Jeffrey Mattox wrote:
> Whenever I have a conversation with a modern cell phone user, it's
> digital-ness is highly obvious because the sound quality is so bad.
For what it's worth, it didn't used to be this way. Back when CDMA was
just getting started in the US, Qualcomm had these QCP-2700 digital
cell phones that really did sound a LOT better than even a landline
phone. of course, that was back when vocoders were set that 13kbps,
which with compression, was ample bandwidth for human speech.
> Well, it not very advanced from my perspective -- it's a step
> backward. And to say that digital "sound quality .. [is] superior" is
> a lie. Because of the low sample rate (is it 8 KHz?), it's not
> possible to reconstruct perfect speech.
The problem here is that when digital started penetrating the market,
cellular rates started coming down and demand for cellular phones
increased. This taxed even the most well-off digital networks, and so
to accomodate everyone, cell carriers began demanding vocoders that
used less bandwidth. 13kbps vocoders gave way to 8kbps, and then even
further to the variable rate vocoders that are in use today (which can
step down as far as 4kbps depending on what sounds are being
transmitted).
> Is this situation ever likely to be improved?
Probably not. In fact, the trend is for MORE compression, using less
bandwidth. Granted the new vocoders being developed are using more
ingenious methods to make this happen, and they promise to improve
sound quality while still reducing their network overhead. We'll see
if that promise is fulfilled.
> Why isn't it possible to sample at a higher rate, compress the
> result in the phone for transmission, and then decompress at the
> base (and do the same thing for the reverse channel)?
Compression IS happening right now. The problem is that there is a
huge difference between lossy compression (the type which is in use
now, which discards certain elements of the audio that the codec
thinks the human ear won't notice is missing much) and type of
compression that you're probably expecting, lossless compression (the
type which does not discard any audio information during compression).
Your modern cell phone has to be able to compress audio
instantaneously. It wouldn't do anyone much good if the phone
captured several seconds of audio, spent a few more seconds
compression that audio using a lossless algorithm, and then sending
it down the line where a further delay is added while the receiving
end decompresses and replays the signal. At present, cell phones
just don't have powerful enough processors to do lossless
compression in real time.
The compromise then is to use a very lossy codec, one that discards
portions of the sound that is presented to it, in the hopes that the
human ear ont he receiving end will not notice that those components
are missing. Although the quality suffers, it does allow a modern
cell phone to quickly compress the audip it needs to transmit in
almost-real time. You'll note, there is still a delay of about a
quarter to a half a second between the time the person actually says
something into a digital cell phone, and the time in which the sound
arrives at the other end, and you'll also note that this delay just
isnt't present when one uses an analog cellular or wireline phone.
> Are all digital cellular phones the same?
No. Different digital cell phones use different codecs. ALL of them
use compression, and NONE of them truly give toll quality speech, but
the differing codecs DO sound different, and it tends to be a matter
of personal preference as to which sounds better. I know some people
who prefer GSM phones, as they feel the sound coming from GSM is more
natural to them. I, on the other hand, find GSM to be very harsh
sounding to my ears ... very raspy, very shrill, and actually rather
painful to listen to after an extended period. I prefer my CDMA
phone, as the sound from CDMA (in my opinion, yours may differ) is
much easier on my ears.
> Aren't many people bothered by the horrible sound quality?
> Jeff
E-mail fudged to thwart spammers.
Transpose the c's and a's in my e-mail address to reply.
Isaiah Beard (sacredpoet@sacredpoet.com)