Subtitle drift and a fixed delay are different problems

A fixed offset keeps its size

If text appears consistently early by roughly the same amount at several positions, one delay may be useful. If the beginning aligns and the end does not, that same delay cannot make both ends correct. A sudden jump after one scene is another pattern.

These are observations, not a diagnosis of the file's exact cause. Timing-rate differences, edits and other source issues require evidence beyond the player control.

Use three small observations

Choose a line near the beginning, one in the middle and one later in the same edition. Record the approximate direction and size of the mismatch. Keep playback speed at its normal value for the comparison.

Avoid measuring from a live feed that changes while you investigate. If only a live source is available, state that limit instead of claiming an exact repeatable test.

Apply the tool that fits the evidence

VLC documents a subtitle delay control in its desktop synchronization tools. Test it on one reference and verify the others. If they diverge, restore the original value and ask for a matching subtitle or appropriate timing repair through the legitimate source.

Repeatedly nudging delay during viewing can make a scene temporarily readable, but it is not proof that the full file has been corrected.

Keep audio timing independent

A subtitle can disagree with speech while speech agrees with the picture. Conversely, moving audio to match an incorrect subtitle can create lip-sync failure. Record each relationship separately when the problem is ambiguous.

This article provides a timing comparison method based on documented player controls. It does not specify an editing operation, validate a downloaded file or claim that an IPTV provider supplies a correctly matched track. The useful result is knowing whether a constant offset is a plausible remedy.

Do not confuse a seek with continuous drift

Check whether the mismatch grows while playing continuously or changes only after seeking. Those are different transitions for a developer to investigate. Record one uninterrupted passage and one seek comparison separately. A delay that appears after every seek should not automatically be described as a file-wide timing-rate problem without further evidence.