FINAL REPORT
NPR Program Automation Advisory Committee

System comment was varied and diverse. The range of responses ran the gamut from proposing the simplest possibility (one single tone) to proposals for elegant digital cueing signals traveling as a datastream (not part of the actual program audio).

In an attempt to make some sense of the apparent disparate needs of the system, it was necessary to return to the original resolution passed by the NPR Board:

The resolution specifically calls upon the PAAC to address the issue of floating internal breaks (cutaways) and programs having nonstandard lengths requiring an "end of message" tone (EOM).

Accordingly, the committee chair proposed that the committee focus on the immediate resolution (a "quick fix"), rather than a more elegant, long-form solution -- which committee members believe will become increasingly necessarily as more as more stations incorporate more sophisticated program automation equipment into their broadcast systems.

There are presently two systems in use today at public radio stations that provide cue tones. Minnesota Public Radio has developed a cue system for use with its Classical 24 Music Service (referred to as C-24). This system uses "super-audible" tones (tones higher in frequency that the program audio). It provides for station cutaways in time units of 20, 60, 90, and 120 seconds. This system allows stations to break from network audio for the selected time period; the system also will start a local audio device (such as a cart machine). According to public radio international, 115 broadcast stations may be using C-24 decoders.

A C-24 system was sent to NPR Engineering by Minnesota Public Radio for evaluation. Test results follow:

In bench testing the C-24 system, both the Encoder and Decoder measured very good.

From the standpoint of passing audio, neither the encoder or decoder should add any significant artifacts to the audio. Both units were constructed to very good engineering standards using very high quality components and manufacturing techniques.

Service testing provided the following results.

a. The system is very robust and operated successfully to a level of -40 below normal program operating level of -4 dbu.

b. Audible observations were:

1. The tone bursts must be operated in the clear (no other audio present) (see note 1) - they are audible in any non over-the-air situation.

2. When the tone burst is applied to a commonly used FM Broadcast processing system (Orban 8100) it produces audible artifacts well within the audio passband. The listener hears clicking sounds several times during the tone burst.

3. Any station incorporating a digital recording system in their broadcast system will have to operate the system at a sampling rate sufficient to allow the reliable recording and reproduction of the data burst centered at 18.6KHz. This would require a sampling rate of at least 44.1Khz, requiring consumption of more disk space for the recording process.

Note 1.The C-24 system has to be operated in the clear (with no other audio present). Should this procedure fail to be followed, the tone burst will not pass through the MUSICAM encoding of the distribution system. In the presence of other audio, the MUSICAM looks at the frequency domain of the C-24 tone burst, apparently determines it is unnecessary information, and eliminates it from the data stream.

Note 2. Minnesota Public Radio/PRI maintain that the C-24 encoder cue tone could be actuated during a musical selection with a reduction to -16 dbm of the audio level in the right channel.

(Ed note: Minnesota Public Radio's General Home Page is http://www.mpr.org/)

This system is capable of three control functions. These functions are made possible by encoding the 25Hz tone in the following manner.

Function 1, Left Channel 25Hz tone

Function 2, Right Channel 25Hz tone

Function 3, Left and Right channel 25Hz tone

Each function provides separate relay isolated relay contact closures to the external equipment. There are three (3) separate relays.

Bench testing the Zephyrus 25Hz system revealed that the encoder and decoder measured almost identically, with two exceptions in the decoder. The THD of the decoder was slightly higher than the encoder (.1% vs. .09%) and the cross talk in the decoder was higher, measuring Left into Right at 58db (vs. 80 dB for the encoder). We did not consider this a problem because the decoder does not have to be installed in the program stream to function. The decoder only needs to be in the program stream to provide hi-pass filtering for removal of the 25Hz tone in program audio. The decoder can operate as a side-chain.

Neither the encoder nor decoder should add any significant artifacts to the audio. Both units incorporate hi-pass filters to remove information in the 25Hz region for this type of system to function reliably. The filters exhibit a flat frequency response down to about 45Hz and then dropping to -25 db at 25Hz. Both Zephyrus units were constructed to good engineering standards using high quality components and manufacturing techniques.

