FR2826771A1 - STUDIO-TYPE GENERATOR COMPRISING A PLURALITY OF SOUND REPRODUCTION MEANS AND METHOD THEREOF - Google Patents

Abstract

The invention relates to a studio-type generator (10) that is intended to produce sounds that can be represented by a series of sound events. The inventive generator (10) comprises: an inlet for receiving (12) digital orders of play for sound pieces; a bank of sound samples (16, 18, 20); means for selecting (14) samples according to the orders of play received; and means for transmitting the selected samples to numerous (22) reproduction means. The aforementioned means are such that, for one same sound and one same digital level for said sound, all the means supply more or less the same analogue output level.

Description

<Desc / Clms Page number 1>

 The invention relates to a method and a system for reading game orders of a musical work. It relates more particularly to a studio-type generator for producing sounds that can be represented by a score, such as musical and / or sung and / or spoken sounds and / or noises.

 By "partition" is meant here a representation of a sequence of sound events to be performed at specific times. "Sound event" means musical notes, and / or lyrics, and / or vocals and / or noises.

 A musical work can be played by electronic instruments. For this purpose, the scores or game orders must be written in a musical computer language understood by these instruments. Currently the most common musical computer language is the midi protocol.

The devices allowing such reproductions of musical works by electronic control are of various types:
Synthesizers are the most common devices. They have various sounds provided by the manufacturer and a midi interface that allows you to play these sounds after reading information of midi musical game orders received either from a piano (keyboard) included in the device, or from an external computer equipped with a midi interface and a midi recording software generally called sequencer. A sequencer constitutes a recorder (such as a tape recorder) of musical game orders.

 Expanders are synthesizer-like devices but do not have a keyboard.

 The samplers also allow you to record sounds or sound samples and have them played by an external sequencer.

 There are also hybrid devices with both synthesizer and sampler function.

 It is also possible to play a musical score recorded in midi format on some pianos sometimes called "disklaviers".

 There are also known synthesizers made in the form of electronic card (s) or plug-in sound card in computers.

 Finally, there are synthesizers in a purely software form. Such synthesizers are, for example, recorded on the hard disk of a computer.

<Desc / Clms Page number 2>

mesure, le jeu de la note, l'arrêt de la note, la hauteur de la note, la longueur de la note. These various devices or software, which use the GM mode (General Midi), thus provide at least 128 different sounds and the main functions of the musical game, namely: the tempo, the signature (4 / 4,3 / 4, etc. .), the location of the note in the

measure, the playing of the note, the stop of the note, the pitch of the note, the length of the note.

 The GM mode can also be used to assign performance parameters such as: vibrato, glissando, damp (without stopping a note, for example the sustain pedal of a piano) and tune (chord). GM mode also allows control of effects such as reverb, chorus (sound width), and echoes.

 However, the GM mode of the midi protocol is, in general, not rich enough to satisfy demanding musicians. Therefore, to this standard protocol, each device or software manufacturer adds other possibilities in the form of an additional "owner" protocol which, because of its proprietary nature, is generally not included by similar devices from other manufacturers.

 In addition, whether or not they work with a proprietary protocol, these various devices have a limited memory, which prevents storage of a large sound library. If one wants to record sound samples of high quality, and therefore with a high sampling frequency and of sufficient duration, it is necessary to have a large memory capacity of a relatively high cost.

 Thus, if a musician wishes to have a composition or a musical arrangement tested on electronic devices, he / she must contact a studio that has a plurality of devices and, in general, mixing means allowing to assign to each sound a particular sound treatment.

 In a studio, one can thus have a large memory capacity and a plurality of devices. However, the use of a studio requires minimal knowledge of the various devices made available. In addition, renting a studio is expensive.

 A first object of the invention is to facilitate the use of the apparatus and software of a studio. A second object of the invention (distinct from the first) is to reduce the cost of using a studio.

<Desc / Clms Page number 3>

 For this purpose, the invention provides according to a first aspect of a studio type generator for producing sounds that can be represented by a score such as musical sounds and / or sung and / or spoken and / or noise, which comprises: - a reception input of numerical orders for the playing of sound pieces, - at least one bank of sound samples, for example sound samples of musical instruments, samples of sung phonemes, and, possibly, sound samples. samples of various sound effects, means for selecting the samples according to the received game orders, and means for transmitting the selected samples to a plurality of reproduction means, these means being such that for the same sound and the same level Digital sound of this, they all provide substantially the same analog output level.

