DE102021000324A1 - Intuitive and acoustically realistic glissando using the keyboard - Google Patents

Abstract

Software evaluates control information from an electronic or virtual keyboard in real time to enable musicians to play a keyboard instrument in an intuitive glissando style. Based on the velocity of the keys, a program decides whether a monophonic glissando style of play or a regular polyphonic style of play should be played for the individual notes struck on the keyboard, and thus allows the musician to play chords and glissando transitions simultaneously on the keyboard to play. For notes that are played in a glissando style, the program uses the sequence of notes played by the musician to decide how quickly the respective glissando should be carried out from one note to the other. The actual sound generation is based on a large number of audio samples. An acoustically realistic glissando tone generation is achieved through a combination of pitch change and adequate crossfading of individual samples.

Description

TECHNICAL AREA

The invention represents a method that enables an intuitive and acoustically realistic glissando or portamento way of playing using a keyboard instrument, hereinafter referred to as a keyboard, for musicians on such an instrument.

BACKGROUND

In music, “glissando” denotes a continuously gliding transition from one pitch to another pitch, in particular often over a larger interval of several semitones or even several octaves.

The term “portamento” is used closely with this style of playing and overlaps with it. The difference in semantics between portamento and glissando is rather subtle and is particularly relevant in the classical notation of an orchestration, where glissando means the technique of gliding note transitions per se, whereas portamento expresses the fundamental necessity of this playing style for certain note transitions. In the context of this invention, this subtle difference can be neglected and both terms considered equivalent.

Occasionally, the term "true legato" is used as an alternative, since a smooth transition in pitch is also meant here, but so quickly that this smooth transition is usually not consciously perceived by the listener, nevertheless is very probably decisive for an acoustically natural perception of the played sequence of notes. The latter therefore plays a role in the synthetic sound generation, e.g. of string instruments based on audio samples, in order to achieve a more realistic sound generation of orchestral sounds.

OVERVIEW

The basic problem that arises with the above-mentioned ways of playing on a keyboard is that due to the discrete and rather rough division of pitches into individual keys on a keyboard, which is usually in semitone steps in well-tempered mood, the keyboard instrument is subject to the direct influence of the musician for smooth pitch transitions. The playing styles mentioned above originally come from musical instruments where the player has direct control of the exact pitch at all times, for example by precisely positioning or moving the finger on a violin string. Technical aids are therefore required to enable the musician to play this way on a keyboard in the most intuitive and at the same time self-determined way.

Apart from the intention of the player whether and which way of playing should be used how exactly, there is also the problem with a keyboard of a sound generation of the glissando way of playing that sounds as convincing as possible, in particular a realistic-sounding glissando imitation of real orchestra instruments by means of a keyboard. A convincingly realistic glissando sound generation is difficult to separate from the problem from [P5].

DETAILED DESCRIPTION

Most keyboards have a spring-loaded pitch bend wheel, which allows the musician to at least partially influence the predefined pitches of the individual keys. The pitch bend wheel varies the pitch of all currently active notes in an interval that is typically +- 2 semitones. This pitch bend wheel is only conditionally suitable for a glissando on a keyboard, for several reasons. On the one hand, the interval is very limited, usually +-2 semitones. This means that in practice a glissando over several semitones or even several octaves cannot be played by the musician. The pitch bend wheel is usually located on the left side next to the keyboard. In order to use it, the musician has to take his left hand off the keyboard and is thus severely limited in the notes that can be played. For example, playing a chord and a glissando sequence at the same time is impossible. In addition, the pitch bend wheel usually springs back into the neutral position after it is released. This can lead to undesirable post-tuning effects, since each note played is usually subject to a certain decay phase (in English "release stage") in the tone generation of the keyboard.

Due to the lack of direct control options for a glissando playing style, the actual, continuous pitch transition for glissando on keyboards is therefore usually indirectly, automatically interpolated for the pitches between the keys used by the musician, i.e. by electronics or a program. This means that the musician continues to hit the respective semitones on the keyboard with both hands, as is usual with a keyboard, and the automatic system performs the smooth pitch transition from the previously played note to the most recently played note on the keyboard in the tone generation. However, there is still the problem that the automation involved must nevertheless learn the intention of the musician when a glissando playing style is desired by the musician and when not, and which glissando time should be used for the automated glissando in detail.

The question of whether a glissando should be played is usually solved by selecting a glissando sound preset. However, this also means that the tone generation automatically performs a (monophonic) glissando playing style for all notes played, until the musician explicitly selects another sound preset. Other solutions use a specific controller, e.g. to be able to switch between monophonic glissando and regular polyphonic playing by pressing a specific button. However, as with the pitch bend wheel, there is also the problem here that the musician has to take his hand off the keypad to press the button. In addition, such a way of playing is very unnatural, i.e. the constant change between glissando and regular polyphonic articulation using a controller, is musically limiting and rather difficult to play, even if, for example, a pedal is used instead of a button.

