AppTek Announces Media Syncing Platform
AppTek has launched the Auto-Sync platform, which leverages AppTek's proprietary neural-network-based automatic speech recognition technology to create time-stamped transcription of audio files. This data is then used to align existing transcripts to the original media asset.
"Auto-Sync is an extremely valuable platform for media producers, creators, and owners globally," said Adam Sutherland, AppTek's CEO, in a statement. "The act of marrying the existing script or transcript to the actual media asset allows for more accurate closed captioning, subtitling, discoverability, and facilitates re-use through more accurate archive searching."
Auto-Sync is available in 14 languages including English, Spanish, French, Mandarin, German, Russian, and Arabic.
Auto-Sync is the latest product from the machine learning/AI labs at AppTek, whose bi-directional long-short-term-memory (BLSTM) modelling has led to drastic improvements in the accuracy and speed of AppTek's ASR.
"Our ASR models outperform any system deployed for media captioning/transcription or call center mining/analytics in quality and speed," said Jintao Jiang, AppTek's chief scientist, in a statement. "The new models are faster, more accurate and easier to train. They represent a true step-change in machine learning and ASR."
"AppTek's scientific team is laser-focused on language technologies built on machine learning and artificial intelligence. It's all we do," Sutherland added. "That means we are able to compete with, and in many cases out-perform, much larger companies and teams. Our technology continues to increase the value of recorded media for our partners in media, telephony, mobility. and government."