Patent Grant
10229663
The present invention relates to an apparatus that allows a player to quietly play a reed instrument, e.g. while practising.
The normal method of playing a reed instrument (e.g. clarinet, oboe, saxophone, bassoon) is well known. The user blows such that the reed vibrates, thus introducing a complex set of tones into the instrument. A resonant cavity is provided, having a plurality of keys. Depending upon which key(s) are depressed, resonance is produced such that a standing acoustic wave is formed that matches the resonance of the cavity. In this way the traditionally known notes are formed.
Typically when practising, it is desirable to reduce the noise output of reed instruments out of courtesy for those in the vicinity.
US 2014/0224100 A1 describes a system for use with bagpipes in which the normal reed is replaced with transducer apparatus comprising a speaker and a microphone. The speaker delivers sound to an air chamber of the bagpipes, the speaker being driven by a test signal comprising a periodic signal consisting of linear chirps, each linear chirp comprising only frequencies above 16 Khz, i.e. outside the audible range. The microphone detects the sound delivered to the air chamber and then the signal played by the speaker is correlated with the signal detected by the microphone to yield the response function of the acoustic system and thereby the musical note played by the instrument.
According to the present invention there is provided a system for representing sounds of a reed instrument according to claim 1.
The use of a pressure sensor enables the control of timing of operation of the system e.g. in the output of sound by the microphone to the air chamber or the output of synthesized musical notes.
Preferably the signal sent by the pressure sensor to the processing unit additionally indicates how hard the user is blowing through the mouthpiece. This can be used to vary volume of the synthesized musical note output or to recognise an octave shift which can be achieved in some reed instruments by the player blowing hard. Also the air pressure variations may be used to modulate the synthesized sounds, e.g. to recognise when the player is applying a vibrato breath input to the reed instrument and in response import a vibrato into the synthesized sounds.
Other preferred features of the system of the invention are set out in claims 3 to 23.
Preferably, the excitation unit is arranged to drive the speaker to produce sound at a volume chosen based on an amount of ambient noise. For example, the volume may be chosen to exceed ambient noise by a predetermined amount. The level of ambient noise may be measured using any known sensor, but is preferably measured using the microphone or by a separate ambient noise microphone measuring noise outside of the instrument. In one embodiment the user can select an operating mode in which the volume of sound produced by the excitation means can be manually selected.
The present invention allows a musician to practice with the system fitted to the reed instrument, without the generation of any significant noises which may disturb people nearby.
The output means may be one or more of: an interface for a computer; a wireless device for exchanging data over short distances using short-wavelength UHF radio waves; a MIDI (musical instrument data interface) connection; an HD protocol interface; and/or a transmitter.
The speaker and microphone may be mounted on a housing, the housing being adapted for attachment to an air chamber of a reed instrument such that the speaker and microphone are in communication with the air chamber. This allows for the system to be easily retrofitted to a musician's instrument. The speaker and microphone may be mounted on an inner surface of the housing in communication with a cavity formed therein, the housing being adapted for attachment to an air chamber of a reed instrument such that the speaker and microphone are in communication with the air chamber. Preferably the housing is adapted for attachment to a mouthpiece of a reed instrument and the housing is arranged to form a barrier between the mouthpiece and the air chamber.
In another preferred embodiment, the speaker and microphone may be mounted on a housing, the housing being adapted for attachment to an air chamber of a reed instrument such that the speaker and microphone are in communication with the air chamber; the housing forms a mouthpiece; a bore extends through the mouthpiece, the bore being separate from the cavity.
In yet another preferred embodiment, the mouthpiece may comprise a tip with an opening in communication with its bore. The mouthpiece comprises a false reed (in place of a normal reed) extending along the mouthpiece and, optionally, arranged to close the tip of the mouthpiece (although this is not essential). The false reed may be rigid so as not to vibrate when the user blows. The false reed has formed therein an air-pressure groove or air-pressure relief passage extending to a bleed hole formed in the false reed. This can be retrofitted onto existing instruments, and the air-pressure relief groove or passage can allow for the ejection of condensed moisture.
The air pressure sensor may be provided in the bore or in the air-pressure relief groove or passage. This allows the system to detect when the user is blowing and only play tones at these times. Additionally, as mentioned above, the strength of the blowing can be factored into the generation of the output signal and/or a vibrato input breath recognised and a vibrato element incorporated in the synthesized musical note.
The processing unit may be arranged to receive the measurement signal, recognise a played note from the measurement signal and then synthesize a corresponding musical note, the synthesis taking account of both the air pressure in the bore and a characteristic of a difference between the sound produced by the speaker and the sound received by the microphone.
