Breakthrough Describing Tone Quality in Guarneris
The following text is an introductory chapter from a book we are writing on tone quality of violins.
The persevering reader will be offered an insight into the complex structure of musical sound, which can be depicted graphically with curves. These graphic images are called spectra, and they can be viewed in real-time on a TV screen or can be studied in printouts.
We recognize the sound of most common musical instruments by their special tone quality, or timbre, but every instrument type can be even more accurately described by their respective audio spectra.
The necessary instrumentation is an audio signal analysis system which is available from commercial sources. It can be inexpensive as an extension of a personal computer, or it can be a more sophisticated and expensive engineering equipment, and the price may be anywhere between $200 and $20,000. We advise individual musicians and music schools to seek professional help in procuring and setting up a sound analysis system.
The importance of visualizing the sound goes well beyond academic interest. This technique provides the most reliable learning tool for the improvement of tone quality, and it also offers useful information on the suitability of the violin, identifying its strengths and weaknesses. It provides the means of rating and ranking violins in an objective manner based on tone quality, which thus can become a measurable commodity and a major factor in pricing violins.
Our studies on violins have shown that each individual note of the musical scale has a different pattern in its spectrum, and the corresponding spectra obtained when great artists were playing their fine violins show many common features. This allows us to recognize a range of acceptable standards for each note, and also explain the trouble spots in the variety of bad notes. In the essay that follows a detailed interpretation of the spectra is given, and it is my hope that the reader would be curious enough to work his way through it. I should concede, however, that a full understanding is not necessarily required, and one can take a minimalistic approach of working only with the simple technical process. Most of us don't know much of automobile mechanics, yet we have mastered the skill of driving.
The task at hand is to play a note into the microphone and try to adjust the changing image on the screen to a standard of excellence. This can be done by trial and error using one's own devices, but some advice can be found in later chapters. The only example here is the spectrum of the open D note played on one of my violins. The fundamental pitch of this D vibrates 294 times per second; it is designated as D 294 Hz (Hz, stands for Hertz, the unit of vibration). Its spectrum is comprised of many peaks, which are known as harmonics or overtones.
The D is a good note to start with because it sounds good on most violins, and after a few failed attempts some degree of gratification can soon be achieved. The goal is to obtain a wave-like pattern with gradual change from peak to peak. The worst situation, the dominance of the 3rd and 5th peaks, must be avoided. The nasal range of 1300 to 1600 Hz must be minimized; while the "operatic range" of 2500 to 3000 Hz should be maximized. Eventually, all notes of a two-octave scale should be studied always keeping in sight the corresponding best spectrum provided later in the text and the appendix. Sooner or later, the student will discover the limitations of his/her violin, which may not sound even remotely like a Strad. The effort is still worthwhile because each violin has many good notes, and even the less satisfying ones can be optimized. Spectra of bad notes come in great varieties; they lack regularity.
A brief discussion of the human singing voice is necessary here because of its relationship to the sound of string instruments. Our perception of tonal beauty in the sound of the violin depends, to some extent, on how close we can come to imitating the singing voice and the elements of the language, such as the various vowels and consonants. Unfortunately, for the vowels of the violin one has to go to languages like French, German, Hungarian and Turkish which offer a variety not found in English. For example, when the note D is sung with the vowel of the French word peu, one obtains the following spectrum which is clearly related to the violin spectrum of the same note. The singer of this note was my 12-year old daughter Katie, who is definitely a non-professional singer.
The singing part of the bowstroke often comes close to a vowel, which will change with frequency. With a little imagination, one can hear vowels like u, ah, ee and some nasal sounds like the French on, un, in and an. The violin is more nasal than the human voice, and the limits of acceptibility are hard to define.
The attack of the bow in initiating a musical note creates a consonant sound which is an integral part of the overall tone quality. The consonant sounds of the violin are not as diverse and clearly articulated as those of the human voice, but they are an essential part of violin playing. Depending on the pitch and the string chosen, the attack of the bow in staccato can cause a sound like d, dz, k, and t. A short clicking sound is preferable to a long mushy and scratchy one, and a persistent sh or s is downright undesirable.
Another similarity between the voice and the violin sound is their vibrancy, which is responsible for much of their emotional impact. In addition to the rhythmic modulation of frequency by the deliberate vibrato, one perceives a somewhat irregular flickering of tone color, whose best visual analogy would be the flickering of candle light. The change of tone color can cause a modification of the vowel, the kind we hear in the diphthongs of speech, but repetitious: wowowowow, ayayayay. Vibrancy is a precious property, which varies a great deal among violins. Good players tend to select violins of outstanding vibrancy. This allows them to develop a personable tone, one which allows identification by the listener.
A good projection of the sound is a major concern of all performers, and it still remains a somewhat elusive entity. There is more to it than just the loudness of sound that can be measured with a sound level meter. Not all frequency ranges project equally well, and they tend to have a variance in different directions. Useful predictions can be made, however, on the basis of the sound spectra of each note. It is desirable to have a strong first peak for reasons of sonority and also comparable strong emissions in the range of 2,500 to 3,300 Hz where the human ear is the most sensitive.
It is far from my intentions to advocate as a goal the uniformity in tone quality; to the contrary, I am in favor of individuality. Pressures already exist for students to sound like Itzhak Perlman. Most of us recognize the variance of individual tastes and the wealth of aesthetic possibilities. It is my hope that these possibilities can now be better explored with the help of new technologies.
| Home | Choices | How To Buy | Biographies | About Us | Tone Quality |
© 2001-2008 Joseph Nagyvary