Multichannel Television Sound

Initial work on design and testing of a stereophonic audio system began in 1975 when Telesonics approached Chicago public television station WTTW. WTTW was producing a music show titled Soundstage at that time, and was simulcasting the stereo audio mix on local FM stations. Telesonics offered a way to send the same stereo audio over the existing television signals, thereby removing the need for the FM simulcast.

Telesonics and WTTW formed a working relationship and began developing the system which was similar to FM stereo modulation. Twelve WTTW studio and transmitter engineers added the needed broadcast experience to the relationship. The Telesonics system was tested and refined using the WTTW transmitter facilities on the Sears Tower.

In 1979, WTTW had installed a stereo Grass Valley master control switcher and had added a second audio channel to the microwave STL (Studio Transmitter Link). By that time, WTTW engineers had further developed stereo audio on videotape recorders in their plant, using split audio track heads manufactured to their specifications, outboard record electronics, and Dolby noise reduction that allowed Soundstage to be recorded and electronically edited. In addition, an Ampex MM1100, 24-track audio recorder was also used for music production and mixing. PBS member stations who wished to deliver Soundstage in stereo were provided with a four-track (left, right, vertical drive, and time code) audiotape that could be synced with the video machines in those cities.

During the FCC approval process, several manufacturers applied to the FCC for consideration. Most notably the Electronic Industries Alliance (EIA) and Japanese EIA asked to be included in order to represent their members in the testing and specification phases of the approval process. WTTW engineers helped set standards for frequency response, separation, and other uses of the spectrum. They also provided program source material used for the testing and maintained the broadcast chain. A 3M 24-track audio recorder was used to allow the selection of 12 different stereo programs for testing. The Matsushita Quasar TV manufacturing plant and laboratory, just west of Chicago, was used as the source for all testing of the competing systems. Following the testing, several questions were raised about the validity of some of the tests, and a second round of testing began.

WTTW installed a Broadcast Electronics prototype stereo modulator in October 1983 and began full-time broadcasting in stereo at that time using the Telesonics system prior to Federal Communications Commission (FCC) rule-making on the BTSC system. MTS was officially adopted by the FCC on 23 April 1984. Following EIA and FCC recommendations, the BE modulator was modified to meet BTSC specifications, and by August 1984 was in full-time use on WTTW.

Sporadic network transmission of stereo audio began on NBC on July 26, 1984, with The Tonight Show Starring Johnny Carson, although at the time only the network's New York City flagship station, WNBC, had stereo broadcast capability;[1] regular stereo transmission of NBC programs began during early 1985. ABC and CBS followed suit in 1986 and 1987, respectively. FOX was the last network to join in 1987, with the four networks having their entire prime-time schedules in stereo by late 1994 (The WB and UPN launched the following season with their entire line-ups in stereo). One of the first television receiving systems to include BTSC capability was the RCA Dimensia, released in 1984.[2]

From 1985 to 2000, the networks would display the disclaimer "in stereo (where available)" at the beginning of stereo programming, sometimes using marketing tags such as CBS's "StereoSound" to describe their institution of stereo service.

The original North American television standards provided a significant amount of bandwidth for the audio signal, 0.5 MHz, although the audio signal itself was defined to extend from 50 Hz to 15,000 Hz. This was centered on the audio carrier signal 4.5 MHz above the video signal, and given 25 kHz on either side of the carrier, using only 15 kHz of it.[6]

This meant the lower and upper 0.2475 MHz of the audio channel was unused. Due to the nature of the NTSC color signals added in the 1950s, the upper parts of the color signal pushed into the lower audio sideband. With the audio signal centered within the 0.5 MHz channel, and the lower 0.25 MHz being partially infringed on by leftover video signal, the upper 0.25 MHz was left largely empty.[6]

MTS worked by adding new signals to the free portion of this upper 0.25 MHz allocation. The original audio signal was left alone and broadcast as it always had been. Under MTS, this is the Main Channel. The actual signal in this channel is constructed by adding together the two stereo channels to produce a signal largely identical to the original monoaural signals and can be received on any NTSC television even without stereo circuitry.[6]

