EP1239453B1 - Method and apparatus for generating sound signals - Google Patents

Abstract

The method involves generating or selecting a time variable signal (rn(t)), generating or selecting an imaging function (f) that forms an image of the signal as a sequence of real numbers and computing a sound signal (o(t)) by imaging the time variable signal using the imaging function. The time variable signal and/or the sound signal is computed in time discrete steps. Independent claims are also included for the following: an arrangement for implementing the method.

Description

The invention relates to a method for generating a sound signal according to the preamble of claim 1. The invention further relates to a device for generating a sound signal according to the preamble of claim 11.

It is known to calculate sound signals using mathematical functions. Thus, for example, a time-varying sound signal, such as a sine, triangle or square wave signal, can be calculated on the basis of the mathematical definition by calculating a digital value representing the amplitude of the sound signal at regular time intervals. The thus generated, digital sound signal is then converted into an analog sound signal and can, for example, with an electroacoustic transducer such as a speaker, be converted into an acoustic, perceivable by the human ear as a sound or as a sound vibration.

Known methods for generating sound signal have the disadvantage that only relatively simple sounds can be generated, or that the sounds can be changed to only a limited extent.

From the publication EP 1 005 017 In addition, a cross-fader is known to which two signals are supplied. With the aid of this fader, the amplitudes of the two signals can be changed and the sum of the two signals can be formed.

This fader is not suitable for a method for generating sound signal, since it can produce no novel sounds.

It is an object of the present invention to propose a method that allows to produce sophisticated sounds.

This object is achieved by a method comprising the features of claim 1 and having a device comprising the features of claim 11. The dependent claims 2 to 10 relate to further advantageous method steps.

The inventive sound signal generation method, which is also referred to as a sound synthesis method, allows to generate a variety of different sounds, both interesting, unusual sounds, as well as natural-sounding sounds. The inventive sound signal generation method allows to produce much more complex sounds than is possible with conventional synthesizers. In addition, the sound signal generation method offers various ways to influence the sounds, such as the tone, in addition to influence. The required for sound signal generation device can be very simple and inexpensive be configured, and comprises in a preferred embodiment, essentially a computer with memory, input and output means such as a keyboard and / or a computer mouse, and a corresponding control program or software for controlling the computer.

The inventive method for generating sound signal essentially comprises two sub-methods. In a first partial procedure becomes a time-variable, preferably digital signal r (t) is generated, which is mapped in a second sub-process using a mapping function f in a time-variable sequence of real numbers, which form a digital sound signal o (t), as an electrical signal or for example via a speaker can be issued.

• dass damit Klänge erzeugbar sind, welche mit anderen bekannten Verfahren nicht erzeugbar sind,
• dass eine Veränderung der Klänge in Echtzeit möglich ist,
• dass der Rechenaufwand zur Berechung des Klangsignals o(t) gering ist,
• und dass die Klänge über eine entsprechend ausgestaltete Schnittstelle interaktiv, zum Beispiel durch eine Handbewegung beziehungsweise über eine Computermaus, veränderbar sind.

The inventive sound signal generation method has the advantages

• that sounds can thus be generated which can not be produced by other known methods,
• that a change of the sounds in real time is possible,
• that the computational effort for the calculation of the sound signal o (t) is low,
• and that the sounds are interactively changeable via an appropriately configured interface, for example by a hand movement or via a computer mouse.

Fig. 1

ein Signalflussdiagramm einer Klangsignalerzeugungsvorrichtung;

Fig. 2

schematisch den Aufbau einer Klangsignalerzeugungsvorrichtung;

Fig. 3

ein erstes Beispiel eines zweidimensionalen, zeitvariablen Signals r(t);

Fig. 4

ein zweites Beispiel eines zweidimensionalen, zeitvariabeln Signals r(t);

Fig. 5

ein erstes Beispiel einer Abbildungsfunktion;

Fig. 6

ein zweites Beispiel einer Abbildungsfunktion;

Fig. 7

Beispiele für Abbildungsfunktionen und zeitvariable Signale;

Fig. 8

ein analoges und digitalisiertes Klangsignal;

Fig. 9

ein drittes Beispiel einer Abbildungsfunktion;

Fig. 10

ein viertes Beispiel einer Abbildungsfunktion.

