FR2845195A1 - METHOD FOR PREPARING AND RETURNING A SOUND SYSTEM REPRESENTING ACOUSTIC CHARACTERISTICS OF A MOTOR VEHICLE, DEVICE FOR IMPLEMENTING IT - Google Patents

Abstract

The invention is characterized by a method for developing a sound library of digitized sounds recorded with a frequency transposition step on the recorded sounds so that they all have the same fundamental frequency F0. It is also characterized by a restitution process in which the sounds are selected according to the current operation of the motor measured by one or more parameters then change of scale of the measurement to the theoretical value of the parameters (the parameter scales n ' being not the same between the recording vehicle and the rendering motor vehicle). During the selection, a calculation can be carried out in order to refine the restitution (in particular weighted combination and / or interpolation). Finally, the selected sounds which are always at the fundamental frequency F0 must be frequency shifted by an interpolation or decimation as a function of a fundamental restitution frequency Fr determined as a function of the measurements. An implementation device with standby means is also considered.

Description

The invention relates to methods for preparing and restoring a

  sound library representing acoustic characteristics of a motor vehicle in the form of sound sequences. It has applications in the automotive field and in recreation by simulating the sounds of an engine or other elements of an automobile. A device for implementation

is also claimed.

  The essence of the invention is to simulate inside the passenger compartment of a motor vehicle the sound of an engine and / or other organs including in particular the brakes, the exhaust, a another type of vehicle and preferably a prestigious or original vehicle. Thus, the driver on board his vehicle will have the impression of being inside a Ferrari or any other type of vehicle that may interest him. 15 In order to reinforce the subjective impression, the simulation is

  based on the actual operating condition of the engine.

  It can also be implemented on or in other types of vehicles than a car and, for example, a truck, a

bus, a motorcycle.

  Devices are known for simulating engine sounds in a car interior. In particular, US Pat. No. 6,356,185 discloses a digital device with a programmable filter as a function of the engine load. Also known from application WO 01/33543 are methods and devices for developing and using a sound library for the purpose of generating sounds as a function of the conditions of use of the engine, a sound synthesis is implemented there. The present invention provides an alternative to these known devices in which sounds are used which are processed before being recorded in the device. These sounds are chosen according to the operating conditions of the engine, in particular the frequency of rotation, this allowing

  also activation or standby of the device.

  In addition, in the invention, a specific processing of the recording sound makes it possible to reduce the amount of data to be stored in memory unlike the devices in which the sounds are recorded raw or with simple filtering. With 5 the same amount of memory, more simulation possibilities are available. This treatment requires calculations upon return. However, these calculations are simplified and reduced by

  report to the processes implementing a synthesis.

  The invention therefore relates to a method of developing a sound library representing the acoustic characteristics of sequences of sounds recorded from at least one reference motor vehicle, each sequence corresponding to operating conditions of the reference vehicle, the conditions being defined by at least one operating parameter, the recorded sounds being intended to be reproduced later on a motor vehicle by selection as a function of the operating conditions of the motor vehicle on which the reproduction takes place, at least one of

  parameters being the frequency of rotation of the motor.

  According to the invention of the method, during the preparation of the sound library: - sequences of sounds corresponding to operating parameters are recorded in the form of digital data by digitization at a sampling frequency Fe, said operating parameters being recorded in parallel, - a frequency analysis of each of said sequences is carried out in order to determine the fundamental frequency of each of the sequences, - a reference frequency Fo is determined from the fundamental frequencies of the sequences, - a frequency transposition is carried out on each of the

  sequences recorded so that they all have the same fundamental frequency substantially equal to the reference frequency Fo in order to form the sound library.

