Subtitle Edit

Subtitle problems are scattered but constant: a downloaded subtitle runs two seconds off, the file is ass but your player only reads srt, your own video needs subtitles from scratch, a translation needs re-timing, or a line is too long to fit on a phone screen.

Subtitle Edit covers all of it. 13,829 stars, MIT licensed, the acknowledged standard tool in subtitling circles, maintained for well over a decade.

Core capabilities:

• Format conversion across more than two hundred subtitle formats — srt, ass, ssa, vtt, sub and assorted broadcast formats — including batch conversion of whole folders.
• Timing adjustment: shift everything earlier or later, scale proportionally (for drift caused by frame rate mismatches), auto-correct from two sync points, or nudge individual lines. This is the most-used feature.
• Waveform and spectrogram views. The audio is drawn as a waveform so speech positions are obvious, and subtitle blocks drag into alignment. The spectrogram distinguishes voice from background more reliably.
• Speech to subtitles. Whisper is integrated and can run local models offline, generating a timed first draft from the video for human correction — several times faster than transcribing from scratch.
• Translation support: source and target displayed side by side with timing preserved, plus machine translation for a first pass.
• Quality checks that find overlaps, durations that are too short, excessive line counts, characters-per-second above reading-speed limits and common OCR errors, with batch fixes.
• OCR from image subtitles: Blu-ray sup and idx picture subtitles convert to text.

To be clear about:

• It is primarily Windows software. The Windows build is what the project promotes; macOS and Linux have historically run it through Mono with a worse experience and slower updates. Cross-platform alternatives include Aegisub (oriented toward styled subtitles) and Gaupol.
• The feature count makes the interface dense. The first launch presents a wall of buttons, but the main path is open video, open subtitle, drag against the waveform; everything else can wait until needed.
• Whisper transcription requires downloading a model of several hundred megabytes to several gigabytes, and speed depends heavily on your hardware (a GPU helps enormously). The draft still needs human review, particularly for proper nouns and sentence segmentation.

If you make videos, translate them, or simply want a downloaded subtitle to line up, this is the strongest free tool in the field.