Exported Video Subtitles Out of Sync with Audio? Subtitle Delay, Frame Rate & Encoding Timeline Diagnosis Guide

2026-11-21 | GPU Repair | 29 minutes | HK GPU Fix Editorial Team

Exported Video Subtitles Out of Sync with Audio? Subtitle Delay, Frame Rate & Encoding Timeline Diagnosis Guide

Key takeaways: Subtitles that sit early or late against the dialogue in an exported file, while the audio matches the picture, are usually a fixed delay, a frame-rate drift, or a subtitle file timed for a different cut. Measure the offset at the start, middle and end, shift, retime or replace the subtitle file accordingly.

A finished export plays back with the picture and the sound in step, but the subtitles are not. A line of dialogue is heard, and the subtitle for that line appears too early or too late — a fraction of a second at first, perhaps, or several seconds by the end of the programme. On the timeline in the editor, the same subtitles sat exactly on the same lines of dialogue. The first instinct is often to nudge the subtitle track on the timeline, to export the whole programme again, or to assume the subtitle file was written wrongly. None of those steps addresses the cause reliably, because the cause depends on the shape of the error, and the shape can only be found by measuring it: a fixed delay, a drift that grows with time, and a subtitle file timed for a different cut are three different faults, and each one needs a different repair.

The distinction that matters most is that subtitles are a third timeline. Sound is one timeline, picture is a second, and subtitles are a third, made of cues, each with its own start time and end time. Sound matching picture proves only that the first two timelines agree. It proves nothing about the third, because subtitle timing can go wrong in places that sound and picture never pass through: the subtitle file's own timestamps, the frame rate the subtitles were timed at, the export step that burns the subtitles into the picture or writes them as a separate stream, and the start timecode of the subtitle track. A subtitle that is wrong in the export may be exactly right on the timeline, and moving the subtitle track on a timeline that was already right will break the timeline to suit a faulty export.

This guide treats the subtitle offset as a measurement problem first and a settings problem second. Three measurements — at the start, in the middle, and at the end of the programme — separate the three shapes from each other. If the offset is the same at all three points, it is a fixed delay, and the whole subtitle file can be shifted by that one amount. If the offset is small at the start and grows steadily toward the end, it is a drift, and shifting will not repair it: the file must be retimed by a frame-rate ratio. If the offset is correct up to one edit point and then jumps suddenly to a different fixed value, the subtitle file was timed for a different cut, and no single shift or retime will repair it. Each test below measures before it moves anything, so the repair is chosen from the measurements, not from a guess.

What shape is this article about?

This article covers only a narrow shape, and six conditions must all hold. First, the subject is a finished exported file: a file that has already been through the export stage. If the subtitles are already out of step on the timeline, the fault sits in the timeline or in the subtitle file itself, and it must be divided at the subtitle-file timestamp layer first; it cannot be called an export fault directly. Second, the subtitles are out of step with the sound, but the sound itself matches the picture. If the sound is already out of step with the picture, that is an audio-video offset, and it belongs to the guide on fixed audio-video offset, identified in this library as article 529, not to this guide. Third, the subtitles do appear — they are visible, and only their timing is wrong. Subtitles that do not appear at all, a burn-in that failed, or a subtitle track that is missing altogether are questions of whether subtitles exist, and they are a different class of fault from whether their timing is right. Fourth, the error must be stated with a direction and an approximate size: whether the subtitles are early or late relative to the sound, and by roughly how much at the start of the programme and at the end. Without measurements, a fixed delay and a progressive drift cannot be told apart. Fifth, the same finished file, played in any player at the same timecode, must show the same offset, in the case of a burned-in file. If only one player shows the subtitles out of step while other players show them correctly, the fault sits in the player or decoding layer, and the finished file cannot be blamed directly. Sixth, the source of the subtitles must be stated: a subtitle track on the timeline, an imported .srt, .ass or .vtt file, an external subtitle file, or burned-in subtitles. Different sources place the responsibility for the timestamps in different places.

Several neighbouring shapes look similar and must not be confused with this one. An audio-video offset, in which the sound sits early or late against the picture, is article 529, where the measurement is the gap between lip movement and sound, and where the words "subtitle" and "caption" do not appear at all. An audio-video drift that grows over time, in which the sound itself drifts away from the picture, belongs to the variable-frame-rate and drift direction of articles 496 and 529, where the measurement object is the sound, not the subtitles. A file whose pitch and speed have both changed belongs to the sample-rate guide, identified as article 504. A recording in which two sources drift apart from each other belongs to the source-clock guide, identified as article 523. A television picture whose lip movement does not match the sound belongs to the television direction, where the television's own picture processing adds delay. Subtitles with wrong words, wrong translation, or wrong line breaks are not timing faults at all, and they are not this guide's shape.

