EP0890166B1 - Digital audio creation method and sound synthesising and/or recording devices using same - Google Patents

Description

The present invention relates to a method for creating a segment of digital sound resulting from mixing digital sound segments components, and devices for synthesizing and / or recording its implementing this process.

By "digital sound segment", or by "sample", we mean a series of sample points. For example, a segment of a second of 16.1-bit stereo sound at 44.1 kHz has 44,100 sample points doubles each occupying 4 bytes (16 bits or 2 bytes for each channel). A segment is not necessarily continuous: the rests can be compressed to take up less memory.

Traditional processes, to create a segment of digital sound resulting from the mixing of component digital sound segments, consist entirely of simultaneous mixing operations or successive. They are used in particular in devices for sound synthesis and in recording devices digital audio.

These existing devices simultaneously mix in real time a number of component segments. The number of these (polyphony in the case of synthesizers and number of tracks simultaneous playback for recording devices digital audio) is limited among other things by the computing power and / or by the speed of the mass memory means.

Existing digital recording systems have also a track reduction function to merge by mixing several segments simultaneously or successively components to free up tracks. But once these segments mixed components, delete or modify one or more of them forced to completely remix the merged track, which requires lots of calculations.

The method and the devices according to the invention make it possible to remedy these limitations and drawbacks.

The invention also aims to enable a device, which may simply consist of a microcomputer such as an entry-level Apple® Power Macintosh® running a integrated music and multimedia creation program, which fills MIDI sequencer and digital audio functions, synthesizer and infinite polyphony and multitimbral sampler, direct disc recorder with infinite number of channels, mixing, editing, effects generator, automatic saturation correction, song recording, sound and post-synchronization movies and multimedia applications.

(a) obtenir, et mixer à l'ancienne version du segment résultant par une première opération numérique, certains des dits segments composants, les informations relatives d'une part à leur obtention et d'autre part à leur position au sein de la nouvelle version du dit segment résultant étant mémorisées,
   mais également l'étape :

(b) obtenir, et démixer de l'ancienne version du segment résultant par une deuxième opération numérique, certains des dits segments composants mixés précédemment, en utilisant les dites informations mémorisées relatives d'une part à leur obtention et d'autre part à leur position au sein de l'ancienne version du dit segment résultant,

   les dites étapes (a) et (b) intervenant selon les besoins de la création du dit segment résultant.The method according to the invention, for creating a digital sound segment resulting from the mixing of component digital sound segments, in fact comprises not only the step:

(a) obtain, and mix with the old version of the resulting segment by a first digital operation, some of the said component segments, the information relating on the one hand to their obtaining and on the other hand to their position within the new version of said resulting segment being stored,
but also the stage:

(b) obtain, and demix from the old version of the resulting segment by a second digital operation, some of the said component segments previously mixed, using said stored information relating on the one hand to their obtaining and on the other hand to their position within the old version of said resulting segment,

said steps (a) and (b) intervening according to the needs of the creation of said resulting segment.

• sa lecture dans une mémoire,
• sa réception via des moyens d'entrée de données,
• son calcul par transposition en fréquence et/ou bouclage partiel d'un échantillon de base en fonction de données MIDI (par "échantillon de base" on entend un échantillon servant de base au calcul de différentes notes d'un même instrument),
• son calcul par mixage d'autres segments,
• son calcul à partir d'équations mathématiques d'oscillateurs,
• son calcul à partir d'un modèle physique d'instrument acoustique,
• son calcul par décompression d'un segment de son numérique comprimé,
• l'application d'un ou plusieurs traitements, tels que modification de l'enveloppe, du volume, du panoramique, ou effets, tels que délai, réverbération, filtres, à un autre segment,
• l'application d'une modulation d'amplitude ou de fréquence à un segment ou à un réglage.

By "obtaining" a segment, we mean for example:

• its reading in a memory,
• its reception via data entry means,
• its calculation by frequency transposition and / or partial looping of a basic sample according to MIDI data (by "basic sample" means a sample used as a basis for the calculation of different notes of the same instrument),
• its calculation by mixing other segments,
• its calculation from mathematical equations of oscillators,
• its calculation from a physical model of acoustic instrument,
• its calculation by decompression of a segment of compressed digital sound,
• the application of one or more treatments, such as modification of the envelope, volume, pan, or effects, such as delay, reverberation, filters, to another segment,
• applying amplitude or frequency modulation to a segment or adjustment.

By "mixing component segments with the old version of the segment resulting "means the action of mixing, for example by addition, these component segments possibly attenuated or amplified to the old version of the resulting segment, the recorded result of this operation constituting the new version of the resulting segment.

By "demixing component segments of the old version of the resulting segment "means the action of removing, for example by subtraction, these possibly attenuated component segments or amplified from the old version of the resulting segment, the result recorded of this operation constituting the new version of the resulting segment.

• Une opération de mixage numérique ne peut donner qu'un résultat possible, contrairement aux opérations analogiques qui sont sensibles aux variations infimes de paramètres tels que vitesse d'un moteur, fréquence d'un oscillateur ou gain d'un amplificateur.
• L'effet d'une opération de mixage numérique, et en particulier de la plus simple qui consiste à additionner les segments composants point-échantillon par point-échantillon, peut être annulé exactement par une autre opération numérique, par exemple la soustraction dans le cas du mixage par addition, pour peu que les segments composants à supprimer et leur position au sein du segment résultant puissent être réobtenus à l'identique et que les retenues soient gérées, et ce même si d'autres opérations du même type sont intervenues entre-temps : dans le cas de l'addition par exemple, (x+y+z) -y = x+z.
• Le démixage numérique exact d'un segment composant est donc possible et prend typiquement le même temps de calcul que le mixage de ce segment : dans le cas du mixage par addition par exemple, le démixage par soustraction prend le même temps de calcul que le mixage.
• Lors d'une modification de la composition du segment résultant, qui peut se traduire par la suppression d'anciens segments composants et/ou par l'introduction de nouveaux, il est fréquent que les segments composants à garder soient plus nombreux que les segments composants à supprimer. Quand c'est le cas, étant donné que le temps de calcul moyen pour obtenir et démixer un segment est le même que pour obtenir et mixer un segment, il sera généralement plus économique en temps de calcul de partir de la version antérieure du segment résultant, de démixer les segments à supprimer et de mixer les segments à introduire, que de partir de zéro et de mixer tous les segments à garder plus les segments à introduire. L'utilisation, lors de certaines phases de la création du segment résultant, au choix de l'utilisateur ou sur décision automatique, de l'étape (b) de démixage caractéristique du procédé selon l'invention, autorise donc un gain de temps de calcul. Ce gain de temps de calcul permet notamment de manipuler en temps réel un plus grand nombre de segments composants : ce nombre n'est en fait pas limité puisque la plupart des segments composants n'ont pas à être réobtenus et remixés.

The foundations and advantages of the process according to the invention are as follows:

• A digital mixing operation can only give a possible result, unlike analog operations which are sensitive to minute variations in parameters such as motor speed, frequency of an oscillator or gain of an amplifier.
• The effect of a digital mixing operation, and in particular of the simplest one which consists in adding the component segments point-sample by point-sample, can be canceled exactly by another digital operation, for example subtraction in the case addition mixing, provided that the component segments to be deleted and their position within the resulting segment can be re-obtained identically and that the holdbacks are managed, even if other operations of the same type have been performed time: in the case of addition for example, (x + y + z) -y = x + z.
• The exact numerical demixing of a component segment is therefore possible and typically takes the same calculation time as the mixing of this segment: in the case of addition mixing for example, subtraction demixing takes the same calculation time as mixing .
• When the composition of the resulting segment is modified, which can result in the deletion of old component segments and / or the introduction of new ones, it is frequent that the component segments to be kept outnumber the component segments to delete. When this is the case, since the average computation time to obtain and demix a segment is the same as to obtain and mix a segment, it will generally be more economical in computation time from the previous version of the resulting segment , to demix the segments to be deleted and to mix the segments to be introduced, than to start from zero and to mix all the segments to keep more the segments to be introduced. The use, during certain phases of the creation of the resulting segment, at the choice of the user or on automatic decision, of step (b) of unmixing characteristic of the method according to the invention, therefore allows a saving of time of calculation. This saving in computation time makes it possible in particular to manipulate a greater number of component segments in real time: this number is in fact not limited since most of the component segments need not be re-obtained and remixed.

