Messages announced over train station loudspeakers are notorious for being unintelligible. It can also be difficult to understand announcements made in airports, at conferences, or in any number of other busy public spaces. Germany's Fraunhofer Institute for Digital Media Technology is trying to do something about it, however. It's developed new audio-enhancement software, which might even find use in smartphones.
Known as ADAPT DRC, the software uses strategically-placed microphones to continuously monitor the ambient sound in the environment. Based on what it hears, it then manipulates the audio of spoken messages, in order to compensate for competing noise. This doesn't mean that it simply turns up the volume, however.
There are already systems that simply make announcements louder when ambient sound levels increase, but this doesn't always make them easier to understand. In fact, higher volume can sometimes distort spoken words, as it causes the speakers to vibrate.
Instead, ADAPT DRC selectively manipulates specific frequencies, and it does so in real time. This includes boosting consonant sounds such as "t" and "p" – because these are short-duration high-frequency sounds, they're particularly prone to being drowned out by other noises.
Additionally, using an existing technique known as Dynamic Range Compression, the software simultaneously boosts the volume of quiet speech sounds while also lowering the volume of loud ones. This results in a message that's delivered at one consistent volume, making it easier to discern.
As mentioned, ADAPT DRC may also find use in smartphones, allowing phone calls to be heard in noisy environments. It may actually be easier than implementing the system in train stations or other locations, as phones already have built-in mics.
Source: Fraunhofer