The processor may generate an output signal by synthesising the sound of a reed instrument, with the frequency of the synthesised sound being based on frequency content of the measurement signal and also based on the air pressure sensed by the air pressure sensor, and with the amplitude of the synthesised sound being based on the air pressure sensed by the air pressure sensor.
The present invention also provides a method as claimed in claim 24 an apparatus for use in such a method as claimed in claim 25.
The present invention further provides transducer apparatus as claimed in claim 26. Such transducer apparatus provides a unit conveniently attachable to a reed instrument in place of a reed which will allow a player to practice playing the reed instrument without the generation of any significant noise which might trouble others in the vicinity. Preferred features of the transducer apparatus are set out in claims 27 to 34. The transducer apparatus can form part of a practice system as claimed in claims 35 and 36. The communication between the transducer apparatus and a laptop, tablet or personal computer or a smartphone allows for a better learning experience for the player practicing playing of the reed instrument, e.g. graphical representations of played musical notes can be compared against graphical representations of ‘ideal’ played musical notes. Also musical scores and training exercises can be presented to the player.
The present invention provides an electronic system for determining a musical note played by a reed instrument as claimed in claim 37, with a preferred feature of this system given in claim 38. The system of both claims allows the sound delivered by the speaker to be a low scarcely audible level, since ambient noise is removed from the measurement signal.
The present invention provides an electronic system for determining a musical note played by a reed instrument as claimed in claim 39, with preferred features of the system given in claims 40, 41 and 42. The system of all three claims uses an exponential chirp which has a lowest frequency in the audible range, corresponding at least approximately to a lowest musical note playable by a reed instrument. In contrast the system of US 2014/0224100 A1 uses a chirp which a linear rather than an exponential chirp and one that only comprises frequencies above 16 Khz, i.e. above the audible range of frequencies. Using a linear chirp means that only a smaller range of frequencies can be included in the chirp and this does not allow for recognition of a shift of frequencies occasioned in a reed instrument e.g. by the use of a register shift key. The prior art uses a high energy signal outside the audible range, whereas the present invention uses a low volume signal including frequencies in the audible range. This can provide the effect of playing a near-silent instrument while providing for reliable musical note recognition.
The present invention provides an electronic system for determining a musical note played by a reed instrument as claimed in claim 43, with preferred features of the system given in claims 44, 45 and 46. The selection of an excitation signal with components corresponding to played notes allows for reliable musical note detection from the measurement signal and allows for use of a filter bank with filters tuned to the relevant musical notes. This can provide the effect of playing a near-silent instrument while allowing for reliable musical note detection.
The present invention provides an electronic system for determining a musical note played by a reed instrument as claimed in claim 47, with preferred features of the system given in claims 48 and 49. The systems claimed employ a feedback arrangement in which the excitation signal is adapted following an initial detection of a played musical note so that it contains frequencies better suited to detection of the played musical note in the measurement signal. This can provide the effect of playing a near-silent instrument while allowing for reliable musical note detection.
The present invention provides a method of practising playing of a reed instrument as claimed in claim 50, with preferred versions of the method set out in claims 51 to 58. Further methods of practising playing of a reed instrument are provided as claimed in claims 59 and 60. The methods allow a player to easily and quickly convert his/her own reed instrument into a version which allows near silent practice.
For a better understanding of the invention, and to show how the same may be put into effect, reference is now made, by way of example only, to the accompanying drawings in which:
While the detailed description will be made with reference to a clarinet, it will be appreciated that this is by way of example only and the present invention can be used with any suitable wind instrument (in particular, a reed instrument).
The acoustics of reed instruments, e.g. clarinet, oboe, saxophone, bassoon are well known. The player provides wind energy such that the reed vibrates thus introducing a variety of tones into the instrument. Depending upon which key(s) are depressed a resonant cavity is produced in the air chamber of the instrument such that a standing acoustic wave is set up matching the resonance of the cavity, and the result is the sound which is recognised aurally as the played musical note. The terms first harmonic and fundamental are often used as alternative terms for the lowest frequency component of the played musical note; i.e. the frequency which is aurally perceived.
With reference to
The clarinet 10 also comprises a barrel 14 (also known as a socket) which is again cylindrical and hollow. The barrel 14 has an outer and an inner diameter substantially similar to those of the mouthpiece 11. A section of the inner diameter of the barrel 14 is removed at a proximal end thereof so as to seal with the tenon cork 13 of the mouthpiece 11.
