In short, NTT has been developing a bunch of cool stuff they never told us about...
This includes:

-3.5G, "HSDPA" technology

-Mouth recognition technology. Not voice recognition, but *mouth* recognition. You don't even have to make a sound.

-3D sound technology

-Receptionless and reception-proof room

-Echoless room

NTT DoCoMo opened the "NTT DoCoMo R&D Center," in Kanagawa Prefecture's Yokosuka City o­n March 2nd to members of the press. The contents included what DoCoMo calls "3.5G," a next-generation communication system, as well as next-generation interfaces, and a heap of other things that were previously not presented to the public. Through this R&D Center pass 800 DoCoMo employees, and including handset makers and other industry related workers, nearly 2,500 people enter and exit.

HSDPA Experiment Facility

HSDPA is a wireless transmission standard to see productization sometime next year, and is being called "3.5G" by DoCoMo. It operates o­n a W-CDMA base, and has a maximum downstream of 14Mbps, though it averages about 2-3Mbps. o­ne of the merits of this system is that existing FOMA base system equipment can be upgraded, by a simple board change and software update, making it a relatively low cost solution.

Another merit of the HSDPA system is the high rate of correspondence. In o­ne second, the cellular unit will communicate with the base station 500 times, making calculations for better reception and such. This means that a better network balance will be achieved across multiple cellular units. By today's standards, HSDPA should have about 3 to 4 times the reception utilization efficiency as existing FOMA standards.

Part of the demonstration process consisted of connecting a PC to the HSDPA trasmission terminal. Using a phony reception system, the PC was put further from the trasmission terminal, and it sustained a throughput of about 2Mbps. When the PC was put right next to the trasmission terminal and the phony reception was turned off, the average throughput was close to 14Mbps. This technology will allow for higher, more efficient and stable data transfer to multiple cellular units - the "transmission diversity effect" was turned o­n and the throughput rose all across the board.

Another demonstration took place where a PC was connected via a FOMA (384kbps) unit to the internet, and also an HSDPA unit. When compared with a Flash animation o­n the FOMA 900i series web site, it took the FOMA unit about 8 seconds to load the animation, but the HSDPA unit took around 3 seconds to load the animation. The demo throughput, however, was limited by some TCP/IP limitations, and the bandwith was 1Mbps.

The final demo was a streaming video using Windows Media Player. It played nice and smooth, with no chop, and throughput sat at around 2-3Mbps.

Mouth Recognition Technology

The next item presented at the R&D Center was the "Kuchi Paku" (Mouth open) system which recognizes syllables by simple mouth movement, and converts them to characters or voice. Sensors are attached to the thumb, index finger, and middle finger, then when held up to the chin, cheek, and throat, will sense the movement of the muscles surrounding the mouth. It currently recognizes the Japanese "a i u e o" vowels with a 90%+ success rate. The technology is still being developed, and consonant recognition and foreign language applications will be researched.

The experimental unit of this consisted of something actually used in the medical field. Obviously, this technology is still in its very early stages, and certainly not ready for market. However, when you think about it, being able to communicate without sound would rapidly expand in the cellular phone realm.

3D sound

Continuing o­n the tour was a very interesting technology called "3D Space Sound Playback Technology." This allows sound to come from a certain angle and distance, with a fixed location, in a room.

Explaining this may get a bit complicated. First of all, there is a headset with a sensor o­n it connected to a PDA. The sensor o­n the headset is for location purposes, and interfaces with another sensor o­n the top of the room. A speaker that does not make any sound is placed in the test room, and the test subject feels as though there is sound coming from it. The entire process is done by the sensor o­n the ceiling and a computer - the sensor o­n the ceiling gets the location data from the sensor o­n top of the headset of the user, and sends it to the computer. The computer computes this data in regards to the sound that is supposedly coming from the speaker, and sends it back to the headset. The result? You think that sound is coming from the speaker, when in reality it's not.
Of course, all of this is done in real time - the virtual "position" of the speaker can even be changed.

The primary applications for this are, of course, for phones. NTT wants to make it so that when you are talking o­n the phone, it seems as though the person is really there. When you are using TV phone, they want to make it seem as though the sound is really coming from their mouth. Again, this technology is still in early development, and there are currently no plans for a physical product or even practical implementation.

Receptionless Room

Continuing, a special room was presented that is used o­nly for experiments. First was a "reception quiet room," which has the effect of cancelling all electromagnetic waves. This means two ways as well - not o­nly do no waves get in, but none get out. Foam urethane-carbon pyramids, line the walls, floor, and ceiling of this 20m wide, 13m tall room.

Since this is the same technology used by stealth fighters, no waves can escape the room.

Echoless Room

Finally, there was a mysterious room terribly different from normal everyday life. This room is used to collect sound data from different computer/cellular terminals, and in order to collect correct data, it had to be made completely echo-proof. The floor has a wire-mesh lattice o­n it, and both the floor and walls are lined with class fiber. Since sound does not bounce at all, even if a person talks in the middle of the room, it sounds as though they are whispering. When the talking stops, there is absolutely no reverberation, and the room goes completely silent.

Inspired by:
http://k-tai.impress.co.jp/cda/article/news_toppage/17870.html
This link has plenty of photographs of inside the center, and the technologies described above. So please take a look if you're interested.

NTT DoCoMo R&D Center:
http://www.nttdocomo.co.jp/corporate/rd/muse/