The shape also excludes two situations that are easy to mistake for an export fault. If a subtitle file was timed for a different cut of the programme — scenes added or removed, or an advertising segment inserted — the subtitles will be correct up to one edit point and then jump to a different fixed offset, and no global shift will repair them; the handling is to retime the cues after the jump separately, or to replace the subtitle file with one timed for the cut that was actually exported. If only one player shows the subtitles out of step, the player layer must be checked first, and a Bluetooth earphone must not be used to measure an offset, because the earphone's own delay corrupts the measurement.

How it works: where subtitle timing comes from, and where it can move

Subtitles are not sound, and they are not picture. They are a third timeline, and each cue carries its own start time and end time, written in seconds. The sound and the picture can agree while the subtitles disagree, because subtitle timing can go wrong in three places independently: the subtitle file's own timestamps, the frame rate the timestamps were calculated at, and the export step that delivers the subtitles. Each of those places produces a different shape of error, and the shape is what the measurements reveal.

1. Fixed delay, progressive drift, and a different cut: three kinds of out-of-sync, three different repairs

External guides agree on three forms of subtitle error, and the repair for each one is different. The table below sets them out, and the diagnostic steps below measure which form is present before any repair is attempted.

FormWhat the measurements showCauseRepair
Fixed delay, also called constant offsetThe start, middle and end are wrong by the same amountA section was trimmed from the start, a logo or title card was added, the subtitle file belongs to a version with a different opening, or the audio carries a delayShift the whole subtitle file by the same number of milliseconds
Progressive driftThe start is right, and the error grows steadily toward the end, in proportion to timeThe subtitle timing was calculated at one frame rate and the file is played at anotherRetime by the frame-rate ratio; do not shift
Different cutCorrect up to one edit point, then a sudden jump to a different fixed value, which then stays unchangedThe subtitle file was timed for another cut, with scenes added or removedNo global repair; retime the cues after the jump separately, or replace the subtitle file with one timed for the exported cut

Confusing the three forms is the most common reason a repair fails. A fixed delay is repaired by shifting, a drift is repaired by retiming, and a different cut is repaired by neither. Shifting a drift will repair one point and make the other points worse, because a shift is a constant and a drift is a proportion, and the two mathematical shapes are fundamentally different. No amount of shifting can substitute for retiming when the error grows with time. The first step must therefore always be to measure the start, the middle and the end, and the guide must never settle for a single statement that "the subtitles are out of sync".

2. Drift is arithmetic, not random: the frame-rate ratio decides how large the error grows

A timestamp is a position written in seconds. A cue that belongs at frame F, timed at frame rate A, has a timestamp of F divided by A. When it is played at frame rate B, the same frame appears at F divided by B. The correction is to multiply every timestamp by A divided by B. The error therefore grows in proportion to time: the start is right, and the end is wrong by a large amount, and that growth is not random. The figures below are calculated from the frame-rate ratios, and they must be read as arithmetic, not as measured results.

Frame-rate pairError per minute, calculatedCorrection multiplier
23.976 to 25, PAL speed-up2.46 seconds, subtitles grow later0.959040
25 to 23.9762.56 seconds, subtitles grow earlier1.042709
25 to 29.979.95 seconds0.834168
23.976 to 29.9712.00 seconds0.800000
24 to 23.976, 1.001x pulldown3.6 seconds per hour, small but accumulating over a long programme1.001000

The 23.976 to 25 pair is the one most often met in practice: subtitles timed at 23.976 frames per second and played at 25 frames per second fall behind by 2.46 seconds every minute, calculated from the ratio, so that a programme of ten minutes is already more than twenty seconds out by the end if the ratio is left uncorrected. The 24 to 23.976 pair is the smallest: 3.6 seconds per hour, calculated, which is easy to dismiss on a short clip and easy to accumulate on a long programme. The figures from the external guides agree with each other on these pairs, and they are given here as calculated values, not as measurements taken from any particular file.