The invention will be described below with reference to several embodiments. Show it

Fig. 1

a signal flow diagram of a sound signal generating device;

Fig. 2

schematically the structure of a sound signal generating device;

Fig. 3

a first example of a two-dimensional, time-variable signal r (t);

Fig. 4

a second example of a two-dimensional, time-variable signal r (t);

Fig. 5

a first example of a mapping function;

Fig. 6

a second example of a mapping function;

Fig. 7

Examples of mapping functions and time-varying signals;

Fig. 8

an analog and digitized sound signal;

Fig. 9

a third example of a mapping function;

Fig. 10

a fourth example of a mapping function.

1 shows schematically the signal flow diagram of a sound signal generating device 1. A signal generator 2 supplies a time-variable, digital signal r (t). In the illustrated embodiment n digital signal generators 2a, 2b, .. 2n are arranged, each having a separate partial signal r yields (t) _i, namely, the sub-signals r ₁ (t), r ₂ (t), ... rn (t) , The signal generators 2a, 2b, .. 2n are connected via electrical signal lines 5 to an imaging device 3. The imaging device 3 is connected via a signal line 6 and a high-pass filter 4 to an output line 7, to which an analog or digital sound signal o (t) is applied. After the high-pass filter 4, a digital-to-analog converter can be arranged so that an analog sound signal is applied to the output line 7, while the high-pass filter is designed as a digital filter. The range of values of the analog sound signal o (t) may be, for example, between -5 volts and 5 volts. The value range of the digital sound signal o (t) can be, for example, between -2 ¹⁵ and 2 ¹⁵ -1. The digital signal r (t) comprising the sub-signals r ₁ (t), r ₂ (t), ... r _n (t) can also be referred to as a multi-dimensional time-variable signal.

In a preferred embodiment, a time-variable, periodic, digital signal r (t) is generated in the signal generator 2, wherein its parameters, such as the waveform, the frequency, the center and / or the amplitude, can be predetermined. Middle is understood to be the offset or the analog DC voltage component. Thus, for example, in each signal generator 2a, 2b, .. 2n an individual partial signal r _i (t) can be generated. In the signal generator 2, however, a time-variable signal r (t) may also be permanently stored, for example in a memory designated as ROM (Read Only Memory).

In the first partial method for generating sound signal, the sub-signals r _i (t), ie the sub-signals r ₁ (t), r ₂ (t), ... r _n (t) are generated. In a second partial method, these partial signals r _i (t) are imaged into a sound signal o (t) using an imaging function f. This mapping function f maps the digital sub-signals r _i (t) as a discrete-time sequence of real numbers, this sequence having a DC component. Preferably, as is customary in digital technology, a digital value is generated at regular time intervals. After the high-pass filter 4 is a DC-free sound signal o (t) on.

If the sound signal generating device 1 comprises two and more sub-signals r _i (t), the time-variable signal can also be used as a vector r (t) representing the sub-signals r ₁ (t), r ₂ (t), ... r _n (t), so that for the sound signal o (t): $$\mathit{O}\left(\mathit{t}\right)=\mathit{f}\left(\overrightarrow{\mathit{r}}\left(\mathit{t}\right)\right)$$

The mapping function f thus has the property, the vector r (t) including the sub-signals r ₁ (t), r ₂ (t), ... r _n (t) in the sound signal o (t) transform. The n-dimensional vector r (t) is thus mapped by the mapping function f into a one-dimensional function o (t) which defines the sound signal or an audio signal.

• die einzelnen Teilsignale r_i(t) bezüglich Frequenz und/oder Amplitude und/oder Phase und/oder Offset veränderbar sind
• und/oder die Abbildungsfunktion f veränderbar ist
• und/oder die gegenseitige Lage der Teilsignale r_i(t) und der Abbildungsfunktion f veränderbar ist.