A distal end of the barrel 14 engages with an upper joint 16 of the clarinet 10. Again a section of the inner diameter of the barrel 14 is removed at the distal end thereof so as to seal with a tenon cork 19 of the upper joint 16. The upper joint 16 is provided with a plurality of tone holes, only two of which are shown at 17A, 17B, over which are mounted tone hole rings and keys 18A, 18B. The keys can either be in an undepressed state 18A, or a depressed state 18B, to uncover or cover the holes 17A, 17B, respectively. The upper joint 16 is then in turn attached to a lower joint and a bell (not shown) to form the completed clarinet. These components define a cylindrical air chamber 15 which extends throughout the clarinet 10.
To play the clarinet 10 a user blows into the mouthpiece 11, causing the reed 12 to vibrate. Standing waves are formed in the air chamber 15, which is shaped such that these correspond to the commonly known musical scale. Opening and closing of the holes 17A, 17B alters the shape of the generated standing wave, and hence the musical note produced.
In a first embodiment of the present invention, the barrel 14 of
The present invention recognises that it often hard for players of reed instruments to practice without unduly disturbing others and so provides an arrangement by which the player can still blow into the mouthpiece 11 and open and close the tone holes (e.g. 17A, 17B) in the normal manner, but without generating sound that will disturb others. Instead the speaker 28 will deliver a largely or totally inaudible sound to the air chamber 15 of the instrument 10, which will be modified by the acoustic function of the air chamber 15 as selected by the player by opening and closing the tone holes (e.g. 17A, 17B), the modified sound then forming part of the sound in the air chamber 15 which is received by the microphone 16, which will output a measurement signal from which can be determined which musical note has been selected by the player of the instrument by the opening and closing of the tone holes 17A,17B. The measurement signal can be then used by the system to produce a sound delivered e.g. by headphones to the player, so that the player can hear the musical note played without the instrument producing a sound which would disturb others. As will be described below, a pressure sensor separate and independent from the microphone can be used to determine when and how hard the player is blowing into the mouthpiece 11 (which will not have a functioning reed), so that the timing and volume of the musical notes delivered as sound, e.g. via headphones to the player, can be varied accordingly.
The apparatus of the first embodiment has an operating mode for playing the instrument in a manner that is substantially inaudible, for instance the apparatus may be arranged to limit the power output of an excitation unit 101 (see
For example, the power output of the speaker 28 may be chosen to be greater or less than the measured ambient sound level by a predetermined amount or by a predetermined factor.
Preferably, when the measurement of ambient sound is taken by the microphone 26 (or by a second ambient noise microphone), the power output of the speaker 28 is chosen to be greater than the measured ambient sound level by a predetermined amount or by a predetermined factor. In such embodiments, the power output of the speaker 28 may be a factor of two or more times the power of the ambient noise received by the microphone 26 (or the second ambient noise microphone).
In this way, the selection of power output can be configured (for a given instrument) such that the sound produced by the speaker 28 is expressed by the reed instrument at a level that will effectively allow the instrument to be played quietly such that it cannot be heard over the sound of the ambient noise.
In a preferred embodiment the apparatus is arranged to excite the speaker 28 such that the frequency of sound produced by the speaker 28 is between 20 Hz and 20 KHz. The excitation signal sent to the speaker 28 preferably comprises a series of exponential chirps. The chirp will preferably excite a selected range of audible frequencies equally. Each chirp is preferably an exponential chirp, sometimes called an exponentially scanned chirp or a geometric chirp, but could be a concatenated set of sine-waves at carefully selected frequencies. In an exponential chirp the frequency of the signal varies exponentially as a function of time: f(t)=f0kt, where f0 is the starting frequency (at t=0) and k is the rate of exponential change in frequency. Unlike a linear chirp, an exponential chirp has an exponentially increasing frequency rate. The exponential chirp will provide equal frequency discrimination to each musical note of the instrument and therefore address the issue that the signal to noise ratio can be higher for some musical notes due to the presence of ambient noise, which could otherwise lead to poor musical note recognition.
The microphone 26 then picks up the acoustic waveform in the air chamber 15, which will contain the waveform output by the speaker 28 modified by the acoustic transfer function of the air chamber 15, such acoustic transfer function being selected by the player of the reed instrument by the opening and closing of tone holes. This signal is passed to the processor 102 (see
Whist it is possible that the invention could be implemented and used with a conventional reed still in place and the user refraining from blowing, it will be more typical that to implement the invention the mouthpiece of the reed instrument will be replaced by a modified mouthpiece which is part of the apparatus of the invention or, more preferably, the regular mouthpiece of the instrument will be modified by removing the regular reed and replacing this with a reed substitute according to the invention, as will be described more fully later. In this manner the user can practice the instrument very quietly without disturbing others within earshot. Optionally, a vent hole is provided either in the modified mouthpiece or in the substitute reed to ensure that the user feels the same resistance to blowing as would be felt with a normal mouthpiece.
