Our journey begins with some of the earliest recording studios. The introduction of reel-to-reel tape machines and the creative engineers that used them brought on some of the fattest sounds imaginable. From the inherent warmth and luscious tape saturation—to the forgotten art of how these machines were manipulated—they are capable of creating a huge range of beautiful, distinctive, and enveloping sounds.
When we began this project of researching the technology, mechanics, and methods used, it was an important reminder that these machines really were the first effects pedals—they just wouldn’t have fit on your pedalboard!
The result of this in-depth study of tape are the sounds and technology found within Deco Tape Saturation & Doubletracker. Pete Celi, our Sound Designer and DSP Engineer illustrates the research and sound design process in the White Paper below.
Studio Reel-to-Reel Tape Dynamics and Double-Tracking White Paper
The inherent warmth of reel-to-reel mixing and mastering tape decks is well known and still highly regarded today. Described as making tracks sound ‘fatter’, ‘fuller’, or ‘punchier’, the intrinsic properties of the record/playback process create a level- and frequency-dependent dynamic response. This allows for transparency at low levels, with a reduction of harsh high frequency spikes on transient attacks and a harmonically rich low end. Running two decks simultaneously creates a host of sonic possibilities that benefit from the tonal enhancements produced in the record/playback process.
Tape Signal Coloring
The professional reel-to-reel machines designed in the 50’s and 60s were high-fidelity units built to accurately record and reproduce full-range audio signals. The NAB and CCIR recording standards in the US and Europe respectively called for a flat frequency response through record/playback process with minimal harmonic distortion. With that in mind, it seems like there’s not much opportunity for ‘coloring’ of the sound. However, the sound was shaped by a number of elements that result in dynamic equalization effects and saturation effects.
Record and Playback Equalization
Although the overall process has a flat response, the individual record and playback operations themselves are far from flat. The magnetic properties of the tape and self-erasure effects from the record head result in a loss of high-frequencies when recording, while the playback process lends a further high-end rolloff due to speed-dependent gap-loss effects. Low frequencies can experience a boost on playback due to a phenomenon known as ‘head bump’ that is dependent on the tape width. So how do we get a flat overall frequency response? Through an EQ process known as Pre-emphasis/De-emphasis.
The natural high-frequency rolloff of the playback heads is partially offset by equalization in the playback amplifiers. The remaining high-frequency loss of the playback process, plus the losses of the recording side, are equalized by boosting the high frequencies prior to recording. The equalization on the recording side is known as Pre-emphasis and the equalization on the playback side is De-emphasis. The overall response of the recording side (tape plus Pre-emphasis) has a boost in high frequencies, while the playback side (tape plus De-emphasis) has a corresponding cut in high frequencies. This scheme helps to reduce tape hiss issues. The low frequencies were also given a small boost/cut treatment to reduce low-frequency mechanical noise and any power-related hum in the NAB standards. The figure below shows the individual and combined frequency responses.
Fig:1 Record, Playback and Combined Frequency Responses
Furthermore, the introduction of noise-reduction circuitry in the mid 60s adds additional dynamic boost at various frequencies when recording, with a corresponding dynamic reduction during playback to further reduce tape hiss. The general ideas is – boost before recording and cut after recording to minimize noise.
Tape saturation characteristics are determined by a combination of many factors, including tape formulation, bias signal and track width. The 1965 NAB specifications stated that 3rd order harmonic distortion should not exceed 3% at peak levels (6dB above nominal level) . Their goal of course was for a sonically accurate record/playback system, and 3% third harmonic distortion on peaks is still very ‘clean’. In an effort to maximize signal-to-noise performance, recordings were generally done at the highest possible levels without noticeable clipping. Depending on the dynamics of the material, the transient peaks might experience much greater than 3% distortion. So, what happens when those VU meters are pushed further into the red?
Under normal operating conditions, a small amount of 3rd order harmonic distortion is present which contributes to some of the low-end warmth of the tape. Third-order harmonics increase appreciably as the recording level increases and fattens up the lows and mids, adding higher odd-order harmonics as the level is further cranked. The high frequencies (already boosted by the pre-emphasis curve) will be pushed harder into limiting while their distortion components will be largely inaudible but will contribute to intermodulation distortion during playback. Along with the tape saturation, signal path components such as transformers or tubes will also impart their sonic signature as they are driven harder based on circuit design parameters. Let’s see what happens when we put the whole system together.
