| Isaiah Beard 2006-01-23, 5:48 pm |
| Quick wrote:
> JK wrote:
>
> What?! What have you been smoking? "Synthesizes the
> human voice"? "and actually only certain languages"?!?!
> What happens if I start talking in one of those other
> languages?
The guy's typing out of his rear-end, I'd ignore him. ;)
The real reason for CDMA's munging of music is that the vocoder
currently is use (EVRC) was designed to compress, and compress again,
all in the name of beating GSM in capacity. All it cares about is a
human voice, which generally has a lot of unique and ever-changing audio
characteristics to it (varying pitch, volume, random pauses, etc.).
Anything repetitive or homogenous (like music or continuous tones) is
perceived as background noise, something that's unnecessary to
conversation, so it's simply filtered out to save bandwidth.
Whether or not that's a bad or good thing is really up to the end user.
Many people don't mind it; others do.
And FWIW, the original QCELP 13k vocoder that was in use long ago on
CDMA actually did a pretty good job of reproducing music. And anyone
who is fortunate enough to still have a working a Qualcomm QCP-2700 or
2750 activated on Verizon or Sprint are probably getting the best call
quality you can get on either CDMA *or* GSM.
However, QCELP wasted bandwidth like crazy, and carriers wanted to cram
more calls into each channel, so it was ditched in favor of what we have
now. And when GSM carriers upgrade to UMTS, the vocoders they use will
probably be just as compression-obsessed.
>
>
> Oh yea... they have to recognize what you are actually
> saying and then synthesize it.... yea, that's the ticket!
Heh, this guy's never echo-tested a GSM call. The delay between CDMA
and GSM is actually about the same.
--
E-mail fudged to thwart spammers.
Transpose the c's and a's in my e-mail address to reply.
|