We've all been at an airport or train station, or even a concert, where there's an announcement and we can't quite make out what it was over the noise. Those days may soon be over thanks to a new speech enhancement technology from Fraunhofer called ADAPT DRC.
Researchers at the Project Group Hearing, Speech and Audio Technology at Fraunhofer Institute for Digital Media Technology developed the ADAPT DRC software in an effort to improve any type of electronic communication, but venues with congregations of people were the primary target.
What happens is that microphones are strategically placed around a venue to constantly monitor the ambient noise level. When the noise gets too loud, the software boosts the speech frequencies instead of the overall volume of the speakers. This keeps the speakers from distorting, making it even more difficult to understand the announcement.
Instead ADAPT DRC strategically boosts the consonant sounds like "P," "T," and "K," which are often spoken quickly, but are really the key to understanding what is being said.
The software also takes into account the parts of the speech signal that are naturally at a different volume and uses an intelligent algorithm called Dynamic Range Compression (the DRC in ADAPT DRC) to boost the intelligibility. This technology is already used on many cell phones.
I always marvel at how crappy some announcement sound systems can be, considering the technology that's available today. Let's hope that ADAPT DRC works as sited and is widely adapted so we don't miss that next flight to AES.
(Photo credit: Dornum72 via Wikipedia)