Calculation time savings can also be made in how to translate a change in composition: for example, to change the volume or pan of a component segment, you better mix or demix only the difference in volume rather just demix the old volume and remix the new one.

• la première opération numérique consiste à additionner à l'ancienne version du segment résultant chaque segment composant à mixer pour obtenir la nouvelle version du dit segment résultant, la dite addition s'effectuant point-échantillon par point-échantillon à partir du point-échantillon correspondant à la position souhaitée du dit segment composant au sein du dit segment résultant,
• la deuxième opération numérique consiste à soustraire de l'ancienne version du segment résultant chaque segment composant à démixer pour obtenir la nouvelle version du dit segment résultant, la dite soustraction s'effectuant point-échantillon par point-échantillon à partir du point-échantillon correspondant à la position du dit segment composant au sein de l'ancienne version du dit segment résultant.

According to a particular embodiment of the method according to the invention, as mentioned previously:

• the first numerical operation consists in adding to the old version of the resulting segment each component segment to be mixed to obtain the new version of said resulting segment, said addition being carried out point-sample by point-sample starting from the corresponding sample point at the desired position of said component segment within said resulting segment,
• the second numerical operation consists in subtracting from the old version of the resulting segment each component segment to be demixed to obtain the new version of said resulting segment, said subtraction being carried out point-sample by point-sample starting from the corresponding sample point at the position of said component segment within the old version of said resulting segment.

Any saturations during additions are advantageously stored in order to avoid saturation errors during the following operations, for example by coding each sample point of the resulting segment on more bits than the segments components, or by maintaining a list of sample points saturated with the resulting segment, mentioning for each his position and the value of the withholding. This list is advantageously searchable by the user, allowing him to correct the volumes of different component segments, corrections can also be automated. We can also lower or cancel in real time the volume of the segment added by memorizing this "forced" drop among the information relating to obtaining the segment, then distribute the drop in volume over all segments contributing to saturation.

• des moyens de contrôle et de calcul,
• des moyens de mémoire associés aux dits moyens de contrôle et de calcul,
• des moyens d'entrée de données associés aux dits moyens de contrôle et de calcul,
• des moyens de sortie de son associés aux dits moyens de contrôle et de calcul,

   fonctionne comme suit :

• les segments de son numérique composants sont, au moins pour une partie d'entre eux, synthétisés par les moyens de contrôle et de calcul,
• les informations relatives à l'obtention et à la position des dits segments composants, constituées d'une part d'informations génériques et d'autre part d'informations spécifiques, sont stockées dans les moyens de mémoire,
• le segment de son numérique résultant, enregistré dans les moyens de mémoire, constitue la version en cours du mixage de segments composants synthétisés et d'autres segments composants éventuels,
• les dits moyens de contrôle et de calcul exécutent, selon les informations reçues via les moyens d'entrée de données, l'étape (a) du dit procédé pour les segments composants éventuels à mixer au segment résultant, et l'étape (b) du dit procédé pour les segments composants éventuels à démixer du segment résultant,
• les dits moyens de contrôle et de calcul envoient le dit segment résultant vers les dits moyens de sortie de son, après traitements globaux non destructifs éventuels.

The digital sound synthesis device according to the invention, which implements the method according to the invention, and which comprises at least:

• means of control and calculation,
• memory means associated with said means of control and calculation,
• data entry means associated with said means of control and calculation,
• sound output means associated with said means of control and calculation,

works as follows:

• the component digital sound segments are, at least for a part of them, synthesized by the control and calculation means,
• the information relating to the obtaining and the position of said component segments, consisting on the one hand of generic information and on the other hand of specific information, are stored in the memory means,
• the resulting digital sound segment, recorded in the memory means, constitutes the current version of the mixing of synthesized component segments and other possible component segments,
• said control and calculation means execute, according to the information received via the data input means, step (a) of said method for the possible component segments to be mixed with the resulting segment, and step (b) of the said process for the possible component segments to be demixed from the resulting segment,
• said control and calculation means send said resultant segment to said sound output means, after possible non-destructive global processing.

The term “sound synthesis device” means in particular: synthesizer, drum machine, sampler, computer provided with sound synthesis software, a sound card for computer, a sound module.

• des moyens de calcul tels qu'un micro-processeur CISC ou RISC, un DSP, ou la combinaison de plusieurs de ces dits moyens de calcul,
• des moyens de contrôle tels qu'un système d'exploitation, un logiciel enregistré, des circuits d'interface avec les divers moyens de mémoire, d'entrée et de sortie.

The means of control and calculation may include, for example:

• calculation means such as a CISC or RISC microprocessor, a DSP, or the combination of several of these said calculation means,
• control means such as an operating system, recorded software, interface circuits with the various memory, input and output means.

• RAM et ROM,
• RAM et mémoire de masse telle que disque dur, disque magnéto-optique, CD-ROM, DAT, disquettes, mémoire flash,
• RAM, ROM et mémoire de masse.

The memory means can be for example:

• RAM and ROM,
• RAM and mass memory such as hard disk, magneto-optical disk, CD-ROM, DAT, floppy disks, flash memory,
• RAM, ROM and mass memory.

The data entry means can be for example: port serial, SCSI, MIDI, MIDI keyboard, MIDI controllers, keyboard alphanumeric, mouse, joystick, data glove, reader removable data carriers such as CD-ROM, floppy disks, disk magneto-optical.

The sound output means can be for example: an output analog sound, digital sound output, one or more speakers, data output means such as serial interface, SCSI, MIDI, mass recording media.

The device can also advantageously include display means such as a screen, LCD display or headset data, sound input means for capturing samples of base, as well as data output means other than sound such as specific information.

The device can for example consist of a microcomputer general purpose, controlled by appropriate software, provided internally or through an additional sound output card CD quality, and connected to a MIDI keyboard. The said microcomputer can for example be a Power Macintosh® model from Apple® equipped with a internal 16-bit 44.1 kHz stereo sound output.

The device can also be made up of entirely hardware dedicated to the application, for example a compact material of the type workstation with the appearance exterior of a musical keyboard fitted with controllers and displays.

The modified parts of the resulting segment can of course be rewritten in memory instead of their previous version in the since all changes are reversible by remixing or selective demixing, but it may be advantageous to allow them to are written elsewhere at the user's choice, for example if wish to try it out which is more likely to be deleted than to be retained.

"Generic information" means information allowing the obtaining of several component segments, such as basic samples or mathematical or physical models of sound sources.

"Specific information" means information such as as MIDI data, model parameters or settings, which combined with generic information define exactly the content of a particular component segment and its position within the resulting segment.

By "other possible component segments" is meant for example segments of recorded digital audio tracks, as we will see later in the context of synthesis devices and sound recording.

The "information received via the data entry means" is for example MIDI commands from a MIDI keyboard, a MIDI controller or a MIDI file entered via a floppy disks, or sequence editing or partition, entries for example using an alphanumeric keyboard and / or a mouse.

By "possible global non-destructive treatments", we mean treatments such as effects, adjustments, or mixing with other segments of its digital potential, applied to the resulting segment in its together according to the needs of the final sound creation, but whose result is not re-recorded in the said resulting segment, the said treatments that can be automated. The so-called other segments of its digital sound may include other segments results obtained by the process according to the invention. Such operation by sub-groups allows real-time modification of the mixing a large number of instruments.

