Introduction

Consumer/Prosumer level Camcorders and Still Cameras (DSLRs and DSCs) require a self-contained media file format to carry captured audio/video footage and metadata. The QuickTime File Format [1] fills this need very nicely, however there is an increased level of interest among camera manufacturers for a specification based-on the standard such as MP4.

This document addresses such interest by defining the format of the timecode and chapter marker tracks in the MP4 File Format [2], borrowing relevant definitions from the QuickTime File Format. In addition Apple recommends using the User Data container for storing non-temporal metadata in an MP4 file with the keys defined in the 3GPP standard [5]. The document also defines additional metadata keys intended for use with video editing applications such as Final Cut Pro X.

MP4 File Format

The MP4 File Format is a multimedia container file format defined as part of the MPEG-4 standard. The official name of the standard is ISO/IEC 14496-14:2003 [2], often times referred to as MPEG-4 Part 14. It defines how to store video and audio streams encoded with certain codecs, such as MPEG-4 Part 2 or MPEG-4 Part 10 (H.264/AVC) and Advanced Audio Coding (AAC). The MP4 File Format is also used to store subtitle text in the MPEG-4 Timed Text format (MPEG-4 Part 17 [3]).

The MP4 File Format has general provisions for carrying temporal metadata using the definitions in the ISO Base Media File Format (MPEG-4 Part 12 [4]), but there is no concrete specification to store timecode information or chapter markers. There is very limited non-temporal or clip metadata (only copyright) defined specifically for MP4.

Timecode and Chapter Markers

Apple recommends that developers and manufacturers use the following two track types to store timecode and chapter markers in an MP4 file:

• Timecode Track

• Chapter Marker Track

These track types are already available in the QuickTime File Format. The MP4 File Format design is similar to the QuickTime File Format, therefore it's natural to borrow these constructs defined in the QuickTime File Format to support the features in the MP4 File Format. It may, however, be appropriate to drop legacy QuickTime specific constructs that are awkward in the context of the ISO specification.

Timecode Track

The timecode track in an MP4 file uses the constructs defined in the QuickTime File Format, with some minor modifications.

• The text formatting information to display the timecode is dropped.

• The Null Media Header box (nmhd) box is used instead of the Generic Media Header box (gmhd).

The rest of the QuickTime Timecode track functionality, such as 64-bit timecode support, is preserved.

class TimecodeSampleEntry(codingName) extends SampleEntry(codingName) {

   const unsigned int(32) reserved = 0;

   unsigned int(32) flags;

   int(32) timescale;

   int(32) frameDuration;

   int(8) numFrames;

   const unsigned int(8) reserved = 0;

}

Chapter Marker Track

The Chapter Marker track stores text chapter markers in an MP4 file as a disabled text track, with a track reference that identifies the track as a chapter marker track. It uses the constructs defined in the QuickTime File Format Specification with minor modifications.

Clip Metadata

For clip (or non-temporal) metadata, Apple recommends using the User Data container and store metadata items as user data items with user data keys, each of which is a four character code. This provides sufficient functionality to carry text-oriented metadata items with language identification. The 3GPP file format [5] defines more than a dozen user data keys, and Apple recommends to follow those definitions. The following three keys are currently supported by Final Cut Pro X:

• titl – The title for the media.

• cprt – The information about the organization holding the copyright for the media file.

• auth – The author of the media.

Appendix A: 3GPP User Data Keys lists the metadata keys defined in the 3GPP Release 12 [5]. Refer to the specification [5] for the definitions of custom data types.

Apple also defines additional user data keys for video applications such as Final Cut Pro X, such as the manufacturer and the model name of the camera. Appendix B: Video Editing User Data Keys contains a list of those user data keys. Note the string value for the creation date key is expected to be in the ISO 8601 date string format [7].

Unlike with the QuickTime Metadata keys, there isn't a safe way to introduce private keys for user data keys. For an MP4 file, private metadata should be stored in a UUID box as defined in the ISO Base Media file format [4].

References

[1] QuickTime File Format Specification, 2012, Apple Inc.

[2] MPEG-4 Part 14: MP4 file format; ISO/IEC 14496-14:2003, International Organization for Standardization

[3] MPEG-4 Part 17: Streaming text format; ISO/IEC 14496-17:2006, International Organization for Standardization

[4] MPEG-4 Part 12: ISO base media file format; ISO/IEC 14496-12:2012, International Organization for Standardization

[5] 3GPP TS 26.244 v12.1.0: 3GPP file format (3GP) (Release 12), 3rd Generation Partnership Project

[6] 3GPP TS 26.245 V12.0.0: 3GPP Timed text format (Release 12), 3rd Generation Partnership Project

[7] Data elements and interchange formats - Information interchange - Representation of dates and times; ISO 8601:2004, International Organization of Standardization

Appendix A: 3GPP User Data Keys

Refer to [5] for definition of custom data types.

Appendix B: Video Editing User Data Keys