3. Burned-in, embedded and sidecar are three different deliveries, and the fault sits in different layers

Burned-in, embedded and sidecar subtitles are three different ways of delivering subtitles, and they must not be written as if they were the same thing. Burned-in subtitles, also called open captions, are drawn into the picture frames themselves, and they cannot be switched off. Their timing is decided by the renderer as it draws each frame at its position on the timeline, so a fault in the rendering layer can make the burned-in version drift across the whole programme while a sidecar version of the same subtitles stays correct. Embedded subtitles, also called closed captions, are encoded as a separate stream — for example CEA-608 or CEA-708 — and the player can switch them on or off. Their timing follows the container's timestamps, so a fault in the muxing layer can make the embedded version wrong while the burned-in version is unaffected. Sidecar subtitles are a separate .srt or .vtt file, delivered alongside the video. If the sidecar version is correct and the burned-in version is wrong, the fault points at the burn-in or rendering layer; if the sidecar version is also wrong, the fault points at the subtitle file itself or at the timeline layer.

Premiere Pro offers burn, embed and sidecar as three independent export options, and an Adobe Community reply states the distinction plainly: "embed = burned in is incorrect… Burned in is rendered into the video appearance itself and CANNOT be turned off." The words "subtitle" and "caption" also need to be used together in this guide, because Premiere Pro uses the word captions and VLC uses the word subtitles, and a reader searching for either word should reach the same diagnosis. A guide that uses only one of the two words will miss half of its readers.

4. How the export layer can create subtitle drift

Four export-layer mechanisms can move subtitles away from the dialogue, even when the timeline itself was correct. The first is Time Tuner or a speed-changing export: Premiere Pro can adjust the programme length at export with Time Tuner, and a community report states that the subtitle timing did not follow the adjustment, leaving the subtitles at the end of a programme of 30 to 60 minutes several minutes out of step, with Premiere Pro and Media Encoder 2022 behaving normally and the 2024 version showing the fault. That is a user report, not a confirmed defect from Adobe, and it must be written as such: it cannot be stated as "Premiere Pro 2024 has this bug". The suggested workaround in that community discussion was to retime the subtitles with the Synchronization and Changed Speed function of Subtitle Edit. The second mechanism is frame-rate conversion: if the export frame rate differs from the sequence frame rate, subtitle timestamps calculated at the old frame rate will drift at the new one, by the ratios in the table above. The third is trimming the start: if the export range starts somewhere other than 00:00:00 and the subtitle track's start timecode does not move with it, every cue is shifted together by the same fixed amount. The fourth is variable-frame-rate material: when variable-frame-rate material is converted to constant frame rate, the timestamps are realigned, and the time base of the subtitle cues can be pulled away with them. That last point is an inference from the variable-frame-rate mechanism described in article 496, which deals with audio-video stutter rather than subtitles, and it must be written as an inference, not as something article 496 itself states.

5. The subtitle track start timecode: a supporting check

A subtitle or caption track has its own start timecode, and it must agree with the sequence start timecode. If a section was trimmed from the start of the programme, the subtitle track's start may not have moved with it, and every cue will then be wrong by the same fixed amount. This is a supporting check, not one of the four new tests below, and it belongs with the fixed-delay form: a start timecode that does not match the sequence start produces a fixed offset, not a drift, and it is repaired by shifting the track start, not by retiming the cues. The import offset described in article 529 concerns the audio timestamps of MXF audio and of MP4 files from consumer cameras, not the start of a subtitle track, and the two must not be confused.

Diagnostic steps: eight tests in order

The order below matters, because each step narrows the fault before the next step is attempted. Only one thing is changed at a time, and the offset is measured again after each change. A short test export is used for every comparison, so that each change can be checked quickly without re-exporting the whole programme for every test. Write every measurement down before moving on: the direction of the error, early or late, and the size of the error in milliseconds at the start, in the middle, and at the end.

Step 1 — Confirm the shape and write down the direction of the error

Confirm that the exported file's subtitles are out of step with the sound for the whole programme, and that the sound itself matches the picture. Then write down whether the subtitles are early or late relative to the dialogue, and by roughly how much at the start of the programme. Also write down the editor and its version, the export preset, the subtitle source — timeline subtitle track, imported .srt, .ass or .vtt file, external subtitle, or burned-in subtitles — and the players used. If the timeline already shows the subtitles out of step, stop: the shape is not this guide's shape, and the subtitle file's timestamps must be divided first.