The sound synthesis device allowing to visualize the result of the synthesis of an instrument in digital audio form, we can advantageously edit this result, and not only the basic sample as in the synthesis devices conventional, which makes it possible to combine MIDI editions and audio. The simultaneous consideration of MIDI and audio allows also feedback from audio to MIDI in order to if necessary, automatically modify the latter to improve the quality of the song, for example to avoid saturations or improve the realism of the synthesis.

• le segment de son numérique résultant est stocké dans des moyens de mémoire de masse,
• les moyens de contrôle et de calcul transfèrent chaque tronçon à modifier du dit segment résultant depuis les dits moyens de mémoire de masse vers au moins un moyen de mémoire tampon, afin d'exécuter le dit procédé, puis transfèrent la nouvelle version du dit tronçon vers les moyens de mémoire de masse.

According to a particular embodiment of the digital sound synthesis device according to the invention:

• the resulting digital sound segment is stored in mass memory means,
• the control and calculation means transfer each section to be modified of said segment resulting from said mass memory means to at least one buffer memory means, in order to execute said process, then transfer the new version of said section to mass memory means.

For real-time operation with segment storage resulting on mass memory, we will use for example two buffer memories: while the current section is played on the sound output from the first buffer as and when update, the new version of the section preceding that in course is transferred from the second buffer to the mass memory then the old version of the section following the one in course is loaded into said second buffer. Then the roles of the two buffers are reversed, and so on. The time diagram of such an exemplary embodiment, operating on the Macintosh® model mentioned above, will be detailed later in reference to figures.

• des moyens de contrôle et de calcul,
• des moyens de mémoire associés aux dits moyens de contrôle et de calcul,
• des moyens d'entrée de données associés aux dits moyens de contrôle et de calcul,
• des moyens d'entrée de son associés aux dits moyens de contrôle et de calcul,
• des moyens de sortie de son associés aux dits moyens de contrôle et de calcul,

fonctionne comme suit :

• les segments de son numérique composants sont, au moins pour une partie d'entre eux, obtenus à partir d'au moins une piste audionumérique enregistrée dans les dits moyens de mémoire,
• les informations relatives à l'obtention et à la position des dits segments composants, constituées d'une part des dites pistes audionumériques enregistrées et d'autre part d'informations spécifiques, sont stockées dans les moyens de mémoire,
• le segment de son numérique résultant, enregistré dans les moyens de mémoire, constitue la version en cours du mixage de segments composants obtenus à partir des dites pistes audionumériques et d'autres segments composants éventuels,
• les dits moyens de contrôle et de calcul exécutent, selon les informations reçues via l'un au moins des dits moyens d'entrée, l'étape (a) du dit procédé pour les segments composants éventuels à mixer au segment résultant, et l'étape (b) du dit procédé pour les segments composants éventuels à démixer du segment résultant,
• les dits moyens de contrôle et de calcul envoient le dit segment résultant vers les dits moyens de sortie de son, après traitements globaux non destructifs éventuels.

The digital sound recording device according to the invention, which implements the method according to the invention, and which comprises at least:

• means of control and calculation,
• memory means associated with said means of control and calculation,
• data entry means associated with said means of control and calculation,
• sound input means associated with said control and calculation means,
• sound output means associated with said means of control and calculation,

works as follows:

• the component digital sound segments are, at least for a part of them, obtained from at least one digital audio track recorded in said memory means,
• the information relating to the obtaining and the position of said component segments, consisting on the one hand of said recorded digital audio tracks and on the other hand of specific information, are stored in the memory means,
• the resulting digital sound segment, recorded in the memory means, constitutes the current version of the mixing of component segments obtained from said digital audio tracks and of any other component segments,
• said control and calculation means execute, according to the information received via at least one of said input means, step (a) of said method for the possible component segments to be mixed with the resulting segment, and the step (b) of said method for the possible component segments to be demixed from the resulting segment,
• said control and calculation means send said resultant segment to said sound output means, after possible non-destructive global processing.

The term “sound recording device” means in particular: device for direct recording to disc or other medium mass, a sampler, a computer with software sound recording and / or editing, a sound card for computer with recording function, module direct recording to disk or other mass media.

The word track is used here, in accordance with general practice in digital audio, by functional analogy with the tracks of a tape recorder. But a track is just a segment of its numeric as defined above, which is not necessarily continuous or as long as the song it is playing in, not more than necessarily positioned at the start of the song. A track is not necessarily fully mixed.

• l'identification du segment de piste dont il provient,
• sa position dans le segment résultant,
• son volume de mixage dans le segment résultant,
• une courbe de volume ("fade-in", "fade-out"),
• son réglage ou sa courbe de panoramique,
• la valeur de sa transposition éventuelle en fréquence, ainsi éventuellement que sa durée en cas d'expansion ou de compression temporelle("time-stretching"),
• des effets ou courbes d'effets à appliquer.

The specific information relating to the obtaining and the position of the component segments can be for example for each:

• identification of the track segment from which it originates,
• its position in the resulting segment,
• its mixing volume in the resulting segment,
• a volume curve ("fade-in", "fade-out"),
• its setting or its panning curve,
• the value of its possible transposition into frequency, as well as possibly its duration in the event of time expansion or compression ("time-stretching"),
• effects or effect curves to apply.

By "information received via at least one of the said means input ", we also hear the sound input modifying the tracks digital audio themselves, that data entry modifying the specific information for obtaining segments components from these tracks. Modification of the segment resulting from the receipt of this information via one at less of the said means of entry can involve the two stages (a) and (b) of the process, for example when re-recording a track or moving a component segment within the segment resulting, or only step (b), for example during the deletion of a component segment or reduction of its volume, or only step (a), for example during the introduction, duplicating or increasing the volume of a component segment.

The device can for example consist of a microcomputer general purpose provided, internally or through a additional card, CD quality CD input and output, controlled by appropriate software. The said microcomputer can by example be an Apple® Power Macintosh® model with an input and an internal 16-bit 44.1 kHz stereo sound output.

The modified parts of the resulting segment can of course be rewritten in memory instead of their previous version in the since all changes are reversible by remixing or selective demixing, but it may be advantageous to allow them to are written elsewhere at the user's choice, for example if wish to try it out which is more likely to be deleted than to be retained.

• le segment de son numérique résultant est stocké dans des moyens de mémoire de masse,
• les pistes audionumériques sont stockées, au moins pour une partie d'entre elles, dans des moyens de mémoire de masse,
• les moyens de contrôle et de calcul transfèrent chaque tronçon à modifier du dit segment résultant, et chaque tronçon de piste audionumérique nécessaire à l'obtention des segments composants éventuels à mixer et des segments composants éventuels à démixer, depuis les dits moyens de mémoire de masse, vers au moins un moyen de mémoire tampon pour le dit segment résultant plus au moins un moyen de mémoire tampon pour les pistes audionumériques, afin d'exécuter le dit procédé, puis transfèrent les nouvelles versions éventuelles des dits tronçons de pistes et la nouvelle version du dit tronçon de segment résultant vers les moyens de mémoire de masse.

According to a particular embodiment of the digital sound recording device according to the invention:

• the resulting digital sound segment is stored in mass memory means,
• the digital audio tracks are stored, at least for a part of them, in mass memory means,
• the control and calculation means transfer each segment to be modified from said resulting segment, and each segment of digital audio track necessary for obtaining any component segments to be mixed and any component segments to be demixed, from said mass memory means , to at least one buffer memory means for said resulting segment plus at least one buffer memory means for digital audio tracks, in order to execute said method, then transfer the possible new versions of said track sections and the new version from said resulting segment segment to the mass memory means.

The mass memory means can for example be constituted one or more hard disks and / or one or more magneto-optical disks.