  In various embodiments of the invention of a process for developing a sound library, the following means used alone or in all their technically possible combinations are used before recording or after recording, the sounds are passed through a filter band pass, - the band pass filter is intended, on the one hand, to eliminate the frequency components which are too high compared to the sampling frequency for digitization, the cut-off frequency of the low pass part corresponding substantially to half or less than the sampling frequency, on the other hand, by the high pass part to eliminate the too low frequency components which cannot be restored by an acoustic amplification and restitution means, the operating state is evaluated by a or more of the following parameters: - starter engaged or not, - position C of the accelerator pedal, 20 - frequency of rotation R of the engine, - the intake vacuum, - the gearbox engaged, - the position of the brake pedal, - the front / rear acceleration or deceleration measured by a sensor, - the rotation of the steering wheel, the lateral acceleration or deceleration measured by a sensor, - the sounds are chosen from the following source types 30 - the sounds of the reference vehicle in general, - the sounds of the engine of the reference vehicle, - the transmission sounds of the reference vehicle, - braking sounds of the reference vehicle, - rolling sounds, in particular screeching of tires, of the vehicle, of reference - exhaust sounds of the reference vehicle, a parameter of sound type being also recorded, - sounds from different source types have the same common reference frequency Fo, - sounds are recorded from several types of reference vehicles, a parameter of ty eg vehicle being recorded, - the frequency analysis is carried out on a frequency curve obtained by a fast Fourier transform 10 (FFT) of the recorded digital sounds, the fundamental frequency corresponding to the main vertex of said curve, - in addition, each sequence is cut into blocks, each block comprising Ne samples and being of duration Te reduced 15 compared to the total duration of the corresponding sequence, Te = Ne / Fe, and the frequency analysis and the transposition are performed on each blocks in such a way that the method corresponds to the following steps: - the sequences of sounds corresponding to operating parameters are recorded in the form of digital data at a sampling frequency Fe, said operating parameters being recorded in parallel, - divides each sequence into n sequential blocks, 25 each block comprising Ne samples for a duration Te = Ne / Fe, - a frequency analysis of each of said blocks is carried out in order to determine the fundamental frequency of each of the blocks, - a reference frequency Fo is determined from the fundamental frequencies of the blocks, - a frequency transposition is carried out on each of the blocks recorded so that they all have the same fundamental frequency substantially equal to the reference frequency Fo, - the number n of blocks for a sequence is greater than or equal to 1 and preferably between 1 and 5, - the reference frequency Fo of the blocks is common to the blocks of a given sequence, different sequences may have different reference frequencies Fo for their respective blocks, - preferably, the reference frequency Fo of the blocks is common to all the blocks of all the sequences , different sequences having for their respective blocks identical reference frequencies Fo, the frequency of reference Fo is calculated by fundamental frequencies of each of the blocks respectively, - the reference frequency Fo is calculated minimum of the fundamental frequencies of sequences or blocks respectively, - the reference frequency Fo is calculated maximum of the fundamental frequencies of sequences or blocks respectively , an average sequence or per take each per take each of the of the - the sampling frequency is greater than or equal to 32KHZ and preferably equal to 44.1KHz, - a digitization resolution of at least 14bits and preferably 16bits,

  the sound library is stored in an electronic memory of the integrated circuit type.

  It is understood that the parallel recording of the parameters corresponding to the operating conditions is intended to subsequently allow the selection of the sounds as a function of the operating conditions of the engine on which the playback takes place. If this recording of the parameters can be continuous, preferably, for the restitution, each sequence or block is associated with a minimum number of parameter values, in particular in the case of blocks, only one parameter value is stored for each parameter of the block defining the operating conditions, this in order to reduce the amount of information associated with each sequence or block of sounds. For example, in memory, in a first set of stored data, the directory, we have for each sequence or block, the values of the parameters and addresses (physical or symbolic address 5: file name for example) of start and end of the sequence or block (we can also have the start address and the size), the other stored data sets, the sound files, corresponding to sequences or blocks of sounds. Thus, for the reproduction, the device determines the current operating condition, searches the directory for a correspondence which will enable it to determine the address of the sound data. In addition, data concerning the frequency transpositions and scaling of the parameters are stored so that the rendering device can correctly calculate the

return conditions.

  the sound sequences can correspond to changing operating conditions of the reference vehicle, for example a progressive depressing of the accelerator pedal so that the engine speed rises during the sequence. In such a case, the block cutting is intended to form blocks which, each in themselves, correspond to relatively stable conditions of the parameters, for example 1st block at 1000 rpm, 2nd block at 1100 rpm, 3rd at 25 1200rpm, etc. We may have to keep only

  some blocks depending on the definition step you want.

  The sequences can also correspond to stable operating conditions. In all cases, a block must correspond to substantially stable operating conditions.

  The choice of the reference frequency Fo makes it possible to optimize the process according in particular to quality objectives

of reproduction.

