The best way to start any discussion on digital levels must be to come to some understanding of a common reference point (e.g. analogue level) and the way in which film recordists work with this level. The best description to emerge at the recent AMPS general meeting was Richard Daniels' 'Window' or what I would prefer to term 'Working Window'. On a Nagra the working window is about 35dB as shown on the Modulometer.
When recording a tone at the head of the tape the operator sets a reference level for this window. If a '0 dB' reference tone is recorded, the working window would be from +10dB to -25dB. Above +10dB the saturation of the tape would act as a limiter; below -25dB the noise from the tape hiss would be unacceptably high. Due to the behaviour of tape, the working window in more severely limited at both high and low frequencies from a 35dB window to 25dB or less.
Up to this point I have not mentioned nanoWebers/metre (nWb/m) or milliMaxwells/millimetre (mMx/mm), which are both measures of the flux density on tape. This has changed over the years as tapes have developed. In the case of analogue this has meant a gradual increase in flux level on tape allowing a gradual widening of the working window. The flux level on tape has changed over time and remains variable from recording engineer to recording engineer, test tape to test tape.
Digital recording has changed this situation. The available working window has widened to at least 50dB and that does not have any limitations at high and low frequencies. The sound recordist is not constrained by his equipment to narrow his working window. He has only an absolute upper limit as set by 0FS (maximum digital level).
Whilst an absolute reference level on digital machines (e.g. -18FS for +4dBm out) may help, they are not a solution. One problem mentioned at the meeting was that it is possible to line up a DAT so that full modulation will clip the pre-amp! Your Nagra replay has a level pot so that you can line up on the reference tone at the head of the tape, not all DAT machines include this. Where necessary an external level control or pads should be provided.
The introduction of digital recording has caused a problem in the transfer bay. The digital tapes received by them can have much wider dynamic range than analogue tapes. Two new types of problem have emerged as follows:
1. The transfer recorder now has to handle the same wide dynamic range as the Nagra in the field. Too often the electronics in the film recorder are not up to this. They do not have Nagra electronics and large peaks that would be naturally compressed by tape compression are clipped by the record electronics.
The solution: Update the electronics and where that is not possible use a limiter to reduce the peaks. I would recommend the same type of limiter as developed by the BBC for use on transmitters and before A-to-D converters, a slow limiter followed by a diode clipper. This has a much better sound than a fast limiter. The limiter side chain should include pre-emphasis to mirror the clip level of the recorder. A suitable limiter is the F601 from Audio Design which I designed some years ago.
2. Where the floor mixer has taken advantage of the wider working window available in the digital format and recorded a wider dynamic range than would have been possible on analogue, the transfer operator may have no possibility of copying this onto the narrower working window of analogue.
Here the solution is more complex. If the final working medium is film then this should not occur since the responsibility for the dynamic range has been passed from the floor mixer to the transfer operator? Where auto-conform is used this responsibility has been passed to the editor or dubbing mixer. The solution must be in good working practice. Good working practice should include pre-testing of the transfer bay by the floor mixer. He should discuss with the transfer operator working levels, tape flux levels and if any compression should be included in the signal path. It is of little use to complain if you have not followed this route.