If a groove 213 is provided (as shown in
The pressure sensor 37 may send a signal to indicate when and/or how hard and/or in what manner (e.g. vibrato) the player is blowing through the passage 213. The substitute reed 212 of
While the embodiment of
The bore 32 is provided with a pressure sensor 37, which sends a signal to the processor 102 (see
A further alternative is shown in
A further version of transducer apparatus according to the present invention is shown in
Whilst the transducer apparatus shown with in
Each of transducer apparatus of
In some embodiments of the invention, a method based upon the application of simple sine tones is used. A stimulus frame comprises tones chosen for each of the possible notes of the clarinet 10 (or other reed instrument). The tones can be applied discretely or contiguously one after another. Each tone may be formed of more than one frequency component. A stimulus-frame comprises the tones arranged in a known order.
The stimulus-frame is applied as an excitation to the loudspeaker 28. Excitation may be carried out periodically, or may commence after an event (such as when the pressure sensor 37 senses the user has blown into the mouthpiece). The microphone 26 picks up the stimulus-frame and the resonances generated and passes this information to the processor 102. The processor applies a filter bank or fast Fourier transform in order to measure the intensity of the received sound signal at different frequencies. From the intensity measurements it is possible to identify the musical note played by the player of the reed instrument.
The processor 102 may use data from the pressure sensor 37 to decide when to initiate the speaker 28, and/or the microphone 26 and/or generation of the output signal, and/or operation of the output means 103. The signal may also be used to alter the characteristics of the output signal generated by the synthesizer 220 (see
A predetermined set of stimulus-frames may be stored in memory 104.
The system may be programmed to learn the response of the instrument 10 to one or each tone within a stimulus-frame. For example, the user may be instructed by a user interface to depress the keys 18 required to play one or more notes (perhaps, all possible notes) in order to characterise the resonance of the instrument 10. Whilst each key 18 is depressed, the excitation unit 101 excites the loudspeaker 28 with a stimulus-frame and the response is received using the microphone 26. The processor 102 can analyse the received response and use this to store a representation of the played musical note in memory 104. In this way, the system can adapt to the particular instrument 10 to which it is applied.
Alternatively, or in addition, the learning process can be used to adapt the stimulus-frame. For example, if the microphone 26 receives sound energy having a primary fundamental frequency (e.g., the lowest received frequency) that is higher than that of a tone transmitted by the speaker 28, the processor may increase the frequency of that tone of the stimulus frame, or all of the tones of the stimulus frame, by a factor equal the ratio of the primary fundamental frequency received by the microphone 26 to the tone that was transmitted by the speaker 28.
Alternatively the processing unit 100 comprising the excitation unit 101, the processor 102, the output means 103 and the memory 104, can generate from the measurement signal sent by the microphone 26 to the processor 102 an output signal comprising a time series of data characterising a difference between the sound produced by the speaker 28 driven by the excitation unit 101 and the sound received by the microphone 26. The excitation signal produces by the excitation unit 101 can be relayed to the processor 102 to allow direct comparison with the measurement signal received by the processor 102 from the microphone 26. The difference is indicative of the acoustic transfer function of the air chamber 15 and this is turn indicates the musical note played by the player; thus the processor 103 can select the musical note played, e.g. by comparing the indicated acoustic transfer function with a series of acoustic transfer functions stored in the memory 104 (each of which would be associated with a particular musical note). The synthesizer 220 (see
When a player is playing the instrument 10 of the embodiment of
When a user is playing the instrument 10 of the embodiment of
The pressure signal 102 can also be used to trigger the excitation of the loudspeaker 28. For example, the excitation may be triggered when the air pressure sensor 37 senses a pressure exceeding a threshold and continued until the pressure drops below a/the threshold. When the stimulus-frame method is used, the stimulus frames may be repeated during the excitation. In embodiments in which the speaker 28 continually produces a repeated sequence of sounds, the processor 102 can use the signal from the pressure sensor 37 to restart the sequence.
The pressure signal also represents the volume of note intended to be played by the user. The processor 102 instructs the output means 103 to synthesize a note having a volume that depends on the sensed pressure.