  The invention also relates to a method of restoring recorded sounds from a sound library according to any one of the preceding characteristics, taken individually or in combination, by selection of recorded digital sounds whose fundamental frequency is substantially equal to the reference frequency Fo , as a function of operating conditions 5 of a motor vehicle on which restitution takes place, the operating conditions being defined by operating parameters and at least one of the parameters being the

  frequency of rotation of said motor.

  According to the restitution method, the operating state of the motor vehicle is determined by measuring at least one current operating parameter and in that, for each operating parameter considered: a theoretical value of the parameter is determined. current operation for the reference vehicle from the measurement of the current operating parameter, - the digitized sounds recorded whose parameter corresponds to the theoretical value are selected, - the fundamental frequency of reproduction Fr of the sounds is calculated as a function of the parameter of current operation, - a frequency shift is carried out of the selected recorded digital sounds so that they have a fundamental frequency substantially equal to the fundamental frequency of restitution Fr, said shift being an interpolation or decimation as a function of the respective values of the fundamental frequency of restitution Fr and the frequency of Fo reference, and in that: - we send all the selected and shifted sounds in

  a means of digital analog conversion, of acoustic amplification and restitution.

  In various embodiments of the invention of the method for developing a sound library, the following means used alone or in all their technically possible combinations are used: a weighting calculation is performed on the sounds to be reproduced, the weighting depending on the measurements of the current parameters, the selection of the digitized sounds recorded comprises in addition a calculation step consisting in carrying out a weighted combination of sounds during which sounds whose parameters are close to the theoretical value are associated, ( in practice numerically equal or close) - the weighting coefficients are calculated, - the calculation step consists in carrying out a weighted combination of sounds during which we associate sounds whose parameter is closest less than or equal to the value theoretical with sounds with the closest parameter greater than or equal to the theoretical value, - the calculation step consists of interpolating the theoretical value of the parameter from the recorded sounds, - we combine sounds from several different source types for reproduction, - we use a sound library with digitized sounds recorded in the form of sequences, the sounds being restored by sequence, (or looping of sequences) - we use a sound library with digital sounds recorded in the form of sequences cut into blocks, and 25 beams for each period of time Te - we determine the operating state of the machine at engine by measuring at least one current operating parameter and in that for each operating parameter considered: - a theoretical value of the current operating parameter for the reference vehicle is determined from the measurement of the current operating parameter, - a weighted combination 35 of n blocks is selected and calculated, the parameters of which are close to d e the theoretical value by calculating a set of n weighting coefficients associated with the n selected blocks then weighting of each of the n blocks by its weighting coefficient obtained and summation of the weighted blocks, - the fundamental restitution frequency Fr is calculated as a function of the current operating parameter, - a frequency shift of the weighted combination is carried out so that it has a fundamental frequency 10 substantially equal to the fundamental restitution frequency Fr, said offset being an interpolation or decimation as a function of the respective values of the frequency fundamental of restitution Fr and of the reference frequency Fo, and in that: - the offset combination is sent in a means of digital analog conversion, of amplification and of acoustic restitution, - the sampling frequency is greater than or equal to 20 32KHZ and preferably equal to 44.1KHz, - a resolution of n is used umeration of at least 14bits and preferably 16bits, - the sound library is stored in an electronic memory of the integrated circuit type, - the reproduced sounds are amplified and the means of acoustic restitution of the sounds is the equipment in loudspeakers of the machine with a motor in which the restitution takes place, - in addition by a standby means, the frequency of rotation of the motor is measured and when said frequency 30 falls below a predetermined threshold, the electrical supply of the means intended for restitution, and when the frequency goes back above said threshold, the power supply is restored, - when the duration of the restitution is greater than the duration 35 of the selected sounds, said restitution is looped in

  restarting playback at the start of the selected sounds.

  (a sequence or one or more blocks is therefore looped, before or after transposition, in particular according to the availability in storage memory of the selected sounds transposed) The sound library and the restitution can therefore be more or less extensive. In a minimal version, the sound library only includes sounds from a single source and a single type of vehicle, in this case the parameters intended to allow the selection of sounds may not include any source type parameter and / or of vehicle types, only the operating parameters, for example engine speed (rotation frequency), then being used. In more advanced versions, several sounds of different source types can be recorded in a sound library with their parameters 15 corresponding source types and the restitution can combine sequences / blocks of several source types. It is understood that in this case, it is preferable that the reference frequency Fo is the same for all the sequences / blocks of all the source types in order to simplify the restitution. In the most advanced version in which the sound library includes sounds coming from several types of vehicles, it is understood that the restitution will only concern a particular type of vehicle, the sequences / blocks being selected according to the parameter

corresponding vehicle type.

