Subtitle formats explained

Every subtitle file stores the same three things: the text, when it starts, and when it ends. The formats differ in what else they can store: styles, positions on screen, and data for TV equipment.

SRT comes from SubRip, a program that read subtitles from DVDs. Today it is the format that almost every platform and player opens. Each subtitle is a number, a start and end time, and one or more lines of text. A blank line ends the subtitle.

11200:00:01,000 --> 00:00:04,2003Who has the Spanish file?452600:00:04,600 --> 00:00:07,0007<i>Mina has it. She sent it back.</i>

• A comma comes before the milliseconds: 00:00:01,000.
• SRT has no official specification. Players agree on the numbers and times, but not on tags such as <i> and <font>.
• The file does not say which text encoding it uses. Save SRT files as UTF-8: YouTube and Vimeo both ask for it.
• YouTube reads SRT as plain text, without its tags. To control position on YouTube, use WebVTT or TTML.

Who asks for SRT

YouTube, Vimeo, LinkedIn, and other video platforms. Most subtitle editors open it, so it is also the usual file to send to a translator or reviewer.

WebVTT (Web Video Text Tracks) is the W3C format for captions on the web. It grew out of SRT, so the two look almost the same. The <track> element in HTML video takes WebVTT files, so every browser can show them.

1WEBVTT2300:00:01.000 --> 00:00:04.2004Who has the Spanish file?5600:00:04.600 --> 00:00:07.000 line:10% align:center7<i>Mina has it. She sent it back.</i>

• The file starts with the word WEBVTT.
• A period comes before the milliseconds: 00:00:01.000. The hours are optional.
• Subtitle numbers are optional.
• Settings after the time move a subtitle: line, position, size, and align. The second subtitle above sits near the top of the picture.
• A STYLE block in the file, or the website's CSS, sets fonts and colors.

Who asks for WebVTT

Video players on websites, through the HTML <track> element. Vimeo accepts SRT too, but recommends WebVTT. YouTube reads WebVTT positions, and bold, italic, and underline.

SubStation Alpha (SSA) and its later version, Advanced SubStation Alpha (ASS), are formats for styled subtitles. A file has named styles for font, size, color, border, and shadow. Tags in braces change one line: its position, color, rotation, or a karaoke highlight. Aegisub, a free editor, is made for this format.

1[Events]2Format: Layer, Start, End, Style, Name, MarginL, MarginR, MarginV, Effect, Text3Dialogue: 0,0:00:01.00,0:00:04.20,Default,,0,0,0,,Who has the Spanish file?4Dialogue: 0,0:00:04.60,0:00:07.00,Sign,,0,0,0,,{\pos(960,120)}Mina has it.

• Times have hundredths of a second, not milliseconds: 0:00:01.00.
• {\pos(960,120)} puts a line at an exact position in the script's resolution.
• YouTube and Vimeo do not accept ASS files. To show ASS styles on those platforms, burn the subtitles into the video.

Who uses ASS

Anime fansub groups, karaoke videos, and editors who burn styled subtitles into the picture. Netflix asks for TTML and US broadcasters ask for SCC, so ASS is not a delivery format for them.

TTML (Timed Text Markup Language) is a W3C format in XML. Its first version was called DFXP (Distribution Format Exchange Profile). The .dfxp file extension still uses that name, and YouTube reads a DFXP file as TTML.

1<tt xmlns="http://www.w3.org/ns/ttml" xml:lang="en">2  <body>3    <div>4      <p begin="00:00:01.000" end="00:00:04.200">Who has the Spanish file?</p>5      <p begin="00:00:04.600" end="00:00:07.000">Mina has it.<br/>She sent it back.</p>6    </div>7  </body>8</tt>

TTML can do a lot, so buyers limit it with a profile. A profile says which parts of TTML a file may use:

• IMSC: the W3C profile for subtitles and captions on the internet. Apple added IMSC captions to HLS streaming in 2017.
• SMPTE-TT: the profile the US FCC accepted in 2012 for captions on TV programs shown online.
• EBU-TT: the profile of the European Broadcasting Union, for exchange and archives.

Who asks for TTML

Netflix, for subtitle and SDH files. Other buyers name a TTML profile in their delivery specification. On 2026.10.13, Netflix accepted IMSC 1.1 only for Japanese.

Our free tools do not read or write TTML. Use your subtitle editor's TTML export, and check the buyer's profile before you deliver.

SCC files carry US TV closed captions, the CEA-608 standard (also called Line 21). The file does not store words. It stores the codes that a TV caption decoder reads.

1Scenarist_SCC V1.02300:00:01;00	9420 9420 94ae 94ae 9470 9470 5768 ef20 6861 7320 f468 e520 d370 616e e973 6820 e6e9 ece5 bf80 942f 942f

• The first line must be Scenarist_SCC V1.0.
• Times count frames at 29.97 frames per second. A semicolon before the frames, as in 00:00:01;00, means drop-frame timecode.
• A caption row holds 32 characters at most.
• Control codes are sent twice, in case the TV signal loses one.

Who asks for SCC

US broadcasters. YouTube names SCC as its preferred format for CEA-608 captions.

Our free tools do not read or write SCC. To make one, use a caption editor with SCC export, and check the result on a CEA-608 decoder.

• SBV (.sbv): SubViewer, a plain text format with no tags and no subtitle numbers. YouTube accepts it.
• EBU-STL (.stl): a binary format for European broadcast TV. You need a subtitle editor to open it.
• TXT: a transcript with no times. It must get times before a player can show it.
• Excel or CSV: not a subtitle format, but a common way to send subtitles to a client or reviewer who does not use a subtitle editor.

The place where the subtitles go decides the format. When you are not sure, ask for the buyer's delivery specification.

Every format keeps the text and the times. A conversion loses only what the new format cannot store.

Keep the original file after you convert. You cannot get styles or positions back from the converted file.