Step 2 — Measure at three points: start, middle and end

This is the first decisive measurement, and its structure is shared with the audio-video offset guide, although the object measured here is different. Pick one line of dialogue near the start, one near the middle, and one near the end, and for each one measure the gap between the moment the subtitle appears and the moment the words are spoken, in milliseconds. The pattern of the three numbers is the diagnosis. If all three are wrong by the same amount, the fault is a fixed delay. If the start is nearly right and the error grows steadily toward the end, the fault is a drift. If the first two are right and the third jumps suddenly to a different fixed value that then stays unchanged, the fault is a subtitle file timed for a different cut. The three-point measurement must be completed before any repair is attempted, because the three forms need different repairs, and a repair chosen from a single measurement is a guess.

Step 3 — Inspect the subtitle file's timestamps

This is the first of the four new tests in this guide. Open the .srt, .ass or .vtt file in a text editor and read the start time of the first cue and the start time of the last cue, then compare those times with the moment the same lines are spoken on the audio track. If the first cue is already wrong by a fixed amount, the fault is a fixed delay, and it sits in the subtitle file layer rather than in the export layer. If the first cue is right and the last cue is wrong by a large amount, the fault is a drift, and it sits in the frame-rate layer. If the subtitle file itself is already wrong, the export settings must not be blamed, and the handling is to correct the subtitle file's timestamps before exporting again. No neighbouring article inspects subtitle file timestamps: the subtitle timing terms in this guide appear nowhere else in the library, and the letters "srt" elsewhere in the library refer to a log file name, SrtTrail.txt, not to a subtitle file.

Step 4 — Measure the offset with the player subtitle-delay control

This is the second of the four new tests. Play the finished file in VLC and use the subtitle-delay control to measure how far the subtitles must be moved before they sit on the dialogue. Pressing G or H moves the subtitles by 50 milliseconds per step, and the Tools menu, under Track Synchronization, accepts an exact value in milliseconds. Write the value down: that value is itself the measurement of the fixed delay. If one value makes the whole programme correct, the fault is a fixed delay. If the value that is correct at the start is wrong by the end, and the subtitles need to be moved further and further as the programme goes on, the fault is a drift. The player control is a measurement instrument, not a repair: it changes what is heard and seen during that playback, and it changes nothing in the file. Writing the measured number down is the diagnosis; changing the subtitle file's timestamps is the repair, and the two must not be confused.

Step 5 — Check the frame-rate retiming

This is the third of the four new tests. Compare the frame rate the subtitles were timed at with the sequence frame rate and the export frame rate — 23.976 against 24, 25, or 29.97 — and calculate the ratio from the table in the mechanism section. Then retime the subtitle file by that ratio, using a two-point synchronisation or a frame-rate conversion, and measure the start, middle and end again. A retime that makes all three points correct confirms a drift; a shift that makes the start correct and the end worse confirms that the fault was a drift all along and that shifting was the wrong repair. The variable-frame-rate guide deals with audio-video stutter, and the sample-rate guide deals with pitch and speed changing together; neither of them touches subtitle timestamps or a retiming ratio, and neither of their tests can be substituted for this one.

Step 6 — Compare a burned-in version against a sidecar or embedded version

This is the fourth of the four new tests. Export the same timeline twice: once with the subtitles burned in, and once with the subtitles delivered as a sidecar .srt file or as an embedded subtitle stream, and compare the two versions in VLC. If the sidecar version is correct and the burned-in version is wrong, the fault points at the burn-in or rendering layer. If both versions are wrong by the same amount, the fault points at the subtitle file or timeline layer, and the burn-in layer is excluded. The two deliveries must be made from the same timeline, with nothing else changed, so that the comparison isolates the delivery layer. The word "burned-in" in this guide means subtitles drawn into the picture; elsewhere in the library the word "burned" appears only in the sense of an offset burned into a file, which is a different meaning and must not be confused with subtitle burn-in.

Step 7 — Compare the timeline preview against the finished file

This is a shared check, already used in the audio-video offset and fully silent export guides. Play the same section on the timeline and in the finished file, and compare the subtitle timing in each. If the timeline is correct and the finished file is wrong, the fault sits in the export layer, and the export settings are the place to look. If the timeline is already wrong, the fault sits in the subtitle file or timeline layer, and the export settings must not be changed to suit a faulty export: moving the subtitle track on a timeline that was already right would break a correct timeline to match a wrong file. The preview comparison decides which side of the export boundary the fault sits on, and it must be made before any export setting is changed.