 With such a studio, the user is freed from tedious adjustments of each means of sound reproduction because the sound volumes of these various means are harmonized by construction.

 In one embodiment, the sound reproduction means are such that the general level of sound provided by each of these reproduction means depends on the level of the range of its dominant frequencies. Indeed, for the same volume, some sounds are perceived as having a higher volume than others. In general, the predominantly acute devices appear to provide more volume than the predominantly serious devices. Thus, in a preferred embodiment, the sound reproducing means are such that the general level of the supplied sound varies in the opposite direction of the dominant frequency range. However, for the highest dominant frequencies, in particular those higher than about 10 KHz, it is not necessary to make the general level dependent on these frequencies.

 Also in order to facilitate the use of an automatic studio, provision is made, in one embodiment, for a predetermined arbitrary mixing by a family of musical and / or sung sounds and, preferably, this predetermined arbitrary mixing is a function of a status signal of each musical sound and / or sung in an orchestration. For example, the leading song will generally correspond to the largest volume.

<Desc / Clms Page number 4>

 This predetermined mix will, for example, take precedence over sound adjustment commands included in the game order file.

 According to one embodiment, the families of sounds comprise at least two of the families taken together comprising: brass, wind, rubbed-type strings, human voices, guitars, basses, drums, sounds synthetics and various noises.

 The situation signals of the sound in the orchestration represent, in one embodiment, situations which are part of the set comprising: the leader song, the response to the lead vocal, the counter-counter-counter, the rhythmic chords, arpeggios, basses, tablecloths or mats and instrument dubbing for one or more orchestral parts.

 In one embodiment, there is provided means for introducing effects to the signals provided by the sound reproducing means and, preferably, means for introducing a predetermined effect to the signals provided by the sound reproducing means, this effect depending on a signal indicative of the type of music. For example, it will be possible to automatically introduce a reverberation effect if the indicative signal represents a so-called typical music. In one embodiment, there is provided a means for making the existence and / or the level of the effect of a sound situation signal depend on the orchestration in a predetermined manner for each orchestral part. .

 The effect may depend on a signal indicative of the type of music and / or a random draw of a sound effect.

 As with the predetermined mix, the predetermined effect is preferably prioritized over effect controls contained in the game orders.

 The sound reproducing means may be made either in software form, for example in the hard disk of a computer, or in the form of a device itself. They can be part of the group including: synthesizers, expanders and samplers.

 Banks of sound samples are, for example, stored in permanent memories.

 In a preferred embodiment, the generator constitutes a server, that is to say that it comprises means for remotely transmitting the signals provided by the

<Desc / Clms Page number 5>

 means of sound reproduction. In this case, it is preferable to provide means for digitizing and compressing the signals to be transmitted. In a particular application, the signals are transmitted to receivers, for example subscriber receivers. A particularly interesting application is the case where there is provided a means of automatic production of musical composition, for example adapted to each subscriber or each type of subscriber.

 It is also advantageous that game orders can be received by telecommunication means. In this case, it can be a server that can be enriched with sound samples and / or sound reproduction means and this automatically. Thus, the generator can also be a site through which musicians or music users pool their sound sample creations and / or their sound reproduction means to enrich the sound bank of the site.

 We recall here that sounds are any type of musical instrument, including the human voice and various sounds.

 In one embodiment, means are provided for mixing the signals from the sound reproducing means. In this case, the mixing means may comprise means for performing a volume balance and / or stereophonic placement and / or confer a general volume.

 The generator comprises, in one embodiment, a volume equalization control means which, for the same digital level of sound at the input of each of the reproduction means, determines the analog output level of each of these reproduction means. sound and which affects at the digital level at the input of each sound reproduction means a weighting allowing all these reproduction means to provide substantially the same analog output level.

 In this case, the generator preferably comprises means for determining the introduction of a new sound reproducing means and for operating said equalization adjusting means.

 The invention also relates to a method for producing sounds that can be represented by a score such as musical and / or sung and / or spoken sounds and / or noises, characterized in that it comprises the following steps;

<Desc / Clms Page number 6>

 receiving digital commands of sets of sound pieces, selecting samples in a bank of sound samples according to the received game orders, and transmitting the selected samples to a plurality of reproduction means, these means being such as for the same sound and the same digital level of this sound, they all provide substantially the same analog output level.

 This method is preferably implemented by means of a generator according to any one of the embodiments of the invention.

