Captions and/or timed text

All timed text assets must be conformed to match the length of the accompanying video prior to delivery to Prime Video. Whenever available, Prime Video prefers to receive captions/SDH (subtitles for the deaf or hard-of-hearing) over subtitles alone, to provide an enhanced viewing experience to customers who are deaf or hard of hearing.

• Captions/SDH: Timed text that includes both spoken dialogue and atmospherics for the deaf and hard-of-hearing.
  
• Subtitles: Timed text assets that provide on-screen text of the program dialogue. Any subtitle, closed caption, or SDH file delivered as part of an MTA package must contain a full translation of all spoken dialogue and narrative text that needs to be understood by the audience, including Forced Narratives.
  

• Forced Narratives: Forced Narratives (also known as Forced Subtitles) translate spoken dialogue and on-screen text that is in a different language from the primary audio content of the program when creative intent requires that the dialogue be understood by the viewer. They are displayed to the customer based on their audio choice, rather than selecting that subtitle language. If a title contains Forced Narrative content, delivery of separate Forced Narrative timed text assets is required for each language available in a multi-track audio (MTA) package. A Forced Narrative file is required for every dubbed audio language being provided. You may also need to provide a Forced Narrative file in the original language of the title if you are providing us with a semi-textless master, where texted elements present in the theatrical or original broadcast version have been removed.

  Important:
  

  • If you are a former Prime Video Direct provider migrated to Slate, you are not able or expected to upload both English caption files and Forced Narratives in English for the same title on Slate. However, English captions are required for all titles in the US locale.
  • The forced narrative language-LOCALE must match the main video mezzanine. If it does not match, the forced narrative won’t play to customers when subtitles are set to the off state. For more information about language-LOCALE settings, see Languages, localization, and genres.
    

Technical guidelines for timed text

Prime Video accepts the following file types for delivery of timed text:

• DFXP Full/TTML (Timed Text Markup Language) with a .dfxp file extension
• Lambda Cap with a .cap file extension (only format that supports discrete Japanese timed text delivery)
• EBU-STL with a .stl file extension
• iTT (iTunes Timed Text) files with a .iTT file extension
• SCC (Scenarist Closed Caption) with a .scc file extension
• SMPTE-TT (RP-2052) with an .xml file extension
• SRT

Frame rates and drop/non-drop frame indication
Prime Video allows a wide range of timed text formats, some of which don’t natively include frame rate or drop/non-drop values. In general, if the time code format is in clock time (such as hh:mm:ss.sss), frame rate or drop/non-drop information isn’t required. If it’s in a frame-based format (such as hh:mm:ss:ff or hh:mm:ss;ff), then you must send both frame rate and drop/non-drop information via the MMC asset manifest.

Note: Frame-based subtitle streams must include frame rate in the manifest/file naming convention. A rate of 29.97 is assumed, when applicable, if this is left blank.

Depending on the specification and namespace used (TTML or TTAF) in DFXP, XML, and ITT files, Prime Video uses either the TTML timebase or TTAF timebase (media and SMPTE only) accordingly for parsing.

The following tables summarize the timed text formats that Prime Video accepts and whether or not that information must be included in the delivery manifest document. Information provided via delivery manifests takes priority over assumed default values found in the header metadata of the timed text file.

Important: All Japanese timed text must be delivered as Lambda Cap.

Timecode format

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

Allowed frame rates

³ Non-drop, SMPTE frame-based subtitle streams with decimal values are drift-corrected with a multiplier of 1.001 after delivery.
⁴ Drop/non-drop info must be included in the manifest/file naming convention for 29.97 and 59.94 frame rates.

Positioning and styling
While Prime Video only supports timed text positioning for Lambda Cap and iTT formats, positioning information should be included when subtitles overlap with onscreen text and graphics. The following table specifies what positioning and styling is supported for each timed text format.

Time code offset
Prime Video doesn’t support time code offsets. All timed text files must be submitted with a 00:00 offset in order to conform to the video mezzanine file.

Character support
Prime Video supports the full UTF-8 character set, except musical notes.

Language-specific style guides
Prime Video has created Language-Specific Style Guides for localization. These guidelines encompass in-market industry standards, local preferences, and timed-text technical specifications. By following these comprehensive guidelines, users can ensure consistency across all content types on Prime Video.

Download Prime Video Localization and Technical Style Guides