About the two kinds of absolute hearing.

(Tone quality recognition and pitch recognition.)

Originally published in Zeitschrift Internationalen Musikgesellschaft, 14, 130-7.

Géza Révész, Budapest.

Translated by Christopher Aruffo
www.acousticlearning.com

From direct observation of the musical scale and through experimental trials I came to the discovery that, excepting intensity and timbre, each pitch has not-- as was generally accepted until now-- a musical characteristic, but rather two musical characteristics which are independent of each other [1].

The one musical characteristic, which I call Quality of the pitch, appears in the scale in a remarkable way.  It appears regardless of the octave, since the same pitch in any octave will return similar feelings.  The other musical characteristic, which I will call Tone Height, appears when playing the scale as the "raising" or "lowering" of the pitch.  If we play a scale through several octaves, the distinction of these two musical characteristics forces itself upon us automatically.  On the one hand we notice a change taking place in a continuous direction, and on the other hand we have the impression that each octave is repeating itself.  What repeats from octave to octave is the Quality; what changes in a constant direction is the Height of the pitch [2].

In many cases, resemblance between tones is due to partial equality, i.e. to overlap of one or several elements.  This is the case here, in that the octaves sound similar in the equivalence of the Quality feature.  The similarity is particular to the octave relationship.  The view that it comes in an indirect way, as Helmholtz imagined when he sought to prove octave equivalence by common overtones, I could disprove; using tones whose overtones have been extinguished, I have shown that an octave similarity appears in exactly the same manner as with overtone-rich tones.  The fact that octaves are not absolutely identical is a consequence of the other characteristic, the differentiation of Height.

The commonality between the octave tones is expressed in the musical names of the tones, since the same pitch in each octave carries the same name.  The difference between them is designated in the name by the number of its octave.  The Quality is the same, the Height its index.  C1 and C2 both have the same name, C, because they have a common Quality; however, they have a different index because they are different with regards to Height.

This view of the scale will be much clearer to a musician, and will correspond more closely with his observations than the usual view, which attributes each tone with the single musical characteristic of pitch.

I have furnished a very important proof of my view.  I made observations of the two characteristics separate from their usual connections, thus proving that the two musical characteristics are independent of each other and materially separable.

In a different experiment I showed that different Qualities appeared to a subject.  A listener imagines pitch in such a way that two characteristics are apparent.  For instance, E and F were both observed in the second octave, and between these two pitches, only the Quality difference was noticed and not a difference in height.

By contrast, a characteristic of different Heights could also be found.  Anyone with normal hearing can convince itself of this with the octave tones.  Under pathological conditions, however, a large set of neighboring tones may possess the same Quality, but all have their normal height.  Thus it occurred that almost all tones from D3 to G#4 had the Quality of D#.  The Height difference was nonetheless expressed so that playing a scale in this region gave the normal impression of rising and sinking.

It is even possible to demonstrate a scale without Quality.  In the highest areas of the audible scale the Qualities are undetectable; that is, one cannot sing these high tones nor recognize them as octave, fifth, or by any other name, but the tones clearly differ from each other in Height.

This much description may be sufficient here.  I refer to my detailed work:  Zur Grundlegung der Tonpsychologie, Leipzig 1913.

These opinions bear close relationships with certain important questions of absolute hearing.

One has generally assumed until now that there are those people who have absolute hearing and those who do not, and a sharp border has been drawn between these two categories.  Naturally, then, these categories were examined accordingly and observations were made of those who possessed absolute hearing in the one specific form-- namely, the ability to name or recall a desired tone.  The remaining subjects have been neglected, despite being the majority of absolute listeners, since the ability was denied to them except in certain moments.

If one examines the people who are purported to have absolute hearing, one discovers the fact that their ability to correctly identify pitch names is confined to a limited region of the musical scale.  Above or below this region, the subjects' answers become more incorrect and subjectively more uncertain, until one finally comes to a threshold at which no more identifications can be made.

