Example Configuration - Advanced
The following diagram shows a more complex configuration, which includes an encoding task and multiple analysis tasks.
- Media Server receives a
Process
action, which starts a new session. The Video ingest engine receives video from the source specified in the action. It produces an image track and an audio track. - The image track is used as the input for the MPEG encoder and Face Detection analysis tasks.
- The audio track is used as the input for the MPEG encoder, and the Speaker ID and Speech-to-text analysis tasks.
- The MPEG encoder writes a copy of the video to disk as a set of MPEG files. It also produces proxy information that is available to the output engine.
- The face detection engine produces a track that contains information about detected faces. This is used as the input for the face recognition analysis engine.
- The output tracks from the analysis engines are used as the input for the IDOL Server output engine. The analysis engines all produce multiple tracks, some of which are not used.
- The IDOL output engine transforms the records produced by the MPEG encoder and analysis engines into documents and indexes the information into IDOL Server.
A Media Server configuration that matches this process is shown below.
[Session] Engine0=Ingest Engine1=MPEGencode Engine2=Face Engine3=FaceRec Engine4=SpeakerID Engine5=SpeechToText Engine6=IDOLOutput [Ingest] Type=Video [MPEGencode] Type=mpeg OutputPath=\\server\folder\file.mpg URLBase=https://www.myserver.com/folder/ Segment=TRUE [Face] Type=FaceDetect [FaceRec] Type=FaceRecognize Input=Face.DataWithSource [SpeakerID] Input=Default_Audio Type=SpeakerID Database=speakers [SpeechToText] Input=Default_Audio Type=SpeechToText LanguagePack=ENUK SpeedBias=Live [IDOLOutput] Type=IDOL Input=FaceRec.Result,SpeakerID.Result,SpeechToText.Result,MPEGencode.Proxy IDOLHost= IDOLPort= Mode=time OutputInterval=60s XslTemplate=./xsl/toIDX.xsl SavePostXML=true XMLOutputPath=./output/idol/