For some instruments 10, the pressure of air provided by the user can also affect the note played. In some embodiments, the synthesizer (220 in
Irrespective of how the microphone 26, speaker 28, and optional air pressure sensor 37, are mounted (i.e. as in the case of
Quiet play: the system may be provided with a quiet operating mode in which the excitation unit 101 is arranged to drive the speaker 28 to produce sound at a volume selected based on a measurement of ambient sound. The measurement of ambient sound may be taken by the microphone 26 (or a separate and independent ambient noise microphone). In this way, the instrument can be “played” by the user (either without blowing, or with the breath redirected as in
Game interface: the output means 103 may be adapted to provide a signal to a computer programmed to challenge the user to play a certain piece of music. The computer may display in real-time the notes played and/or score the ability of the user to play the piece of music, based on timing and/or frequency of the signal produced by the microphone 26. This may optionally also apply the quiet operating mode.
Virtual orchestra: the output means 103 may be adapted to provide a signal to a communications device (e.g., an internet connection). The communications device may receive signals from other such devices and/or other types of instrument and synthesize the sound of a plurality of instruments playing simultaneously. Again, this may optionally also apply the quiet operating mode.
The transducer apparatus 200 has a printed circuit board 204 on which is mounted various electronic components which together provide the processing unit (217 in
As can be seen in
The arm 205 provides a housing for a speaker 208 and a microphone 209, as can be seen in
The reed replacement section 203 has an air passage that extends therethrough from an inlet 211 shown in
The transducer apparatus 200 is also provided with an ambient noise microphone 214 which faces outwardly of the apparatus 200 and which receives ambient sound surrounding the apparatus 200. The ambient noise microphone 214 produces an ambient noise signal which is relayed to the electronic signal processing unit (217 in
Batteries 215 and 216, preferably rechargeable, are provided on the printed circuit board 204 to power the electronic components on the board 204. Also a wireless transmitter 218 is provided to wirelessly transmit an output signal from the transducer apparatus 200, e.g. to the be received by a receiver of wireless headphones.
In use the transducer apparatus 200 will be mounted on the mouthpiece 201 of the reed instrument in place of a reed. The player will then blow through the inlet 211 of the apparatus while manually operating keys of the reed instrument to open and close tone holes of the instrument and thereby select a note to be played by the instrument. The blowing through the inlet 211 will be detected by the pressure sensor 212 which will send a pressure signal to the processing unit provided by the electronics on the printed circuit board 204. The processing unit (100,217), in response to the pressure signal indicating blowing of the player, will activate the excitation unit (101,222) of the processing unit (100, 217) to output an excitation signal to the speaker 208, which will then output sound to the air chamber 15 of the reed instrument. The frequency and/or amplitude of the excitation signal can be varied by the excitation unit (101,222) having regard to the pressure signal output by the pressure sensor 212, so as to take account of how hard the player is blowing. Also air pressure variations measured by the pressure sensor 212 may be used to modulate the synthesized sounds, e.g. to recognise when the player is applying a vibrato breath input to the reed instrument and in response import a vibrato into the synthesized sounds. The frequency and/or amplitude of the excitation signal can also be varied by the excitation unit (101,222) having regard to the ambient noise signal output by the ambient noise microphone 214, e.g. to make sure that the level of sound output by the speaker 208 is at least greater than preprogramed minimum above the level of the ambient noise.
The microphone 209 will receive sound in the air chamber 15 and output a measurement signal to the processing unit (217 in
The transducer apparatus as described above has the following advantages:
The invention as described in the embodiment above introduces an electronic stimulus by means of a small speaker 208 built in the transducer apparatus 200, placed near the connection of the mouth-piece to the remainder of the instrument. The stimulus is chosen such that the resonance produced by depressing any combination of key(s) causes the acoustic waveform, as picked up by at least one small microphone, e.g. the microphone 209 described above, preferably placed close to the stimulus provided by the speaker 208, to change. Therefore analysis of the acoustic waveform, when converted into an electric measurement signal by microphone 208, and/or derivatives of the signal, allows the identification of the intended note associated with the played key positions.
The stimulus provided via the speaker 208 can be provided with very little energy and yet with appropriate processing of the measurement signal, the intended note can still be recognised. This can provide to the player of the reed instrument the effect of playing a near-silent instrument.