O. Abraham supplies a good example in his publication on absolute tone consciousness [3].  He judged the tones from F# to F4, from F# to A#1, and from F#4 to E#5.  Throughout these ranges he was able to correctly identify the tones with errors no larger than a semitone upward or downward.  The tones A#1 and E#5 represent sharp borders with Abraham, because the inaccuracy of the judgments suddenly jumps to 80%, so that nearly all tone judgements between A#1 and D#2 deviate by a third, and with the exception of the first two tones in the high area (thus with all tones between d5 and d6), he made errors of even a fourth or a fifth.  I found something similar with two test subjects talented with absolute hearing, one of whom was the young composer Ervin Nyiregyházy, whose absolute hearing is probably the most accurate and extensively developed as could be found.

Already, this curious feature of absolute hearing demands an explanation, i.e. that there should be any kind of region for absolute hearing (that absolute pitch only develops within a range) outside which the accuracy of judgment decreases, and that the area where incorrect judgments occur is further divided into districts of greater and lesser accuracy.

How can it be explained that people equipped with absolute hearing may judge particular tones correctly or approximately correctly, but others with large inaccuracies?

The most evident answer is that the difference in judgment is based on a difference in the judgment criterion, so that the observer in one case relies on other characteristics of the pitch feeling which are different from those they use in a different situation.  I say now that these pitch judgments are actually based on two judgment criteria, and that one must differentiate two kinds of absolute hearing according to these two criteria.

I now return to my opinions of the two musical characteristics of the tones, in order to join together the statements I have made.

It is clear that each judgment is based on one or the other of these characteristics, or on both simultaneously; thus either on the Quality or the Height of the tone, or both.  I differentiate these judgments as either tone Quality recognition and pitch Height recognition.

The consistently accurate judgments of people equipped with so-called absolute hearing are based on the ability to recognize tone Quality.  The feeling for Quality is somewhat individualized for each person.  Tones which do not have Qualities, e.g. very high tones or most unmusical noises, have no musical individuality, and are therefore often confused with neighboring tones.  Observers indicate the fact that tones appear to them as individually distinct; e.g. all F-tones share a common Quality which differentiates them from all G-tones, etc., so that they never mistake pitches for each other.

It is very strange that the ability has this consistent result, with which for instance an F is never confused for a G or even an A-flat, regardless of the octave.  The behavior of the observer can be explained by saying that they use another judgment criterion in the regions further from the middle octave, and that criterion is the Height.  If the judgment of the observer depended entirely on the Quality characteristic, this would demonstrate that the feeling for tone does not receive any individuality from its Height characteristic.  Just as one may see in the visible spectrum that individual colors appear with different brightness, their individuality is not determined by this brightness, so too the individual tone qualities of Height do not provide individuality; it is by its Qualitative characteristic that a pitch is placed in the musical scale.

However, each tone does have a unique Height, and to that extent the Height is something which describes it uniquely.  The observation merely shows that the impression of its Height individuality does not come as with the Quality.  It behaves thereby in a similar way as a series of grays, which lead from black to white; each individual grey has its own unique brightness, but to us the grays will simply appear "gray" and be remembered as such, unless we attempt to compare between them.  It is questionable whether we are capable of remembering the specific brightnesses of these grays.  Thus we may remember the individual Height of a sound, but we do not use it for identification, because in making a judgment with absolute hearing, it is completely sufficient to know the approximate position of the tone.  If we know the Quality of the tone, we can recognize from its height whether the tone most appropriately belongs to the second octave, or third, or whichever.  We do not need to know the precise height of a C2; as soon as we recognized the Quality of a C, it is enough to know that its Height is typical of the second octave.  I get along well with only a general feeling for Height areas.

We may thus imagine that the region of absolute hearing where the judgments are exact are occurring through evaluation of Quality, although naturally Height plays a role (for octave determination).  In the areas where judgment errors occur, height alone is the determining factor.

The difference between these two kinds of absolute hearing is evident in the psychological behavior of the observer, i.e. in the manner by which they formulate their judgment.  If the observer recognizes the tones from their Quality, they will provide a rapid, certain judgment; the names of the tones appear directly, effortlessly, as though it were a psychologically passive activity.  If, however, the same observer must recognize tones by their Height, then this usually occurs only after some deliberation, with lesser subjective certainty, and as an active judgment; i.e. the judgment is won by conscious activity, which can even be noticeably arduous.  The attempts are more instructive than long explanations.

The two kinds of absolute hearing can be demonstrated in the same pitch area, under the circumstance that tones of certain timbres are usually better recognized than others.  Usually, sung tones cause special difficulty, while piano and violin tones tend to be easily recognized.  With different timbres of sound, one may very easily demonstrate that the observer judges sung tones by Height, while judging piano and violin tones by Quality.

