In last month’s episode I explored the genesis of the first song uttered by a computer, Daisy Bell, and how that song ended up in 2001: A Space Odyssey. In this installment on the history of speech synthesis I’ll track the use of the vocoder in popular music on up to its implementation into the DMR radios that are currently a big buzz in the ham community.
In 1968 synth wizard Robert Moog built the first solid state vocoder. Two years later Moog built another musical vocoder, working with Wendy Carlos. This was a ten-band device inspired by Homer Dudley’s original designs. The carrier signal came from a Moog modular synthesizer. The modulator was the input from the microphone. The brilliant application of this instrument made its debut appearance in Stanley Kubrick’s film A Clockwork Orange, where the vocoder sang the vocal part from the fourth movement of Beethoven’s Ninth Symphony, the section titled “March from a Clockwork Orange” on the soundtrack. It’s something I could sit down and listen to on repeat over and over while enjoying a fine glass of moloko velocet. This was the first recording made with a vocoder and I find it interesting that the two earliest uses of speech synthesis for music ended up in films made by Kubrick. The song “Timesteps”, an original piece written by Wendy, is also features on the soundtrack. She had originally intended to include it as a mere introduction to the vocoder for those who might consider themselves “timid listeners” but Kubrick surprised Wendy by its inclusion in his dystopian masterpiece.
Coming down the road in 1974 was the classic album Autobahn by the German krautrockers Kraftwerk. This was the first commercial success for the power-station of a group. Their previous three albums had been highly experimental, though well worth an evening of listening. Kraftwerk’s contribution in the popularization of electronic music remains huge. Besides using commercial gear such as a Minimoog, the ARP Odyssey, and EMS Synthi AKS, Kraftwerk were dedicated homebrewers of their own instruments. Listening to the album now I can imagine the band soldering something together in the back of a Volkswagen Westfalia as they cruise down the highway at 120 km/h on to their next gig.
Three years later in 1977 Electric Light Orchestra released the album Out of the Blue, much to the delight of discerning listeners everywhere. There is nothing quite like the music of ELO to lift me up out of the melancholy I often find myself in during the middle of winter when spring seems far away. “Mr. Blue Sky” and “Sweet Talking Woman” are songs that toggle the happy switches in my brain. When I hear them things brighten up. This is in no small part to the judicious use of the vocoder. ELO was in love with the vocoder and it can be found littered across their recordings. (As a bit of a phone phreak another favorite cut is “Telephone Line”.)
During the 1980’s the vocoder started being used in the early hip-hop and rap groups. Dave Tompkins, author of How to Wreck a Nice Beach: The Vocoder from WWII to Hip-Hop notes the echo of history in the vocoders use alongside two turntables for the SIGSALY program and how DJs use two turntables to mix and scratch phat beats while a rap MC will drop lyrics over top of the sounds being produced by the vinyl, sometimes processing those vocals through the vocoder. The use of the vocoder continues to present times on hip-hop and jazz fusion albums such as Black Radio (1 & 2) from Robert Glasper Experiment.
While the vocoder was enjoying great success in the entertainment industry, its use in telecommunications was still ticking away, though a bit quieter, in the background. Since 1970’s most of the tech in this area has focused on linear-predictive coding (LPC). It is a tool used for representing the spectral envelope of a digital signal of speech in compressed form, using the information from a linear predictive model and is a powerful speech analysis technique. When it came out the NSA were among the first to get their paws on it because LPC can be used for secure wireless with a digitized and encrypted voice sent over a narrow channel. The early example of this is Navajo I, a telephone built into a briefcase to be used by government agents. About 110 of these were produced in the early ’80s. Several other vocoder systems are used by the NSA for encryption (that we are allowed to know about).
Phone companies like to use LPC for speech compression because it encodes accurate speech at a low bit rate, saving them bandwidth. This had been Homer Dudley’s original intention with his first vocoding experiments back in the 1930’s. Now LPC has become the GSM standard protocol for cellular networks. GSM uses a variety of voice codecs that implement the technology to jam 3.1 kHz of audio into 6.5 and 13 kbit/s of transmission. Which is why to my ear, smart phones, for all the cool things they can do with data, apps and GPS, will never sound as good with voice as an old school toll call on copper wires. LPC is also used in VoIP.
LPC has also been used in musical vocoding. Paul Lansky created the computer music piece notjustmoreidlechatter using LPC. A 10th order derivative of LPC was used in the popular 1980s Speak & Spell educational toy. These became popular to hack by experimental musicians in a process known as circuit bending, where the toy is taken apart and the connections re-soldered to make sounds not originally intended by the manufactures. This technique was pioneered by Cincinnati maker and musician Q. Reed Ghazala into a high art form. Reed’s experimental instruments have been built for Tom Waits, Peter Gabriel, King Crimson’s Pat Mastalotto, Faust, Chris Cutler, Towa Tei, Yann Tomita, Blur and many other interesting musicians. And not so interesting ones (to me) such as Madonna. A future edition of The Music of Radio will cover his work in detail, but a lot can be found on his website anti-theory.net.
Finally vocoders are utilized in the DMR radios that are currently gaining popularity among hams around the world. In Ohio the regional ARES groups are being encouraged to utilize this mode as another tool in the box. DMR is an open digital mobile radio standard. DMR, along with P25 phase II and NXDN are the main competitor technologies in achieving 6.25 kHz equivalent bandwidth using the proprietary AMBE+2 vocoder. This vocoder type uses multi-band excitation to do it’s speech coding. Besides it’s use in DMR the AMBE+2 is also used in D-Star, Iridium satellite telephone systems, and OpenSky trunked radio systems.
From what I’ve heard I didn’t really care for the audio quality of DMR, as on cell phones. My ears would rather dig through the mud of the HF bands than listen to the way speech is compressed in these modes. I think the vocoder is better suited to musical studios where it can be used for aesthetic effects. However with the push to use these in ARES, and needing something to play with at OH-KY-IN’s digital night on the fourth Tuesday of the month, I do plan on taking the plunge into DMR. And when I do I will know that every time I have a QSO using the DMR platform I will be taking part in a legacy starting with Homer Dudley’s insights into human vocal system as a carrier wave for speech. A legacy that stretches across the fields of telecommunication, cryptology and popular music.
Chip Talk: Projects in Speech Synthesis by David Prochnow, Tab Books, 1987.
…and some other research on the interwebs.
This piece was originally published in the April 2017 issue of the Q-Fiver.