Analogously, a controller such as a rotary control is often offered so that the musician can adjust the glissando time directly and supposedly steplessly while playing. With the latter approach, however, there is again the problem of an unnatural and extremely difficult way of playing for the musician. Occasionally, the glissando time can be controlled by the velocity of the keys, but the problem from [P9] whether and when a glissando should be played remains unsolved.

Originally, early electronic keyboards achieved a glissando effect by adjusting the fundamental frequency of their oscillators. However, this early sound generation, based purely on oscillators, was not very suitable for realistically reproducing the sound of real instruments such as those of an orchestra. For this reason, tone generation based predominantly on samples has become generally accepted for this purpose. A large number of digital sound recordings (in English "samples"), for example real orchestra instruments, are used. With sample-based sound generation, a glissando effect is usually achieved by adjusting the playback speed of the samples. This variable factor in the playback speed of the samples is called "pitch". High-resolution and therefore very fine pitch changes are achieved using suitable interpolation algorithms for the samples, which is also referred to as "resampling". The problem with pure pitch-based glissando tone generation is that the pitch change always means a shift in all frequencies of the sample. Decisive for the human acoustic perception are the so-called "formant frequencies", which are also shifted with such a pitch change of the samples in contrast to the pitch adjustment of a real instrument or a human singing voice, for example. As a result, when the pitch is changed beyond a certain value, a clearly unnatural sound character arises. The existing solutions of a sample-based glissando sound generation are therefore also perceived as clearly unnatural and unrealistic with a glissando interval of several semitones or more.

These previously mentioned approaches for glissando using the keyboard, in particular from [P7] to [P11], are currently state of the art. In summary, it can be said that there are currently solutions for individual sub-areas, but so far no approach that solves all the challenges of glissando using the keyboard simultaneously and convincingly.

The invention described here uses the velocity of the respective key to decide whether a glissando should be played or not. With a low velocity, a glissando style of playing is performed by the tone generator for the note struck. In this case, an automated, gliding pitch transition from a previously played glissando note to the most recently played glissando note is interpolated by the tone generation. On the other hand, with a high velocity, a regular polyphonic way of playing is assumed for that key by the tone generation. In this case, the note is output acoustically without glissando by the tone generator, i.e. the pitch is not smoothly interpolated. This solution aspect of the invention has a decisive advantage: the musician can now simultaneously combine notes with glissando and notes without glissando at the same time, and without having to take his hands off the keyboard or use another unnatural aid when playing. The musician can, for example, grab a chord with one hand at any time and at the same time play a gliding sequence of notes with the other hand ( 1.1 and 1.2 ), or instead of a chord, play a sequence of single notes with hard phrasing, such as an arpeggio ( 2.1 and 2.2 ). Since the two-part division of the attack speed range used here is only a very rough one If the musician needs to exert influence, this approach can be played very easily and intuitively by the musician.

The invention keeps a cursory history of previously and currently played keys, as well as their velocity, to solve your problems. The played keys are separated from each other in the history in glissando and non-glissando notes. In the following, the non-glissando notes are henceforth referred to as "polyphonic"notes; in the context of this invention, this term is to be understood independently of the actual number of notes. A glissando progression is preferably performed between the set of glissando notes in the history, and not between the set of notes thus previously classified as polyphonic. This prevents, among other things, that a chord played with one hand would have an undesired influence on a glissando progression that is possibly being played with the other hand in a time-overlapping manner ( 1.1 and 1.2 ). Otherwise, this could lead to the glissando sequence possibly drifting several octaves into the chord range played, which would not correspond to the natural intention of the musician. This solves the same problem, for example, when playing a hard bass line with the left hand and a time-overlapping soft glissando sequence with the right hand (similar to 2.1 and 2.2 ).

An explicit exception to the separate consideration of the history between glissando notes and polyphonic notes described above is made when starting a new glissando sequence. When starting a completely new glissando sequence, the history of the polyphonic notes is exceptionally accessed if there is no suitable previously played glissando note. This allows the musician, for example, to play a fast, hard sequence of notes in a lower octave, pause briefly after releasing all the keys, and then, by softly hitting a single key in an upper octave range, automate a glissando to that note over the octaves in between let the sound generation run ( 3.1 ). This behavior is not erratic, because the musician still has full control over whether he wants this glissando course in this case. If the musician instead wanted a new glissando progression from the new note after the rest, and not a glissando progression from before the rest, he would simply play the first note after the rest with a faster velocity, and the following ones Notes of the glissando sequence with a slow velocity ( 3.2 ). This also remains intuitively accessible with maximum control for the musician.