  The present invention finally relates to a device for implementing the above methods and in particular for the restitution from the sound library of digitized sounds recorded and transposed to a reference frequency Fo. The device comprises at least one computer means of the microprocessor, microcontroller or digital signal processing (DSP) type, program memory for execution according to the restitution method, sound memory comprising the sound library, means of interpolation / decimation, said interpolation / decimation means preferably being software and at least one digital converter

analog (DAC).

  In various embodiments of the invention of the device, the following means used alone or in all their technically possible combinations are used: - the device also comprises a standby means whose operation is linked to the frequency of rotation of the motor, - the standby corresponds to an interruption of the supply to the playback device, the standby means continuing to be supplied, - the interruption means is an electromagnetic relay controlled by the setting means in standby, the rest position, lifted off, corresponding to the interruption, - the restitution is carried out in stereophony, the recording

  being itself stereophonic, two CAN being employed.

  Thus, the invention is characterized during the development of the sound library by a frequency transposition step on the recorded sounds so that they all have the same fundamental frequency Fo. During the restitution, the sounds are selected according to the current operation of the motor measured by one / several parameters then change of scale of the measurement to theoretical value of the parameter (s) (the scales of parameters not being the same between the vehicle 25 of the recording and the engine with restitution engine). During the selection, a calculation can be made in order to refine the restitution (in particular weighted combination or interpolation). Finally, the selected sounds which are always at the fundamental frequency Fo must be offset frequently by an interpolation 30 or decimation as a function of a fundamental restitution frequency Fr determined as a function of the measurements (or, rather, of

  the theoretical value of the parameter (s) derived therefrom).

  The present invention will be better understood on reading the following of a nonlimiting exemplary embodiment o: FIG. 1 corresponds to a block diagram of a device for the restitution of a sound library, FIG. 2, an electronic diagram of a Figure 3 represents a functional diagram of implementation of a restoring device on standby

of the invention.

  In Figure 1, which schematically shows a device usable for simulating the sound of an engine in a motor vehicle from a sound library, there is a controller 4, which can be a microcontroller, a microprocessor or a signal processor digital (DSP), connected by addressing and data buses to memories. A read only memory 6 for storing the operating program. The ROM also contains the sound library (digital samples and parameters) and any configuration data intended to establish a relationship between the parameters on recording and the current parameters, for example the range of frequency (speed) of the motor which can be different between the registration car and the return car. The controller 4 is also associated with random access memory 7 and with a mono or stereo digital or analog converter 8 (DAC) (or even more) as the case may be. An amplifier 9 is provided at the output of the DAC. A means for adjusting the sound level being able to be implemented by software by logical shift operations of the digital data, a conventional potentiometer type adjustment means has not been shown In the Figure two parameters for evaluating the operating state engine speed and load are considered. The load C of the engine 2 is evaluated by the position 30 of the accelerator. The engine R speed 1 (rotation frequency) is evaluated by the engine rotation speed (specific sensor). The corresponding signals 1 and 2 are sent to the controller 4 via a module 3 for shaping the signals. The signal shaping module 3 transforms the analog signal of the engine speed R into a filtered logic signal whose frequency translates the state of the engine. The engine R speed 1 and the engine load C 2 are evaluated regularly over time and determine the selection of sounds from the sound library. In the case of sounds in sequences 5 cut into blocks, in a simplified version, a block from all the blocks stored in the read-only memory is selected. In a more advanced version, a weighted combination of blocks is used for the restitution. Indeed, the sound library is physically limited and all the operating conditions of the reference vehicle could not be recorded. It is therefore necessary to use sounds corresponding to parameters close to the theoretical value which must be restored to produce sounds corresponding to current operation. For example, for a parameter which is the motor rotation frequency, the reference vehicle goes from 800 revolutions / min to 9000 revolutions / min, the engine of the motor vehicle on which the restitution takes place ranges from 500 revolutions / min to 4000 rpm and it is therefore necessary, from a current measurement for example equal to 3000 rpm, to determine a theoretical value corresponding to that of the reference vehicle. For this we perform a

  conversion which gives a theoretical value at around 6658tours / mn. The sound library includes, for example, sequences / blocks at ... 5000, 5500, 6000, 6500, 7000 ... revolutions / min.

