Captions and Subtitles for Playable Assets

Deliver an Accessible Viewing Experience

~3 min read October 28, 2026

In this guide, you’ll learn how to prepare and deliver captions and timed text for Prime Video. By following these steps, you can ensure your content is accessible to all viewers. This includes those who are deaf or hard-of-hearing. You’ll also avoid common delivery issues.

Watch first

Why Timed Text Matters

1. Captions and subtitles help your content reach a wider audience.
2. Prime Video prefers Captions/SDH over subtitles alone.
3. They provide a richer experience that includes dialogue, speaker IDs, and atmospherics like sound effects and music cues.

Know Your Timed Text Types

Before you deliver, make sure you understand the three types of timed text:
1. Captions/SDH: Includes spoken dialogue and audio cues (for example, [door slams], [music playing]). This ensures usability for the intended audience of these assets: individuals who are deaf or hard-of-hearing.
2. Subtitles: Provides on-screen text of plot-pertinent dialogue. This includes all spoken dialogue and narrative text the audience needs, including Forced Narratives.
3. Forced Narratives: Also called Forced Subtitles. These translate dialogue or on-screen text in a different language from the primary audio. They appear when the creative intent requires the viewer to understand it. They display automatically based on the customer’s audio choice, not their subtitle selection.

[ICON] Quick Tip

Quick Tip

Every subtitle, closed caption, or SDH file in an MTA package must contain a full transcription of plot-pertinent dialogue.

For further guidance on textless, semi-textless, and non-subtitled masters, see: Video Requirements

Forced Narrative Requirements

1. A separate Forced Narrative file is required for each dubbed audio language in your MTA package.
2. You may also need a Forced Narrative in the original language. This applies if you’re providing a semi-textless master. Texted elements from the theatrical or broadcast version will have been removed.
This is important for former Prime Video Direct providers migrated to Slate. English captions are still required for all titles in the US locale. You can submit both English captions and English Forced Narratives for the same title. These assets serve different purposes and are not interchangeable.

[ICON] Quick Tip

Quick Tip

A separate Forced Narrative is needed for each audio language in your package, including dubbed audio. The Forced Narrative language-LOCALE must match the associated audio track. If it does not match, the Forced Narrative will not play when subtitles are set to “off.” On-screen text, location cards, and graphics are not translated in dubbed audio tracks. A Forced Narrative covers these elements for each language. Learn more in Languages, Localization, and Genres.

Accepted File Formats

Prime Video accepts the following timed text file types:

Format

File Extension

DFXP Full / TTML

.dfxp

Lambda Cap

.cap

EBU-STL

.stl

iTT (iTunes Timed Text)

.itt

Scenarist Closed Caption (SCC)

.scc

SMPTE-TT (RP-2052)

.xml

SRT

.srt

Get Frame Rates Right

Some timed text formats don’t natively include frame rate or drop/non-drop values. Here’s the rule of thumb:

1. Clock time format (for example, hh:mm:ss.sss): Frame rate info is not required.
2. Frame-based format (for example, hh:mm:ss:ff or hh:mm:ss;ff): Include frame rate and drop/non-drop information in your MMC asset manifest.
3. The frame rate of your timed text must match the frame rate of the mezzanine file.

[ICON] Quick Tip

Quick Tip

If frame rate is left blank for frame-based streams, a rate of 29.97 is assumed where applicable.

Timecode Format Support by File Type

File

Media Frame1

SMPTE Frame2

Clock (ms)

EBU-STL

✓

SCC

✓

Lambda Cap

✓

✓

DFXP

✓

✓

✓

SMPTE-TT

✓

✓

iTT

✓

✓

✓

SRT

✓

1. Drop/non-drop doesn’t apply to media frame-based subtitle streams.
2. SMPTE frame-formatted events are converted to clock time (milliseconds) before being paired with videos. SMPTE timecode tracks embedded in videos are not referenced.

Allowed Frame Rates by File Type

File

23.976 ND

23.98 ND

24 ND

25 ND

29.97 D & ND

30 ND

50 ND

59.94 D & ND

EBU-STL

✓

✓

✓

SCC

✓

✓

✓

✓

✓

✓

✓

✓

Lambda Cap

✓

✓

✓

✓

✓

✓

✓

✓

DFXP

✓

✓

✓

✓

✓

✓

✓

✓

SMPTE-TT

✓

✓

✓

✓

✓

✓

✓

✓

iTT

✓

✓

✓

✓

✓

✓

✓

SRT

1. Non-drop, SMPTE frame-based streams with decimal values are drift-corrected with a 1.001 multiplier after delivery.
2. Drop/non-drop info must be included in the manifest/file naming convention for 29.97 and 59.94 frame rates.

Positioning and Styling

Prime Video supports timed text positioning for Lambda Cap and iTT formats only. Still include positioning information whenever subtitles overlap with on-screen text or graphics.

File

Italic

Bold

Underline

Unicode

Positioning

EBU-STL

✓

✓

✓

SCC

✓

✓

✓

Lambda Cap

✓

✓

✓

DFXP

✓

✓

✓

SMPTE-TT

✓

✓

✓

✓

iTT

✓

✓

✓

✓

✓

SRT

✓

[ICON] Quick Tip

Quick Tip

By delivering high-quality captions and timed text, you’re not just meeting a requirement. You’re making your content discoverable and enjoyable for millions of additional viewers. Accessible content performs better and reaches further. Review Prime Video’s subtitling guidelines to learn about our principles for subtitles.

Language-Specific Style Guides

Want to ensure consistency across all your content? Prime Video has created Language-Specific Style Guides. These cover in-market industry standards, local preferences, and technical specifications for timed text.

Download Prime Video Localization and Technical Style Guides

Know Before You Go

No timecode offsets: Submit all files with 00:00 offset so they align with the video mezzanine.

Conform your timed text to the accompanying video before delivery. Ensure there is no offset or drift between your timed text and the video.

All Japanese timed text must be delivered as Lambda Cap.

In this guide



Internal Server error! Please try again
Your session has expired

Please sign in to continue

Sign In
edit