Omnia 6 Use and Installation Guide – Version 1.00a
47
47
Finally, The New Frontier” on the Omnia Tech Info web page at www.omniaaudio.com .
When using the AES/EBU outputs to drive a digital STL or exciter, your Omnia-6fm’s sampling rates must
be set to match the lowest internal operating sampling rate used in any device which follows the Omnia-6
in the audio chain. You will find these options in the Output Menu; available selections are 32kHz,
44.1kHz, 48kHz, and 96kHz. Failure to select the proper output sampling rate will result in generation of
overshoots.
Pre-Emphasis: Where To Insert?
Most digital exciters can generate pre-emphasis within their DSP stereo generator. This is unnecessary
when using an Omnia-6fm, which provides not only the proper pre-emphasis for any region in the world,
but maintains excellent peak control of the pre-emphasized signal so that no overshoots occur. Using pre-
emphasis in the digital exciter instead of that built into your Omnia-6 will result in degradation of
performance. Here is some background in the context of operating in a complete digital environment:
• Most exciter manufactures recommend the use of their own pre-emphasis network. If this is done,
incoming audio data from the processor must be de-emphasized before it reaches the exciter. Why
perform the emphasis function twice? Also, and more importantly, if the pre-emphasis phase delay
is not exactly the same as the previous de-emphasis in the processor, overshoots will occur.
Therefore, it is far more advantageous to employ pre-emphasis only once—in the Omnia-6fm—and
set the exciter to accept a flat input, which greatly reduces the chance of overshoots being
generated. The Omnia6-fm should always
be the standard!
• We understand why the exciter manufacturers provide these pre-emphasis options; it has to do with
digital codec STLs. Research has shown that pre-emphasized audio routed through a codec is not
very friendly sounding. Hence the de-emphasis in the processor, and second pre-emphasis in the
exciter. This is, in our opinion, a poor solution to the problem. If you must use a codec STL, locate
the processor at the transmitter. Doing so completely eliminates any pre-emphasis/codec
abnormalities in the processed audio, and also allows the processor to have the best possible
electrical coupling to the exciter (be it analog or digital).
Hopefully this discussion has served to illuminate some of the issues involved in creating an all-digital
transmission path. Should you experience any problems with overshoot or added distortion after this system
is in place, you should look to the function of the digital exciter and its stereo encoder. As a reference, you
can easily make signal comparisons with the analog outputs from the Omnia-6fm system, especially the
composite output. Viewing the signal on an oscilloscope should reveal that the output is well controlled,
with no overshoots, providing a clean baseband spectrum.