The problem of the variable glissando time is solved in the invention by evaluating the note history described above and thereby automatically adjusting the glissando time. In the case of fast-played note sequences, a fast glissando time is automatically selected by the automatic system, while in the case of slow note sequences, a slow glissando time is selected. Although automation is involved, the player still retains control over the resulting glissando speed, and since the glissando time is not governed by some abstract and time-alien source such as the position of a knob or similar, it is temporal is derived directly from the timing of the musician's played notes, this represents an intuitive and easily playable, yet self-determined approach. Controllers such as knobs are provided to fine-tune this behavior, for example to limit the minimum and maximum glissando time of the automatic if necessary , or to regulate the influence of the played sequence of notes on the glissando time ( 5 ). In contrast to existing approaches with, for example, a rotary control that would directly control a glissando time, the approach presented here has the decisive advantage that these controls only have to be used sporadically for fine adjustment by the musician, for example once when preparing for a complete musical piece , but not or very rarely during a piece of music.

In addition, the invention uses an optional controller (e.g. a simple button) that lets the musician decide, if required, whether the determined glissando time should apply to the entire interval of the glissando, or be proportional to the size of the glissando interval ( 5 ). If we assume that the player would play note sequences consistently at a constant speed, then in the first case a glissando would always be carried out in the same time by the tone generation, regardless of whether it is a glissando of a semitone or even a glissando over several octaves would. In this case, the total glissando time would always be constant, but the glissando speed would depend on the interval. In the second case, on the other hand, the glissando time would be constant per semitone and the glissando speed would instead always be constant, and accordingly the sound generation would complete the glissando in an overall shorter time for small intervals and in a longer time for larger intervals.

The invention described here is also based on tone generation using samples. Thereby a realistic and universal sound spectrum is given. In contrast to existing solutions, the actual glissando effect is not only achieved by changing the pitch of the samples, but also by combining it with sample crossfades.

In order to achieve this, the tone generation first receives the core variables of the glissando sequence to be played by determining as already described in [P13] to [P17], in particular the information from which note to which other note the respective glissando sequence by the tone generation should be interpolated, and with which final speed or time the glissando interval should be carried out. The tone generator then divides this total glissando interval into an appropriate number of intermediate intervals. Finally, these individual intermediate intervals are played sequentially by the tone generator until the desired overall interval is finally completed. Each intermediate interval is assigned a sample appropriate to the pitch range of the intermediate interval. The aim of dividing into sufficiently small intermediate intervals and assigning a suitable sample for each intermediate interval is that the tone generation can carry out the entire glissando without major pitch changes, in order to avoid unnatural sound effects due to general frequency shifts.

Within an inter-interval, the sample associated with the inter-interval is slid from the bottom to the top of its inter-interval by pitch change. The transition from one intermediate interval to the next is achieved by fading the sample of the previous intermediate interval to the sample of the next intermediate interval. This crossfade is done with an envelope curve that is individually tailored to the samples in order to avoid erasing or boosting individual frequency ranges of the samples. A simple linear envelope, for example, is often not acoustically appropriate in practice.