Step 8 — Confirm with a short test export, changing only one thing

For every change, re-export a short test section only, and change only one thing at a time: the shift, the retiming, or the delivery method, never two or three together. After each change, measure the offset at the start, middle and end again, and write the three numbers down next to the change that produced them. Re-exporting with the same settings will produce subtitle timestamps identical to the last export, so a second export on its own repairs nothing; only a changed setting changes the result. Once a short test export measures correct at all three points, export the full programme with those settings and measure the full file once more, because a short test confirms the direction of the change and the full export confirms the result.

TestWhat is measured or changedWhat the result means
Subtitle file timestampsFirst and last cue start times against the audio trackFile already wrong: subtitle file layer; file right: export layer
Player subtitle-delay controlMilliseconds needed in VLC, G or H in 50 ms stepsOne value correct throughout: fixed delay; value grows: drift
Frame-rate retiming checkSubtitle timing frame rate against sequence and export frame ratesRetiming corrects all three points: drift confirmed
Burned-in against sidecar or embeddedSame timeline, two deliveriesSidecar correct, burned-in wrong: rendering layer; both wrong: subtitle file or timeline layer
Three-point measurement (shared)Offset at start, middle and endSame throughout: fixed delay; growing: drift; sudden jump: different cut

Boundaries: what this article does not cover

An audio-video offset, in which the sound itself sits early or late against the picture, is article 529. Its measurements are the gap between lip movement and sound in milliseconds, and its boundary table contains no subtitle row; the measurement object is different, and its tests must not be written as subtitle tests. An audio-video drift that grows over time, in which the sound itself drifts away from the picture, belongs to the variable-frame-rate and drift direction of articles 496 and 529, where the measurement object is again the sound. A file whose pitch and speed have both changed is article 504, where the fault is in the sample rate. Subtitles that do not appear at all, a burn-in that failed, or a subtitle track that is missing altogether are questions of whether subtitles exist, and they are a different class of fault from whether their timing is right. A file whose subtitles are wrong in only one player, while other players show them correctly, belongs to the player or decoding layer. A recording in which two sources drift apart from each other is article 523, where the fault is in the source clocks. A television picture whose lip movement does not match the sound belongs to the television direction, where the television's own picture processing adds delay. Subtitles with wrong words, wrong translation, or wrong line breaks are not timing faults, and they are not this guide's shape.

Two further points sit outside this guide even though they feel close to it. The variable-frame-rate effect on subtitle timestamps is an inference from article 496, which deals with audio-video stutter, and it must be written as an inference, not as something article 496 itself states. The Time Tuner subtitle report is a community user report, not a confirmed defect from Adobe, and it must be written as a user report, with the 2022 version behaving normally and the 2024 version showing the fault according to that report.

Common misconceptions: eight claims that mislead

  1. "Out of sync means drift." Wrong. A fixed delay and a progressive drift are measured differently: three measurements that are wrong by the same amount mean a fixed delay, and an error that grows toward the end means a drift. The two need different repairs.
  2. "Shifting the subtitles repairs it." Wrong. Shifting repairs only a fixed delay. A drift needs retiming by a frame-rate ratio, and shifting a drift will repair one point and make the other points worse.
  3. "Moving the subtitle track on the timeline repairs it." Wrong. If the timeline is correct and the finished file is wrong, the fault sits in the export layer. Moving the subtitle track on a correct timeline breaks the timeline to suit a faulty export.
  4. "The player subtitle-delay control is the repair." Wrong. The VLC G and H keys, or the Track Synchronization value, are a measurement made during playback. The file itself is unchanged. Writing the measured number down is the diagnosis; changing the subtitle file's timestamps is the repair.
  5. "Exporting again repairs it." Wrong. Exporting again with the same settings produces subtitle timestamps identical to the last export. Only a changed setting — a shift, a retiming, or a different delivery method — changes the result.
  6. "Burned-in, embedded and sidecar subtitles are the same thing." Wrong. The three are different deliveries, and the fault sits in different layers for each. The Adobe Community reply quoted above states that embedded is not the same as burned-in, because burned-in subtitles are rendered into the picture itself and cannot be switched off.
  7. "Subtitles must be wrong because the editing was wrong." Wrong. An import offset, a frame-rate conversion, Time Tuner, or trimming the start can each create subtitle drift, and the timeline itself may have been entirely correct.
  8. "If another player plays the subtitles correctly, the finished file is fine." Wrong. The file must be compared across all players and on a second device. If only some players are correct, the fault is in the player layer, not in the finished file, and a single-player check proves nothing.

