For decades, engineers in telephony have been creating ever more elaborate CODECs attempting to minimize transmission bandwidth, while increasing the fidelity of audio reproduction at the receiving end. The MOS (Mean Opinion Score) has traditionally been used to grade these improvements. It measures on a scale from 0 to 5 the subjective appeal of sound reproduction, with 5 meaning that there is no audible difference between the original sound and its reproduction. However, much of this effort is a bit wrong-headed, since they have been applying music reproduction criteria, where maximum fidelity is demanded by listeners, to a situation for voice communications where clarity of speech is more important than how faithful it sounds. And when people are speaking on a phone, their listening environment is generally less than ideal due to environmental noise. So the listener normally has to boost the volume. High fidelity rendering is pointless in noisy environments, but can be satisfying in quiet conditions. There hasn’t been any real alternative to boosting the volume in noisy situations because the physics of hearing hasn’t been well understood.
Voice communication demands clarity of speech, and that isn’t the same thing as full-spectrum high fidelity audio. Makers of telephone systems and radio systems have long known that most of the information in speech is contained in a relatively narrow frequency bandwidth from around 350 Hz to 3500 Hz. Crowded radio spectrum regions benefit greatly by restricting the transmission bandwidth of voice to this narrow range, to accommodate more channels for communications. Old fashioned POTS (Plain Old Telephone System) benefited from narrowed transmission bandwidths to help overcome dispersive effects of transmitting audio over long distances by wire. Of course, now that the era of digital transmission is with us, these former audio bandwidth restrictions no longer seem necessary. But let’s look again… By applying a high-fidelity music standard to voice communications, especially while listening in noisy environments, it is wasteful of battery power and does little to improve the clarity of communications. About 3/4 of audio amplification power is expended on frequency regions that do nothing to aid voice communications. And the noisier the listening environment, the more power is expended when the listener pushes the cell-phone to maximum volume levels in an attempt to overcome the surrounding noise. Fidelity goes out the window with lots of ambient noise. You couldn’t hear a pin drop in those situations, no matter how well the phone can reproduce that sound. Without understanding the physics of hearing and speech there really isn’t much else to do apart from raising the volume. Perhaps you could add a little treble EQ, but then in quiet environments the listener gets an abrasive experience, and they would much prefer the high-fidelity solution without EQ or excessive volume.
At Acudora we have the benefit of nearly a decade of privately funded research into the physics underlying human hearing. Our research gave us a model to show how to overcome ambient noise in the listening environment, while simultaneously conserving battery power through noise-aware, adaptive, frequency selective amplification. Our technology leverages psychoacoustics as well, showing us how to achieve clarity in voice communications without wasting power on useless frequency regions. Tests show that under comparable noisy listening conditions, our vDBM technology improves the MOS score by 350% over that obtained from simple volume boost. And it does so while expending only 1/4 as much power on the audio amplifier. Battery drain trials show that we can extend battery life by more than 200% in typical situations (often times nearly 300%). Adaptive processing means that when you do finally obtain a quiet listening place, you can enjoy full-spectrum high fidelity of HD Audio, if your phone offers it, as vDBM processing becomes completely transparent. Being noise-aware means that it gracefully curtails its actions as the environment becomes more quiet, and becomes more aggressive when the ambience grows much louder. Frequency selective processing is important because the noise of a crowded room is different from the street noise of traffic. You can hold a clear conversation while walking from the quiet of your closed-door office and onto the street in the midst of rush hour traffic — without ever touching the volume control. Think of vDBM as a new kind of CODEC, aimed at the transmission channel that exists between the speaker and the listener’s brain. That’s the final barrier to clear communication. And that’s the channel most suitably addressed by our vDBM technology.