For real-time operation, for example, two groups of buffers in RAM: each group includes a buffer for the resulting segment plus a buffer by digital audio track necessary to obtain the segments possible components to be mixed and possible component segments to demix, plus a sound input buffer: while simultaneous sections are changed (that of the resulting segment by the process, and possibly those of the tracks by the input means of sound) in a buffer group, new versions of previous sections are transferred from the other group of buffers to mass memory then older versions subsequent sections are transferred from mass memory to said other buffer group. Then the roles of the two groups are reversed, and so on. The time diagram of a as an example of implementation, operating on the Macintosh® model above, will be detailed later with reference to the figures.

According to a particular embodiment of the synthesis device sound or digital audio recording according to the invention, for at minus some of the changes in the resulting segment, the means of control and calculation execute said process in real time, and send the said resulting segment in real time as updating it to said sound output means, after possible non-destructive overall treatments.

Existing real-time music synthesizers are limited by polyphony, i.e. in the number of sounds they can play at the same time, by the fact that they synthesize and mix individually all the sounds in real time, and that their power calculation is limited.

The synthesizer according to the invention processes the sounds already synthesized in previous steps as a single digital audio track that it simply reads in his memory. He can therefore devote most of its computing power in synthesis and mixing or demixing of sounds to add or remove. In one use typical these are the sounds of the current instrument, for example the piano, while the resulting segment contains all the already recorded instruments of the song, for example the drums, the bass, guitars, violins, and any earlier version some piano.

For a given computing power therefore for a given cost, the synthesizer according to the invention therefore schematically has polyphony per instrument roughly equal to the total polyphony of a existing synthesizer, and the equivalent of total polyphony unlimited: the number of different simultaneous sounds can be arbitrarily large.

In fact we can even consider that polyphony by instrument itself is not limited to the extent that the instrument can be played in several passes, or completed in deferred time if you have reaches the limits of the computing power available during the game in real time.

Existing recording devices do not allow the simultaneous playback of more than a certain number of tracks depending on computing power and data transfer speed from the recording medium. To read more tracks simultaneously, it is necessary to merge more than one and this time-consuming operation must be repeated at whenever a component track of the merged track is changed.

For computing power and transfer speed of data allowing an existing device to read a few tracks simultaneously, the device according to the invention in its real-time version can play an arbitrary number of tracks simultaneously, even during the registration or modification of one of them, this number being limited only by the capacity of its supports recording. If we store the tracks on supports removable recording media such as magneto-optical discs or removable hard disks, this number is in practice not limited.

• les informations permettant l'obtention d'au moins le début des segments composants sont stockées dans des moyens de mémoire rapide pour au moins un sous-ensemble des dits segments composants,
• en fonction des besoins de la création, les moyens de contrôle et de calcul peuvent modifier le dit sous-ensemble par transfert de données depuis les moyens de mémoire de masse vers les dits moyens de mémoire rapide.

According to another particular embodiment of the sound synthesis or digital audio recording device according to the invention:

• the information enabling at least the start of the component segments to be obtained is stored in rapid memory means for at least a subset of said component segments,
• according to the needs of the creation, the control and calculation means can modify the said subset by transferring data from the mass memory means to the said fast memory means.

Said mass memory means can for example be consisting of a CD-ROM containing basic samples.

Advantageously, even for a game in real time, we will not store not in quick memory (RAM) the generic information of all instruments, but only those of the current instrument since these are the only ones to which the synthesizer according to the invention has need to access quickly. It differs in this from existing synthesizers which individually synthesize all sounds to play at the same time and therefore need to access quickly to the generic information of all instruments. The subset mentioned above will therefore be the subset of sounds possible of the current instrument, said current instrument being user selectable.

For a given RAM size therefore for a given cost, the synthesizer according to the invention therefore schematically has a size of RAM per instrument equal to RAM size for all instruments of an existing synthesizer, and the equivalent of a RAM size for all instruments equal to the size of mass memories devoted to generic information. If we use removable mass memory such as CD-ROM samples of base, this RAM size for all instruments, and therefore the number of instruments (multi-timbre of the synthesizer), are not not limited: a song can contain an arbitrary number great of arbitrarily long sounds. In fact we can even consider that the RAM size per instrument is not itself limited in since the instrument can be played in several passes.

Only the information allowing the beginning of the sounds to be obtained (approx. 0.5 s) must absolutely be in quick memory for a game in real time: during these 0.5 s we have time to get the next section in mass memory. Regarding a recording device or a device for both synthesis and recording according to the invention, this technique allows in particular instant manual triggering of any track regardless of its size.

• le segment de son numérique résultant, enregistré dans les moyens de mémoire, constitue la version en cours du mixage de segments synthétisés et de segments obtenus à partir de pistes audionumériques enregistrées.

The digital sound synthesis and recording device according to the invention, which implements the method according to the invention, operates as follows:

• the resulting digital sound segment, recorded in the memory means, constitutes the current version of the mixing of synthesized segments and of segments obtained from recorded digital audio tracks.

• d'une RAM libre de 8 Mo pour stocker les échantillons de l'instrument en cours. Cela correspond donc à une taille de RAM par instrument de 8 Mo, bien supérieure à celle disponible dans des dispositifs de synthèse sonore classiques équipés de 32 coûteux Mo de RAM. Ces derniers doivent en effet partager leur RAM entre tous les instruments, par exemple au nombre de 16, ce qui ne laisse que 2 Mo par instrument.
• d'un lecteur de CD-ROM pour lire des CD-ROM d'échantillons du commerce ou des CD audio,
• d'un disque dur de 1 Go pour stocker segments résultants, pistes audio, et échantillons de base de l'utilisateur dont ceux enregistrés ou synthétisés par lui au moyen du dispositif, ce dernier permettant avantageusement la conversion de segments résultants en échantillons de base,
• d'un lecteur de disques magnéto-optiques amovibles pour archiver ces divers éléments et/ou les transporter,
• d'un lecteur de disquettes permettant l'import de fichiers MIDI du commerce par exemple, et l'export de fichiers d'informations spécifiques générés par le dispositif selon l'invention pour archivage ou transport des morceaux sous forme compacte,
• d'un microphone stéréophonique pour numériser du son acoustique en tant que piste ou échantillon de base.

The device can for example consist of a general-purpose microcomputer provided, internally or by means of an additional card, with a CD quality input and output, controlled by appropriate software, and connected to a MIDI keyboard. Said microcomputer can for example be a Power Macintosh® model from Apple® provided with an internal 16.1 44.1 kHz stereo sound input and output. It is equipped for example:

• 8 MB free RAM to store samples of the current instrument. This therefore corresponds to a size of RAM per instrument of 8 MB, much greater than that available in conventional sound synthesis devices equipped with 32 expensive MB of RAM. The latter must indeed share their RAM between all the instruments, for example the number of 16, which leaves only 2 MB per instrument.
• a CD-ROM drive to read CD-ROMs of commercial samples or audio CDs,
• a 1 GB hard disk to store the user's resulting segments, audio tracks, and basic samples, including those recorded or synthesized by him using the device, the latter advantageously allowing the conversion of resulting segments into basic samples,
• a removable magneto-optical disc player to archive these various elements and / or transport them,
• a floppy disk drive allowing the import of commercial MIDI files for example, and the export of specific information files generated by the device according to the invention for archiving or transporting pieces in compact form,
• a stereo microphone to digitize acoustic sound as a basic track or sample.

La figure 1 illustre le procédé selon l'invention.

La figure 2 représente le dispositif de synthèse de son selon l'invention.

La figure 3 représente le diagramme de temps d'un mode particulier de réalisation de ce dispositif.

La figure 4 représente le dispositif d'enregistrement audionumérique selon l'invention.

La figure 5 représente le diagramme de temps d'un mode particulier de réalisation de ce dispositif.

La figure 6 représente le dispositif de synthèse de son et d'enregistrement audionumérique selon l'invention.

The accompanying drawings illustrate the invention:

FIG. 1 illustrates the method according to the invention.

FIG. 2 represents the sound synthesis device according to the invention.

FIG. 3 represents the time diagram of a particular embodiment of this device.

FIG. 4 represents the digital audio recording device according to the invention.