 Other features and advantages of the invention will become apparent with the description of some of its embodiments, this being done with reference to the accompanying drawings in which: FIG. 1 is a flowchart showing a server conforming to FIG. Figure 2 is a diagram showing a use of the server shown in Figure 1, and Figure 3 is a diagram illustrating a feature of the server shown in Figure 1.

 The system that will be described in connection with the figures mainly uses a server connected to a telecommunications system such as the Internet network to which can connect a number of users with personal computers.

 The server is intended to receive game orders of compositions or musical arrangements that are transmitted to it by computers and this server is able to render these game orders in compressed audio form or not.

Game order data includes:
1) The designation of a channel.

 2) For each channel, the designation of a sound sample that is either a sample of a musical instrument, or a sound sample, or a phoneme in the case of singing generation.

<Desc / Clms Page number 7>

 This sample designation includes, in the first place, an identification of a sound database in which the sample is located or an identification of a synthesizer or sampler. In the case of a databank, it may be of the midi type or of a particular type with a code specific to the studios, especially to online studios.

 The sample designation includes, secondly, an identification code of the sample concerned in the corresponding bank or synthesizer (or sampler).

 3) The musical playing orders, ie the musical parameters, namely the pitch of the notes, the locations of the notes to be played in the measures of the piece, the velocity eVou the volume of the notes, their durations, etc. .

 4) Mixing orders of the different selected sounds: the general volume, the stereophonic pan, and the volume level for each chosen instrument or sound.

 5) Digital or analog sound processing commands for each instrument or sound chosen. In analog processing order, reference is made here to analog devices having digital controls, for example apparatus for performing reverberations.

 To satisfy these orders of play, it is expected that the server is equipped with eVou devices software to transform the game orders into audio sounds to be broadcast to the applicants. These eVou software devices are: a) At least one midi sequencer intended to record and read the received game orders. b) A plurality of synthesizers, samplers and mixed synthesizer-amplifiers together forming a sound bank comprising, on the one hand, sounds of a plurality of musical instruments, on the other hand, a plurality of noises and, finally, phonemes allowing to reconstitute human voices. c) A mixing apparatus of the digital type allowing to adjust for each of the channels: the volume, of the various channels, that is to say their equilibrium so as to create, the stereophonic location and the equalization (adjustment of the frequencies sound of the way).

<Desc / Clms Page number 8>

 d) At least one processing device for each sound for performing digital effects. It is also possible, as indicated above, to control analog devices via the midi protocol. However, these will, of course, be equipped with a midi interface.

 The effects produced by these processing devices are, in particular: the reverberation to simulate the resonance of a room (the resonance depends on the dimensions of the room), the chorus widening in frequencies the sound by acting on the phases, the delay or echoes allowing the repetition of a sound, the spacing widening the range of stereophony, the harmonizing effect transposing towards higher and / or lower values of the given frequencies, the enhancer (effect allowing to create a "presence" sound by cutting certain frequencies and increasing others, which allows a marked separation of the various sounds), the compressor effect making it possible to level the different levels of sound volumes, the limiting effect clipping the volumes, etc.

 To take into account the difficulties of use of this server constituting an online studio, according to a feature of the invention, provision is made for predetermined settings of the mix.

 These predetermined settings consist in assigning to each large family of sounds, and possibly to their subgroups, a volume that is a function of the situation of the sound in the orchestral part that is intended for it.

 In the example, the various families of sounds and their subgroups are: - the brass (mono or poly subgroups), - the winds, - the stitched strings (allowing pizzicati) or the bowed strings, - the voice human (subgroups: men, women and children), - guitars, - basses (subgroups: acoustic or synth-tick), - percussion, and - various noises.

 The different parts of the orchestra are in an example: - the leader song, - the response to the leader song, - counterpointing it in counterpoint,

<Desc / Clms Page number 9>

 - rhythmic chords of a first type such as chords for rhythm guitar or accompaniment guitar, - rhythmic chords of a second type such as chords for rhythmic piano or piano accompaniment, - arpeggios, - the bass (in particular a double bass or an electronic bass), - the tablecloths or rugs, ie the notes held continuously and respecting the harmonic evolution of the piece, and - the instrument doublings for one or more parts orchestra. As an illustration, if the orchestral part is the leader song, it can be played by two different instruments, such as a trumpet and a saxophone.