character list

• 1.1 zeigt ein Beispiel bei dem der Musiker zunächst mit der linken Hand im Bassbereich einen Dreiklang spielt und gleichzeitig mit der rechten Hand in einer höheren Oktave die Note E. Die linke Hand schlägt hierbei die drei Noten des Akkords mit hoher Anschlagsgeschwindigkeit (Dynamikangabe „ff“) auf dem Tastenfeld des Keyboards an, hingegen die rechte Hand die Note E mit geringer Anschlagsgeschwindigkeit (Dynamikangabe „pp“). Als nächstes betätigt der Musiker mit der rechten Hand das nächst höhere C auf dem Keyboard, ebenfalls mit geringer Anschlagsgeschwindigkeit („pp“). Die Tonerzeung führt nun automatisiert ein Glissando von der Note E hin zur Note C durch. Obwohl im ersten Schritt bereits 4 Noten aktiv waren, erkennt die Automatik von welcher Note genau der Musiker das Glissando wünscht, da in der Historie der Dreiklang der linken Hand zum Zweck eines Glissandos durch die Automatik ignoriert werden kann, denn der Dreiklang wurde mit hoher Anschlagsgeschwindigkeit durch den Musiker gespielt.
• 1.2 zeigt ein Beispiel mit nur geringer Abweichung zum Beispiel aus 1.1: Der Dreiklang der linken Hand wird hier in 1.2 mit etwas zeitlicher Verzögerung gespielt, also genau zwischen den zwei Noten der rechten Hand. Auch hier kann die Automatik verlässlich das durch den Musiker beabsichtigte Glissando erkennen und in der Tonerzeugung dementsprechend interpolieren. Der Musiker ist also keiner zeitlichen Synchronisierungsanforderung zwischen linker und rechten Hand unterworfen.
• 2.1 verdeutlicht den in 1.2 beschriebenen Freiheitsgrad des Musikers. In 2.1 spielt der Musiker eine Melodie (ein Arpeggio in Staccato Spielweise) mit der rechten Hand, während er getrennt mit der linken Hand eine Abfolge von mehreren miteinander verbundenen Glissandi spielt. Der Musiker schlägt in diesem Beispiel also 4 Tasten mit der linken Hand auf dem Keyboard an, und die Tonerzeugung interpoliert automatisch die gleitende Tonhöhenanpassung zwischen diesen 4 Noten, so das sie für den Zuhörer wie eine durchgängig gezogene Note erscheint, dessen Tonhöhenverlauf lediglich durch diese 4 gewählten Noten wie vermeintlich einem Finger auf einer Violinsaite geführt, erklingt. Die zeitliche Überschneidung der gespielten Noten zwischen linker und rechter Hand, und die dennoch durch die Automatik korrekte separate Behandlung wird in diesem Beispiel besonders deutlich.
• 2.2 verdeutlicht den Vorteil der vorgestellten Lösung gegenüber anderer möglichen Lösungen noch mehr. In diesem Beispiel spielt der Musiker eine Abfolge von Glissandi Überkreuz. D.h er spielt mit der rechten Hand wie im vorherigen Beispiel eine eigenständige Melodie im mittleren Oktavbereich des Keyboards. Jedoch spielt er in diesem Beispiel jetzt mit der linken Hand zunächst eine Note unterhalb (links von) der rechten Hand, anschließend jedoch greift die linke Hand über die rechte und spielt also oberhalb (rechts von) der rechten Hand ihre zweite Note. Die Tonerzeugung erkennt aufgrund der gewählten Anschlagsdynamik, dass die obere Note zur vorherigen, unteren Note gehört, und nicht zur den Noten der rechten Hand, und spielt dementsprechend automatisch ein Glissando das die Noten der rechten Hand durchquert, ohne dass hier eine unerwünschte Wechselwirkung zwischen den beiden getrennt beabsichtigten Notensätze der Hände entstehen würde. Anschließend wiederholt sich diese Überquerung der Hände und der generierten Glissandi noch zwei mal. Eine alternativ denkbare Lösung durch positionsabhänge Betrachtung würde in diesem Beispiel scheitern.
