Extensive Services Through Automatic Speech Recognition
Because of its unique technology VETAIL-X.COM is able to transcribe audio and video files. In this process an automatic speech recognition software (ASR) recognizes spoken language and puts out written text. Previous speech recognition software used to be speaker-dependent and had to be trained for each individual speaker. Our speaker-independent software is able to transcribe every speaker as well as recognize different speakers. Through the transcription the actual content of a video is extracted and made searchable.
Step 1: The automatic speech recognizer transcribes the audio or video file. Step 2: Places and personal names found in the text are extracted (e.g. Angela Merkel, Germany, Europe). Step 3: The transcribed text is categorized (e.g. politics -> election (Score: 19 %), politics -> government -> national government/federal government (Score: 16 %). Step 4: Keywords are extracted (e.g. FDP, SPD, Westerwelle, left-wing party). Step 5: For video files we automatically create a snippet of the video sitemap. Step 6: With video files we can automatically generate a video overlay that displays the spoken text in the respective time slot (subtitles). Different speakers are marked with different colors. This is also a further step towards web accessibility. Step 7: Using similar keyword combinations, we check what terms for video integration can be found in Google search results and use these as additional keywords to optimize the video for the Google video search. Advantages
- The transcribed videos are accessible to people with hearing impairment.
- The actual video content is made searchable.
- The displayed “similar videos” are more similar indeed, as the similarity is based on the actual content instead of manually added tags.
- The users stay longer on the video platform which results in more exposure for the marketer.
- Additional readable text for the search engines ensures more long-tail traffic.
- Overlay banners can be booked and displayed specifically for the words spoken (analogous to AdWords).
- Automatic video categorization
- Automatically generated video sitemaps for Google
- More video integration within the universal search through optimized meta daten
- Automated tagging using automatically found keywords, names and places