 A simple example helps to better understand the principle of this preset. We first consider a string sound, such as violins, forming a leader song. At this string sound, we assign a high volume, for example between 90 and 127 in the standard midi. On the other hand, if this string sound is assigned to the tablecloth part, the general volume will then be lower, for example between 30 and 60 in the midi standard.

In one embodiment, the volume decrease rates for each orchestral part are as follows: - lead vocal 0% - lead vocal response 10% - counter-vocal counterpoint 14% - rhythm chords 1 14% - chords rhythmic 2 18% - arpeggios 13% - bass 10% - percussions 30% - tablecloths or carpets 60%
In other words, the leader song will have the maximum volume and the volume of the response to the leading song will be 10% less than the volume of the leader song, etc.

<Desc / Clms Page number 10>

 If instrument dubbing is planned, in this example, the volume of the leader song is decreased by 12% and the dubbing of the lead song is reduced in the same proportion.

 According to yet another aspect of the invention, the preset consists of assigning a sound effect to certain styles of music and to specific parts of the orchestra.

For example, if the style of music is "hard-rock", it will apply a chorus effect on the lead vocals and arpeggios if the chosen sound is a guitar called Fuzz (with distortion). As an example again, if the style is "musette", we will apply a slight reverb effect on the accordion.

Moreover, the same effect can be applied to several orchestral parts but the volume (or existence) of this effect may vary according to the nature and / or identity of these parts of the orchestra. By way of example, we consider a so-called typical music such as bossa-nova in which a reverberation effect is applied. In this case, the reverb applied automatically on each part of the orchestra will have the following MIDI level (0 to 127): - leader song 40 - response to the leader song 30 - counterpoint counterpoint 30 - rhythm chords 1 20 - rhythmic chords 2 20 - arpeggios 10 - bass 0 - percussions 30 - tablecloths or carpets 100
Of course, these various presets can be combined or modulated according to indications provided by the applicant. For example, the reverb effect indicated above may be limited to specific locations in the piece of music, such as a measure and / or chorus or movement.

 It should be noted that the volumes to be considered are, as in the norm midi, on the one hand, the general volume of a given sound which is generally indicated at the beginning of the piece and, in any case, whenever the we change the sound and, on the other hand, the volume of each note of this sound.

<Desc / Clms Page number 11>

 If the user, through an appropriate command, selects the predetermined or preset setting, the mixes and effect settings that may possibly be included in the game order file sent by the requestor are inhibited. In other words, these effect setting controls included in the received midi file will not be taken into account by the sequencer of the studio that will apply its own predetermined settings (presets).

 It is also possible to provide that the user can, through an interface, make mixing and effects settings online or disconnected from the network after receiving the sound information.

 After processing the received midi files, the audio result is transferred, preferably in real time to the requester. The transmission is, for example, in compressed form.

In the receiver at the user, the audio data is read either in deferred time, that is to say by downloading, or in real time, the reading is thus carried out at the same time as the download continues (usually called "streaming'l
It should be noted that the broadcasting can not only be carried out using a telecommunication network but also in a conventional manner, by cable or radio link. Thus, the virtual studio can be a kind of radio station broadcasting permanently, possibly on demand and subscription, sound programs. In this case, it is possible to use an automatic composition technique such as that described in PCT application WO 00/17850 in the name of René Louis BARON for "Automatic musical generation method and device". In this hypothesis, the sound environments received, and for which the user has subscribed a subscription, may correspond to his activity. In addition, automatic generation has the advantage of not being subject to payment of copyright or being subject to inexpensive payments.

 Figure 1 is a flowchart showing a server 10 according to an embodiment of the invention. This server has an input 12 receiving midi files in situ or from remote applicants online and the input 12 is connected to a multiplexer 14 for selecting the midi musical information to distribute them

<Desc / Clms Page number 12>

 between the various sound sample banks, namely a musical instrument sample bank 16, a sound sample bank 18, a sung phoneme sample bank.

 In this example, these sample banks and the multiplexing are made by software on the hard disk of a server computer.

 The sound samples are transmitted to devices that can be realized in software but which, in the example shown, are partially analog devices, namely samplers and synthesizers.