• 3.1 zeigt ein Beispiel der in [P15] beschriebenen Ausnahmebehandlung. In den vorherigen Beispielen waren die Glissandi-Noten von den polyphonen Noten klar durch die Anschlagsgeschwindigkeit voneinander trennbar. In 3.1 spielt der Musiker mit der linken Hand zunächst ein kurzes, tiefes A mit hoher Anschlagsgeschwindigkeit auf dem Keyboard (eine polyphone Note). Anschließend lässt er diese Taste los und setzt eine kurze Pause ein. In diesem Augenblick sind also keine Tasten mehr gedrückt und keine Note aktiv hörbar. Nach Verstreichen der Pause betätigt der Musiker mit der rechten Hand nun ein hohes F mit geringer Anschlagsgeschwindigkeit. Mangels passender Glissando Note in der Historie führt nun die Automatik eine Rückfallbehandlung durch und wählt aus der Historie eine polyphone Note anstatt einer Glissando Note. Die Tonerzeugung führt dementsprechend nun ein Glissando von fast 4 Oktaven vom tiefen A hin zum hohen F aus. Zu beachten sei, dass in diesem Beispiel der Musiker tatsächlich nur 2 Tasten auf dem Keyboard betätigt hat. Die in 3.1 markierte dritte, mittlere Note wird durch die Tonerzeugung automatisch generiert um das Glissando nach der Pause wie gewünscht durchzuführen. Eine andere Besonderheit in diesem Beispiel ist, dass die polyphonen und die Glissandi-Noten auch hier nun nicht mehr klar rein positionell von einander trennbar wären. D.h. auch hier zeigt sich die Stärke der Erfindung, dass sie stets in der Lage ist beliebige Über-Kreuz-Spielweisen intuitiv und zuverlässig zu ermöglichen.
• 3.2 zeigt eine Abwandlung der Intention des Musiker im Bezug auf 3.1. Im vorherigen Beispiel aus 3.1 war die Intention des Musikers nach der Pause ein Glissando von der ersten zur zweiten Taste zu spielen. In 3.2 betätigt der Musiker die exakt gleichen zwei Tasten auf dem Keyboard wie in 3.1, ebenfalls mit Pause dazwischen, nur sei hier nun die Intention des Musikers nach der Pause eben kein Glissando von der ersten zur zweiten Note zu spielen, sondern stattdessen nach der Pause ein neues Glissando vom oberen F zum darunterliegenden D. Dies erreicht der Musiker indem er die zweite Taste (das hohe F) mit hoher anstatt mit geringer Anschlagsgeschwindigkeit betätigt, und anschließend die darauf folgende dritte Note (das darunterliegende D) mit geringer Anschlagsgeschwindigkeit.
• 4.1 bis 4.3 veranschaulichen den tonerzeugenden Aspekt der Erfindung, wobei als beispielhafte Grundlage wie in 4.1 dargestellt, die Intention des Musikers ein Glissando von der Note e im Bassbereich hinauf zur Note c', also ein Glissando um fast einer Oktave sein soll. Das Intervall dieses Glissando ist bereits hinreichend groß, im Sinne das andere Lösungen hier keinen akustischen Realismus mehr erreichen würden. Der Musiker spielt die beiden Noten auf dem Keyboard mit geringer Anschlagsgeschwindigkeit und die Tonerzeugung der Erfindung setzt dies akustisch wie folgt beschrieben um. Die Überlegenheit der Erfindung zeigt sich bei noch größeren Intervallen noch deutlicher, aber für die rein prinzipielle Betrachtung der technischen Vorgehensweise ist dieses Intervall als Beispiel ausreichend. Noch größere Intervalle werden gleichwertig wie hier beschrieben behandelt, nur dass die Anzahl der automatisch eingefügten Zwischenintervalle größer wäre.
• In 4.2 sei angenommen, dass für jeden Halbton im beabsichtigten Glissando-Intervall des Beispiels ein passendes Sample der Tonerzeugung zur Verfügung stehen würde. D.h. für jeden Halbton des Glissando-Gesamtintervalls (eine kleine Sexte) gäbe es jeweils ein Sample das ursprünglich in der gleichen Tonhöhe des jeweiligen Halbtons aufgenommen wurde. In diesem Fall würde die Tonerzeugung vorzugsweise das Gesamtintervall von einer kleinen Sexte automatisch in insgesamt 8 Zwischenintervalle von je einer kleinen Sekunde (also je ein Halbton) aufteilen. In diesem Beispiel spielt die Tonerzeugung zu jedem Zeitpunkt 2 Samples simultan ab. Die Tonerzeugung beginnt mit dem Abspielen der beiden Samples für die Noten e und f, wobei anfänglich nur das Sample e, nicht aber das Sample f hörbar ist. Nun blendet die Tonerzeugung kontinuierlich im Zeitraum des ersten Zwischenintervalls vom Sample e zu Sample f über. Gleichzeitig passt die Tonerzeugung den Pitch beider Samples kontinuierlich an. Der genaue Pitch-Wert beider Samples unterscheidet sich hierbei. Der exakte Pitch-Wert wird aus dem Alternationsquotienten des jeweiligen Samples berechnet, um akustisch genau die Tonhöhe des Samples zu erreichen, welche zum jeweils aktuellen Zeitpunkt des Glissando Verlaufs benötigt wird. Am Ende des ersten