  A weighted combination of the 25 sequences / blocks is then carried out at 6500 and 7000 revolutions / min. In the most advanced mode, an interpolation is made from the available sounds. After this step, the sounds are always at a reference frequency F0 and it is necessary to perform a frequency shift to correspond to the current operating mode. After these calculations, the data is sent to the digital-analog converter 8, the analog signals thus reconstituted being amplified and filtered by the amplifier 9 which includes a filter and sent to an acoustic reproduction means 10 of the loudspeaker type. Depending on the quality level 35 required, the available computing power and / or the

  memory capacity, the resolution of the converters can be chosen between 8 and 24 bits. Preferably 16 bits are used. The acquisition frequency is preferably 441 00Hz.

  However, it is envisaged that the frequency of acquisition and of restitution will be lower and, for example, up to 8,000 Hz also depending on the level of quality required, the computing power and / or the memory capacity. Likewise it is envisaged

  a stereophonic device or not.

  A standby device depending on the engine speed 10 R enables the restitution device to be switched off when the engine of the vehicle in which the device is installed is stopped. When the engine speed R is zero, the restitution device is de-energized so that the electrical consumption of the system used is almost zero, that is to say reduced to the consumption of said standby device. A perfectly explicit electronic diagram alone is given in FIG. 2 with a low-consumption CMOS integrated circuit of the stable mono flip-flop type, an input of which is connected to a clipping amplifier receiving a filtered signal (RICI) 20 corresponding to the rotation frequency. of the motor. At each pulse of this signal the flip-flop is reset to zero. If there is no reset, after a time defined by the R6C3 components, a switchover occurs and the KI relay is switched, which cuts the power to the playback device. In other embodiments of the invention, the amplifier 9 can be omitted, the output of the digital-analog converter being connected to an existing sound device of the motor vehicle and for example to a car radio. In another less advanced form of the invention, the measurement of the load 30 C of the engine 2 can be omitted, only the speed R of the engine 1 being

measured or vice versa.

  FIG. 3 represents a general functional diagram of the method of the invention which, for reasons of simplification, considers only a single sound source depending on a single parameter, the engine speed R. It is understood that the method which there is shown can be extended to any number of source types and parameters. On this same Figure, R is the engine speed, S the sound source, B1 ... B, are n sound blocks (sounds digitized and transposed to the reference frequency Fo), E1 ... The n sound samples are

  corresponding to the blocks and C1 ... Cn are n weighting coefficients. The configuration table 11 corresponds to the set of weighting coefficients of the blocks of the sequences. For each operating state of the engine, a set of n weighting coefficients of the n blocks is calculated.

  Thus in the case where the operation of the engine is measured by the engine speed, the values of the weighting coefficients will be referenced as a function of the corresponding values of R. The values of these weighting coefficients Ci (i = 1 to n) selected in the configuration table 11 as a function of R are sent to a module 12 of additive synthesis or they are associated with the samples El (i = 1 to n) of the corresponding sound blocks for weighting and the set of weighted blocks is added, allowing d '' obtaining in the time domain a time evolution of fundamental frequency Fo. Thus, in this embodiment of Figure 3, from n selected blocks of unit duration Te, we obtain by

  weighting and adding the equivalent of a single block of duration Te.

  A frequency shift (transposition) is then operated, calculated according to the operating state of the motor. If the operating state of the motor requires a sound reproduction of a fundamental frequency greater than Fo, the previously obtained sound sequence is decimated, in the opposite case, it

is interpolated.

