Digital Measurement SEI Audio File

From Engineering Client Portal

Revision as of 07:31, 22 June 2024 by Admin (talk | contribs) (→‎Header Record)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)

Engineering Portal / Digital / Digital Streaming Measurement / Digital Measurement SEI Audio File

The Nielsen SEI server-to-server solution (Digital Measurement SEI Content Viewing) needs audio files for each watermarked source that requires crediting.

Audio Compression Type Tested ideal Compression Rate
AC3 Stereo 192 Kbps or higher
AC3 5.1 384 Kbps or higher
Enhanced AC3 192 Kbps
MPEG2 audio 192 Kbps or higher
AAC with ADTS headers
  • If source audio is stereo, include both Left and Right channels
  • If source audio is 5.1, take L + .707C into Left and R + .707C into Right and deliver stereo
  • Audio sampling rate should be 32kHz, 44.1kHz or 48kHz

S3 Bucket and Prefix Naming Convention

DTVRAudio_<partnerid>_<AssetId>_<UTCStart>_<UTCEnd>

Name Description
partnerid Abbreviation provided by Nielsen for each provider or publisher
AssetId in-house assetId
UTCStart time in 32-bit unsigned int UTC in seconds from first second of audio captured
UTCEnd time in 32-bit unsigned int UTC in seconds from first second of audio captured
  • Files should be placed in S3 buckets in a mutually agreed upon folder structure.
  • 15 minutes of audio does not have to align with hour boundary. For example for broadcast from 1 AM to 2 AM, 4 audio files each with 15 minutes of audio from 1:00 AM to 1:15 AM, 1:15 AM until 1:30 AM, 1:30 AM until 1:45 AM and 1:45 AM until 2:00 AM, but if the encoder turns on at 1:07, deliver through 1:22, etc. The preference is hourly bound but not required.
  • If difficult to achieve exactly 15m increments, err towards shorter. I.e. 10m instead of 20m
  • Avoid small, fragmented audio files. Watermark decoding algorithms require runway. Too short of a file will result in lost audio codes
  • If file times/audio overlap, ingestion will overwrite the portion of the older file
  • Precision is important for UTC start/end file naming convention. Use UTC timestamp of Server time/encoder
  • The file should not be encrypted.
  • One copy per stream only.

Audio Asset Metadata

S3 Bucket and Prefix Naming Convention

DTVRAssetMetaData_<partnerid>_<AssetId>_<UTCStart>_<UTCEnd>

Name Description
partnerid Abbreviation provided by Nielsen for each provider or publisher
AssetId in-house assetId
UTCStart time in 32-bit unsigned int UTC in seconds from first second of audio captured
UTCEnd time in 32-bit unsigned int UTC in seconds from first second of audio captured

Header Record

Parameter Description Mandatory Specified by Format
appid Nielsen-provided server application identifier Yes Nielsen Example: A487421B-XXXX-YYYY-8343-E3BBB66E44F2
creationtime File creation time Yes Client Format: Unix timestamp in 32-bit unsigned int in seconds

Example: 1577858505

assetid In-house id used for a video asset Yes Client Format: alphanumeric

Example: VID123456789

recordcount # of records in body of file Yes Client Format: numeric

Body

Parameter Description Mandatory Specified by Format
audiofilename Audio file name Yes Client Example: aacFile1.aac
referencestart Start of in-house reference time for the audio file Yes Client

Format: Unix timestamp in 32-bit unsigned int in seconds

  • For VOD, parameter should reflect offset from the start of the asset
  • For Livestream, parameter should reflect in-house time
referenceend End pf in-house reference point for the audio file Yes Client

Format: Unix timestamp in 32-bit unsigned int in seconds

  • For VOD, parameter should reflect offset from the start of the asset
  • For Livestream, parameter should reflect in-house time

SLA

The files must be delivered into the proper S3 bucket within 2 hours of the broadcast. For example files from 1:00 AM to 1:15 AM must be delivered before 3:00 AM.