 All of these devices is represented by block 22. Each of these devices has a MIDI input and the MIDI file allows you to select samplers or synthesizers or this selection is performed automatically. Each of these devices of the set 22 allows the reading and the play of the values of the midi musical parameters and the assembly is transmitted on an audio output interface 24, for example analog, and this output interface 24 is connected to a device. 26, for example in software form, allowing the audio processing to achieve the musical effects: compressor, limiter, reverb, chorus, delay, enhancer, spacing, harmonizer, etc. After this effect-generating processing device, mixing 28, also preferably in software form, is provided, which consists of adjusting the various volumes, including the general volume, and performing the stereophonic placement, ie the position occupied. by a sound (part of orchestra) in the stereophonic range of 180 (from left to right). At the output of the mixing means 28, there is provided an interface 30, which is connected to a block 32 for compression and digitization and the digitized signal is transmitted, for example over the air or by telephone.

 FIG. 2 is a diagram showing an example of use of the server according to the invention in which this server 10 receives midi files from a computer 100 and sends the audio files to this computer, which comprises an amplifier 110 and high 112.

 According to yet another aspect of the invention, an equalization of the volume of the sounds of all the apparatuses of the studio is provided which makes it possible to avoid tedious manual adjustments.

<Desc / Clms Page number 13>

 It has indeed been found that for the same volume adjustment according to the midi protocol, the same sounds do not provide the same level of sound volume, this volume may vary from one device to another. Moreover, the same synthesizer provides sounds of various natures that do not have the same level of reproduction.

 These differences in sound volumes can create unwanted or even unsightly effects.

 This is why, it is expected that a volume equalization of sounds that is performed prior to the installation of each device, including synthesizers, in the studio, including a server.

Equalization is performed as follows:
Each sound of different devices is played automatically on several octaves under the control of a program. This program assigns a general volume of a set midi value for all devices, preferably a midday value of between 64 and 127. Advantageously, for each note, a high mid-value volume, for example 127, is adopted so as to take into account the additional sounds that are produced by the synthesizers when the note is played heavily.

 These various sounds controlled by the general volume midi values and note volumes indicated above are recorded and the sound volumes of all these sounds are determined.

 The output level of a music generating apparatus, including a synthesizer, is expressed in volts. Usually, the output level of such a device is of the order of 250 mV.

To perform the equalization, we measure, for the same volume midi, the output level of all devices and we determine the lowest level at which we assign the highest midi value, namely 127. The midday values assigned to other devices will have the following value: 127x VI
VM
In this formula, Vi is the volt level of the output signal of the concerned apparatus and VM is the output volt level of the apparatus providing the maximum signal. Since the midi values are integer values, if the value indicated

<Desc / Clms Page number 14>

 above does not provide an integer, one will select the integer closest to (or an integer close to) the value provided by the formula above.

The table below is a simplified example showing the measurement results and the corresponding (automatic) settings.

<Tb>
<Tb>

Its <SEP> Bank <SEP> Level <SEP> M <SEP> Volume <SEP> noon
<tb> 00 <SEP> 01 <SEP> 0, <SEP> 24 <SEP> 127
<tb> 01 <SEP> 01 <SEP> 0, <SEP> 30 <SEP> 102
<tb> 02 <SEP> 01 <SEP> 0, <SEP> 25 <SEP> 105
<Tb>

When a new device is inserted into the server, a new setting must be made.

 According to another embodiment of the invention, which can be used independently of the equalization described above, but preferably in combination, the volume of each sound is assigned a weighting factor which depends on the dominant frequency of that sound. his. Preferably, the volume decreases as the frequency increases. For example, if, for example, a double bass and a violin are played simultaneously at the same level, the ear has the impression that the violin playing in the highs has a predominant level; this is why, we reduce the volume of treble compared to serious.

 To determine weighting factors, one can proceed empirically so that the result is most pleasing to the ear.

 Figure 3 is a diagram illustrating an example of weighting. In this diagram, the predominant F frequencies were indicated on the abscissa and the midi V volumes on the ordinate. Curve 120 represents the weighting assigned to the midday volume. The maximum midi volume, ie 127, is applied to the predominantly low frequencies and the midday volume is decreased as the frequency increases. However, as shown in step 122, from about 10,000 Hz, it is no longer necessary to apply weighting as a function of frequency in this example.