FIG. 5 represents the time diagram of a particular embodiment of this device.

Figure 6 shows the sound synthesis and digital audio recording device according to the invention.

Referring to Figure 1, the current version of the segment resulting (1) is mixed, in this case by addition, with the component segment (2) and the component segment (3). We then obtain a new resulting segment (la). Then according to the needs of creation of the resulting segment we decide to demix, so in the case present to subtract, the component segment (2) from the resulting segment (the). The resulting segment (1b) is thus obtained strictly equivalent to mixing the resulting segment (1) and the segment component (3).

• des moyens d'entrée de données (6) tels qu'un clavier MIDI, un clavier alphanumérique et un écran.
• des moyens de contrôle et de calcul (4) tels qu'une unité centrale Power Macintosh® munie d'un logiciel approprié,
• des moyens de sortie de son (7) tels que le circuit audio interne et un casque stéréophonique,
• des moyens de mémoire (5) comprenant par exemple de la RAM et un disque dur, pour stocker le segment résultant (1), les informations génériques (8) et les informations spécifiques (9).

With reference to FIG. 2, the sound synthesis device according to the invention comprises:

• data input means (6) such as a MIDI keyboard, an alphanumeric keyboard and a screen.
• control and calculation means (4) such as a Power Macintosh® central unit equipped with appropriate software,
• sound output means (7) such as the internal audio circuit and a stereo headset,
• memory means (5) comprising for example RAM and a hard disk, for storing the resulting segment (1), the generic information (8) and the specific information (9).

• en RAM :
  • les informations génériques (8) c'est-à-dire les échantillons de base de l'instrument en cours, chargées au préalable depuis le disque dur.
  • les informations spécifiques (9) correspondant à la partition de l'instrument en cours, chargées au préalable depuis le disque dur. Ces dernières sont donc remplacées en RAM au fur et à mesure des entrées de données (14), mais seulement après avoir été prises en compte dans le calcul des segments composants à démixer. A l'issue de l'enregistrement en cours, elles seront sauvegardées sur le disque.
  • deux mémoires tampon du segment résultant, notées MTSR1 et MTSR2, pouvant contenir chacune par exemple quelques secondes de son audionumérique. MTSR1 et MTSR2 sont divisées en portions de l'ordre de quelques millisecondes, ces subdivisions étant notées MTSR1(i) et MTSR2(i), i variant de 1 pour la première subdivision à p pour la dernière. On a donc par exemple MTSR1(1), MTSR1(2)... MTSR1(p). La longueur et donc le nombre des subdivisions n'ont pas été respectées dans la figure pour des raisons de clarté.
  • la liste des retenues correspondant aux saturations éventuelles, comprenant pour chacune sa valeur et sa position au sein du segment résultant. La version antérieure de cette liste a été chargée au préalable depuis le disque dur. A l'issue de l'enregistrement en cours, la nouvelle version de cette liste sera sauvegardée sur le disque.
• sur disque dur :
  • les informations génériques (8) c'est-à-dire les échantillons de base de tous les instruments,
  • les informations spécifiques (9) correspondant à la partition de chacun des instruments,
  • le segment résultant (1),
  • la liste des retenues.

With reference to FIG. 3, the time diagram represents a period of an operating phase of a particular mode of the sound synthesizer device, operating in real time. In the operating phase described below, the real-time play of the current instrument replaces the previous version in real time, both in terms of digital audio in the resulting segment and in terms of specific information. It should be noted that the disk read or write operations, as well as sound output, are performed asynchronously, that is to say in parallel with the other operations.
This device includes:

• in RAM:
  • generic information (8), that is to say the basic samples of the current instrument, loaded beforehand from the hard disk.
  • the specific information (9) corresponding to the partition of the current instrument, loaded beforehand from the hard disk. The latter are therefore replaced in RAM as and when the data inputs (14) are made, but only after having been taken into account in the calculation of the component segments to be demixed. At the end of the recording in progress, they will be saved on the disc.
  • two buffer memories of the resulting segment, denoted MTSR1 and MTSR2, which can each contain, for example, a few seconds of its digital audio. MTSR1 and MTSR2 are divided into portions of the order of a few milliseconds, these subdivisions being denoted MTSR1 (i) and MTSR2 (i), i varying from 1 for the first subdivision to p for the last. So we have for example MTSR1 (1), MTSR1 (2) ... MTSR1 (p). The length and therefore the number of subdivisions have not been respected in the figure for reasons of clarity.
  • the list of deductions corresponding to any saturation, including for each its value and its position within the resulting segment. The previous version of this list was previously loaded from the hard drive. At the end of the current recording, the new version of this list will be saved on the disk.
• on hard drive:
  • the generic information (8), that is to say the basic samples of all the instruments,
  • the specific information (9) corresponding to the partition of each of the instruments,
  • the resulting segment (1),
  • the list of deductions.

• la synthèse de la partie du segment correspondant à cette subdivision,
• s'il est à mixer, l'addition de cette partie au contenu de la subdivision, y-compris la gestion des retenues correspondant aux saturations éventuelles.
• s'il est à démixer, la soustraction de cette partie du contenu de la subdivision, y-compris la gestion des retenues correspondant aux saturations éventuelles.

The calculation (12) or (13) of an MTSRx subdivision (i) comprises time management and updating of the state of the component segments to be synthesized as a function of time, of the data inputs (14) and of the specific information (9), and, for each component segment to be synthesized:

• the summary of the part of the segment corresponding to this subdivision,
• if it is to be mixed, the addition of this part to the content of the subdivision, including the management of deductions corresponding to any saturation.
• if it is to be demixed, the subtraction of this part of the content of the subdivision, including the management of the deductions corresponding to possible saturations.

For i = 1 to p-3, the end of the game (11) of a subdivision MTSR2 (i) triggers the game (11) of MTSR2 (i + 1), the calculation (13) of MTSR2 (i + 2), and the data entry record (14) (information specific) which will be taken into account in the calculation (13) of MTSR2 (i + 3) and following subdivisions.

The end of the game (11) of the MTSR2 subdivision (p-2) triggers the game (11) of MTSR2 (p-1), calculation (13) of MTSR2 (p), and recording (14) the entry of data (specific information) which will be taken into account in the calculation (12) of MTSR1 (1) and the subdivisions following.

The end of the game (11) of the MTSR2 subdivision (p-1) triggers the game (11) of MTSR2 (p), calculation (12) of MTSR1 (1), and recording (14) the entry of data (specific information) which will be taken into account in the calculation (12) of MTSR1 (2) and the subdivisions following.

The end of the game (11) of the MTSR2 subdivision (p) triggers the game (10) of MTSR1 (1), calculation (12) of MTSR1 (2), recording (14) data entry (specific information) to be taken taken into account in the calculation (12) of MTSR1 (3) and subdivisions following, and saving (17) the MTSR2 buffer in the 2m block of the resulting segment (1), the end of the latter recording triggering the loading (18) of the 2m + 2 block of the segment resulting (1) in MTSR2.

For i = 1 to p-3, the end of the game (10) of a subdivision MTSR1 (i) triggers the game (10) of MTSR1 (i + 1), the calculation (12) of MTSR1 (i + 2), and the data entry record (14) (information specific) which will be taken into account in the calculation (12) of MTSR1 (i + 3) and following subdivisions.

The end of the game (10) of the MTSR1 subdivision (p-2) triggers the game (10) of MTSR1 (p-1), calculation (12) of MTSR1 (p), and recording (14) the entry of data (specific information) which will be taken into account in the calculation (13) of MTSR2 (1) and subdivisions following.

The end of the game (10) of the MTSR1 subdivision (p-1) triggers the game (10) of MTSR1 (p), calculation (13) of MTSR2 (1), and recording (14) the entry of data (specific information) which will be taken into account in the calculation (13) of MTSR2 (2) and subdivisions following.