This method has the advantage, in contrast to conventional sound generation methods used in synthesizers, that there are various possibilities for influencing the sound, in which:

• the individual sub-signals r _i (t) are variable with respect to frequency and / or amplitude and / or phase and / or offset
• and / or the mapping function f is changeable
• and / or the mutual position of the sub-signals r _i (t) and the mapping function f is variable.

An advantage of the novel synthesis method lies in the hitherto unknown division of the control or influencing the sound parameters.

In the case of a one-dimensional time-variable signal r ₁ (t), the fundamental frequency of the sound signal o (t) preferably depends solely or essentially on the fundamental frequency of the signal r ₁ (t). For a multi-dimensional, time-variable signal r (t) depends both the fundamental frequency of the sound signal o (t) and other frequency components of the respective fundamental frequency of the individual sub-signals r ₁ (t), r ₂ (t), ... r _n (t) from.

It may prove to be an advantage that the spectrum of the output signal through the frequencies of the signal r (t) can be specified.

The spectrum of the sound signal o (t) is in addition to the properties of the signal r (t) essentially determined by the mapping function f. For example, indicates the signal r (t) and / or the mapping function f discontinuities or discontinuities, so is a sound signal o (t) with a demanding frequency spectrum expect. The frequency spectrum of the sound signal o (t) thus also depends on the properties of the imaging function f, in particular on the spatial spectrum of the imaging function f.

The two sub-methods for generating the sound signal o (t) will be described below with reference to several embodiments of a sound signal generating device 1 with two signal generators, i. explained for the special case of n = 2.

In the first partial method, a partial signal r _i (t) is generated.

Am Signalgenerator 2a ist die Frequenz ω und am Signalgenerator 2b die Frequenz ω und der Faktor k einstellbar. In diesem Ausführungsbeispiel weisen beide Teilsignale dieselbe Frequenz ω auf.In the first exemplary embodiment, the signal generator _{2 a} generates the digital, time-discrete partial signal r ₁ ( t ) = sin (ω t ) and the signal generator ₂ b generates the digital, time-discrete partial signal $r_2\left(t\right)=\cos (\omega t)+\frac{t}{k}\mathrm{,}$

At the signal generator 2a, the frequency ω and the signal generator 2b, the frequency ω and the factor k is adjustable. In this embodiment, both partial signals have the same frequency ω.

For the time-variable signal r (t) in vector representation applies $$\overrightarrow{\mathit{r}}\left(\mathit{t}\right)=\left(\sin \left(\mathit{.omega.t}\right).\cos \left(\mathit{.omega.t}\right)+\frac{\mathit{t}}{\mathit{k}}\right)$$

Let the mapping function f be for the transformation of two partial signals $$\mathit{f}(\mathit{x}\mathit{.}\mathit{y})=\frac{{\mathit{x}}^2{\mathit{<figure-callout\; id="2"\; label="y"\; filenames="imgf0001.png,imgf0004.png"\; state="\{\{state\}\}">y</figure-callout>}}^2}{1+({\mathit{x}}^2+{\mathit{<figure-callout\; id="2"\; label="y"\; filenames="imgf0001.png,imgf0004.png"\; state="\{\{state\}\}">y</figure-callout>}}^2{)}^2}$$

gilt somit: $o\left(t\right)=f\left(\overrightarrow{r}\left(t\right)\right)=f\left(x,,y\right)=\frac{x^2{y}^2}{1+{\left(x^2+y^2\right)}^2}$

In the second sub-procedure the time-variable signal becomes r (t ) mapped by the mapping function f to the sound signal o (t).
With x = r ₁ (t) = sin ( ωt ) and $y=r_2\left(t\right)=\cos (\omega t)+\frac{t}{k}$

thus: $O\left(t\right)=f\left(\overrightarrow{r}\left(t\right)\right)=f\left(x,,y\right)=\frac{x^2{\mathrm{<figure-callout\; id="2"\; label="y"\; filenames="imgf0001.png,imgf0004.png"\; state="\{\{state\}\}">y</figure-callout>}}^2}{1+{\left(x^2+{\mathrm{<figure-callout\; id="2"\; label="y"\; filenames="imgf0001.png,imgf0004.png"\; state="\{\{state\}\}">y</figure-callout>}}^2\right)}^2}$

