GPU Picture and Sound Out of Sync on the TV? TV Video-Processing Delay, Audio Path & Sync-Setting Diagnosis Guide

2026-11-03 | GPU Repair | 18 minutes | HK GPU Fix Editorial

GPU Picture and Sound Out of Sync on the TV? TV Video-Processing Delay, Audio Path & Sync-Setting Diagnosis Guide

Key Takeaways

When the picture on the TV is clear but the lips move out of step with the sound by a steady amount, the two signals are taking paths with different delays: this guide reads the direction of the mismatch first, then works through one variable at a time to separate TV picture processing from the audio path.

What Exactly Counts as This Symptom

This guide deals with a narrowly shaped symptom, and the shape has to be matched before any setting is changed. The graphics card is connected to the television by HDMI, the picture is normal, and the sound is present and clean. Nothing is missing and nothing is distorted. The only problem is timing: the movement of the lips on screen and the sound that belongs with it do not line up, and the gap stays the same, in the same direction, for the whole programme.

Several lookalike situations must be ruled out first, because each one leads to a different diagnosis. A picture with no sound at all is not a sync problem; it belongs with audio-output diagnosis. Sound that crackles, breaks up, or drops out, especially sound that only degrades when a game is putting the system under load, is not a steady timing gap either; it belongs with load-related audio diagnosis. A picture that goes black, drops resolution, or loses a feature such as HDR when a receiver is added to the chain is a chain-capability problem, not a sync problem. And a stream that stutters, buffers, or skips is a network or streaming problem, not a lip-sync problem. If the mismatch comes and goes, or arrives together with stuttering, buffering circles, or broken sound, stop: that is not the steady timing gap this guide covers.

The steadiness of the gap is itself the most important clue. A gap that is constant, in one direction, across a whole programme points at two paths with different fixed delays. A gap that changes from moment to moment points somewhere else entirely. Note the direction before touching anything, because the direction decides which side of the chain to investigate first.

The Mechanism: Two Paths, Two Delays

A television does not put an HDMI picture on screen the instant it arrives. After the signal reaches the television, the picture passes through processing: upscaling, motion smoothing or interpolation, noise reduction, HDR tone mapping, and conversion between the frame rate of the content and the refresh rate of the panel. Every one of those steps takes time, and the more processing is applied, the later the picture appears.

Sound travels a different route with its own delay. If the television speakers are used, the sound path is short. If the sound is sent back out of the television through ARC or eARC to an amplifier or a soundbar, the sound has to travel an extra return link and then be decoded at the receiving end. If the sound goes by Bluetooth, it has to be encoded, transmitted, decoded, and buffered before it is heard. Each of those steps adds delay on the sound side only.

When the two delays differ, lips and sound fall out of step. Neither side is necessarily faulty. A television applying heavy picture processing is doing what it was designed to do. An amplifier decoding a return-channel signal is doing what it was designed to do. The mismatch is the difference between two normal delays, and that is why the diagnosis in this guide is about locating which side carries the larger delay, not about finding a broken part.

Reading the Direction First

Direction is the first piece of evidence, and it must be read correctly, because reading it backwards sends the whole investigation to the wrong side.

If the lips move first and the sound arrives afterwards, the sound side is late. Either the sound path is long, or the picture processing is light, or both. If the sound arrives first and the lips move afterwards, the picture side is late: the television picture processing is holding the picture back while the sound has already arrived.

The second pattern is the one most often misread. Hearing the sound first does not mean the sound is too fast. Sound does not speed up. It almost always means the picture has been slowed down by television processing. Treating sound-first as a sound problem leads to adjusting the wrong side of the chain, so the direction must be written down in plain words, lips first or sound first, before any test is run.

Step One: Direction and Scope, Without Changing Anything

The decisive first step changes no setting at all. Write down three things: which side comes first, whether every programme shows the same mismatch, and whether every sound output shows it. Test the television speakers and any external sound equipment separately, and try a different programme and a different application. This baseline is the reference for every later test, and every later result is judged against it.

