Send vMix the session log and the exact time of the event if its status or log is unclear. Contact your internet provider with the time and connection evidence if the outbound route appears implicated. Use YouTube’s current help pages when the Live Control Room identifies an ingest or format error. Do not share a stream key in a public support post; treat it as a credential and replace it if it may have been exposed.
For overnight operation, decide in advance who can see an alert, what evidence to preserve, and what action is acceptable if the stream stops. A restart may restore a session, but repeated restarts without recording errors can erase the distinction between a brief recovery and a recurring fault. If the main burden is keeping a prerecorded file on air while your computer is off, StreamNeo can remove the need to keep a local vMix machine running for that particular workflow; it is YouTube-only and does not replace diagnosing a live vMix production.
Before committing, compare the operating options on the pricing page. When the file and channel are ready, start free — 24-hour trial, no card.
FAQ
Does an amber vMix status prove my internet connection is the problem?
No. vMix says an amber or orange status after red often indicates network congestion, but that is a clue to investigate, not a diagnosis. Compare the status and log with the YouTube health message and network evidence at the same time.
Should I increase the vMix Network Buffer to stop disconnects?
Only consider it when the evidence suggests variable network speed rather than a complete outage or a rejected connection. vMix says a larger buffer can help when maximum network speed is unreliable, but it adds latency and cannot repair a failed route or provider outage. Change it alone and test the delay.
What should I send vMix support?
Preserve the newest session log from C:\ProgramData\vMix\streaming, note the time and time zone, and include the status and exact error text. Also mention whether the local output remained correct and what YouTube Live Control Room showed at that time. Do not include your stream key in a public message.
Is vMix the right way to run a prerecorded channel for 24 hours?
It can be part of a live production workflow, but continuous prerecorded playout is a distinct use case. Compare the operating requirements and consult YouTube’s current official setup guidance before choosing a method. No approach should be treated as a guarantee that a stream will never disconnect.