The end of the game (10) of the MTSR1 subdivision (p) triggers the incrementation of m, the set (11) of MTSR2 (1), the calculation (13) of MTSR2 (2), recording (14) of data entry (information specific) which will be taken into account in the calculation (13) of MTSR2 (3) and subsequent subdivisions, and registration (15) of the buffer memory MTSR1 in the block 2m-1 of the resulting segment (1), the end of this last recording triggering the loading (16) of the 2m + 1 block of the resulting segment (1) in MTSR1.

• des moyens d'entrée de données (6) tels qu'un clavier alphanumérique et un écran,
• des moyens de contrôle et de calcul (4) tels qu'une unité centrale Power Macintosh® munie d'un logiciel approprié,
• des moyens d'entrée de son (19) tels que le circuit audio interne et un microphone,
• des moyens de sortie de son (7) tels que le circuit audio interne et un casque stéréophonique,
• des moyens de mémoire (5) comprenant par exemple de la RAM et un disque dur, pour stocker le segment résultant (1), les pistes audionumériques (20) et les informations spécifiques (9).

With reference to FIG. 4, the digital audio recording device according to the invention comprises:

• data input means (6) such as an alphanumeric keyboard and a screen,
• control and calculation means (4) such as a Power Macintosh® central unit equipped with appropriate software,
• sound input means (19) such as the internal audio circuit and a microphone,
• sound output means (7) such as the internal audio circuit and a stereo headset,
• memory means (5) comprising for example RAM and a hard disk, for storing the resulting segment (1), the digital audio tracks (20) and the specific information (9).

• en RAM :
  • les informations spécifiques (9) correspondant aux traitements de la piste en cours, chargées au préalable depuis le disque dur. Ces dernières sont donc remplacées en RAM au fur et à mesure des entrées de données (14), mais seulement après avoir été prises en compte dans le calcul des traitements de l'ancienne version à démixer. A l'issue de l'enregistrement en cours, elles seront sauvegardées sur le disque.
  • deux mémoires tampon du segment résultant, notées MTSR1 et MTSR2, pouvant contenir chacune par exemple quelques secondes de son audionumérique. MTSR1 et MTSR2 sont divisées en portions de l'ordre de quelques millisecondes, ces subdivisions étant notées MTSR1(i) et MTSR2(i), i variant de 1 pour la première subdivision à p pour la dernière. On a donc par exemple MTSR1(1), MTSR1(2)... MTSR1(p). La longueur et donc le nombre des subdivisions n'ont pas été respectées dans la figure pour des raisons de clarté.
  • deux mémoires tampon d'entrée son MTES1 et MTES2, de même longueur que MTSR1 et MTSR2, et subdivisées de façon identique.
  • deux mémoires tampon de pistes audionumériques MTPA1 et MTPA2, de même longueur que MTSR1 et MTSR2, et subdivisées de façon identique.
  • la liste des retenues correspondant aux saturations éventuelles, comprenant pour chacune sa valeur et sa position au sein du segment résultant. La version antérieure de cette liste a été chargée au préalable depuis le disque dur. A l'issue de l'enregistrement en cours, la nouvelle version de cette liste sera sauvegardée sur le disque.
• sur disque dur :
  • les pistes audionumériques (20) c'est-à-dire les divers enregistrements,
  • les informations spécifiques (9) correspondant aux traitements de chacune des pistes,
  • le segment résultant (1),
  • la liste des retenues.

With reference to FIG. 5, the time diagram represents a period of an operating phase of a particular mode of the digital audio recording device, operating in real time. In the operating phase described below, the recording in progress replaces the previous version of the digital audio track in real time. The resulting segment and the specific information are also replaced in real time, according to the recording in progress and the data entries (25). It should be noted that the disk read or write operations, as well as sound output, are performed asynchronously, that is to say in parallel with the other operations. This device includes:

• in RAM:
  • the specific information (9) corresponding to the processing of the current track, loaded beforehand from the hard disk. The latter are therefore replaced in RAM as and when data entries (14) are made, but only after being taken into account in the calculation of the processing of the old version to be demixed. At the end of the recording in progress, they will be saved on the disc.
  • two buffer memories of the resulting segment, denoted MTSR1 and MTSR2, which can each contain, for example, a few seconds of its digital audio. MTSR1 and MTSR2 are divided into portions of the order of a few milliseconds, these subdivisions being denoted MTSR1 (i) and MTSR2 (i), i varying from 1 for the first subdivision to p for the last. So we have for example MTSR1 (1), MTSR1 (2) ... MTSR1 (p). The length and therefore the number of the subdivisions have not been respected in the figure for reasons of clarity.
  • two sound input buffers MTES1 and MTES2, the same length as MTSR1 and MTSR2, and subdivided identically.
  • two buffer memories of digital audio tracks MTPA1 and MTPA2, the same length as MTSR1 and MTSR2, and subdivided identically.
  • the list of deductions corresponding to any saturation, including for each its value and its position within the resulting segment. The previous version of this list was previously loaded from the hard drive. At the end of the current recording, the new version of this list will be saved on the disk.
• on hard drive:
  • the digital audio tracks (20), that is to say the various recordings,
  • the specific information (9) corresponding to the processing of each of the tracks,
  • the resulting segment (1),
  • the list of deductions.

• le traitement de la partie de l'ancienne version de la piste correspondant à cette subdivision et la soustraction du résultat du contenu de la subdivision, y-compris la gestion des retenues correspondant aux saturations éventuelles.
• le traitement de la partie de la nouvelle version de la piste correspondant à cette subdivision et l'addition du résultat au contenu de la subdivision, y-compris la gestion des retenues correspondant aux saturations éventuelles.

The calculation (23) or (24) of an MTSRx subdivision (i) includes time management and updating of the state of the tracks to be processed as a function of time, of the sound inputs (26,27), data inputs (25) and specific information (9), and, for each track to be processed:

• the processing of the part of the old version of the track corresponding to this subdivision and the subtraction of the result of the content of the subdivision, including the management of the deductions corresponding to any saturation.
• the processing of the part of the new version of the track corresponding to this subdivision and the addition of the result to the content of the subdivision, including the management of the deductions corresponding to any saturation.

For i = 1 to p-3, the end of the game (22) of a subdivision MTSR2 (i) triggers the game (22) of MTSR2 (i + 1), the calculation (24) of MTSR2 (i + 2), recording (27) of sound in MTES2 (i + 3), recording (25) data entry (specific information) to be taken taken into account in the calculation (24) of MTSR2 (i + 3) and subdivisions following.

The end of the game (22) of the MTSR2 subdivision (p-2) triggers the game (22) of MTSR2 (p-1), calculation (24) of MTSR2 (p), recording (26) sound in MTES1 (1) and recording (25) of data entry (specific information) which will be taken into account in the calculation (23) of MTSR1 (1) and the following subdivisions.

The end of the game (22) of the MTSR2 subdivision (p-1) triggers the game (22) of MTSR2 (p), calculation (23) of MTSR1 (1), recording (26) sound in MTES1 (2) and recording (25) of the data entry (specific information) which will be taken into account in the calculation (23) of MTSR1 (2) and subsequent subdivisions.

The end of the game (22) of the MTSR2 subdivision (p) triggers the game (21) of MTSR1 (1), calculation (23) of MTSR1 (2), recording (26) sound in MTES1 (3), recording (25) of data entry (specific information) which will be taken into account in the calculation (23) of MTSR1 (3) and subsequent subdivisions, and the recording (32) from the MTSR2 buffer in the 2m block of the resulting segment (1), the end of this last recording triggering the loading (33) of block 2m + 2 of the resulting segment (1) in MTSR2, the end of this last load triggering the recording (34) of MTES2 in the block 2n of the digital audio tracks (20), the end of the latter recording triggering the loading (33) of the 2n + 2 block of the tracks digital audio (20) in the MTPA2 buffer.

