Turn WebVTT (.vtt) subtitles into SMI (.smi) format in a few clicks—fast conversion with no software installation.
.vtt file or click "Add Files" to browse. Parsing happens in your browser — the cues never leave your device. Batch is supported, so a whole folder of WebVTT exports can be converted in one pass..smi, Microsoft SAMI 1.0) is preselected. The converter wraps your cues in the required <SAMI><Head><SAMIParam>…<Body> scaffolding, defines a single .ENUSCC language class, and emits each cue as a <SYNC Start=...> block with the timestamp converted from VTT's HH:MM:SS.mmm to SAMI's milliseconds-since-start integer.<b>, <i>, <u>, and <font color> tags pass through verbatim — SAMI is HTML-derived and renders them natively. WebVTT-only constructs (<c.classname>, <v Speaker>, <ruby>, <lang>, cue settings like line: and position:, embedded STYLE and REGION blocks) are dropped because SAMI 1.0 has no equivalent grammar..smi file downloads instantly. No upload, no sign-up, no watermark, no file-size cap beyond what your browser can hold in memory.WebVTT is the modern W3C standard for the HTML5 <track> element — it lives natively inside browser-based players. SAMI (Synchronized Accessible Media Interchange), released by Microsoft in June 1998 and last formally documented as MSFT:1.0 in February 2003, predates HTML5 by a decade. The two formats solve the same problem (timed text alongside video) but expect completely different host environments. When a legacy Windows-era player, a Korean media library, or a Windows Media Player skin refuses your .vtt, transcoding to .smi is usually the fix.
.smi extension and loads it automatically as the closed-caption track. .vtt is not recognised. Anyone maintaining an old WMV/ASF library on Windows needs SAMI for sidecar captions..smi files; replacing them with .vtt breaks the automatic load..smi sidecars (VLC's SAMI demuxer has shipped since at least 2008). For users on those players who only have .vtt files from a web download, conversion makes the existing library "just work."SAMIFileName, SAMILang, and SAMIStyle properties for in-page captions. Internal corporate portals built around that stack still consume SMI.Class mechanism lets a single .smi carry English, Korean, and Japanese tracks distinguished by lang: en-US, lang: ko-KR, etc. WebVTT typically expects one file per language. Consolidating several .vtt files into one multi-language .smi is a common archival workflow.| Property | VTT (WebVTT) | SMI (SAMI 1.0) |
|---|---|---|
| Standardised by | W3C (webvtt1 Candidate Recommendation Draft) |
Microsoft, published Jun 1998; spec updated Feb 2003 |
| Current status | Active web standard | Legacy; superseded by TTML/WebVTT for new work |
| File header | Required: WEBVTT on line 1 |
Required: <SAMI>…</SAMI> document with <Head> and <Body> |
| Underlying grammar | Plain-text cue blocks | HTML-like markup with CSS in <Style> |
| Timestamp format | HH:MM:SS.mmm --> HH:MM:SS.mmm |
<SYNC Start=ms> integer milliseconds from media start |
| Inline styling | <b> <i> <u> <c> <ruby> <v> plus CSS classes |
Most HTML 4.01 inline tags (b i u font basefont strong em) |
| Block styling | STYLE blocks with ::cue selector |
CSS in <Style Type="text/css"> block in the head |
| Positioning | line: position: align: vertical: size: cue settings |
None — relies on <Sync> block CSS classes |
| Multi-language in one file | No (one .vtt per language) |
Yes — Class .ENUSCC {lang: en-US} etc. |
| Native browser support | <track kind="subtitles" src="...vtt"> |
None (no major browser parses SAMI) |
| Player support | Chrome, Firefox, Safari, Edge, VLC, MPV | Windows Media Player, VLC, MPC-HC, KMPlayer, PotPlayer, GOM Player |
| Typical MIME / extension | text/vtt / .vtt |
application/smil (informal) / .smi or .sami |
| Encoding | UTF-8 mandated | Any (UTF-8, EUC-KR, CP949, Shift-JIS — declared in <META> or assumed) |
| VTT element | Behavior in SMI output |
|---|---|
WEBVTT header |
Removed; replaced with <SAMI><Head> scaffolding |
NOTE comment blocks |
Discarded |
STYLE blocks (embedded CSS) |
Dropped — replaced with a minimal <Style> paragraph rule |
REGION blocks |
Dropped (SAMI has no region concept) |
| Cue identifiers (numeric or string) | Dropped — SAMI uses <Sync Start=> only |
Cue settings (line position align vertical size) |
Stripped — SAMI has no equivalent on the cue line |
<b> <i> <u> inline tags |
Preserved verbatim — SAMI accepts HTML 4.01 inline tags |
<font color="..."> |
Preserved (legal HTML in SAMI) |
<c.classname> class spans |
Class name dropped, inner text retained |
<v Speaker> voice tags |
Text retained; speaker label dropped (no SAMI equivalent for the <v> attribute) |
<ruby> annotations |
Dropped — Ruby is not part of SAMI 1.0 |
| Timestamps | Recomputed: 00:01:23.456 → Start=83456 milliseconds |
| End timestamps | Encoded as a blank <Sync> block at the cue's end time (SAMI uses appearance times, not durations) |
| UTF-8 text | Preserved as UTF-8 in the output (declared in the <Style> block) |
& < > entities |
Preserved |
SAMI was designed for the Windows Media DirectShow stack, where timestamps are most naturally expressed as elapsed milliseconds since the media's start. The <SYNC Start=83456> syntax means "display this paragraph 83.456 seconds in." WebVTT, by contrast, was modelled after SRT and uses a human-readable wall-clock notation. The converter does the arithmetic (hours*3600000 + minutes*60000 + seconds*1000 + millis) so the playback timing is byte-equivalent — only the encoding differs.
