A Listening Evaluation of Discrete vs Integrated Circuit Audio Preamplifiers in Stringed Musical Instruments

R.M. Mottola

Copyright © 2003 by R.M. Mottola

[Originally published in the Journal of Musical Instrument Technology #23, 2003]

Abstract

Audio preamplifier circuitry is increasingly used in electric guitars and electric bass guitars.  A general preference for circuits built using discrete solid state components vs integrated circuit operational amplifiers appears to be building within the user community for these instruments.  A double blind experiment was conducted to determine if users showed a preference between a JFET discrete component preamp and an opamp preamp following a listening evaluation.  Results indicate no clear preference for either type.

Acknowledgments

The author wishes to thank Jim Mouradian of Mouradian Guitars, Cambridge, MA for assistance in the calibration of the instrument used in this experiment.

The author also wishes to thank Rich Appleman, Bass Chair, Berklee College of Music, Boston, MA and his staff for their assistance in providing a subject population possessing high critical listening abilities.

Introduction

For most of their history electric guitars and basses contained only passive devices in the instruments themselves. These consisted of a magnetic pickup or pickups and passive volume and tone (treble attenuation) controls. Active circuitry is increasingly found in these instruments, particularly in electric basses. These circuits serve a number of functions. Simple buffer preamps drive the capacitance inherent in the cable used to connect the instrument with a power amplifier and thus increase high frequency output. Some preamp circuitry is used to also provide gain to boost the signal from magnetic pickups designed for less high frequency attenuation and consequently lower output. Instrument manufacturers have also been including active equalization controls in their on-board circuitry.

This article will focus on active circuitry used in electric basses due to the prevalence of such circuitry in those instruments. A cursory and informal examination of bass players' opinions on active preamps, as revealed in personal discussions with musicians and in conversations appearing in topic specific Internet discussion groups, indicates that at least some users consider that preamp circuitry built using discrete (usually Field Effect) transistors are superior in sound to those built using integrated circuit operational amplifiers. This preference appears to be reinforced (or perhaps precipitated) by advertisements by commercial instrument manufacturers and aftermarket circuitry suppliers that present as a selling point the fact that their preamps are made using discrete active components.

A general explanation for the preference of discrete circuitry in electric bass preamps seems to relate to the signal distortion characteristics of the two types of circuits. In this view opamp circuits are considered to distort in a less "musical" manner and thus not sound as good. Distortion characteristics of opamps operated outside their designed dynamic range do appear in the audio literature[1]. But nothing in the literature could be found which indicates that bass players show a preference for discrete component preamps based solely on critical listening evaluation. Some research does exist in the domain of consumer audio which indicate that basic circuitry differences do not result in appreciably detectable preferences[2]. The experiment described was an attempt at such an evaluation in the domain of onboard preamp circuitry in stringed musical instruments.

Figure 1. The test electric bass guitar with two preamps and an A/B switch.

Methods

Double blind evaluations are the gold standard in medical research and are used increasingly in audio research as well[3][4]. A double blind listening evaluation of a purpose-built electric bass was performed as follows.

Preparation of the Test Instrument

A solid body electric bass guitar was built specifically for this experiment (figure 1). The instrument contains a single magnetic pickup (Bartolini MME), one discrete JFET buffer preamp, one integrated circuit opamp buffer preamp, and a selector switch to switch between the two. The selector switch is the only control on the bass guitar.

The selection of the two preamp circuits was considered critical and proved to be problematic. Ideally each circuit would be representative of its type (discrete JFET, opamp) but as there is so much variation possible within each of these types that goal cannot be practically achieved without additional investigation. The idea to select representative preamps from those made by manufacturers of OEM and aftermarket preamps was abandoned, as it was desirable that the results of this experiment should not be misconstrued to be a product endorsement. In the end the circuits selected were two that appeared on an Internet website containing such circuits (ftp://ftp.ee.ualberta.ca/pub/cookbook/audio/guitar/preamps.gif).

The schematic for the discrete component preamp is shown in figure 2. As can be seen the circuit is a typical JFET common source design and is in no way unusual. Figure 3 shows the opamp circuit, this a textbook example of a single power supply amplifier. The opamp is a TL07. That improvements are possible in both circuits is obvious. At the very least a transistor and an opamp with lower noise specs could be substituted for those used, but for the purposes of this experiment, and considering that circuits ideally representative of their types would be difficult to achieve within the scope of this experiment, these two circuits were deemed adequate. A slight modification was made to the opamp circuit as posted for this experiment. The "gain set" resistor was replaced with a trim pot of the same value so that the gain of this circuit could be slightly adjusted such that the gain of the two preamps could be matched.

