Source content "preprocessing"

You need to work with your watermarking vendor to produce watermarked content from your unmarked, source material.

This process is referred to as "preprocessing," and it requires that your watermarking vendor perform the following:

  • Convert source material content into "A" and "B" segments/fragments (variants).
  • Store the "A" and "B" variants on your origin server using different "A" and "B" paths.
  • Generate a manifest file without any indication of the "A" and "B" variants.

Convert and store using the "filename-prefix AB naming" method

The recommended method for preprocessing is to generate A and B variants of your source content, by including "a" and "b" in the filename of each variant segment/fragment, and then storing it in the applicable directory in your origin. This is also the only method that currently supports self-service implementation of watermarking in your AMD property via Property Manager.

Below is an example that uses a presentation called "baseball," with a master playlist, a media playlist, and several TS segments.

<base_path>/baseball/master.m3u8This contains relative paths to variants: 1080-media.m3u8, 720-media.m3u8

<base_path>/baseball/1080-media.m3u8This references relative paths to segments: /1080/segment_1.ts

<base_path>/baseball/1080/a.segment_1.ts

<base_path>/baseball/1080/a.segment_2.ts

<base_path>/baseball/1080/b.segment_1.ts

<base_path>/baseball/1080/b.segment_2.ts

<base_path>/baseball/720-media.m3u8This references relative paths to segments: /720/segment_1.ts

<base_path>/baseball/720/a.segment_1.ts

<base_path>/baseball/720/a.segment_2.ts

<base_path>/baseball/720/b.segment_1.ts

<base_path>/baseball/720/b.segment_2.ts

Separate audio and video segments

If you prepared your content with separate audio and video segments using the same file extensions, Akamai will attempt the A to B variant switch with audio as well as video. So, your watermarking vendor must prepare A and B variants for the audio segments as well, even though the A to B switch for the audio has no watermarking benefit.

If the audio segments use a different file extension, they must reside in the default location specified in the manifest because the A to B switch is bypassed for them. This is one of the primary benefits of the “filename-prefix AB naming” approach when preprocessing content.

Generate the manifest file

The original manifest file—before any preprocessing—is what is used by your client application (player) to request content. You don't need to modify the manifest file to obtain variant content. The Akamai edge uses the watermarking token (WMT) that your vendor creates to do this throughout the process, and ensures proper delivery of appropriate "A" or "B" segment.