In a preferred method, the calculation of r ₁ (t), r ₂ (t) and o (t) takes place digitally in time-discrete steps, and also simultaneously or simultaneously or substantially simultaneously. At a given time t1, the value of r ₁ (t1) and r ₂ (t1) is initially charged and subsequently determine directly the value of o (t1). Then, the value of the discrete-time time t1 + .DELTA.T is calculated by initially calculating the value of r ₁ (t1 + At) and r ₂ (t1 + At), and then immediately the value of o (t1 + At) is calculated. Thus, the values of r ₁ (t), r ₂ (t) and o (t) are constantly calculated, wherein the signals r ₁ (t), r ₂ (t) can be changed by interventions on the signal generator _{2 a, 2} b, where the effect achieved can be heard immediately via the sound signal o (t). In addition, the mapping function f can be changed by, for example, changing a factor in a mapping function, or by exchanging a given mapping function with another predetermined mapping function. The effect achieved thereby can also be heard immediately via the sound signal o (t). The sound signal o (t) can thus be interactively changed and adjusted as desired and personal taste.

The special case n = 2 has the particular advantage that the time-variable signal r ( t ) as well as the mapping function f are graphically representable.

FIG. 3 shows the time profile 18 of the time-variable signal r (t) = (sin (1.01 ωt ), cos (ω t )) with starting point 18a and end point 18b, wherein in the abscissa the sub-signal r ₁ (t) = x = sin (1.01ω t ) and in the ordinate the sub-signal r ₂ (t) = y = cos (ω t ) is shown.

mit Anfangspunkt 18a und Endpunkt 18b, wobei in der Abszisse das Teilsignal ${\mathrm{r}}_{\mathrm{1}}\left(\mathrm{t}\right)\mathrm{=}\mathrm{x}=\sin (\omega t)-\frac{t}{170}$

und in der Ordinate das Teilsignal ${\mathrm{r}}_2\left(\mathrm{t}\right)\mathrm{=}\mathrm{y}=\cos (\omega t)+\frac{t}{200}$

dargestellt ist.FIG. 4 shows the time profile 18 of the time-variable signal $\tilde{\mathit{r}}\left(t\right)=\left(\sin (\omega t)-\frac{t}{170}.\cos (\omega t)+\frac{t}{200}\right)$

with starting point 18a and end point 18b, wherein in the abscissa the sub-signal ${\mathrm{r}}_{\mathrm{1}}\left(\mathrm{t}\right)\mathrm{=}\mathrm{x}=\sin (\omega t)-\frac{t}{170}$

and in the ordinate the sub-signal ${\mathrm{r}}_2\left(\mathrm{t}\right)\mathrm{=}\mathrm{y}=\cos (\omega t)+\frac{t}{200}$

is shown.

beziehungsweise der Term $\frac{t}{200}$

die Mitte, auch als Offset oder Gleichspannungsanteil bezeichnet.At the same time the term forms $\frac{t}{170}$

or the term $\frac{t}{200}$

the middle, also referred to as offset or DC component.

mit dem Faktor a(t) multipliziert wird. Über den Faktor a(t) kann somit beispielsweise ein Anklang, ein Ausklang oder eine Anschlagstärke aufmoduliert werden.The time-variable signal r (t) can additionally be changed, for example, with respect to the amplitude, for example by the sub-signal ${\mathrm{r}}_{\mathrm{1}}\left(\mathrm{t}\right)\mathrm{=}\mathrm{x}=\mathrm{a}\left(\mathrm{t}\right)\sin (\omega t)-\frac{t}{170}$

is multiplied by the factor a (t). By way of the factor a (t), it is thus possible, for example, to modulate an attack, a conclusion or a velocity.