Zwischenintervalls ist das Sample e nun nicht mehr hörbar, stattdessen ist vollständig und ausschließlich das Sample f hörbar. Sobald das erste Zwischenintervall von der Tonerzeugung vollständig abgeschlossen wurde, geht es in die Behandlung des nächsten Zwischenintervalls über, d.h. es spielt nun genau die beiden Samples f und Sample f# simultan ab. Die Vorgehensweise wiederholt sich. D.h. zu Beginn des zweiten Zwischenintervalls ist ausschließlich das Sample f hörbar, nicht aber das Sample f#, und die Tonerzeugung blendet im Zeitraum des zweiten Zwischenintervalls entsprechend vom Sample f hin zu Sample f# über, mit gleichzeitig kontinuierlicher Anpassung des Pitch-Wertes beider Samples. Dieser Vorgang wiederholt sich so oft bis letztendlich das Gesamtintervall und damit die Tonhöhe der Note c' erreicht wurde.
• 4.3 zeigt im Grunde das gleiche Beispiel, jedoch unter leicht unterschiedlichen Gegebenheit. Auch in 4.3 sei das in 4.1 dargestellte Ziel des Musikers ein Glissando von der Note e hinauf zur Note c' zu spielen. Jedoch sei nun angenommen, dass nicht für jeden Halbton ein eigenständiges Sample der Tonerzeugung zur Verfügung stehen würde. Dies ist eine realistisches Szenario, da ein eigenständiges Sample pro Halbton die mögliche Speicherkapazität vieler Produkte übersteigen würde bzw. dies aus anderen ökonomischen Gründen oft nicht möglich ist (z.B. verfügbares Budget für Orchester-Aufnahmen). In diesem Szenario teilt die Tonerzeugung automatisch das Gesamtintervall in eine geringere Anzahl von Zwischenintervallen auf als im vorherigen Szenario beschrieben. Die genaue Anzahl der Zwischenintervalle, deren Größe und zeitlichen Länge wird abhängig vom zur Verfügung stehenden Sample-Materials automatisch durch die Tonerzeugung gewählt. In diesem Beispiel sei für die anfängliche Note e kein Sample vorhanden. Die Tonerzeugung wählt daher stattdessen als Alternative das nahe liegende Sample des eigentlich darunterliegenden d# und beginnt dessen Wiedergabe mit einem dementsprechend nicht neutralem Pitch-Wert um bereits zu Beginn die Tonhöhe e damit zu erreichen. Für das Ziel des ersten Zwischenintervalls steht jedoch ein Sample für f zur Verfügung also werden diese beiden Samples im ersten Zwischenintervall abgespielt und kontinuierlich überblendet, mit kontinuierlicher Pitch-Änderung beider Samples, wie bereits im vorherigen Abschnitt genau beschrieben. Auch bei den darauf folgenden Zwischenintervallen muss die Tonerzeugung mangels passender Samples Kompromisse eingehen und wählt alternativ vorhandene Samples mit möglichst naher Tonhöhe aus. In diesem Beispiel ist es möglich, dass nicht zu jedem Zeitpunkt immer genau zwei Samples simultan spielen. Die Tonerzeugung wählt automatisch eine angemessene Abgrenzung der Zwischenintervalle gemäß der verfügbaren Samples.
• 5 zeigt beispielhafte optionale Kontrollmöglichkeiten die dem Musiker angeboten werden könnten um das beschriebene Verhalten der Erfindung zu justieren, z.B. auf oder an einem Keyboard oder aber einer grafischen Benutzeroberfläche einer Software. Es wäre z.B. denkbar drei Drehregler und einen Wechselschalter dem Musiker anzubieten. Die drei Drehregler würden den Einfluss und die Grenzen der automatisch aus der zuvor durch den Musiker gespielten Notensequenz ermittelten Glissando-Zeit optional regulieren, falls denn erwünscht. Der Schalter hingegen würde der Tonerzeugung mitteilen, ob die final ermittelte Glissando-Zeit pro Halbton oder aber für das vom Musiker gespielte Gesamtintervall gelten soll. Es sei zu erwähnen, dass dies lediglich ein Beispiel möglicher Controller darstellen soll. In der Praxis würden je nach Anforderung leicht abweichende Konfigurationen in Produkten verwendet werden. Z.B. wäre es denkbar zum Vorzuge der Intuitivität die Anzahl der Drehregler zu reduzieren bzw. deren Funktionalität zusammenzufassen. Auch die genaue Form bzw. Ausführung der Controller sei lediglich beispielhaft. Es wären durchaus andere Kontrollmöglichkeiten wie z.B. kapazitive Touch-Felder, Touch-Screens, Pedale, Schieberegler u.ä. denkbar.