For i = 1 to p-3, the end of the game (21) of a subdivision MTSR1 (i) triggers the game (21) of MTSR1 (i + 1), the calculation (23) of MTSR1 (i + 2), recording (26) the sound in MTES1 (i + 3), recording (25) data entry (specific information) to be taken taken into account in the calculation (23) of MTSR1 (i + 3) and subdivisions following.

The end of the game (21) of the MTSR1 subdivision (p-2) triggers the game (21) of MTSR1 (p-1), calculation (23) of MTSR1 (p), recording (27) sound in MTES2 (1) and recording (25) of data entry (specific information) which will be taken into account in the calculation (24) of MTSR2 (1) and subsequent subdivisions.

The end of the game (21) of the MTSR1 subdivision (p-1) triggers the game (21) of MTSR1 (p), calculation (24) of MTSR2 (1), recording (27) sound in MTES2 (2) and recording (25) of data entry (specific information) which will be taken into account in the calculation (24) of MTSR2 (2) and subsequent subdivisions.

The end of the game (21) of the MTSR1 subdivision (p) triggers the increment of m and n, the set (22) of MTSR2 (1), the calculation (24) from MTSR2 (2), the recording (27) of sound in MTES2 (3), the data entry record (25) (information specific) which will be taken into account in the calculation (24) of MTSR2 (3) and subsequent subdivisions, and registration (28) of the buffer memory MTSR1 in the block 2m-1 of the resulting segment (1), the end of this last recording triggering the loading (29) of the block 2m + 1 of the resulting segment (1) in MTSR1, the end of the latter loading triggering the recording (30) of MTES1 in the block 2n-1 digital audio tracks (20), the end of the latter recording triggering the loading (31) of the 2n + 1 block of the tracks digital audio (20) in the MTPA1 buffer.

• des moyens d'entrée de données (6) tels qu'un clavier MIDI, un clavier alphanumérique et un écran,
• des moyens de contrôle et de calcul (4) tels qu'une unité centrale Power Macintosh® munie d'un logiciel approprié,
• des moyens d'entrée de son (19) tels que le circuit audio interne et un microphone,
• des moyens de sortie de son (7) tels que le circuit audio interne et un casque stéréophonique,
• des moyens de mémoire (5) comprenant par exemple de la RAM et un disque dur, pour stocker le segment résultant (1), les informations génériques (8), les pistes audionumériques (20) et les informations spécifiques (9).

Referring to FIG. 6, the sound synthesis and digital audio recording device according to the invention comprises:

• data input means (6) such as a MIDI keyboard, an alphanumeric keyboard and a screen,
• control and calculation means (4) such as a Power Macintosh® central unit equipped with appropriate software,
• sound input means (19) such as the internal audio circuit and a microphone,
• sound output means (7) such as the internal audio circuit and a stereo headset,
• memory means (5) comprising for example RAM and a hard disk, for storing the resulting segment (1), the generic information (8), the digital audio tracks (20) and the specific information (9).

The operation of the sound synthesis device and of digital audio recording according to the invention, in a mode particular implementation where it works in real time, can be also described as above with reference to FIG. 5, with the following details in relation to a recording device only:

• en RAM :
  • les informations génériques (8) c'est-à-dire les échantillons de base de l'instrument en cours, chargées au préalable depuis le disque dur.
  • les informations spécifiques (9) correspondant aux traitements de la piste en cours et/ou à la partition de l'instrument en cours, chargées au préalable depuis le disque dur. Ces dernières sont donc remplacées en RAM au fur et à mesure des entrées de données (14), mais seulement après avoir été prises en compte dans le calcul des segments composants à démixer. A l'issue de l'enregistrement en cours, elles seront sauvegardées sur le disque.
  • deux mémoires tampon du segment résultant, notées MTSR1 et MTSR2, pouvant contenir chacune par exemple quelques secondes de son audionumérique. MTSR1 et MTSR2 sont divisées en portions de l'ordre de quelques millisecondes, ces subdivisions étant notées MTSR1(i) et MTSR2(i), i variant de 1 pour la première subdivision à p pour la dernière. On a donc par exemple MTSR1(1), MTSR1(2)... MTSR1(p). La longueur et donc le nombre des subdivisions n'ont pas été respectées dans la figure pour des raisons de clarté.
  • deux mémoires tampon d'entrée son MTES1 et MTES2, de même longueur que MTSR1 et MTSR2, et subdivisées de façon identique.
  • deux mémoires tampon de pistes audionumériques MTPA1 et MTPA2, de même longueur que MTSR1 et MTSR2, et subdivisées de façon identique.
  • la liste des retenues correspondant aux saturations éventuelles, comprenant pour chacune sa valeur et sa position au sein du segment résultant. La version antérieure de cette liste a été chargée au préalable depuis le disque dur. A l'issue de l'enregistrement en cours, la nouvelle version de cette liste sera sauvegardée sur le disque.
• sur disque dur :
  • les informations génériques (8) c'est-à-dire les échantillons de base de tous les instruments,
  • les pistes audionumériques (20) c'est-à-dire les divers enregistrements,
  • les informations spécifiques (9) correspondant aux traitements de chacune des pistes et à la partition de chacun des instruments,
  • le segment résultant (1),
  • la liste des retenues.

In the operating phase described below, the current recording replaces in real time the previous version of the digital audio track concerned, and / or the real time play of the current instrument replaces in real time the previous version of this instrument. The resulting segment and the specific information are also replaced in real time, depending on the current recording and the data entries (25). It should be noted that the disk read or write operations, as well as sound output, are performed asynchronously, that is to say in parallel with the other operations.
This device includes:

• in RAM:
  • generic information (8), that is to say the basic samples of the current instrument, loaded beforehand from the hard disk.
  • specific information (9) corresponding to the processing of the current track and / or the partition of the current instrument, loaded beforehand from the hard disk. The latter are therefore replaced in RAM as and when data inputs (14) are made, but only after having been taken into account in the calculation of the component segments to be demixed. At the end of the recording in progress, they will be saved on the disc.
  • two buffer memories of the resulting segment, denoted MTSR1 and MTSR2, which can each contain, for example, a few seconds of its digital audio. MTSR1 and MTSR2 are divided into portions of the order of a few milliseconds, these subdivisions being denoted MTSR1 (i) and MTSR2 (i), i varying from 1 for the first subdivision to p for the last. So we have for example MTSR1 (1), MTSR1 (2) ... MTSR1 (p). The length and therefore the number of the subdivisions have not been respected in the figure for reasons of clarity.
  • two sound input buffers MTES1 and MTES2, the same length as MTSR1 and MTSR2, and subdivided identically.
  • two buffer memories of digital audio tracks MTPA1 and MTPA2, the same length as MTSR1 and MTSR2, and subdivided identically.
  • the list of deductions corresponding to any saturation, including for each its value and its position within the resulting segment. The previous version of this list was previously loaded from the hard drive. At the end of the current recording, the new version of this list will be saved on the disk.
• on hard drive:
  • the generic information (8), that is to say the basic samples of all the instruments,
  • the digital audio tracks (20), that is to say the various recordings,
  • specific information (9) corresponding to the processing of each of the tracks and the partition of each of the instruments,
  • the resulting segment (1),
  • the list of deductions.

• le traitement de la partie de l'ancienne version de la piste correspondant à cette subdivision et la soustraction du résultat du contenu de la subdivision, y-compris la gestion des retenues correspondant aux saturations éventuelles.
• le traitement de la partie de la nouvelle version de la piste correspondant à cette subdivision et l'addition du résultat au contenu de la subdivision, y-compris la gestion des retenues correspondant aux saturations éventuelles,

et/ou, pour chaque segment composant à synthétiser :

• la synthèse de la partie du segment correspondant à cette subdivision,
• s'il est à mixer, l'addition de cette partie au contenu de la subdivision, y-compris la gestion des retenues correspondant aux saturations éventuelles.
• s'il est à démixer, la soustraction de cette partie du contenu de la subdivision, y-compris la gestion des retenues correspondant aux saturations éventuelles.