The scope result already narrows the field. If only one programme or one application is out of step while everything else lines up, the content or that application is the strongest suspect, and the chain as a whole is probably fine. If every programme shows the same mismatch in the same direction, the chain itself is the strongest suspect, and the investigation continues to the tests below. A vague memory of the mismatch is not enough for this comparison: direction and scope must be recorded before anything is changed, because later tests are meaningless without a baseline to compare against.

Step Two: The Content Comparison

Using the same television and the same chain, play a different programme that is known to be normal, or play the same programme through the television built-in application and compare it with the version played through the graphics card HDMI connection. Change only the content or the source, and nothing else.

If changing the content or the source lines the sound up again, the problem sits with the original programme or the original application, not with the chain. Some programmes are produced with a timing error already in them, and no chain adjustment can correct a programme that was made out of step. The correct response is to leave the household settings alone rather than retuning the whole chain for one faulty programme. If every content and every source shows the same mismatch in the same direction, the chain itself is confirmed as the area to investigate, and the next steps apply.

Step Three: Reset Every Sync Offset to Zero

Players and applications commonly offer an audio delay or sync-offset control, and televisions commonly offer a lip-sync setting of their own. These controls do only one thing: they push one side artificially later so that it catches up with the other side. They do not remove the delay difference in the chain, and they do not repair anything. They mask the problem.

That masking is exactly why this step must come before any television-mode change or any sound-path change. An old offset value left over from a different application, a different programme, or a different sound path will contaminate every later test result, because the test will be measuring the chain plus the old offset rather than the chain itself. Set every audio delay, sync offset, and television lip-sync control to zero first, noting the original values so they can be restored if needed, and then test again.

If the sound lines up after the reset, the previous offset value was the whole story: a human-made adjustment, not a chain fault, and no further chain investigation is needed. If the mismatch remains after the reset, the delay difference is genuinely in the chain, and the investigation continues. Any offset applied later, after diagnosis, is at most a final fine adjustment for a specific arrangement, never the first step and never a substitute for locating the cause.

Step Four: The Television Processing Comparison

This step targets the sound-first pattern, where the picture side is suspected of being late. Switch the television to game mode or computer mode, or switch off picture processing items one at a time, such as motion smoothing, noise reduction, and upscaling-style processing, changing only one item per test. Then watch the same programme again and compare against the baseline.

Game mode and computer mode work by bypassing processing: they switch off or sharply reduce the television picture processing, so the picture delay shortens. If switching to game mode lines the sound up, television picture processing delay is the main cause. That conclusion does not mean the television is faulty; it means the processing was the slower side of the chain.

If changing the processing makes no difference at all, television picture processing is not the main cause, and the investigation moves to the sound path. This negative result is just as informative as a positive one, because it stops further time being spent on television settings and redirects attention to the side that actually carries the delay.

There is a trade-off to state honestly. Game mode improves sync by removing processing, and removing processing means some picture treatments are no longer applied. Choosing between picture treatment and sync is a matter of preference for a given programme, not a question with one correct answer, and this guide does not recommend one side of that trade-off over the other. The point of the test is location, not a verdict on how the television should be set permanently.

Step Five: Swap the Sound Path One at a Time

This step targets the lips-first pattern, where the sound side is suspected of being late. Using the same programme and the same television, try each sound route in turn, changing only one route per test: the television speakers, then the amplifier or soundbar through ARC or eARC, then optical, then Bluetooth. After each change, note whether that route lines up or not.

Each result has a specific meaning. If the television speakers line up but the external equipment does not, the external sound path, with its return link, decoding, and buffering, is the delay source, and that conclusion does not imply any fault in the graphics card audio output. If only the Bluetooth route is out of step while every other route lines up, the Bluetooth path itself, with its extra encoding, transmission, decoding, and buffering, is the strongest suspect. That is a property of the Bluetooth route, not evidence that the television or the graphics card is faulty, and the sensible response is to use a different sound route rather than to treat the graphics card as needing repair. If every route shows the same mismatch in the same direction, stop circling the sound paths and return to the television processing comparison or the content-type split below, because the delay difference is not being created by the choice of sound route.