Several drawings follow for a better understanding of the invention and to show how a musician can concretely implement individual glissando playing styles using a keyboard with the invention. The examples are written in musical notation. It should be noted that the dynamic indications "pp" and "ff" in the context of these examples are not to be understood as volume indications, but rather as attack speeds that the musician would play on the keyboard for the individual notes in order to achieve the tonal result in to achieve the examples with the invention. The dynamic specification "pp" means a low velocity and "ff" accordingly a high velocity of the key to be pressed on the keyboard.

• 1.1 shows an example in which the musician first plays a triad with his left hand in the bass range and at the same time plays the note E with his right hand in a higher octave. on the keyboard, while the right hand plays the note E with a low velocity (dynamics indication "pp"). Next, the musician presses the next higher C on the keyboard with the right hand, also with a slow velocity ("pp"). The tone generation now automatically performs a glissando from the note E to the note C. Although 4 notes were already active in the first step, the automatic recognizes from which note exactly the musician wants the glissando, since in the history the triad of the left hand for the purpose of a glissando can be ignored by the automatic, because the triad was played with a high velocity played by the musician.
• 1.2 shows an example with only slight deviation from example 1.1 : The left-hand triad is used here in 1.2 played with a slight delay, i.e. exactly between the two notes of the right hand. Here, too, the automatic system can reliably recognize the glissando intended by the musician and interpolate accordingly in the tone generation. The musician is thus not subject to any temporal synchronization requirement between the left and right hands.
• 2.1 clarifies the in 1.2 described degree of freedom of the musician. In 2.1 the musician plays a melody (an arpeggio played staccato) with the right hand while separately playing a sequence of several connected glissandi with the left hand. In this example, the musician hits 4 keys on the keyboard with his left hand, and the tone generator automatically interpolates the sliding pitch adjustment between these 4 notes, so that it appears to the listener as a continuously drawn note whose pitch progression is only determined by these 4 selected notes as if supposedly guided by a finger on a violin string. The temporal overlapping of the notes played between the left and right hand, and yet by the automatic correct separate treatment is particularly evident in this example.
• 2.2 clarifies the advantage of the presented solution compared to other possible solutions even more. In this example, the musician plays a series of crossed glissandi. This means that he plays an independent melody in the middle octave range of the keyboard with his right hand, as in the previous example. However, in this example he now plays a note below (to the left of) the right hand with his left hand, but then the left hand reaches over the right hand and thus plays its second note above (to the right of) the right hand. Due to the selected touch dynamics, the tone generation recognizes that the upper note belongs to the previous, lower note and not to the notes of the right hand, and accordingly automatically plays a glissando that crosses the notes of the right hand without any unwanted interaction between the notes both separately intended scores of the hands would arise. This crossing of the hands and the generated glissandi is then repeated two more times. An alternative conceivable solution using position-dependent consideration would fail in this example.
• 3.1 shows an example of the exception handling described in [P15]. In the previous examples, the glissandi notes were clearly separable from the polyphonic notes by velocity. In 3.1 the musician first plays a short, low A with the left hand on the keyboard with high velocity (a polyphonic note). He then releases this button and pauses for a moment. At this moment, no keys are pressed and no notes are actively audible. After the pause has elapsed, the musician now hits a high F with his right hand at a low velocity. In the absence of a suitable glissando note in the history, the automatic system now carries out a fallback treatment and selects a polyphonic note from the history instead of a glissando note. Accordingly, the tone generation now performs a glissando of almost 4 octaves from low A to high F. It should be noted that in this example the musician actually only pressed 2 keys on the keyboard. In the 3.1 The marked third, middle note is automatically generated by the tone generator in order to carry out the glissando as desired after the pause. Another peculiarity in this example is that the polyphonic and glissandi notes could no longer be clearly separated from one another purely positionally. This means that the strength of the invention is also evident here, in that it is always able to intuitively and reliably enable any crossed ways of playing.
• 3.2 shows a modification of the musician's intention in relation to 3.1 . In the previous example 3.1 the musician's intention after the break was to play a glissando from the first to the second key. In 3.2 the musician presses exactly the same two keys on the keyboard as in 3.1 , also with a pause in between, but here the intention of the musician is not to play a glissando from the first to the second note after the pause, but instead to play a new glissando from the upper F to the D below after the pause. The musician achieves this by the second key (the high F) is played at high velocity instead of low, and then the third note that follows (the D below it) at low velocity.
• 4.1 until 4.3 illustrate the sound-producing aspect of the invention, using as an example basis as in 4.1 shown, the intention of the musician is a glissando from the note e in the bass range up to the note c', i.e. a glissando of almost an octave. The interval of this glissando is already large enough that other solutions would no longer achieve acoustic realism here. The musician plays the two notes on the keyboard with slow velocity and the tone generation of the invention converts this acoustically as described below. The superiority of the invention becomes even clearer with even larger intervals, but this interval is sufficient as an example for a purely fundamental consideration of the technical procedure. Even larger intervals are treated equally as described here, except that the number of intermediate intervals automatically inserted would be larger.
• In 4.2 assume that for each semitone in the intended glissando interval of the example, an appropriate sample of tone generation would be available. That is, for each semitone of the total glissando interval (a minor sixth), there would be a sample that was originally recorded at the same pitch of the respective semitone. In this case, the tone generator would preferably automatically divide the total interval of a minor sixth into a total of 8 intermediate intervals of a minor second each (i.e. a semitone each). In this example, the tone generator plays 2 samples simultaneously at any point in time. The tone generation starts with playing the two samples for the notes e and f, initially only the sample e but not the sample f is audible. Now the sound generation fades continuously from sample e to sample f in the period of the first intermediate interval. At the same time, the tone generation continuously adjusts the pitch of both samples. The exact pitch value of both samples differs here. The exact pitch value is calculated from the alternation quotient of the respective sample in order to acoustically achieve the exact pitch of the sample that is required at the current time of the glissando process. At the end of the first intermediate interval, sample e is no longer audible, instead sample f is fully and exclusively audible. As soon as the first intermediate interval has been completely completed by the sound generation, it proceeds to the treatment of the next intermediate interval, ie it now plays exactly the two samples f and sample f# simultaneously. The procedure is repeated. This means that at the beginning of the second intermediate interval, only sample f is audible, but not sample f#, and the tone generation fades over from sample f to sample f# during the second intermediate interval, with simultaneous continuous adjustment of the pitch value of both samples. This process is repeated until finally the total interval and thus the pitch of the note c' has been reached.
• 4.3 shows basically the same example, but under slightly different circumstances. Also in 4.3 be that in 4.1 The musician's goal is to play a glissando from the note e up to the note c'. However, it is now assumed that an independent sample of the tone generation would not be available for each semitone. This is a realistic scenario, since a separate sample per semitone would exceed the possible storage capacity of many products or this is often not possible for other economic reasons (eg available budget for orchestra recordings). In this scenario, the tone generator automatically divides the overall interval into a smaller number of intermediate intervals than described in the previous scenario. The exact number of intermediate intervals, their size and length of time is automatically selected by the tone generation depending on the available sample material. In this example, there is no sample for the initial note e. As an alternative, the tone generator therefore selects the nearby sample of the actually underlying d# as an alternative and begins its playback with a correspondingly non-neutral pitch value in order to reach the pitch e right from the start. However, for the target of the first intermediate interval, one sample is available for f, so these two samples are played and continuously crossfaded in the first intermediate interval, with continuous pitch change of both samples, as already described in detail in the previous section. Due to a lack of suitable samples, the tone generation also has to make compromises in the subsequent intermediate intervals and selects alternatively available samples with the closest possible pitch. In this example it is possible that not exactly two samples are playing simultaneously at all times. The tone generation automatically chooses an appropriate delimitation of the intermediate intervals according to the available samples.
• 5 shows exemplary optional control options that could be offered to the musician to adjust the described behavior of the invention, for example on or at a keyboard or a graphical user interface of a software. It would be conceivable, for example, to offer the musician three knobs and a toggle switch. The three knobs would optionally regulate the influence and limits of the glissando time, automatically determined from the sequence of notes previously played by the musician, if desired. The switch, on the other hand, would tell the tone generator whether the finally determined glissando time should apply per semitone or for the entire interval played by the musician. It should be noted that this is only intended to represent an example of possible controllers. In practice, depending on the requirement, slightly different configurations would be used in products. For example, it would be conceivable to reduce the number of rotary controls or to combine their functionality for the benefit of intuitiveness. The exact form and design of the controller is also only an example. There would certainly be other control options such as capacitive touch fields, touch screens, pedals, sliders, etc. conceivable.

