AppTek's Omnifluent™ Media and Translate software can accurately and automatically identify the language of source audio and text in over 30 languages and multiple file formats. This Language Identification capability is a particularly useful quality control procedure for broadcasters with multi-language feed environments to verify the language of audio tracks or subtitle files prior to play out. Available for local installation, our Language Identification software requires minimal hardware resources, offers fast response times, and our RESTful APIs make integration simple.
Omnifluent™ Media uses AppTek's Automatic Speech Recognition (ASR) technology to automatically transcribe broadcast quality audio in 11 languages and multiple accents of English and in 4 languages for telephony quality data, with new languages continually in development. Powered by language and acoustic specific models, Omnifluent Media can be tailored to a client's unique audio setting for optimal performance. Languages available for immediate use with broadcast quality audio: Arabic, Brazilian Portuguese, Traditional Chinese, Dari, English, Farsi/Persian, French, Italian, Pashto, Spanish, and Urdu. Languages available for immediate use with telephony quality audio: Arabic (including Modern, Iraqi and Egyptian dialects), English, Spanish, and Farsi.
Powered by AppTek's Automatic Speech Recognition (ASR) technology, Omnifluent™ Media can segment telephony and broadcast quality audio at the phrase level, providing users with valuable time-coding information for transcribed speech that allows transcripts to be synchronized with the source audio and makes the audio text fully searchable. We can export the segmented audio as an XML file or as a caption file in the following formats: .cap, .pac, .scc, .srt. The latter allows the user to align the text with audio in a video clip as captions.
AppTek's innovative, patent-pending machine translation software, branded Omnifluent Translate, is a Hybrid Machine Translation engine, that uses both rule- and statistical-based engines to improve translation meaning and word accuracy. Built from over 20 years of continual research and development by AppTek's skilled team of computational linguists and software engineers, Omnifluent Translate represents a leading-edge, breakthrough in language translation technology licensed to federal and local government agencies and major commercial clients.
Through our Omnifluent Translate software, AppTek clients can create fully automated foreign language subtitles in near real-time for 24 languages. The software synchronizes the subtitles with the original source audio and outputs multiple caption file formats (.cap, .pac, .scc, and .srt). In order to achieve the highest translation quality standards, AppTek offers its software as a tool for professional translators to leverage, making them three times more productive compared to traditional translation methods with no compromise to accuracy. Additionally, AppTek can provide customers a cost-effective end-to-end subtitling solution that utilizes AppTek's language technology and linguistic services.
Using our Omnifluent Media software, AppTek customers can create fully automated, same-language captions for offline content in 11 languages in near real-time. The software automatically synchronizes captions to the original source audio and outputs multiple caption file formats (.cap, .pac, .scc, .stl, and .srt). In order to achieve the highest transcription quality standards, AppTek offers its software as a tool to professional transcribers, making them more efficient compared traditional transcription approaches while maintaining verbatim recognition accuracy. And if a transcript for a video already exists, users can accurately align the text with the audio through Omnifluent Media without the usual laborious task of text time-coding. Additionally, AppTek can provide customers a cost-effective end-to-end captioning solution that utilizes AppTek's language technology and linguistic services.
Omnifluent Monitor software enables 24/7 monitoring of live video content. Omnifluent Monitor recognizes speech from any variety of speakers in up to 11 languages and transcribes speech in near real-time to create fluent transcripts. The software automatically time stamps and indexes all of the data to provide a rich repository of searchable terms, which enables users to find segments of speech within a file quickly and accurately, regardless of speaker, language, or keyword.
Our unmanned solution captures and indexes telephone conversations in near real-time, making it immediately searchable and accessible. Once audio is captured and indexed, AppTek's powerful suite of Knowledge Management tools facilitate intelligent content search capabilities for users including categorization, clustering and automated summarization for government and commercial organizations alike to maximize their multimedia assets.
A Web Crawler feature captures textual Web content, such as chat windows and blogs, which are less scripted and less grammatical than newswire. It also captures Flash video/audio from Web sites, including user-gnerated content (UGC) and processes them as it does broadcast content.
This feature of AppTek's Monitoring software, automatically highlights text to mark entities, such as people, places, organizations, dates, and times. The system can identify all proper nouns in multiple languages such as human names, organizations, and geographic places.
Our software has bilingual dictionaries that allow users to search in multilingual text simultaneously. With this capability, the user can type a search word or phrase in one language and retrieve the matching words with successive part of speech in other languages using the morphology analyzer as well as the large-size dynamic bilingual lexicons.