Digital betacam, DV, DAT, XDCAM, non-linear recorders, DVD, all record digital audio with substantially better specifications than was possible with analogue recorders in years gone past.
With digital recorders level is generally shown on a ‘digital full scale’ meter which counts down from 0dB to -70dB or lower. Recordings can be beautifully clean with a wide dynamic range and a flat frequency response, provided that the level never attempts to pass 0dB (when 0dB is reached terrible clipping happens instantly).
To ensure that clipping doesn’t occur, a level of -18dB (or -20dB in many quarters) has been adopted as reference level on many recording systems. This gives a substantial margin for loud transients to be captured on location without the likelihood of reaching the 0dB maximum (though its not unusual for peaks to come within a fraction of a dB of the maximum, especially if any kind of automatic gain control is used). When a programme is mixed and laid back to tape, peaks will generally be contained so that they don’t extend more than 8 to 12dB above the reference level (different broadcasters have different stipulations).
If your sound studio or edit suite uses digital recorders, but also uses analogue wiring and audio processing, then it’s important that your equipment can cope with the maximum possible level without clipping. By way of example - we have been supporting an Avid edit suite which uses an analogue audio mixer and an active attenuation amp before feeding the unbalanced audio inputs on a Mojo breakout box. We found that some location DV video tapes had extremely high audio levels approaching 0dB (shot by a one-person crew using AGC for the camera mic level), these audio peaks were clipping in the analogue equipment before reaching the mojo, and we had to had to ‘back off’ the gain in certain places to keep within the equipment’s maximum level capabilities.
However arranging lots of headroom is not enough, signal-to-noise ratio also needs to be considered (in this digital age noise levels are more significant than ever). If reference level on your source tape is -18dB, then it’s not possible to receive a signal that’s more than 18dB louder than reference (not possible as the source tape cannot record louder than 0dB). It’s possible that some device in your signal path can handle 26dB above reference without distortion (ie: 8dB of additional headroom that will never be required), however it’s likely that that device will have a higher noise floor than is potentially possible. If it were possible to ‘wind the settings back’ so that the device offered say 19dB of headroom, then the noise level would also be wound back providing 7dB improvement to the signal to noise ratio.
CPR is now the proud owner of a Neutrix audio alignment system which enables clipping levels and signal-to-noise levels to be read with ease. We’ve developed an efficient approach to aligning signal paths and like to document path and specifications in a CAD diagramme. If you’d like assistance with your system then please give us a call.