A forcefully compelling example is one in which a subject was to judge the tones according to their height or their quality [4].  In this experiment, a part of the scale was, because of detuning, so constituted that a large number of neighboring tones had the same Quality but still possessed their normal Height.  The subject was informed of the two different criteria.  One receives from this situation two different judgments, one due to Qualitative absolute hearing and another due to Height recognition.  The absolute Quality judgments always naturally result in the same quality (in one case, always G#, thus always objectively wrong judgments), while the absolute Height judgments were approximately objectively correct.  Thus, in a certain case, all tones between C3 and B3 were judged to have the Quality of C#, but the heights were recognizably different as the table shows.

Objective tone

Tone judgment

= Height judgment

= Quality judgment

From all past observations, it can be inferred that, apart from those who can identify tones with their absolute hearing, all people are nonetheless capable of pitch recognition.  The ability which until now has been called absolute hearing is the Qualitative absolute hearing.  The difference between people with total absolute hearing and those with absolute hearing within a range is that each range of musical sound has still another Quality specific to that range, and Qualitative recognition may thus extend only over this range.

People equipped with regional absolute hearing are the link between the two categories of people:  between those who have absolute hearing according to the current definition, and those who fall short of the mark.  I have said in my experiments that pitch recognition occurs in very many people; I have found it in all subjects I have examined to this point.  And, perhaps excluding completely unmusical people, I am convinced that all people are able to judge tones by absolute Height with the aid of their pitch recognition ability.

I conducted the following experiment with both musical and unmusical subjects.  I played different tones at the piano and requested the observer to indicate which piano tone had been sounded (by pressing the corresponding key).  After one response, further searching was not permitted.  It thus resulted that even those who had never played music and had only a very rough concept of the placement of tones along the keyboard nonetheless identified middle tones within a third or a fourth; some tones in the middle octave were even more precisely identified, typically not deviating by more than a major second (or the rare minor third).  This behavior is identical to the responses of those with partial absolute hearing in those areas where there Qualitative absolute hearing does not extend.

Furthermore, my perspective of absolute hearing makes sense of additional observations.

It very easily explains why people with absolute hearing will make octave errors.  The attention of these observers is given to Quality above all, by which the individual tone is identified, so that the Height characteristic is neglected.  If they are placed into the situation where their recognition cannot be based on Quality (such as the detuned case described above) then they quickly learn the height characteristic and do not make octave errors.  Furthermore, it is not so strange that most people who do not have Qualitative absolute hearing find a greater similarity between C and C# than they do between C and E or C and F# (the Height difference is between C and C# is smaller than between C and E), but those with Qualitative absolute hearing usually can detect no similarity between C and C#.  The similarity between the octave tones is completely noticeable to the Qualitative listener from their identical Quality, and with its neighboring tones the similarity of Height is overwhelmed by the difference of Quality.  For those who are judging by Height, the neighboring sounds are more similar to each other because of the relative smallness of the heights between them.

To summarize briefly:  most people can deliver absolute pitch judgments.  Some, who until now were the only ones to be said to "have absolute hearing", provide their judgments based particularly on the recognition of Quality, while the others will use Height recognition exclusively.

Further questions regarding absolute hearing are raised in connection with this.  For example, how is it possible that many people with qualitative absolute hearing can determine Quality only in the middle range of musical sound, and in the other areas recognize only Height, even though in these other areas the Quality remains the same?  This perspective also introduces new problems for melody and intervals, and other acoustic features too numerous to list here, as they are thus moved into a new light; here, I must refer to my soon-to-be-released work on the foundations of tonal psychology.

Lastly, I want to briefly address the criticism published by Professor Reimann in this magazine (1912, v. 8), who mainly spoke against what I have just written about absolute hearing.  I would like to note at the start that Professor Reimann did not engage with my philosophies, rather, his conclusions and criticism are drawn from some rather vaguely formulated remarks.  I could only infer that he does not agree with my views; why he claimed his disagreements were in opposition to my work is still not clear to me.  The introduction of the single musical characteristic of Quality he wrongly ascribed to Stumpf.  Stumpf referred to the consonance and precision of the extended intervals, and said that these should possess the same degree of fusion as the narrow intervals.  If this were to be attributed to identical Quality, then Stumpf would have been the one to introduce the concept; however, he specifically rejected those tenets (Konsonanz und Dissonanz, p.45).

