Peter Santos comes across as a pretty mild-mannered executive, until he starts trying to come up with a metaphor for how crummy today’s cellular phones sound relative to how good he knows they could sound.  “What we have in voice today is a 13-inch black-and-white TV set.  We have an opportunity to have a 60-inch set with 1080p high-definition,” he says, his voice rising with a mixture of frustration and excitement. “It’s within our grasp.”

For seven years Santos has run Audience Inc, a chip-maker that specializes in translating the human voice into digital bits and back to voice again without garbling the conversation. Audience chips are now in over 40 million mobile phones, mainly to help to filter out the background noise. (That way, a bus going by doesn’t drown out your conversation.)  Along the way Santos has learned a lot about why cell phones calls sound so bad to the human ear.

While noise suppression helps, there’s only so much Santos has been able to do given limitations of today’s cellular networks.  Those networks have two glaring quality flaws.  They’re slow, adding a noticeable lag between when one caller talks and the other hears their words, leading to overlapping talk. They are also short on capacity, encouraging wireless carriers to smoosh phone calls into tiny digital streams and in the process lose all the little audio nuances that make a conversation sound lifelike.

Now a new high-def era is finally dawning, says Santos.  Verizon (VZ) and AT&T (T) have said this month that they plan on launching “Voice over LTE” in 2012 and 2013 respectively.  LTE, the most popular “4G” standard, is capable of crystal-clear calls with no noticeable lag. “The joke in the industry used to be is that wideband cellular is two years away—and always has been,” says Santos. “Our view is that now that has changed.”

LTE systems will carry more digital bits at lower cost—and the more bits a phone call uses the better the call sounds. U.S. cellular system typically smush a call into 6 or 7 kilobits per second, leading to mediocre sound under the best of circumstances. (Looked at in terms of audible frequencies, today’s phones only pass along frequencies between 300 hertz and 3500 hertz, a fraction of the pitches the ear can detect.  But by doubling or tripling the bits devoted to each call, new “wideband” phones could pass along all audible frequencies.)

Santos is already building chips for this new high-def world, including a new one released last week.  He says that as we get the improved phones there will be even more need for clever noise-supression technologies. That’s because in the new, higher frequencies that LTE phones will carry, background noise is an relatively bigger problem.

In the meantime, Audience’s chips will help today’s cell phone sound better  with new tricks to paper over the network problems.  One example: the chip lets your phone fill in the missing frequencies in today’s networks, creating the illusion of greater bandwidth.  But that’s only a short-term fix, until hi-def handsets take over.

Daniel Beringer, an industry consultant and expert in hi-def telephony, agrees that will happen faster than people expect.  He says that companies like Google (GOOG), Skype, Orange, Audience and others are laying the groundwork by creating HD capable technologies. Right now, most HD capable phones connect to regular phones and must default to the old quality standards.

As more people get HD capable phones, a feedback loop starts, and the number of HD conversations will soar, Beringer predicts.  Once people hear what they are missing, demand will explode he predicts.  “Each person that gets it becomes the salesperson for their family,” he says. “It’s just like putting on a pair of glasses for the first time. There’s a huge hearing impairment you didn’t really notice before.”

Adds Peter Santos: “There’s a tipping point. It’s very close.”