brought to you by Lloyd's satellite constellations, which profiles Iridium
I took a call from an Iridium satellite phone being used in an urban environment in the UK this morning, and compared with terrestrial calls from the same person; the result was pretty much as described in the Washington Post article of 23 November 1998, which grumbled about dropped calls and voice quality.
We had three or four minutes of clear conversation, mostly about the voice quality. The caller's voice sounded far more nasal than it did via land lines (apparently I sounded deeper). There was slight breaking/mangling noticeable in long words, a constant (synthesised to assure people the call was still in place, I think) background hiss when the caller wasn't speaking (comfort noise), and slight silences between that 'background' hiss and the clipped starts of words. Usable, but not likeable.
(I'm a pretty lousy test subject for judging call quality anyway; I only use telephones when I have to, rarely speak over cellular, gave out lots of ones on a 1-5 scale before they stopped asking me to listen to the codecs developed here, and listen to Radio 4 on longwave because it sounds better. Given that, the call quality was better than I was expecting...)
This period was followed by three minutes that were completely useless; although we were both talking continuously, I was hearing around one mangled syllable every ten seconds or so, while the caller was apparently hearing complete silence. Eventually this came to an end and I got to hear an Australian voice say 'your call could not be completed at this time', followed by what I think was the Spanish equivalent, followed by the same message in a bunch of other languages I couldn't identify. Quite odd to get that particular message, considering that it wasn't me who made the call.
The caller described other calls made as similar; three minutes or so of useful conversation, then it falls off and drops. According to the Washington Post article:
Iridium LLC chief executive Edward Staiano [..] attributed that to a hole in the network caused by the 66th satellite, which was launched Nov. 8 but is not yet operational.
Either that 66th satellite is getting about an awful lot, or there may be problems with satellite handover or routing changes. According to the statistics in the article, only 11-12% of calls are dropped at present. That 1-in-10 chance doesn't gel with the caller's experience - but the calls were made from an urban environment.