The calculation (23) or (24) of an MTSRx subdivision (i) includes time management and updating of the state of the tracks to be processed and / or the component segments to be synthesized, as a function of time, of the sound inputs (26,27), data inputs (25) and specific information (9), and, for each track to be processed:

• the processing of the part of the old version of the track corresponding to this subdivision and the subtraction of the result of the content of the subdivision, including the management of the deductions corresponding to any saturation.
• the processing of the part of the new version of the track corresponding to this subdivision and the addition of the result to the content of the subdivision, including the management of the deductions corresponding to any saturation,

and / or, for each component segment to be synthesized:

• the summary of the part of the segment corresponding to this subdivision,
• if it is to be mixed, the addition of this part to the content of the subdivision, including the management of deductions corresponding to any saturation.
• if it is to be demixed, the subtraction of this part of the content of the subdivision, including the management of the deductions corresponding to possible saturations.

The sound synthesis and recording device digital audio according to the invention allows a person to realize, instrument by instrument and recording by recording, a complex piece comprising many instruments and recordings, while allowing him to return to the any instrument or recording to change the score, settings, editing or sound content.

This device can therefore simply consist of a microcomputer such as an Apple® Power Macintosh® running a unique program fulfilling the functions of MIDI sequencer and digital audio, polyphony synthesizer and sampler and infinite multitimbral, direct disk recorder with number of infinite channels, mixing, editing, effects generator, correction automatic saturation, song recording.

In order to compact the file or folder relating to a whole song keeping the resulting segment, we can delete a track audio, or more if they do not overlap in time: to recover them we can always demix from the resulting segment other instruments or tracks superimposed on them.

Obviously, as in existing systems, keep that specific leads and information and rebuild the resulting segment in due time. Such a reconstruction will also solve any problem of hardware or software origin affecting the resulting segment.

The method and the devices according to the invention allow also other applications such as PA, including postsynchronization and dubbing, of films or applications multimedia with music, sound effects, dialogues, special effects etc, using for example Apple® QuickTime® software tools for synchronizing sound with image and / or other media.

They can also be used as part of learning composition and / or interpretation musical.

Claims (9)

1. Method for creating and updating a resultant segment of digital sound (1, 1a, 1b) composed of mixed component segments of digital sound, to be used in devices for generating and/or editing synthesized sounds and/or sounds obtained from recorded digital audio tracks, comprising the following step :

   (a) obtaining, and, by a first digital operation, mixing into the current version (1) of the said resultant segment, certain (2,3) of the said component segments in order to compute the updated version (1a) of said resultant segment - the informations about (on the one hand) obtaining them identically and about (on the other hand) their sample-point accurate position within the said resultant segment being memorized;
   characterized in that it also comprises the following step :

   (b) reobtaining identically, and by a second digital operation, mixing out of the updated version (1a) of the said resultant segment, certain (2) of the said component segments mixed in previously in order to compute the reupdated version (1b) of the said resultant segment - by using the said memorized informations relating on the one hand to their obtaining and on the other hand to their position within the said resultant segment,

   and in that the said steps (a) and (b) take place according to the changes in the composition of the said resultant segment.
2. Method according to claim 1, characterized in that :

   the first digital operation consists in adding to the said current version (1) of the said resultant segment each component segment to be mixed in (2,3) in order to compute the updated version (1a) of the said resultant segment, the said addition taking place sample-point by sample-point starting off from the sample-point corresponding to the desired position of the said component segment within the said resultant segment;

   the second digital operation consists in subtracting from the said updated version (1a) of the said resultant segment each component segment to be mixed out (2), in order to compute the reupdated version (1b) of the said resultant segment, the said subtraction taking place sample-point by sample-point starting off from the sample-point corresponding to the position of the said component segment (2) within the said resultant segment.
3. Apparatus of synthesis of digital sound, comprising at least :

   control and computing means (4),

   memory means (5) associated with the said control and computing means (4)

   data input means (6) associated with the said control and computing means (4),

   sound output means (7) associated with the said control and computing means (4),

      characterized in that :

   It uses the method according to any of the claims 1 or 2,

   The component segments of digital sound (2,3) are, at least in part, synthesized by the control and computing means (4);

   The informations relative to the obtaining identically and to the sample-point accurate position of the said component segments are stored in the memory means (5),

   the resultant digital sound segment (1), recorded in the memory means (5), constitutes the current version of the mixing of synthesized component segments and other possible component segments;

   The said control and computing means (4) carry out, according to information received from the data input means (6), step (a) of the said method for possible component segments to be mixed into the resultant segment (1), and step (b) of the said method for the possible component segments to be mixed out of the resultant segment (1);

   The said control and computing means (4) send the said resultant segment (1) toward the said sound output means (7), after optional global non-destructive processing.
4. Apparatus of synthesis of digital sound according to claim 3, characterized in that:

   the resultant segment (1) of digital sound is stored in mass storage devices,

   the control and computing means (4) transfer each section to be modified of the resultant segment (1), from the said mass storage devices, toward at least one buffer memory so as to carry out the said method, then transfer the new version of the said section toward the mass storage devices.
5. Apparatus of digital sound recording, comprising at least :

   control and computing means (4),

   memory means (5) associated with the said control and computing means (4),

   data input means (6) associated with the said control and computing means (4),

   sound input means (19) associated with the said control and computing means (4),

   sound output means (7) associated with the said control and computing means (4),

   characterized in that :

   It uses the method according to any of the claims 1 or 2,

   The component segments (2,3) of digital sound are, at least in part, obtained from at least one digital audio track (20) recorded in the said memory means (5),

   The data relating to the obtaining identically and to the sample-point accurate position of the said component segments, composed partly of the said recorded digital audio tracks (20) and partly of specific data (9), are stored in the memory means (5);

   The resultant segment (1) of digital sound, recorded in the memory means (5), constitutes the current version of the mixing of component segments obtained using the said digital audio tracks (20) and other possible component segments;

   The said control and computing means (4) carry out, according to the data received via at least one of the said input means (6,19), step (a) of the said method for the optional component segments to be mixed into the resultant segment (1), and step (b) of the said method for the optional component segments to be mixed out of the resultant segment (1);

   The said control and computing means (4) send the said resultant segment (1) toward the said sound output means (7), after optional non-destructive global processing.
6. Apparatus of digital sound recording according to claim 5, characterized in that:

   The resultant segment (1) of digital sound is stored in mass storage devices;

   The digital audio tracks (20) are stored, at least in part, in mass storage devices;

   the control and computing means (4) transfer each section to be modified of the said resultant segment (1), and each digital audio track (20) section necessary for obtaining the optional component segments to be mixed in and the optional component segments to be mixed out, from the said mass storage devices, toward at least one buffer memory means for the said resultant segment (1) plus at least one buffer memory means for the digital audio tracks (20), in order to carry out the said method, and then transfer the possible new versions of the said track sections and the new version of the said section of the resultant segment toward the mass storage devices.
7. Apparatus according to any of the claims 3 to 6 characterized in that, for at least part of the modifications of the resultant segment (1), the control and computing means (4) carry out the said method in real time, and in real time send the said resultant segment (1) as it is updated toward the said sound output means (7), after optional global non-destructive processing.
8. Apparatus according to any of the claims 3 to 7 characterized in that :

   The data allowing the obtaining of at least the beginning of the component segments are stored in rapid memory means, at least for a subset of the said component segments;

   In accordance with the creation's requirements, the control and computing means (4) can modify the said subset by transfer of data from the mass storage devices toward the said rapid memory means.
9. Apparatus of digital sound synthesis and recording characterized in that it features both the characteristics of any of the claims 3,4,7,8 concerning its synthesis functions, and the characteristics of any of the claims 5 to 8 concerning its recording functions, and in that the resultant segment (1) of digital sound, recorded in the memory means (5), constitutes the current version of the mixing of synthesized segments and segments obtained from recorded digital audio tracks (20).

Images