A second channel is then added, the Stereo Subchannel, centered at 31.468 kHz and extending 15 kHz on either side. This left a small gap between the Main and Stereo signals, known as the Pilot, at the 15.734 kHz. This signal is also known as "H", or "1H", and its frequency is selected to be a harmonic of the video's horizontal scan signal so that it can be accurately recreated from the video signal using a phase locked loop. If there is any signal present at the 1H frequency, the television knows a stereo version of the signal is present.[6]

The Stereo Subchannel consisted of the same two audio signals, L and R, but mixed out of phase to produce the "L–R" signal, or "difference". This signal is sent at a higher amplitude and encoded using dbx compression to reduce high-frequency noise. To lower total average power, the carrier is not sent (which means there is no always-on signal at that frequency). On reception, the receiver uses the video signal to create the Pilot, and then examines that frequency to see if there is any signal present. If there is, the difference signal is extracted by filtering out the signal between 1H and 3H into a separate channel, and the carrier is re-created by adding 2H to this. This signal is then decompressed from its dbx format, and then fed, along with the original Main, to a stereo decoder.[6] FM stereo radio works in the same fashion, differing mainly in that the equivalent to the H signal is 19 kHz, not 15.734.

SAP, if present, is centered on 5H, but encodes a monaural signal at lower audio fidelity and amplitude.[6] The PRO signal is likewise encoded at 7H. A signal using all four channels extends only to about half of the available bandwidth in the original audio upper sideband.

In ideal circumstances, MTS Stereo is about 1.5 dB better in performance than standard VHF FM stereo. As with VHF FM stereo, a certain amount of crosstalk is usually encountered, limiting stereo separation. Stereo information is dbx-encoded to increase the signal-to-noise ratio (SNR) at low levels, to aid in noise reduction: typical SNR is better than −50 dB.

The original specifications called for a brick wall elliptical filter in each of the audio channels prior to encoding. The cutoff frequency of this filter was 15 kHz to prevent any horizontal sync crosstalk from being encoded. Manufacturers of modulators, however, used lower cutoff frequencies as they saw fit. Typically, they chose 14 kHz although some used filters as low as 12.5 kHz. The elliptical filter was chosen for having the greatest bandwidth with the lowest phase distortion at the cutoff frequency. The filter used during EIA testing had a characteristic that was −60dB at 15.5 kHz. As transformer audio coupling was common at that time, the lower frequency limit was set to 50 Hz although modulators without transformer inputs were flat down to at least 20 Hz.

Typical separation was better than 35 dB. However, level matching was essential to achieve this specification. Left and Right audio levels needed to be matched within less than 0.1 dB to achieve the separation stated.

Maintaining the phase stability of the horizontal sync is essential to good audio in the decode process. During transmission, the phase of the horizontal sync could vary with picture and chroma levels. ICPM (Incidental Carrier Phase Modulation), a measure of transmitted phase stability, needs to be less than 4.5% for best audio sub channel decoding. This was more of a problem with UHF transmitters of the day. Multi-cavity klystron amplifiers of that time typically had an ICPM above 7%.

Because of the use of dbx companding, every TV device that decoded MTS originally required the payment of royalties—first to dbx, Inc., then to THAT Corporation which was spun off from dbx in 1989 and acquired its MTS patents in 1994; however, those patents expired worldwide in 2004.[7] Though THAT now owns some patents related to digital implementations of MTS, a letter from THAT to the National Telecommunications and Information Administration in 2007 confirms that no license is required from THAT for all analog and some digital implementations of MTS.[4]

MTS has been adopted in:

• United States for NTSC
• Canada for NTSC
• Mexico for NTSC
• Chile for NTSC
• Japan for NTSC-J
• Brazil for PAL-M
• Taiwan for NTSC (Full switchover of FTA terrestrial to DVB-T completed on 30 June 2012, and for pay TV cable to DVB-C completed during 2014–2017. MTS became historical from these dates.)
• Argentina for PAL-N
• Philippines for NTSC

• NICAM
• Zweikanalton
• EIAJ MTS