FIG. 5 shows the course 20 of the imaging function in a three-dimensional representation $$\mathit{f}(\mathit{x}\mathit{.}\mathit{y})=\frac{{\mathit{x}}^2{\mathit{<figure-callout\; id="2"\; label="y"\; filenames="imgf0001.png,imgf0004.png"\; state="\{\{state\}\}">y</figure-callout>}}^2}{1+({\mathit{x}}^2+{\mathit{<figure-callout\; id="2"\; label="y"\; filenames="imgf0001.png,imgf0004.png"\; state="\{\{state\}\}">y</figure-callout>}}^2{)}^2}$$

FIG. 6 shows in a three-dimensional representation the course 20 of a further imaging function
$f=f\left(x,,y\right)=\mathit{Section}\left(\frac{\sin \left(x\right)}{1+{\left(x+\mathrm{<figure-callout\; id="2"\; label="y"\; filenames="imgf0001.png,imgf0004.png"\; state="\{\{state\}\}">y</figure-callout>}\right)}^2}\right)$

The special case n = 2 has the further advantage that, for example, the representations of FIGS. 3 and 5 can be represented one above the other. This is an optical indication of the interaction between the time-varying signal r ( t ) and the mapping function f possible. The time-variable signal r (t) and the mapping function f can be changed by changing parameters, causing a change in the visual display. This results in the advantage that the user is not only acoustically audible to the generated sound signal o (t), but also visually displays the interaction of the sub-steps, which gives the advantage that the sound signal o (t) easier, more user-friendly and differentiated is adjustable. In a particularly advantageous embodiment, the parameters of the time-variable signal r ( t ) and the mapping function f are interactively changed via the visual display, for example by displaying the time-variable signal r ( t ) is shifted with respect to the representation of the mapping function f in the xy plane, which results in a change of the parameters and thus a change of the sound signal o (t).

As a mapping function f, not only could a mathematical function be used, as shown in FIG. 5 or 6, but an arbitrary three-dimensional surface. The mapping function f shown in FIGS. 5 and 6 is calculated digitally, the values being calculated precisely at the intersections of two straight lines running in the x and y directions. These digitally represented values of the mapping function f are also referred to below as a three-dimensional surface. The values between the Intersections are preferably interpolated in each case. An imaging function f, for example, is also an image, for example a photographic, digitized image whose color values or gray scale values form the value f of the imaging function f.

Fig. 2 shows schematically a sound signal generating device 1 for carrying out the inventive method. The illustrated embodiment shows a computer with a microprocessor 11, interfaces 13, user interfaces 14, memory 15, and a digital signal processor (DSP) 16, all of which communicate via a common data bus 12. The required parameters for specifying or for calculating the partial signals r _i (t) and the mapping function 2 are input via the user interface 14. The partial signal r _i (t) is then calculated by means of the DSP software executable in the memory 17 by the digital signal processor 16, and mapped to the sound signal o (t) via the mapping function f. With the aid of the software stored in the memory 17, a digital high-pass filter can also be realized so that the sound signal o (t) calculated by the digital signal processor 16 is supplied to the electrical signal line 7 via the interface 13, for example as an audio signal.

Fig. 7 shows with equations 1 to 7 further examples of mapping functions f, and with equations 8 to 13 further examples of time-variable signals r (t). The illustrated examples apply to the special case n = 2, ie a sound signal generating device 1 with two independent signal generators 2a, 2b. As shown in FIG. 1, the sound signal generating apparatus 1 may comprise an arbitrary number of n signal generators 2a, 2b, .. 2n, so that the vector describing the time-variable signal r (t) r (t) has n dimensions. Corresponding to this, the mapping function f is designed to be the n-dimensional vector r (t) again into a one-dimensional sound signal o (t).

8 shows the time profile of an analog sound signal g (t) as well as the temporal course of the same, digitized sound signal g [n], which consists of a sequence of digital support values which are spaced by the regular time duration ΔT. In Fig. 8, the values g [0], g [1] and g [5] are specifically indicated.