The identification of the intended notes preferably gives rise to the synthesis of a musical note, typically, but not necessarily, chosen to mimic the type of reed instrument played. This electronic sound synthesis will be carried out by the sound synthesizer 220 provided on the printed circuit board 204. The synthesized sound will be relayed to headphones or other electronic interfaces such that a synthetic acoustic representation of the notes played by the instrument is heard by the player. Electronic processing can provide this feedback to the player in close to real-time, such that the instrument can be played in a natural way without undue latencies. Thus the player can practice the instrument very quietly without disturbing others within earshot.
The mouthpiece 201 of the instrument is modified by use of the transducer apparatus 200 to replace the reed typically mounted on the mouthpiece 201 of the reed instrument. The player expresses air into a small aperture provided by the inlet 211 to a passage which ends in a permanently open vent hole providing the outlet 213 to the outside of the instrument, typically in the vicinity of a junction between the mouthpiece 201 and a remainder of the reed instrument. The purpose of the vent hole is preferably two-fold; to mimic the normal playing air-pressure experienced by the player; and to provide a path for condensed moisture egress. Alternatively a second vent hole may be provided which is sealed until opened via a small key to allow for the ejection of condensed moisture. The dimensions of the or each vent hole are chosen to mimic the normal range of pressures exerted when playing a conventional instrument.
As mentioned above the air pressure within the passage between the inlet 211 and outlet 213 is detected by the pressure sensor 212. Typically an analogue signal representing the measured pressure is provided to the electronic processing unit shown as 100 in
The electronic processing unit (100,217) will use one or more of a variety of well-known techniques for analysing the measurement signal in order to discover a transfer function of the resonant cavity provided by the air chamber 15 of the reed instrument, and thereby the intended note, working either in the time domain or the frequency domain. These techniques include application of maximum length sequences either on an individual or repetitive basis, time-domain reflectometry, swept sine analysis, chirp analysis, and mixed sine analysis.
An embodiment based on the consecutive application of simple sine tones will now be described, but alternate processing methods may be used.
In the preferred embodiment the stimulus signal sent to the speaker, e.g. speaker 208, will be a stimulus-frame comprised of tone fragments chosen for each of the possible musical notes of the instrument. The tones can be applied discretely or contiguously following on from each other. Each of the tone fragments may be comprised of more than one frequency component. The tone fragments are arranged in a known order to comprise the stimulus-frame. The stimulus-frame is applied as an excitation to the speaker (e.g. 208) typically being initiated by the player blowing into the instrument (as detected by the pressure sensor 212). A signal comprising a version of the stimulus-frame as modified by the acoustic transfer function of the air chamber (as set by any played keys and resonances generated thereby) is picked up by the microphone 209. The time-domain measurement signal is processed, e.g. by a filter bank or fast Fourier transform (fft), to provide a set of measurements at known frequencies. The frequency measures allow recognition of the played note, either by comparison with pre-stored frequency measurements of played notes or by comparison with stored frequency measurements obtained via machine learning techniques. Knowledge of ordering and timing within the stimulus-frame may be used to assist in the recognition process.
The stimulus-frame typically is applied repetitively on a round-robin basis for the period that air-pressure is maintained by the player (as sensed by the pressure sensor 212). The application of the stimulus frame will be stopped when the pressure sensor 212 gives an pressure signal indicating that the player has stopped blowing and the application of the stimulus frame will be re-started upon detection of a newly timed note as indicated by pressure sensor 212. The timing of a played note output signal, output by a component of the processing unit (217 in
It is desirable to provide the player with low-latency feedback of the played note, especially for low frequency notes where a single cycle of the fundamental frequency may take tens of milliseconds. A combination of electronic processing techniques may be applied to detect such notes with low latency by applying a tone or tones at different frequencies to the fundamental such that the played note may still be detected from the response.
On some reed instruments the played note is changed by means of one or more register or octave-key(s) opening at least one additional ‘vent’, or alternatively by ‘over-blowing’ (i.e. the player blowing at a significantly higher pressure) such that a harmonic sounds rather than the fundamental. Over-blowing may be detected by the pressure sensor 212 through the additional air-pressure exerted. Use of a register or octave-key causes the resonant frequency of the fundamental to move slightly without significantly affecting the frequency of the higher harmonics and thus provides a basis for recognition through the measurement signal provided by the microphone 209. Alternatively the position of the register or octave-key could be detected via a variety of conventional methods, e.g. by use of a magnetic switch or a micro-switch.
In a further embodiment the excitation signal sent to the speaker 208 is an exponential chirp running from 20 Hz to 20 kHz. The signal will include a lowest frequency in the range 20 Hz to 200 Hz. This signal excites the air chamber of the reed instrument via the loudspeaker on a repetitive basis, thus forming a stimulus-frame. The starting frequency of the scan is chosen to be below the lowest fundamental (first harmonic) of the instrument, roughly 150 Hz in the case of a Bflat clarinet.
