The large amount of data made ​​available by the media today makes it difficult for the selection and analysis of relevant aspects. The data ‘raw’ must be organized in structures that help to effectively extract information of interest, to relate them, and to draw the appropriate conclusions. This management is even more difficult when you have to examine media files (audio-video) to analyze the ‘words’ in them in relation to their ‘context’. VAMP is a system designed to organize, analyze and represent, in the most convenient way, file archives, media, and facilitate the main purpose of the analysts: to move from words (data) to information. With VAMP you can:

  • find the exact quotes and comments in multimedia archives, performing phonetic search with a certain degree of accuracy;
  • perform research on automatic transcriptions of audio / video files, specifying not only the ‘keywords’, but also semantic parameters as arguments and “sentiment” (“polarity” positive / negative / neutral) of the texts or their sections;
  • automatically build the “indices” of the topics of educational broadcasts or information
  • monitor,on demand or periodically, the presence of “key words” in television and radio programs;
  • display alternating speakers and non-speech content (music, advertising);
  • analyze the relationships between the user’s query, the documents, the speakers and topics through a series of “maps” built automatically;
  • see the automatic transcriptions of each document, with the text to audio sync, the ‘entity’ detected (names, places, organizations) and the ‘polarity’ of each section of the document.


  • Speaker Diarization: segmentation corresponding to the interventions of the different speaker or context switches.
  • Scene Change Detection: video shot change detection .
  • Voice Recognition: automatic transcription of the text with specialized vocabularies.
  • Sentiment Analysis: analysis of the positive / negative / neutral polarity of the texts or their sections.
  • Topics Identification: automatic detection of arguments and their indexing (‘summary’).