Fig. 9 shows the third example of a mapping function f. In an xy plane, a meandering track 21 is shown. The track 21 has a continuous series of mutually equally spaced points, with each of these points being assigned a support value g [n]. In FIG. 9, for example, the values g [0], g [1] and g [5] are specially marked. All support values g [n] of the sound signal g (t) shown in FIG. 8 are assigned along the track 21 in this way. These support values g [n], similar to those shown in three dimensions in FIGS. 5 or 6, form an imaging function f 3 that extends three-dimensionally over the xy plane. As already described, this mapping function can be used to obtain from a time-variable signal r (t) via the mapping o ( t ) = f ( FIG. r (t) ) to generate a sound signal o (t). In the mapping of the signal r (t) with the mapping function f shown in FIG. 9, the following peculiarity results: If the signal r (t) is chosen such that its values in the xy plane coincide exactly with those in FIG corresponds to the output points o (t), the digitized signal g (t) and the digital signal g [n]. Thus, it is possible to use the mapping function the digital sound signal g [n] produce. If the signal r (t) is selected to be slightly offset with respect to the track 21, as shown in FIG. 9 with the dashed line 22, the output signal o (t) has similarities to the sound signal g [n]. on. Depending on the distance between the track 21 and the track 22 determined by the signal r (t), the sound distortion of the sound signal g [n] will be of different magnitude. This method thus makes it possible to reproduce a sound signal g [n] as recorded as original sound, or to vary it by changing the parameters of the signal r (t) in a variety of ways.

The course of the track 21 may be defined in an x-y plane in a variety of ways. For example, Fig. 10 shows a spiral track 21 along which the digital values of the sound signal g [n] are plotted. The digital values are preferably arranged along the track 21 at equidistant intervals and thus define a three-dimensional surface consisting of a multiplicity of support values, or the imaging function f. In order to reproduce the original sound signal g [n], the signal r (t) for the mapping function f of the spiral track 21 shown in FIG. 10 must follow.

Claims (11)

1. A method of generating a sound signal (o(t)), comprising the steps:

   - generating or selecting a time variable signal (r(t)), wherein the time variable signal (r(t)) is formed as a vector (r(t)) comprising at least two time-variable part signals (r_i(t)), which part signals are each described by a function having at least one periodic component;

   - generating or selecting a non-linear imaging function (f) which images the time variable signal (r(t)) as a sequence of real numbers, wherein this imaging does not correspond to a linear combination of the time variable part signals (r_i(t));

   - and calculating the sound signal (o(t)) in that the time variable signal (r(t)) is formed by the imaging function (f).
2. A method in accordance with claim 1, characterised in that the time variable signal (r(t)) and/or the sound signal (o(t)) is/are computed in time discrete steps.
3. A method in accordance with any one of the preceding claims, characterised in that the sound signal (o(t)) is filtered using a high pass filter.
4. A method in accordance with any one of the preceding claims, characterised in that the wave shape and/or the frequency and/or the offset and/or the amplitude can be varied for each time variable part signal (r_i(t)).
5. A method in accordance with any one of the preceding claims, characterised in that the time variable signal (r(t)) and/or the imaging function (f) is/are shown graphically and can in particular be changed interactively.
6. A method in accordance with any one of the preceding claims, characterised in that the imaging function (f) is configured as a two-dimensional function and defines a three-dimensional area in space.
7. A method in accordance with claim 6, characterised in that the imaging function (f) is configured as a digital image whose colour values or grey scale values define the three dimensional area in space.
8. A method in accordance with claim 6 or claim 7, characterised in that the time variable signal (r(t)) comprises two time variable part signals (r_i(t)); in that the time variable signal (r(t)) is shown extending in a plane by the two time variable part signals (r_i(t)) each defining one dimension of the plane; and in that the time variable part signals (r_i(t)) and the imaging function (f) are shown together and can be mutually displaced interactively to vary the sound signal (o(t)) in this manner.
9. A method in accordance with any one of the claims 6 to 8, characterised in that a track (21) extending in an x-y plane is generated; and in that the imaging function (f) is formed by a digital sound signal (g[n]) whose values are entered along the track (21).
10. A method in accordance with claim 9, characterised in that the time variable signal (r(t)) is selected to extend along the track (21) such that the sound signal (g[n]) is generated via the imaging function (f).
11. An apparatus for the carrying out of the method in accordance with any one of the preceding claims, comprising a computer (11), software which, when operated on the computer, causes the computer to carry out all the steps of this method, as well as input and output means (13, 14) with which at least one electroacoustic converter for the acoustic output of the sound signal (o(t)) can be connected in a signal transmitting manner.

Images