No. WebVTT's line:, position:, align:, vertical:, and size: cue settings have no counterpart in SAMI 1.0 — SAMI relies on CSS paragraph styles declared in <Style> and Class definitions, applied to entire <P> blocks rather than per-cue. The converter drops the cue settings and emits each cue as a plain <P Class=ENUSCC>. If your VTT relied on positioning to avoid covering on-screen text, you will need to re-author the layout in the SAMI <Style> block manually.
Three common causes. First, the .smi file must have the same basename as the media file and live in the same folder (MyVideo.wmv + MyVideo.smi). Second, Windows Media Player only auto-loads .smi, not .sami. Third, captions must be enabled under Play → Lyrics, Captions, and Subtitles → On if Available. SAMI files with <META> declaring a non-default encoding can also fail silently on older WMP builds — UTF-8 with a BOM is the safest choice for cross-version compatibility.
Yes — that's one of SAMI's defining features. The converter emits one Class block per detected language. If your source VTT is English-only, the output .smi has a single .ENUSCC class with lang: en-US. To add Korean, run a separate VTT-to-SMI conversion on the Korean .vtt, then merge the two <P> blocks under the same <Sync> timestamps and define both .ENUSCC and .KRCC classes in the header. KMPlayer, GOM Player, and PotPlayer will then surface a language switcher.
UTF-8 by default, declared in the SAMI <Style> block. This works in every modern player. Legacy Korean SAMI files were often saved as EUC-KR or CP949 because that's what Windows-XP-era Korean Windows used; if you're authoring .smi for an old PotPlayer or GOM Player build that ignores the BOM, you can resave the output as EUC-KR in a text editor. Modern KMPlayer, VLC, and MPC-HC detect UTF-8 automatically.
Yes. VLC's SAMI demuxer has shipped since the 0.x series in the late 2000s and is still in current builds — it loads .smi sidecars the same way it loads .srt. The implementation is lenient about case (<SAMI> vs <sami>) and tolerates malformed CSS, but it does require the basic <SAMI><Body><Sync Start=...> skeleton. Files emitted by this converter are tested against VLC, MPC-HC, KMPlayer, and Windows Media Player.
WebVTT supports <ruby> annotations natively; SAMI 1.0 does not. If your VTT uses ruby for Japanese furigana or Chinese pinyin, the converter strips the ruby base/text wrapping and keeps the base characters only. For Japanese SAMI workflows that need furigana, the convention is to inline the reading in parentheses after the kanji — that's lossy compared to VTT's structured ruby, but it's the only option SAMI 1.0 permits.
For new work, no — WebVTT for browsers and TTML / IMSC1 for broadcast and OTT deliveries are the modern defaults. Convert to SMI when you have a specific legacy reason: a Windows-era video archive with WMP playback, a Korean drama library that ships .smi sidecars, a CD-ROM-era multimedia title, or a corporate intranet built around the old Windows Media Player ActiveX control. For any other workflow, prefer .vtt (HTML5) or .ttml/.dfxp (timed-text broadcast).
Because SAMI 1.0 positions captions through CSS rules on <P Class> paragraphs in the document <Style> header, not per-cue. WebVTT's line:0 (top of screen) or position:90% align:end (bottom-right) cue settings are simply dropped. To re-create per-cue positioning in SAMI you would need to define multiple Class styles (.TopCC, .BottomRightCC, etc.) with different margin-top and text-align values, then assign each <P> to the appropriate class. That's a manual re-authoring step the converter doesn't attempt.