It should be noted that on many reed instruments the opening associated with the register key is physically small in relation to the other key openings. This has the effect of the opening being largely transparent to high frequencies since the phase of the waveform reverses before significant sound energy can escape through the small hole. It is important that the bottom scan frequency of the chirp signal provided by the stimulus-frame sent to the microphone is at least as low as the lowest fundamental frequency of the instrument, e.g. ˜150 Hz on a standard Bflat clarinet.
The sound present in the air chamber 15 is sensed by the microphone 209 and assembled into a frame of data lasting exactly the same length as the exponential chirp excitation signal (which provides the stimulus-frame). Thus the frames of microphone data and the chirp are synchronised.
An FFT is performed upon the frame of data in the measurement signal provided by the microphone 209 and a magnitude spectrum is thereby generated in a standard way.
The transducer apparatus in this embodiment preferably has a training mode in which the player successively plays all the notes of the instrument and the resultant magnitude spectrum of the measurement signals provided by the microphone are stored correlated to the notes being played. Preferably the transducer apparatus is provided with a signal receiver as well as its signal transmitter and thereby communicates with a laptop, tablet or personal computer or a smartphone running application software that enables player control of the transducer apparatus. The application software allows the player to select the training mode of the transducer apparatus. Typically the memory unit (104, 219) of the apparatus will allow three different sets of musical note data to be stored. The player will select a set and then will select a musical note for storing in the set. The player will manually operate the relevant keys of the instrument to play the relevant musical note and will then use the application software to initiate recording of the measurement signal from the microphone 209. The transducer apparatus will then cycle through a plurality of cycles of generation of an excitation signal and will average the measurement signals obtained over these cycles to obtain a good reference response for the relevant musical note. The process is then repeated for each musical note played by the instrument. When all musical notes have been played and reference spectra stored, then the processing unit (217 in
Rather than use application software on a separate laptop, tablet or personal computer or smartphone, the software could be run by the electronic processing unit (100, 217) of the transducer apparatus 200 itself and manually operable controls, e.g. buttons, provided on the transducer apparatus 200, along with a small visual display, e.g. LEDs, that provides an indication of the selected operating mode of the apparatus 200, musical note selected and data set selected.
An accelerometer 221 (see
When the transducer apparatus 200 is then operated in play mode a pre-stored training set is pre-selected. The selection can be made using application software running on a laptop, tablet or personal computer or on a smartphone in communication with the transducer apparatus. Alternatively the transducer apparatus 200 could be provided with manually operable controls to allow the selection. The magnitude spectrum is generated from the measurement signal as above, but instead of being stored as a training set it is compared with each of the spectra in the training set (each stored spectrum in a training set representing a single played note). A variety of techniques may be used for the comparison, e.g. a least squares difference technique or a maximised Pearson second moment of correlation technique. Additionally machine learning techniques may applied to the comparison such that the comparison and or training sets adjusted over time to improve the discrimination between notes.
It is convenient to use only the magnitude spectrum of the measurement signal from a simple understanding and visualisation perspective, but the full complex spectrum of both phase and amplitude information (with twice as much data) could also be used, in order to improve the reliability of musical note recognition. However, the use of just the magnitude spectrum has the advantage of speed of processing and transmission, since the magnitude spectrum is about 50% of the data of the full complex spectrum. References to ‘spectra’ in the specification and claims should be considered as references to: magnitude spectra only; phase spectra only; a combination of phase and amplitude spectra; and/or complex spectra from which magnitude and phase are derivable.
In an alternative embodiment a filter bank, ideally with centre frequencies logarithmically spaced, could be used to generate a magnitude spectrum, instead of using a Fast Fourier Transform technique. The centre frequencies of the filters in the back can be selected in order to give improved results, by selecting them to correspond with the frequencies of the musical notes played by the reed instrument.
Thus the outcome of the signal processing is a recognised note, per frame (or chirp) of excitation. The minimum latency is thus the length of the chirp plus the time to generate the spectra and carry out the recognition process against the training set. The processing unit (217 of
As with the other preferred embodiments, the recognised note is synthesized immediately and fed back to the player via wired headphones. Alternatively the synthesized musical note may be transmitted to be used by application software running on a laptop, tablet or personal computer or smartphone or other connected processor. The connection may be wired or preferably wireless using a variety of means, e.g. Bluetooth®. Parameters which are not critical to operation but which are useful, e.g. the magnitude spectrum, may also be passed to the application software for every frame. Thus the application software can generate an output on a display screen which allows the player to see a visual effect in the frequency spectrum of playing deficiencies of the player e.g. a failure to totally close a hole. This allows a player to adjust his/her playing and thereby improve his/her skill.