Professor Reimann does not disprove the numerous proofs which I offered to support my opinion; he did not even discuss them.  I must conclude that he missed the actual subject of my work.  This is already clear from the fact that he places at the front of his report some of the facts which I had placed at the end of mine; not by any means the main evidence for my philosophies, but evidence which I clearly stated to be valuable confirmation of the main evidence.  He also did not disprove my assertion that there are two kinds of absolute hearing to be differentiated.  He does not refute any of what I actually said, but speaks about things about which I said not one word in my report.  Thus Professor Reimann speaks in detail about the meaning of relative versus absolute hearing.  I know the advantages and significance of relative hearing, which Professor Reimann emphasizes in his criticism, and I think it is interesting to study its role in musical manipulation.  However, I said nothing about it.  It seems that Professor Reimann did not want to object to my work, but to anyone who has absolute hearing, since I cannot otherwise comprehend his bitter statements.

He has mistakenly taken offense at the names my test subjects used for the tones.  If my test subject called a tone "G-sharp", this was not to differentiate it from "A-flat".  It was simply redundant to say "G-sharp or A-flat".  I was not concerned with whether an absolute listener makes a distinction between these two.  What he says about the value of pathological cases, that one can draw from them no conclusions about the normal hearing functions, I cannot take seriously.  I would have to speak the greatest banalities if I wanted to refute that opinion.  Don't pathological cases play the greatest role in the arguments for Helmholtz's color theory and his resonance hypothesis?  And even then, if someone should succeed in developing an opinion entirely from normal conditions, it is nevertheless always an essential demand that the opinion be in agreement with all pathological observations.  If a pathological observation were to contradict the normal observation, the theory would have to be abandoned or appropriately modified.  Pathology controls psychological and physiological theories quite strictly, and without indulgence.

Now something material.

Professor Reimann contradicts my indication that the two tones of an interval in the lower areas have a smaller Height (distance) than those in the middle.  He contends that he notices the opposite.  I would not deny his personal observation; however, I must stress that every observation I have collected is the opposite of his.  All my test subjects and even Professor Stumpf [5] who reported on that detail always found the same as I maintained.  Also, here is a small demonstration that contradicts Reimann's observation.  If I present the tones B1-C0 sequentially, then an observer will regard the interval as a minor second.  C0-C# is understood to be a minor second in repeated attempts, thus demonstrating that the direction of the height measurement can be reversed.  However, H0-C2 or C2-C#1 are never judged as seconds.  This is explained simply by the fact that the distance between the first pairs is smaller than the others.  The detailed treatment of this experiment can be found in Chapter 9 of my Grundlegung der Tonpsychologie, "Intervals".  But apart from all that, it is totally unimportant whether the distance of an interval is larger in the lower or higher tones, if they are perceived to be different.  Professor Reimann's observation, if it is correct, does not contradict my own, but rather provides further support for my opinions.

One statement of Professor Reimann that is correct is that the tones of a bass drum do not lack Quality.

I want to make a final remark, which is that my report was intended only to be a short summary of my investigations.  In a short report, only the most important results can be mentioned, and therefore it is not surprising that these results could be misunderstood.  I hope that mistakes and misunderstandings of this kind will no longer be possible after the appearance of my detailed work.


[1] G.Révész, Nachweis, daß in der sog. Tonhöhe zwei voneinander unabhängige Eigenschaften z'u unterscheiden sind. Nachrichten der Gesellsch. der Wissensch. zu Göttingen. Math.-physik. Klasse 1912.

[2] Mit ähnlichen Anschauungen begegnen wir uns bei Brentano, Mach, Drobisch, Lotze und M. Meyer.

[3] Sammelbände der IMG. 1902.

[4] P. v. Liebermann und G. Révész, Experimentelle Beiträge zur Orthosvmphonie und zum Falschhören. Zeitschrift für Psychologie, Bd. 63. 1912.

[5] Stumpf, Konsonanz und Dissonanz. Beitr. zur Akustik u. Musikwissensch. Leipz. 1898. S. 68.