Last time, I talked about the use of pre-emphasis (and de-emphasis) in the cutting of disc masters used in the creation of vinyl LPs. The RIAA curves dramatically alters the frequency contour on both the record and reproduction phases — and hope that the electronics are mirror images of each other.
It turns out that the use of pre-emphasis and de-emphasis is standard operating procedure in analog tape recording too. Tape is inherently noisy. The signal to noise ratio of a first generation tape reaches to about 60-72 dB without noise reduction. Electrical engineers figured out a long time ago that adding a pure ultrasonic tone to the program signal improved the frequency response, reduces distortion, and lowers the noise level — but not by a lot. The frequency of the bias is higher than the audio band (typically 50 kHz or higher) making it inaudible during normal playback (it can be heard when scrubbing a tape during editing).
So analog tape recording uses the same pre-emphasis scheme employed by other systems (FM radio, and vinyl LPs). The upper frequencies are boosted during recording and then reduced by the same amount during playback. Analog tape machines need to be calibrated and aligned on a regular basis. In fact, it was my job as a young second engineer to clean, demagnetize, calibrate, and test all of the tape machines in the studio prior to every session — that meant everyday. I got quite good at it.
Unlike the single RIAA EQ curve for vinyl LPs, analog tape EQ standards vary depending on the country and the speed of the tape. There are NAB (National Association of Broadcasters) for the United States and another set of standards for European tapes (CCIR/DIN). Makers of tape copies make tapes available in either standard.
The whole concept of purposefully distorting the sound in order to achieve better fidelity has merit. However, there linearity of the frequency response is entirely dependent on matching the pre-emphasis with the de-emphasis curves during recording and playback. In professional equipment with regular maintenance, this isn’t really a problem but inexpensive and poorly maintained consumer equipment can’t be counted on to deliver a flat frequency response.
When it comes to capturing and reproducing frequencies without unwanted deviations from a linear response, analog means are barely up to the task. No matter what designers and manufacturers have tried to do (custom EQ curves like Nagra Master), there are always challenges. And don’t forget this is without Dolby noise reduction, which introduces an additional pair of encode and decode conversions.