View Single Post
Old 3rd February 2003
  #8
Lives for gear
 
groundcontrol's Avatar
 

Eduardo, there is no fast and easy answer as to how to relate a PT signal level to an analog equipment level if you only use the meters in PT because they are "peak" reading meters and each and every type of signal has a different "peak-to-average" ratio.

In digital land, the way signal levels have typically been monitored is to relate them to the absolute loudest signal the system is able to produce and pass on to the converters. This is referred to as 0dBFs (Full Scale). That is when your signal hits 0, that's it, there is no getting higher and hard clipping of the waveform follows.

In the analog world the concept used is different. We talk of a "reference" level of 0VU (Volume Units) which is then referenced to an equivalent signal level and voltage measurement.

For example in the pro audio world that reference is 0VU = +4dBm = 1.228V compared to the consumer standard of 0VU = -10dBV = 0.316V. That means that if we were to feed a sine wave that reads 1.228V on an AC voltmeter to an analog processor with it's input and output level controls set so it's not adding or subtracting any gain (this is called "unity gain"), it's VU meter (which is an averaging meter that relates to how the ear perceives loudness and not the instantaneous absolute peak value of the signal) should read zero. Since a sine wave has NO peaks whatsoever, this same signal would read -18dB (Fs) when fed to a factory calibrated 192 i/o. Hence we say that it is calibrated to 0VU = +4dBm = -18dBFs.

Now, one important notion in the analog world is that of "headroom". This is the amount a signal can exceed 0VU in a particular piece of equipment without incurring "clipping" or distortion and is expressed as +something dB. (It is, in fact, the actual maximum AC voltage an equipment can accept at its input, pass through its circuit and produce at its output without clipping of the waveform.) This vary greatly from unit to unit with the best equipment out there (GML, Cranesong and the likes) having headroom somewhere near +30dB! That means that under normal circumstances these pieces can't really be clipped (unless you add a ludicrous amount of gain like in an equaliser for instance).

How should you calibrate your converters?

This is dictated in part by the headroom of the equipment you are using. If you are using equipment with 12dB of headroom for example, calibrating your system to 0VU = -18dBFs would be less than optimal since these pieces will not produce peaks reading more than -6 in PT before clipping the signal. So they will never clip your converters. On the other hand, why use a calibration level of 0VU = -12dBFs if your equipment is able to pass undistorted peaks of +30dB?

The type of music you record also comes into play in deciding how to calibrate your system since, heavily compressed pop/rock instruments or already "produced" synthsounds and samples most probably will not contain the same peaks (or have an as "high" peak-to-average ratio should we say) than uncompressed jazz, classical or acoustic instruments.

Also, your position is different if you do a lot of tracking of "raw" instruments to be processed with plug-ins inside PT compared to doing mostly mixing with analog hardware inserted on already more produced and "controlled" tracks.

Since I own some high headroom equipment and do both a lot of tracking and mixing of varying musical genres, and since I don't like re-calibrating my system every other song (also since I like overdriving some pieces like some Neve modules at times...), I presently stick with 0VU = -18dBFs and don't fret too much if every other track is not hitting close to zero on PT's meters. (I don't think that it's a good idea to hit PT's summing buss with too much level when mixing anyway but that's another totally different (and much discussed!) can of worms...)

To come back to your example, my guess would be that this pair of converters could probably be calibrated somewhere around 0VU = -14dBFs. When mixing (presuming that you use a pro quality unit with relatively high headroom), I would first put the compressor inserted across your mixbuss into bypass and check the input level (what you're feeding it) on it's VU meters and try to build my mix so that it normally reads around 0VU to +2 with peaks maybe hitting +3 before compression taking place. Again presuming that once you engage the compression the output level would normally kick back a bit and stays around 0VU. For a typical overcompressed pop mix, the peaks are rarely more than 12 to 14 dBs over the average level before final limiting so this should theoretically work.

Of course there are more subtleties to using analog gear that I can get into here. For instance, many engineers find that different pieces have different "sweet spots" as far as hitting them with different levels and will use the way certain equipment starts to behave in an sonically interesting and flattering way before crapping out when pushed a bit. By contrast, certain pieces (cheapy mixers for example) are not happy with "hot" levels and are better operated conservatively. So, in the end, it's still back to your ears and what they are telling you...

Like our beloved Slipperman would say: HO HO HO!!!

Have fun!