  We will now describe the method for calculating the weighting coefficients. For each of the blocks of the sound library are determined a lower limit restitution frequency (minimum frequency transposition), an upper limit restitution frequency (maximum frequency transposition) as well as, between these two limits, a range (or value) of frequencies maximum coefficient return (C0 = 100%). In the case where the frequency of restitution calculated according to the operating conditions of the motor vehicle is lower than the lower limit or is higher than the upper limit 5, the calculated coefficient is zero (Ci = 0) and between these two limits, outside the range with maximum coefficient (CI = 100%), the coefficient is calculated by linear extrapolation. Thus in the case of an example with three blocks, the following table is considered in which each box of the table has a (or a range of) frequency and its associated coefficient in percentage (except in the case where another box allows deduct the value as for block 1 x min frequency where we deduct 100% for

  any frequency below 1000Hz.

  frequency frequency range frequency min max block 1 - less than 1000Hz / 100% 2000Hz / 0% block 3 100OHz / 0% 2000Hz / 100% 3000Hz / 0% block 3 2000Hz / 0% greater than 3000Hz / 100% 15 For a frequency with a value calculated at 1600 Hz, we have: block 1: Ci = (2000 - 1600) / (2000 - 1000) 40% block 2: Ci =. (1600 - 1000) / (20M - 1000) = 60 % 20 block 3: Ci (frequency lower than 2000 Hz) = 0% For a restitution frequency calculated at 4000 Hz, we have: block 1: Ci (frequency higher than 2000 Hz) = 0% block 2: Ci (higher frequency at 2000 Hz) = 0% 25 block 3 Ci (frequency higher than 3000Hz) = 100% Depending on the available and / or selected blocks, combinations of blocks may be more complex than in this

example.

  In the time domain, the sampling is done in a manner compatible with that used for the creation of the configuration table, which is identical: the same sampling frequency Fe is used during creation and during restitution / simulation. During the implementation of the methods of the invention, the restitution is preferably carried out by using a temporary storage memory area (buffer or "buffer") of the first in, first out type and / or by using a system with 10 double buffers so that the calculations can be done during the reproduction of the sound by the digital analog converter 10 without the latter having to wait for data and therefore without

  risk of interruption of the restitution.

Claims (15)

