Advertisement

Here's what high-end smartphone speakers could sound like in 2014 (ears-on)

Not all smartphone speakers are created equal. Last year, we were blown away by a prototype NXP audio chip that boosted volume (but not distortion) through the use of feedback circuitry. The chip monitored a speaker's behavior in real-time so that it could be pushed to the limit without creating crummy sound or being physically torn apart. Others were impressed too: Motorola has since used the nine-volt TFA9890 in the Moto X and the Droid Ultra. These phones have single speakers, but are almost able to match the volume and sound quality of a good stereo system, such as HTC's BoomSound audio in the One family of smartphones (which rely on two five-volt NXP-controlled speakers). So, what comes next? That'd be the second-gen TFA9895, which makes some gentle but noticeable improvements, and which should start arriving in high-end smartphones by the time we get to Mobile World Congress next year. Fortunately, we don't have to wait that long to hear it, because NXP just invited us to an exclusive ears-on.

Compared to the difference we heard last year, the TFA9895 is a subtle beast. It's not designed to push voltage further, but to improve sound quality regardless of how much total power is sent to the speaker. It does this by applying feedback to different frequency bands separately, so that parts of the audio that risk causing distortion can be dampened down without suppressing speaker output as a whole. If the vocals in a track suddenly get loud, for instance, they can be controlled on their own, without the backing instruments also having to get quieter -- in other words, different elements in the track don't sound like they're competing against each other.

And yes, listening to a demo board with stereo 5V speakers, we could hear the difference. In some casual blind tests, we managed to identify the improved device three out of four times -- with the one failed test being Avril Lavigne's Sk8er Boi song about a boy who skates, which is already a heavily compressed track where most frequencies have the same volume. We were left with the feeling that this chip will be extremely useful for watching TV and movies, where separation of different strands of the soundtrack is crucial to the experience, so bring on 2014.