Step Six: An Alternative Arrangement After Diagnosis

Once the delay source has been located, one workable rearrangement is to connect the graphics card directly to the television for picture, and send sound back from the television to the receiver using the television ARC or eARC return channel. This changes the sound route rather than the picture route, and it is borrowed from receiver-chain practice, where the same arrangement keeps a receiver out of the video path.

It must be described accurately. It is a chain rearrangement, not a graphics card repair, and the sound formats that pass through it are limited by what the television ARC or eARC capability supports. Not every format can be assumed to pass, and no format should be promised. A purely silent picture, with the picture itself fine, still belongs with audio-output diagnosis rather than with this arrangement.

Step Seven: The Content-Type Split as a Closing Confirmation

As a closing confirmation, not as an opening step, try different kinds of content on the same chain: a local file, a streamed programme, a live broadcast, and programmes at different frame rates. If only one kind of content is out of step, the encoding, frame-rate conversion, or application handling of that kind of content is the suspect. If every kind of content shows the same mismatch in the same direction, the problem is at the chain level, and the earlier steps have already located which side carries the delay. There is no need to keep working through content types once the chain-level result is established.

TestWhat the result provesWhat it does not prove
Direction and scope recorded firstWhich side is late, and whether the problem is chain-wide or limited to one programme or applicationWhich device is responsible, without the later single-variable tests
Content comparison lines upThe original programme or application carried the timing error, not the chainAnything about the chain, which was not the problem in that case
Offset reset lines upA previous human-made offset value was the whole causeThat the chain is fault-free in every arrangement, only that it lines up with offsets at zero
Game mode lines upTelevision picture processing delay is the main causeThat the television is faulty; processing delay is normal behaviour
Game mode changes nothingTelevision picture processing is not the main causeThat the sound path is at fault, without the sound-path swap test
Television speakers line up, external sound does notThe external sound path, with its return link and decoding, is the delay sourceThat the graphics card audio output is faulty
Only Bluetooth is out of stepThe Bluetooth encoding and buffering route is the delay sourceThat the television or the graphics card is faulty
Every sound route shows the same directionThe delay difference is not created by the choice of sound routeWhich side of the chain carries the delay, without the processing comparison

Boundary Cases: When It Is Not This Sync Problem

First, no sound at all is not a sync problem, whatever the picture looks like. Second, sound that crackles, breaks up, or stutters, especially sound that only degrades under gaming load, is a load-related audio problem, not a steady timing gap. Third, a picture that goes black, drops resolution, or loses a feature when a receiver is added to the chain is a chain-capability problem, not a sync problem. Fourth, a mismatch that comes and goes, or arrives together with stuttering, buffering circles, or broken sound, belongs with streaming, network, or load diagnosis, and must not be masked with a sync offset. Fifth, a single programme or a single application out of step while everything else is normal is a content or application problem, and the household settings should be left alone. Sixth, a Bluetooth-only mismatch while every other route lines up is a property of the Bluetooth route, and the response is to change the sound route, not to treat the graphics card as faulty.

Common Misconceptions

Misconception one: the graphics card is not powerful enough, so the lips fall out of step. Load on a graphics card produces stuttering, dropped frames, and broken sound, not a steady timing gap in one direction for a whole programme. Load-dependent symptoms belong with load-related audio diagnosis.

Misconception two: a more expensive or newer HDMI cable will fix the sync. A cable does not perform picture processing and does not create a steady timing gap. Cable problems show up as flicker, black screens, or dropped modes, not as lips out of step.

Misconception three: adjusting the player sync offset counts as fixed. An offset only compensates for one arrangement. Change the application, the programme, or the sound route, and the same offset can be wrong again, or even wrong in the opposite direction. Offsets must be reset to zero before diagnosis, and they can only ever be a final fine adjustment afterwards.