Claims (3)

Evaluation of the velocity or an equivalent alternative control information (e.g. pressure or contact surface of a finger) of individual keys on a physical or virtual keyboard in order to enable the musician to choose between glissando notes and polyphonic (non-glissando) notes intuitively and easily at any time in individual cases . Collection of a history of the notes already played on the keyboard by an automated one Derivation of core quantities for a glissando tone generation using a keyboard. 2.1. Bifurcation of history into glissando notes and polyphonic (non-glissando) notes eg according to claim 1 , to enable simultaneous polyphonic glissando playing on a keyboard, i.e. simultaneous or time-overlapping playing of a glissando sequence and several individual notes such as chords or arpeggios, without undesired interaction between the set of polyphonic notes and the set of glissando notes. 2.2. Fallback consideration of polyphonic (non-glissando) notes in history at the beginning of a new glissando sequence in the absence of an appropriate glissando note in history to combine with claim 1 to enable the musician in a new glissando sequence to choose an immediate glissando sequence introduction up to the first note of the new glissando sequence. 2.3. Evaluation of the note sequence in the history in order to achieve an automated, temporally adequate derivation of the final glissando time while maintaining the musician's control in an intuitive way. 2.3.1. Controllers like knobs allow the musician to set the minimum and maximum glissando time and influence of in claim 2 .3 mentioned function can limit or regulate. 2.3.2. Controller such as a button so that the musician can choose, if required, whether the determined glissando time should be interpreted for the total interval of the respective glissando sequence or instead proportionally to the total interval by the tone generator and processed accordingly. Realistic sample-based glissando sound generation through a combination of pitch change and sample crossfading. 3.1. Division of the overall glissando interval into appropriate, as small as possible intermediate intervals with adequate assignment of a suitable sample for each of the individual intermediate intervals, with the aim of always requiring only small pitch changes in the entire glissando tone generation in order to consistently avoid an unnatural sound character. 3.2. Sequential treatment of the generated intermediate intervals in the glissando tone generation according to claim 3 .1 to ultimately achieve the desired total glissando interval. 3.3. Performing the glissando tone generation of the in claim 3 .1 specified intermediate intervals by automated, continuous pitch change according to the respective intermediate interval claim 3 .1 assigned samples. 3.4. Reconciliation of the individual intermediate intervals according to claim 3 .1 by sample crossfading. 3.4.1. Automatic selection of an adequate envelope for the in claim 3 .4 mentioned crossfading to prevent both erasure and boosting of individual frequency ranges of the samples during crossfading.

Images