Figure 2. Discrete JFET guitar preamp circuit.

The circuits were built on a piece of perfboard and mounted within the shielded control cavity of the bass. The gains of the preamps were matched using a 100 mV p-p, 1 kHz sine wave signal from a signal generator, and an oscilloscope. The selector switch was loosely mounted and the control cavity of the bass closed. The physical orientation of the selector switch was then randomly changed and a switch label with markings "A" and "B" was attached and the switch mounting nut tightened. These last steps were done so that the builder of the bass would not know which switch position corresponded to which preamp, thus double blinding the experiment.

Listening Evaluation

Fourteen (14) subjects were recruited to critically evaluate the preamps. All subjects were professional electric bass players. Three (3) were also builders of musical instruments. Eight (8) of the subjects were members of the bass faculty at the Berklee College of Music in Boston. Each subject was asked to play the test instrument and to critically evaluate the tone of the instrument with the selector switch in both positions. Each subject spent less than five minutes in this effort. After playing the instrument each subject was asked to complete a questionnaire containing a multiple choice question evaluating the tone of the instrument. The choices were:

A sounds better;
B sounds better;
They both sound the same;
They sound different but neither one is better;

Figure 3. Opamp guitar preamp circuit.

Each subject played the instrument through a power amplifier of his/her choice. Five different power amplifiers were used for the experiment. Most subjects left the tone controls of the power amplifiers in a "flat" position, but some set them as they would for their personal style of playing. All subjects played in a variety of styles, making use of the entire range of the instrument. A few subjects attempted to evaluate the tone under extreme conditions - for example turning down bass and turning up treble tone controls and then scratching the strings to elicit a tone rich in upper harmonics.

Subjects were isolated from each other during the evaluation and completion of the questionnaire to control for sociogenic bias.

Results

The experiment was unblinded and the questionnaires tabulated with the following results. Two (2) subjects found the tone of the discrete FET preamp preferable. Two (2) subjects found the tone of the opamp preamp preferable. Five (5) subjects found the sound identical in both preamps. Five (5) subjects found the two preamps sounded different but that neither one was superior in tone. As the data are so obviously distributed normally about the mean no additional statistical evaluations were performed. Subjects that indicated a preference also indicated that perceived differences were very subtle.

Conclusions

The majority of subjects (8) found no preference between the two circuits and preference was evenly divided among the remainder. As a result it is reasonable to conclude that it is unlikely that either the discrete component circuit or the opamp circuit tested would result in any sonic advantage in a bass guitar preamp intended to appeal to a large user base. From the results of this experiment it is tempting to extend this conclusion to include all such simple preamps of these two classes. Doing so would depend on the extent to which the compared circuits are representative of discrete JFET and opamp preamp circuits in general. One factor in favor of considering the tested circuits as representative of their types is that the two circuits were chosen effectively at random.

As is usual in such preliminary investigations, the results of this experiment suggest topics for future work. Given the normal distribution of the data, perceived preferences may need to be validated, possibly by repeated sampling of subjects declaring a preference. Sampling of a wider range of circuits may indicate the degree to which the two circuits evaluated are representative of their types.Finally, the nature of listening evaluations in the domain of on-board circuitry in musical instruments should be investigated with an eye toward determining if such evaluations are less sensitive to the detection of differences than in the domain of consumer audio. In the former case there is considerable feedback in the process of generation of sound, and players may consciously or unconsciously modify playing techniques in real time to effect a desired tone. If this is the case then the ability to detect some otherwise audible differences in circuitry would be muted in this domain, and such differences would be considered irrelevant for all practical purposes.

References

1. Hamm, R. "Tubes vs Transistors - Is There an Audible Difference?", Journal of the Audio Engineering Society Vol. 21, #4, p.267, May 1973

2. Masters, I. G. and Clark, D. L., "Do All Amplifiers Sound the Same?", Stereo Review, pp. 78-84, January 1987

3. Shanefield, D., "The Great Ego Crunchers: Equalized, Double-Blind Tests", High Fidelity, pp. 57-61, March 1980

4. Toole, F. E., and Olive, S. E., "Hearing is Believing vs. Believing is Hearing: Blind vs. Sighted Tests, and Other Interesting Things", 97th AES Convention (San Francisco, Nov. 10-13, 1994)