Claims (26)

A studio-type generator (10) for producing sounds that can be represented by a score such as musical and / or sung and / or spoken sounds and / or noises, characterized in that it comprises: - a receiving input of numerical orders for the playing of sound pieces; - a bank of sound samples; - means for selecting the samples according to the received game orders; and - means for transmitting the selected samples to a plurality (22). ) means of reproduction, these means being such that for the same sound and the same digital level of the sound, they all provide substantially the same output analog level.  2. Generator according to claim 1, characterized in that the sound reproduction means are such that the general level of sound provided by each reproduction means depends on the level of the range of its dominant frequencies (120).  3. Generator according to claim 2, characterized in that the sound reproduction means are such that the general level of the provided sound varies in the opposite direction of the level of the dominant frequencies.  4. Generator according to claim 2 or 3, characterized in that the general level of sound provided by each means of sound reproduction is independent of the level of the dominant frequency range for frequencies above about 10 KHz.  5. Generator according to any one of the preceding claims, characterized in that it comprises a mixing means (28) for imparting a predetermined mix to the signals provided by the sound reproducing means. <Desc / Clms Page number 16>  6. Generator according to claim 5, characterized in that the predetermined mixing is provided by family of sounds and is a function of a situation signal of each sound in an orchestration.  7. Generator according to claim 6, characterized in that the families of sounds comprise at least two of the families taken together comprising: brass, wind, strings type rubbed, human voices, guitars, bass , percussion, synthetic sounds and various noises.  8. Generator according to claim 6 or 7, characterized in that the sound situation signals in the orchestration represent situations that are part of the set comprising: the leader song, the response to the leader song, the counter-vocals in counterpoint, rhythmic chords, arpeggios, basses, tablecloths or mats and doublings of instruments for one or more orchestral parts.  9. Generator according to any one of claims 5 to 8, characterized in that it comprises means for the predetermined mix to be prioritized over sound adjustment orders included in the music game order file.  10. Generator according to any one of the preceding claims, characterized in that the sound reproduction means are part of the group comprising: synthesizers, expanders and samplers.  11. Generator according to any one of the preceding claims, characterized in that it comprises means for mixing the signals from the sound reproducing means.  12. Generator according to claim 11, characterized in that the mixing means comprise means for performing a volume balance and / or stereophonic placement and / or confer a general volume. <Desc / Clms Page number 17>  13. Generator according to any one of the preceding claims, characterized in that it comprises means for introducing effects to the signals provided by the sound reproducing means.  14. Generator according to any one of the preceding claims, characterized in that it comprises means for introducing at least one predetermined effect to the signals provided by the sound reproduction means, this effect depending on a signal indicative of the type of music. and / or a random draw of a sound effect.  15. Generator according to claim 14, characterized in that it comprises means for making a predetermined manner depend on the existence and / or the level of the effect of a sound situation signal in the orchestration.  16. Generator according to claim 14 or 15, characterized in that it comprises a means for the predetermined predetermined effect with respect to effect controls included in game orders.  17. Generator according to any one of the preceding claims, characterized in that the banks of sound samples are stored in permanent memories.  18. Generator according to any one of the preceding claims, characterized in that it comprises means for remotely transmitting, by a communication network, the signals provided by the sound reproducing means.  19. Generator according to claim 18, characterized in that it comprises means for digitizing and compressing the signals to be transmitted.  20. Generator according to claim 18 or 19, characterized in that it comprises means for broadcasting the sound signals to receivers.  21. Generator according to any one of the preceding claims, characterized in that the input data of musical games are in midi format. <Desc / Clms Page number 18>  22. Generator according to any one of the preceding claims, characterized in that it comprises means for receiving game orders issued remotely.  23. Generator according to any one of the preceding claims, characterized in that it comprises a means for automatic production of musical compositions boring the numerical order of play of sound pieces.  24. Generator according to any one of the preceding claims, characterized in that it comprises a volume equalization control means which, for the same digital level of sound at the input of each of the reproduction means, determines the analog output level of each of these sound reproduction means and which affects at the digital level to the input of each sound reproduction means a weighting allowing all these reproduction means to provide substantially the same output analog level.  25. Generator according to claim 24, characterized in that it comprises means for determining the introduction of a new sound reproduction means and for operating said equalization adjustment means.  26. A process for producing sounds that can be represented by a score such as musical and / or sung and / or spoken sounds and / or noises, characterized in that it comprises the following steps: - one receives numerical orders of sets of sound pieces, - samples are selected in a bank of sound samples according to the received game orders and, - the selected samples are transmitted to a plurality (22) of production means, these means being such that for a given same sound and the same digital level of this sound, they all provide substantially the same output analog level, said sound production method being preferably implemented by means of a generator (10) according to any one of the claims 1 to 25.