BTSC Stereo Audio for Analog TV

In my last column, I introduced the BTSC sound standard for analog NTSC television and the AC-3 sound standard for digital ATSC television. This time, I'll look a little bit deeper into the BTSC system and explore some important consumer implementations and issues.

Until the mid-1980s, television audio was nice and simple - it was mono. Audio sent to a television transmitter was used to frequency-modulate the aural carrier located 4.5 MHz above the visual carrier. In 1984, the Broadcast Television Systems Committee, a subcommittee of the Electronic Industries Association (EIA), evaluated various systems for delivering stereo audio over NTSC television channels.

The committee recommended one system - now commonly known as the BTSC system - to the FCC, which adopted it for use in the United States. In addition to the U.S., the BTSC system is used by Canada, Taiwan, Argentina, Brazil and others.

Similar to the FM stereo system used in radio, the BTSC system uses a matrix to create sum and difference versions of the two audio channels, rather than transmitting the Left and Right channels discretely. The channels are added and subtracted to produce two new signals, L+R and L-R.

The L+R is used to frequency-modulate the 4.5 MHz aural carrier: the L-R signal is used to modulate a subcarrier. The L-R signal is "companded" (COMPressed during transmission, expANDED after reception) using a system developed by dbx (THAT Corp., which was formed by senior engineers and managers of dbx, has been the licensor of dbx-TV Noise Reduction since 1989) to provide noise performance similar to that of the main carrier.

As can be seen in Fig. 1, this matrix scheme importantly allows BTSC-capable receivers to produce stereo audio, and also protects all viewers with mono television sets, which - in 1984 - was just about every single one. The mono sets will receive the main carrier just like normal and - because it has been modulated with L+R audio - the viewer will hear all the audio from the Left and Right channels.

(click thumbnail)Figure 1 - The BTSC system maintains compatibility with mono receivers by carrying the audio as sum and difference (L+R and L-R) pairs. The matrix sections have been separated from the rest of the processing for clarity.
The system works great, assuming one big thing: the audio must be in-phase between the two channels. If the 6 o'clock mono news show is sent to the Left channel, then phase-flipped 180 degrees and sent to the Right channel, stereo viewers will hear audio - but it will be a little strange and have sort of a hollow sound. Unfortunately, mono listeners will hear nothing. Again - because the mono receiver reproduces only L+R audio - if one channel is out of phase with the other, the channels will cancel when added together.

You might be thinking that it is pretty hard to make such a basic mistake. Well, mistakes happen, and evidently often enough that a few companies made a tidy business building audio phase correction boxes.

LITTLE BOXES

These boxes would look at the relative amplitude of the L+R and L-R signals. If L+R=0 percent and L-R=100 percent, then the Left and Right channels must be out of phase. The box would then invert the phase of the Left or Right channels, which would bring things back to L+R=100 percent and L-R=0 percent - and all was good.

Things got sticky when networks began distributing stereo audio. There were now instances where the L+R and L-R did not equal 100 percent and 0 percent, respectively. Some of these phase-corrector boxes would end up making bad decisions as they were set to be more sensitive to differences in amplitude between the L+R and L-R signals.

The situation got even more difficult when it was discovered that Surround Sound could be carried by the BTSC system. Although there were a couple of systems from other companies, they have mostly disappeared. I am most familiar with the Dolby Surround system and - as it seems to be the last matrix encode/decode system standing - it is the system I will briefly describe. For readers desiring additional detailed information about the system, there are a number of good papers available at www.dolby.com/tech/; I recommend one in particular, called "Dolby Surround Pro Logic Decoder Principles of Operation."

Dolby Surround encodes four channels of audio into a mono, stereo and - therefore - BTSC-compatible, two-channel signal; after decoding, it produces four channels of audio. Without getting into too much detail, the system uses amplitude and phase techniques to combine the Center and Surround channels with the Left and Right channels to produce the surround-encoded signal called Lt (Left total) Rt (Right total). The Lt Rt signal is basically a stereo signal containing audio from four different inputs. When summed to mono, the left, center and right audio is still fully intact - however, the surround information nicely cancels out. I say nicely because, in many cases, the surround information contains sounds like crowd noise - and if it did not cancel when summed to mono, it could potentially overpower announcers or other important audio.

IT'S A TOUGH JOB

The job of the surround decoder is to take the LtRt and reproduce Left, Center, Right and Surround. In very basic terms, it works as follows: If the audio is all Lt, then the decoder sends the signal to the Left output. If it is all Rt, the audio goes to the right output. If Lt and Rt are equal and in-phase, the audio is sent to the Center output. If Lt and Rt are equal but out-of-phase, the audio is sent to the Surround output.

Fig. 2 shows Left, Center, Right and mono Surround audio encoded into LtRt signals and applied to the BTSC encoder. After reception and BTSC decoding, a Pro Logic decoder turns LtRt back into Left, Center, Right and mono Surround. Because LtRt sounds just like conventional stereo, consumers listening to stereo televisions will hear stereo audio. Because the surround-encoded signal is also mono-compatible, receivers without BTSC decoders will reproduce Lt+Rt (mono) audio.

(click thumbnail)Figure 2 - Surround Sound delivered via the BTSC system
Back to the sticky situation with the phase corrector, which is now faced with not only signals that are stereo and contain some L-R information, but also surround-encoded audio that purposefully contains lots of L-R information from time to time. In fact, it can be so much larger than the L+R signal that a phase corrector will think that the audio is out of phase and try to correct it.

Unfortunately, here is what happens when the audio is "corrected." The bases are loaded, the batter hits a zinger to left field and the crowd goes wild. The announcer is giving the play-by-play description as the action unfolds. At home, you are listening to the game on your surround system in the living room while your spouse listens to the small mono set in the kitchen. As the crowd goes wild, the phase corrector at the transmitter sees the large L-R and decides that something is wrong and flips the phase of one of the stereo channels (in this case Lt or Rt).

Your decoder in the living room now sees the announcer out of phase and sends the audio to the surrounds, while the crowd noise is now in phase and is sent to the Center channel. Your wife has it a little better - she is now hearing all crowd and no announcer! Beware of phase correctors as they may correct your audio incorrectly.

There is one more "gotcha" that has cropped up in the last few years. Many new stereo television sets contain audio circuitry that claims to enhance the sound and give more of a "Surround Sound" from just two speakers. In many cases they are simply manipulating the L-R audio to increase the apparent separation of the audio.

The results may be acceptable for some stereo programs, but for surround-encoded programs, the results can be objectionable. Quite a few viewers of the 2000 Sydney Olympics called during the Dolby Surround-encoded opening ceremonies to report that they could not hear the commentators. It turns out that these viewers were listening on stereo televisions and had the "surround boost" feature turned on full; when they disabled it, the audio was fine.

Next time, we will delve into the world of audio for ATSC digital television. As you might recall, the ATSC defined the audio portion of digital television as using the AC-3 system. Like its analog predecessor, AC-3 - or Dolby Digital, as it is more commonly known - has features that provide compatibility for systems that have less speakers than there are audio channels, but it does not matrix the audio to accomplish this. It also provides the ability to reproduce more channels than are being sent. Stay tuned!