Service testing of the Zephyrus 25Hz system demonstrated reliable signaling at its normal operating level of -5 db below the 0 db (+4 dbu) program operating level. The decoder (as delivered from the manufacturer) operated reliably to about -12 db with the 25Hz tone in the clear and in the presence of program material.

The on air artifacts of the system are:

1. Hearing the 25Hz tone if you do not use the decoder in the program path.

2. The possible loss of some very low frequency elements of the program content due to the hi-pass filtering.

3. The 25Hz system functioned without failure through the MUSICAM encoding of the distribution system, both in the clear and in the presence of other program audio.

(Ed note: Zephyrus home page can be found at: http://www.big-z.com/)

Dual Sensor System, This system is capable of just 2 control functions. These functions are made possible by encoding the 25Hz tone in the following manner.

Function 1, Left Channel 25Hz tone

Function 2, Right Channel 25Hz tone

In the event of a 25Hz signal is present on both channels, both left and right relays in the decoder will operate.

Each function provides separate relay isolated contact closures to the external equipment. There are two (2) separate relays.

Note 3. Conex makes a "DTA-1" Dual Tone Adapter Board that provides three (3) outputs. Conex provided NPR with the two (2) output version for testing.

Bench Testing the Conex 25Hz system the encoder and decoder yielded the following results. THD, IMD and noise measurements were virtually the same, acceptable. Frequency response of the encoder was flat between about 60Hz and 15KHz. Due to the hi- pass filter the low end response fell off rapidly below 60Hz as expected. The high end was down 3 db at 60 Khz where as the other units tested were flat to beyond 100KHz. The Maximum Output Level (MOL) of the encoder was measured at +24 dbu. The decoder was measured at +20 dbu. Both units output level capability are below that expected of modern equipment. We next measured the cross talk of the units. The encoder was marginal at about 60 db L>R and 65 db R>L. The encoder's crosstalk performance is not acceptable. At 100Hz it measures about 60 db in both channels and 35 db at 10KHz. Construction style of the encoder and decoder was vastly different. The encoder is built as a table top device and would be expected to involve increased maintenance due to normal movements of a table top device as well as less than optimum quality of the front panel switches. The decoder is a rackmount device with modular construction.

Due to the poor crosstalk performance of the encoder and only dual channel function of the decoder we did not test the Conex Systems unit in the broadcast chain and thus, did not assess audible artifacts. We would expect degradation of low end to be slightly more noticeable with the Conex equipment as delivered because the filters have a higher rolloff point (60Hz). Bench testing of the Conex Systems demonstrated that the functions operate as advertised.

(Ed note: Conex Electro-Systems, Inc home page is at http://www.pacificrim.net/~conex/)

1. The 25 Hz system can be used for an end of message tone for programs of irregular lengths. The C-24 system provides cutaways for a variety of time durations, but does not specifically provide end of message tones.

2. The 25 Hz system will operate with program audio simultaneously. This means that a cue tone could be actuated during a musical selection without disruption of program audio. Minnesota Public Radio/PRI maintain that the C-24 encoder cue tone could be actuated during a musical selection with a reduction to -16 dbm of the audio level in the right channel.

3. The 25 Hz system is robust; its cue tones can be recorded on reel-to-reel, RDAT, and digital hard disk systems operating at 32 Khz sampling rates. The C-24 system super-audible cue tones can be recorded on reel-to-reel, RDAT, and digital hard disk -- if the sampling rate is 44.1 kHz or greater.

4. The cost of implementation for the 25 Hz system appears to be the least expensive of all options.

5. The 25 Hz cue tone technology is not proprietary. It, therefore, will allow stations the option of selling the equipment at a later date, should this system be replaced by a more sophicaticated cueing system in the future.

The committee then turned its attention to the need for a more sophisticated cueing system -- for the next generation of automation systems. The committee recommends that the NPR Board charge an appropriate committee with the responsibility for developing the requirements for a more sophisticated system, realizing that the specification and implementation process will take several years to become operational -- and could require a significant financial investment.