Pre-Emphasis + Saturation + De-Emphasis = Dynamic Fatness
This is where the magic happens. In the absence of saturation and signal compression (when the signal level or recording level, or both, are low), the pre-emphasis and de-emphasis combine to preserve the the frequency and amplitude response of the input signal. When the tape is saturated (during peak transients or when the record level is cranked up), the effective high boost of the pre-emphasis is diminished as the tape cannot record larger signals. This results in a ‘flattening’ of the response of the recorded signal on tape. On the playback side, the playback heads and de-emphasis curve will roll-off the high frequencies, resulting in an overall spectrum that has reduced high end.
Fig 2a: High Bandwidth Signal Input Spectrums
Fig 2b: Spectrums Recorded onto Tape (Blue dash is Pre-Emphasis EQ)
Fig 2c: Output Spectrums After Playback (Red dash is De-Emphasis EQ)
From a tonal standpoint this means that harsh high-frequency peak transients, e.g. from hard strums or pick-attacks, are fattened by not only the dynamic amplitude-limiting of tape saturation and compression, but also by the effective dynamic frequency-limiting of the playback process. This, along with the 3rd-harmonics of the lower frequencies serves to add punch and fullness to the sound in a subtle but effective manner. Turning up the recording level intensifies this effect and rewards dynamic playing.
Double Tracking Effects
Double-Tracking effects are achieved by combining two copies of the same signal in various manners, creating effects that include thru-zero-flanging, comb-filtering, chorusing, doubling, echo, and split-stereo effects. While the legendary Les Paul pioneered the techniques and concepts of double-tracking and multi-tracking effects, the effects were popularized in music recorded in Memphis and London in the 50s and 60s.
How it was Done in the Studios
The input signal was routed to two decks simultaneously. The 1st deck records the input signal, along with the playback signal signal from the 2nd deck. The tape speed and distance between the record and playback heads of the 2nd deck determines the echo time. This arrangement allowed for real-time live-performance echo to be created and recorded for vocals and guitarists as they tracked their parts in the studio.
Fig 3: Slapback Echo
The echo time was fixed based on the tape deck’s physical head distance, but most decks allowed for a half-speed option, so the delay time could be doubled by running the machine at the slower setting.
Inverting the phase of the echo deck creates a subtle difference in the low-end response that can sound more natural in some applications, as the sound-wave is inverted in a physical reflection.
Thru Zero Flanging and Comb-Filtering
How it was done in the Studios
After recording the original signal, it was played back while the sync signal (record head output) was sent to the 2nd deck’s input. Both deck’s playback outputs were mixed together while the lag deck was sped up and slowed down, creating a dramatic sweeping comb-filter effect as the lag deck crosses through ‘zero delay’ time relative to the reference deck.
Fig 4: Through-Zero Flanging
Because both playback heads were being monitored, the effect had an inherent delay from input to output and hence tape-flanging was not a real-time effect, meaning not able to be used in a live performance situation.
For static comb-filtering effects, the decks are run with a constant offset lag between them.
Through-zero flanging was typically used on a full-range mix as the high frequency content intensifies the effect of the high frequency comb-filters, which is where the action is as the lag approaches ‘zero’. Similarly, it sounds great with heavily distorted guitar signals.
Another set of tonalities can be had by inverting the phase of the lag deck so that the signals experience a cancellation instead of a reinforcement as the delay approaches zero.
Tape-Chorus and Doubling
How it was done in the Studios
The chorus and doubling effects used slightly longer delay times that do not cross through-zero but still required using the sync signal from the reference deck to send to the lag deck. Chorus will typically run the 2nd deck at a lag time between 10ms and 30ms, with slight manual manipulation of the deck speed to modulate the delay time for movement in the sound. Doubling effects would be achieved around 40ms to 60ms with or without additional speed modulation.
Fig 5: Tape Chorus and Doubling
A phase inversion of the lag deck will produce a different response most noticeable in the low end. Experimentation determines which polarity sounds ‘better’ in a given environment.
Stereo Splitting Effects
Double tracking techniques were also used to create stereo effects from a single track using the same methods as above, but with the output of one deck going to the Left channel and the other going to the Right channel. Instead of being summed electronically through a mixing console, the decks are ‘mixed in air’ acoustically. This creates a very different effect as the three-dimensional acoustics of sound and reflections means that the complete cancellation and reinforcement of electronic comb-filtering does not occur when the signals are mixed acoustically. With both decks in sync at zero lag, a slight modulation of the speed of the lag deck creates a subtle panning movement between the channels due to the precedence effect. Increasing the lag time creates a spacious stereo chorus. Further increases in the lag time result in natural doubling and slap effects where the echo is physically displaced from the source.
Fig 6: Stereo Splitting Effects
With zero lag and no modulation, a phase-invert switch on the lag deck acts as a channel phase switch which can come in handy with dissimilar cabs, or amps or speakers that may be wired out of phase. At longer lag times, a phase inversion will give a subtle change in tonality and stereo image.