  1. Method for developing a sound library representing the acoustic characteristics of sequences of recorded sounds of at least one reference motor vehicle, each sequence corresponding to operating conditions of the reference vehicle, the operating conditions being defined by at least one operating parameter, the recorded sounds being intended to be reproduced subsequently on a motor vehicle by selection according to the operating conditions of the motor vehicle on which the reproduction takes place, at least one of the parameters being the frequency of rotation of the motor, characterized in that, during the preparation of the sound library: - sequences of sounds corresponding to operating parameters are recorded in the form of digital data by digitization at a sampling frequency Fe operating parameters being recorded in parallel, - a frequency analysis of each of said sequences in order to determine the fundamental frequency of each of the sequences, - a reference frequency Fo is determined from the fundamental frequencies of the sequences, - a frequency transposition is carried out on each of the recorded sequences so that they have all the same fundamental frequency substantially equal to the frequency of   Fo reference to form the sound library.   2. Method according to claim 1, characterized in that the operating state is evaluated by one or more of the following parameters, starter engaged or not, - the position C of the accelerator pedal, the rotation frequency R of the motor , 35 - intake vacuum, gearbox engaged, - brake pedal position, - front / rear acceleration or deceleration measured by a sensor, - steering wheel rotation, - lateral acceleration or deceleration measured by a sensor. 3. Method according to claim 1 or 2 characterized in that the sounds are chosen from the following source types 10 - the sounds of the reference vehicle in general, - the sounds of the engine of the reference vehicle, - the sounds of the transmission of the reference vehicle, - the braking sounds of the reference vehicle, - the rolling sounds, in particular screeching of the tires, of the reference vehicle - the escape sounds of the reference vehicle,   a sound type parameter being further recorded.   4. Method according to any one of the claims   above, characterized in that, in addition, each sequence is divided into sequential blocks, each block comprising Ne samples and being of a duration Te reduced compared to the total duration of the corresponding sequence, Te = Ne I Fe, and the frequency analysis and the transposition are carried out on each block so that the method corresponds to the following steps - the sequences of sounds corresponding to operating parameters are recorded in the form of digital data at a sampling frequency Fe, said operating parameters being recorded in parallel, - each sequence is divided into n sequential blocks, each block comprising Ne samples for a duration Te = Ne / Fe, - a frequency analysis of each of said blocks is carried out in order to determine the fundamental frequency of each of the 35 blocks, - a reference frequency Fo is determined from the fundamental frequencies of the blocks, - a frequency transposition is performed on each of the recorded blocks so that they all have the same fundamental frequency 5 substantially equal to the reference frequency Fo 0   5. Method according to any one of the claims   above, characterized in that the reference frequency Fo is calculated by an average of the fundamental frequencies of each of the sequences or blocks respectively and in that, preferably, the reference frequency Fo of the blocks is common to all the blocks of all the sequences, different sequences having for their respective blocks frequencies of identical Fo reference.   6. Method according to any one of the claims   above, characterized in that the sampling frequency Fe is greater than or equal to 32KHZ and preferably equal to 44.1KHz, and a resolution of   digitization of at least 14bits and preferably 16bits.   7. Method according to any one of the claims   previous, characterized in that the sound library is stored   in an electronic memory of the integrated circuit type.   8. Method for restoring recorded sounds by selecting digitized sounds recorded as a function of operating conditions of a motor vehicle on which the   restitution, the operating conditions being defined by operating parameters and at least one of the parameters being the frequency of rotation of said motor, characterized in that the sounds are in a sound library 30 produced according to any one of the claims   above, the digitized sounds having a fundamental frequency substantially equal to the reference frequency Fo,. and in which the operating state of the motor vehicle is determined by measuring at least one current operating parameter and in that for each operating parameter considered: - a theoretical value of the current operating parameter for the reference vehicle is determined from the measurement of the current operating parameter, - the digitized sounds recorded whose parameter corresponds to the theoretical value are selected, - the fundamental frequency of reproduction Fr of the sounds is calculated as a function of the current operating parameter, - a frequency shift of the digitized recorded sounds selected is carried out so that they have a fundamental frequency substantially equal to the fundamental frequency of reproduction Fr, said offset being an interpolation or decimation in function tion of the respective values of the fundamental restitution frequency Fr and of the reference frequency Fo, and in that all the sounds selected and shifted are sent in a digital to analog conversion means,   acoustic amplification and restitution.   9. Method according to claim 8, characterized in that a weighting calculation is carried out on the sounds to be reproduced, the weighting depending on the measurements of the current parameters. 10. Method according to claim 8 or 9, characterized in that when the duration of the restitution is greater than the duration of the selected sounds, said restitution is looped in   restarting playback at the start of the selected sounds.   11. The method as claimed in claim 8, 9 or 10, characterized in that the selection of the recorded digital sounds further comprises a calculation step consisting in carrying out a weighted combination of sounds during which sounds are associated with   the parameters are close to the theoretical value.   12. Method according to any one of claims 8 to   11, characterized in that a sound library is used with digitized sounds recorded in the form of sequences cut into blocks, and for each time period Te: - the operating state of the motor vehicle is determined by measuring d '' at least one current operating parameter 5 and in that for each operating parameter considered: - a theoretical value of the current operating parameter for the reference vehicle is determined from the measurement of the current operating parameter, - one selects and calculates a weighted combination of n blocks whose parameters are close to the theoretical value by calculating a set of n weighting coefficients associated with the n selected blocks then weighting of each of the n blocks by its weighting coefficient obtained and summation of the weighted blocks, - the fundamental restitution frequency Fr is calculated as a function of the current operating parameter, - o n performs a frequency shift of the weighted combination so that it has a fundamental frequency substantially equal to the fundamental frequency of restitution Fr, said shift being an interpolation or decimation as a function of the respective values of the fundamental frequency of restitution Fr and the reference frequency Fo, and in that the offset combination is sent in a digital analog conversion, amplification and acoustic reproduction.   13. Method according to any one of claims 8 to   12, characterized in that, in addition, by a standby means, the frequency of rotation of the motor is measured and when said frequency falls below a predetermined threshold, the electrical supply of the means intended for restitution is cut off , and when the frequency goes above said   threshold, power is restored.   14. Restitution device for the implementation of a   method of restoring a sound library, characterized in that the method is according to any one of claims 8 to 13 and that the sounds of the sound library being digitized recorded and 5 transposed to a reference frequency Fo, the device   comprising at least one computer means of the microprocessor, microcontroller or digital signal processing processor (DSP) type, program memory for execution according to the restitution method, sound memory 10 comprising the sound library, interpolation means / decimation, said interpolation / decimation means preferably being software, and at least one digital converter analog (DAC).   15. Device according to claim 14, characterized in that it further comprises a standby means the operation of which is linked to the frequency of rotation of the motor, the standby corresponding to an interruption in the supply of the actual restitution device, the means of setting   standby continuing to be powered.