FAQ

My exported subtitles are late, but the audio matches the picture. Is the subtitle file wrong?

Not necessarily. Measure the offset at the start, in the middle, and at the end first. If all three are wrong by the same amount, the fault is a fixed delay, and the subtitle file may simply need to be shifted by that amount — or the export range may have trimmed the start. If the error grows toward the end, the fault is a drift, and the subtitle file needs to be retimed by the frame-rate ratio. If the subtitle file's own timestamps are already wrong when read in a text editor, the fault sits in the subtitle file layer; if they are right, it sits in the export layer.

How do I measure how far out the subtitles are?

Play the finished file in VLC and use the subtitle-delay control. Pressing G or H moves the subtitles by 50 milliseconds per step, and the Tools menu, under Track Synchronization, accepts an exact value in milliseconds. Note the value that makes the subtitles sit on the dialogue: that value is the measurement of the fixed delay. Measure at the start, in the middle, and at the end, because a value that is correct at the start and wrong at the end means the fault is a drift, not a fixed delay.

The subtitles are right at the start and wrong by seconds at the end. What does that mean?

It means the fault is a progressive drift, most often a frame-rate mismatch. The subtitle timing was calculated at one frame rate and the file is played at another, so the error grows in proportion to time. Shifting the whole file will not repair it: it will repair one point and make the other points worse. Compare the subtitle timing frame rate with the sequence and export frame rates, retime by the ratio — for example 25 divided by 23.976, about 1.0427 — and measure all three points again.

The subtitles were correct until one scene, then suddenly jumped out of step. What happened?

The subtitle file was probably timed for a different cut of the programme, with scenes added or removed, or an advertising segment inserted. A sudden jump to a different fixed value, which then stays unchanged, is the signature of a different cut, and no global shift or retiming will repair it. The cues after the jump must be retimed separately, or the subtitle file must be replaced with one timed for the cut that was actually exported.

The sidecar subtitles are correct, but the burned-in subtitles are wrong. What does that tell me?

It tells you the fault sits in the burn-in or rendering layer. The subtitle file itself and the timeline layer are correct, because the sidecar version proves it, and the rendering step that draws the subtitles into the picture frames is the layer that moved them. Export the two versions from the same timeline with nothing else changed, so that the comparison isolates the delivery layer, and check the export settings for the burn-in step, including any Time Tuner or speed adjustment applied at export.

I moved the subtitles in VLC and they look correct now. Is the file repaired?

No. The VLC subtitle-delay control changes only that playback, and the file itself is unchanged. The measured value is the diagnosis, and it must be written down and then applied to the subtitle file's timestamps — by shifting the whole file for a fixed delay, or by retiming it for a drift — before a short test export is made. Play the new export in VLC with the delay control back at zero, and measure the start, middle and end again to confirm the repair.

If the fault still cannot be divided

Write down the subtitle source, whether the subtitles are early or late, the measured offset in milliseconds at the start, in the middle, and at the end, the first and last cue timestamps read from the subtitle file, the subtitle timing frame rate and the sequence and export frame rates, which delivery was used — burned-in, embedded, or sidecar — and the result of each short test export, and bring that written record with a copy of the finished file and the subtitle file if it is available. The drift figures in this guide are calculated from the frame-rate ratios, not measured from any particular file, and they must be read as arithmetic. The Time Tuner report is a community user report, not a confirmed defect from Adobe, and the variable-frame-rate effect on subtitle timestamps is an inference from the variable-frame-rate mechanism, not a statement from the guide that describes it. The exact panel names for the subtitle and caption tracks differ between versions, and the descriptions in this guide follow the external guides named in the research notes; they should not be read as guaranteed to match every version. Bring the written record described above to Room 1209, Dali Plaza, Shek Mun for inspection. The telephone number is 2186 8010 and WhatsApp is 6558 6806. There is no Mong Kok shop, and there is no on-site service. The site's published three-tier fee is HK$400–780, HK$480–1,380, and HK$680–2,800, an overall range of HK$400–2,800, with the actual fee quoted after inspection.

Related articles

Related services

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