Misconception four: hearing the sound first means the sound is too fast. Sound does not speed up. The far more common explanation is that the picture has been slowed down by television processing. Misreading the direction sends the investigation to the wrong side.

Misconception five: every application being out of step must mean the television is broken. Every application travelling through the same chain and the same sound route will show the same chain-level delay difference. Only the game-mode comparison and the sound-path swap can say which side carries the delay.

Misconception six: game mode is bad because the picture treatment changes, so it should never be used. That is a trade-off, not a right-or-wrong question. Game mode removes processing and improves sync at the cost of some picture treatment, and the choice between them depends on what matters more for a given programme.

Misconception seven: changing the cable, the television mode, and the offset all at once, and getting a lined-up picture, is good enough. A recovery with several changes teaches nothing about which change mattered, and the fault will return the next time the content or the sound route changes. One variable at a time is slower, but it is the only way the result can identify a cause.

When to Stop and Seek Advice

Stop experimenting with settings if any of the following appears. The picture itself is faulty, with black screens, flicker, or dropped modes, which is a different problem from sync. The sound is missing entirely, or it crackles and breaks up rather than sitting at a steady offset. The mismatch comes and goes with stuttering or buffering, which points at streaming or load rather than at a fixed delay difference. Or the television or sound equipment shows signs of a hardware problem beyond settings, such as a port that never responds or intermittent power loss.

Bring four items to any inspection: the recorded direction and scope baseline, the result of the content comparison, the result of the offset reset, and the result of each single-variable test, stating which television mode and which sound route lined up first. Those four let advice start from the chain directly, without repeating tests already done.

Readers nearby are welcome to bring their records to HK GPU Fix for advice, at Room 1209, Dali Plaza, Shek Mun, telephone 2186 8010, WhatsApp 6558 6806. Describing the direction of the mismatch, which sound route was in use, and which single change first lined the picture up is enough for a quick judgement on the next step.

Frequently Asked Questions

Why are the lips and the sound out of step when the picture itself is perfect?

Because picture and sound travel different paths with different delays. The television processes the picture, which takes time, while the sound travels its own route, which takes a different amount of time. A perfect picture and clean sound can still be out of step when those two delays differ.

Does sound arriving first mean the sound is too fast?

No. Sound does not speed up. Sound arriving first almost always means the picture has been slowed down by television picture processing, and the investigation should start on the picture side rather than the sound side.

Why should all sync offsets be reset to zero before testing?

Because an offset is a human-made compensation for one particular arrangement, not a measurement of the chain. An old offset left in place will contaminate every later test, since the test will be measuring the chain plus the offset rather than the chain itself.

Does switching to game mode mean the television is faulty?

No. Game mode bypasses picture processing, which shortens the picture delay. If that lines the sound up, it shows that picture processing delay was the slower side. Processing delay is normal behaviour of the television, not a fault.

Why is only the Bluetooth route out of step?

Because the Bluetooth route adds its own encoding, transmission, decoding, and buffering steps, which other routes do not have. A Bluetooth-only mismatch is a property of that route, and the sensible response is to use a different sound route rather than to treat the television or the graphics card as faulty.

Will a better HDMI cable fix lip-sync?

No. A cable does not perform picture processing and does not create a steady timing gap. Cable problems appear as flicker, black screens, or dropped modes, and none of those is the steady offset this guide covers.

Is adjusting the player sync offset a repair?

No. An offset masks the delay difference for one arrangement. Change the application, the programme, or the sound route, and the same offset can be wrong again. Offsets are a final fine adjustment after diagnosis, never the first step.

What if only one programme is out of step?

Then the programme or the application carrying it is the strongest suspect, not the chain. Leave the household settings alone and check other programmes first, because retuning the whole chain for one faulty programme will only create new mismatches elsewhere.

What should I bring if I seek advice?

Bring the recorded direction and scope baseline, the content comparison result, the offset reset result, and the result of each single-variable test, stating which television mode and which sound route lined up first. That record lets advice start from the chain rather than repeating the whole diagnosis.

Related articles

Related services

If you need Chip-level GPU repair, see these service pages: