When to Trim Losslessly and When to Re-encode
Every trimming tool does one of two things: copy the compressed video across without touching it, or decode and rebuild it. The first is fast and lossless but constrained to keyframes. The second is frame-accurate but slower and costs a compression generation.
Knowing which you want takes about ten seconds of thought, and saves considerably more than that.
Choose lossless when
You are removing waste. Dead air before someone starts talking, fumbling at the end, a long unusable stretch in the middle. Landing a fraction of a second off is irrelevant, and you keep the original quality exactly.
The source is precious. Archive footage, an original recording, anything you cannot re-shoot. Every re-encode discards detail permanently. If you do not need frame accuracy, do not pay for it.
The file is large and you are in a hurry. Stream copying a 4 GB file takes seconds, because nothing is being compressed. Re-encoding the same file can take many minutes.
You will edit it again later. Each lossy generation compounds. Staying lossless through intermediate steps means only the final export costs you anything.
Choose re-encoding when
The exact frame carries meaning. Cutting on a beat, on an action, on a specific word. Keyframe snapping will not respect any of those.
You need the file smaller. A stream copy preserves the original bitrate, so a lossless trim only shrinks the file in proportion to the duration you removed. If you need a real size reduction, you must re-encode at a lower bitrate.
The format has to change. Converting H.265 to H.264 for compatibility, or changing resolution, both require decoding and re-encoding by definition.
The output must start cleanly. Some lossless implementations include frames before your cut point and mark them not to be displayed, which can produce a brief freeze or black frame in less careful players. Re-encoding guarantees a clean first frame.
A workflow that avoids most problems
Start from the original file, not from a previous export.
Do your rough removal losslessly. Strip the dead air and the mistakes while paying nothing in quality.
Do precise work only where it matters. If two or three cuts need to land exactly, re-encode for those and leave the rest alone.
Export once, at the end, in the format the destination needs. Every intermediate export you avoid is a compression generation you do not spend.
The mistake worth avoiding
Re-encoding repeatedly at high quality settings feels safe and is not. “High quality” only means the loss is small each time; it is still loss, and it accumulates. Four rounds of careful re-encoding will visibly degrade footage that would have survived a single pass untouched.
If you find yourself exporting, reopening the export, editing, and exporting again, go back to the source instead. It is almost always faster and always better.
Telling which one your tool used
Watch the clock. A stream copy of a large file finishes in seconds and barely troubles your CPU. A re-encode takes proportionally longer and makes your machine work.
Or compare bitrates. If the output bitrate matches the input, nothing was re-compressed.