Phonexia Language Identification (LID) speech technology automatically recognizes the language and dialect of a speaker.
Cut through large amounts of audio files quickly with the automatic identification of a language and easily locate the speech recordings containing the language of your interest.
Categorize and segment speakers automatically based on their language and dialect so that you can make better business decisions and greatly improve customer experiences.
Offer impressive voicebot experiences by making sure that voicebots can seamlessly recognize the language of a speaker and communicate with them in the same language.
What People Are Saying
Phonexia has a close relationship with the Brno University of Technology, which helps it maintain a strong commitment to innovative technology and forward-looking strategies. Phonexia's focus on innovation will make it a prominent global player. With over 15 years on the market, Phonexia is committed to being a leader in the global voice biometrics industry.
We have been selling Phonexia products in the Asian market for over 3 years now, and we can clearly say that Phonexia products are by far the best ones in the global market. We have encountered some local and global competitors, but we can easily push them off the table by showing how strong and affordable Phonexia products are. Phonexia’s strength comes from a mixture of dedicated genius engineers and down-to-earth, smart business experts who can draw big pictures.
Technology Details
• Is text- and channel-independent
• Applies state-of-the-art channel compensation techniques, verified by NIST evaluation, compatible with the broadest range of audio sources possible: GSM/CDMA, 3G, VoIP, landlines, satellite phones, etc.
• New languages can be added to the system without any assistance from Phonexia—at least 20 hours of audio recordings are recommended for new language training
• WAV or RAW (PCM unsigned 8 or 16 bits, IEEE float 32-bit, A-law or Mu-law, ADPCM), FLAC, OPUS; 8 kHz+ sampling (other audio formats are automatically converted)
• Recommended Speech Length for Identification: 5+ seconds
• XML/JSON format with all results
• Results files with a logarithm of probabilities scoring (-∞;0)
• Percentage metric scoring <0-100%>
Approximately 20x faster than real-time processing on 1 CPU core with the most precise model—for example, a standard 8 CPU core server processes 3,840 hours of audio in one day of computing time.
Oromo, Albanian, Amharic, Arabic_Egypt, Arabic_Gulf, Arabic_Iraqi, Arabic_Levantine, Arabic_Maghrebi, Arabic_MSA, Assamese, Azerbaijani, Bangla_Bengali, Belarusian, Bulgarian, Burmese, Cebuano, Chinese_Cantonese, Chinese_Mandarin, Chinese_Min_Nan, Chinese_Wu, Chuvash, Czech, Dari, Dutch, English_American, English_British, English_Indian, Farsi, French, Georgian, German, Greek, Guarani, Haitian_Creole, Hausa, Hindi, Hungarian, Indonesian, Italian, Japanese, Kazakh, Khmer, Kirundi_Kinyarwanda, Korean, Kurdish, Lao, Lithuanian, Luxembourgish, Macedonian, Ndebele, Pashto, Polish, Portuguese, Punjabi, Romanian, Russian, Serbo-Croat-Bosnian, Shona, Slovak, Slovenian, Somali, Spanish_American, Spanish_European, Swahili, Swedish, Tagalog, Tamil, Telugu, Thai, Tibetan, Tigrignya, Tok_Pisin, Turkish, Ukrainian, Urdu, Uzbek, Vietnamese, Zulu
Schedule a demo with our experts and see for yourself how Phonexia Language Identification technology can help your business.
We will show you a free online demo tailored to your needs, showcasing Phonexia speech and voice recognition technologies and the ways you can use them to achieve your project’s goals.
