High-Quality Speech Coding for Mobile

DoCoMo just announced that it has developed speech coding technology, in collaboration with their research labs based in Palo Alto, which will enable better voice quality for mobile phones, yet only requires the low-level computing power of conventional mobile technologies. This new approach apparently targets a wider frequency range of 50Hz - 16kHz, which is approximately the full range of the human voice. By comparison, the frequency range of legacy services is limited to 300Hz - 3.4kHz. There’s a few fancy graphs on this .PDF to demonstrate the performance range.

According to the company press release:

The technology employs speech compression that is matched to the specific characteristics of human auditory perception, such as the fact that distortion in loud sounds is relatively imperceptible. Moreover, the technology enhances sound quality to a remarkably high degree and processes coding with great efficiency.

The resulting voice quality is so realistic that conversations via mobile phones sound as if the other person were in the same room. The technology will be especially helpful in mobile applications where voice quality is particularly critical, such as teleconferencing and remote education.

DoCoMo’s “hTc Z” handset will be used to demonstrate a working version of the technology at “Telecom Summit Wireless Technology Park 2007″ at the Pacifico Yokohama exhibition complex in Yokohama from April 4 to 5, 2007. [We’ll try to attend for a test-drive — Eds]