Google Voice has one critical flaw, and that is it has an inherent "processing delay" that it introduces into the voice path.

The delay is slightly longer than the delay most cell phones have talking to another cell phone. When you add the Google Voice delay in, it's almost an unbearable 1/3 to 1/2 a second.

I've used it from my land line calling calling out because of the free calling feature, and for that the delay is tolerable. But I can't justify having it forward to my cell phone because if anyone calls me from a cell phone, the combined lag makes the conversations really hard to have.

You can't talk to bus drivers, and that's been banned.

I'm suspicious of the quality of his results, especially his assertion that kids actually *preferred* the MP3 encoded results.

There are MANY explanations for why this could be occuring. For one, many MP3 encoders apply a low-pass filter to encoded data to smooth out artifacts. It's not clear that this professor's audio playback equipment is uniformly able to reproduce higher frequency sounds - very few people have that kind of equipment who aren't professional audio recording engineers.

I would believe it's just as likely that his "uncompressed" audio files actually SOUND WORSE because they are creating distortion in the playback equipment that the MP3 filtered files are not.

This reminds me of a story I heard once on NPR about audio engineers who worked on live audio feeds for music shows. They found that most people had crappy quality radios - they got the best response from supposed audiophiles when they applied a low-pass filter below 9KHz before it went out over the air waves.