In a further embodiment of the invention an alternate method of excitation signal generation and processing the measurement signal is implemented in which an excitation signal is produced comprising of a rich mixture of frequencies, typically harmonically linked. The measurement signal is analysed by means of a filter-bank or fft to provide a complex frequency spectrum. Then the complex frequency spectrum is run through a recognition algorithm in order to provide a first early indication of the played note. This could be via a variety of recognition techniques including those described above. The first early indication of the played note is then used to dynamically modify the mixture of frequencies of the excitation signal in order to better discriminate the played note. Thus the recognition process is aided by feeding back spectral stimuli which are suited to emphasising the played note. The steps are repeated on a continuous basis, perhaps even on a sample by sample basis. A recognition algorithm provides the played note as an additional output signal.
In the further embodiment the content of the excitation signal is modified to aid the recognition process. This has parallels with what happens in the conventional playing of a reed instrument in that the reed provides a harmonic rich stimulus which will be modified by the acoustic feedback of the reed instrument, thus reinforcing the production of the played note. However, there are downsides in that a mixture of frequencies as an excitation signal will fundamentally produce a system with a lower signal to noise ratio (SNR) than that using a chirp covering the same frequencies, as described above. This is because the amplitude at any one frequency is necessarily compromised by the other frequencies present if the summed waveform has to occupy the same maximum amplitude. For instance if the excitation signal comprises a mixture of 32 equally weighted frequencies, then the overall amplitude of the sum of the frequencies will be 1/32 of that achievable with a scanned chirp over the same frequency range and this will reflect in the SNR of the system. This is why use of a scanned chirp as an excitation signal, as described above, has an inherent superior SNR; but the use of a mixture of frequencies in the excitation signal which is then enhanced might enable the apparatus to have an acceptably low latency between the note being played and the note being recognised by the apparatus.
With suitable communications, application software running on an device external to the instrument and/or the transducer apparatus may also be used to provide a backup/restore facility for the complete set of instrument data, and especially the training sets. The application software may also be used to demonstrate to the user the correct spectrum by displaying the spectrum for the respective note from the training set. The displayed correct spectrum can be displayed alongside the spectrum of the musical note currently played, to allow a comparison.
Since the musical note and its volume are available to the application software per frame, a variety of means may be used to present the played note to the player, These include a simple textual description of the note, e.g. G#3, or a (typically a more sophisticated) synthesis of the note providing aural feedback, or a moving music score showing or highlighting the note played, or a MIDI connection to standard music production software e.g. Sibelius, for display of the live note or generation of the score.
The application software running on a laptop, tablet or personal computer or smartphone in communication with the transducer apparatus and/or as part of the overall system of the invention will allow: display on a visual display unit of a graphical representation of a frequency of a played note; the selection of a set of data stored in memory for use in the detection of a played note by the apparatus; player control of volume of sound output by the speaker; adjustment of gain of the pressure sensor; adjustment of volume of playback of the synthesized musical note; selection of a training mode or a playing mode operation of the apparatus; selection of a musical note to be learned by the apparatus during the training mode; a visual indication of progress or completion of the learning of a set of musical notes during the training mode; storage in the memory of the laptop, tablet or personal computer or smartphone (or in cloud memory accessed by any of them) of the set of data stored in the on-board memory of the transducer apparatus, which in turn will export (e.g. for restoration purposes) of set of data to the on-board memory (104, 219) of the transducer apparatus 200; a graphical representation, e.g. in alphanumeric characters, of the played note; a musical note by musical note graphical display of the spectra of the played notes, allowing continuous review by the player; generation of e.g. pdf files of spectra. The application software could additionally be provided with feature enabling download and display of musical scores and exercises to help those players learning to play an instrument.
Whilst above the identification of a played note and the synthesis of a musical note is carried out by electronics on-board to the transducer apparatus, these processes could be carried out by separate electronics physically distant from but in communication with the apparatus mounted on the instrument or indeed by the application software running on the laptop, tablet or personal computer or smartphone. The generation of the excitation signal could also occur in the separate electronics physically distant from but in communication with the apparatus mounted on the instrument or by the application software running on the laptop, tablet or personal computer or smartphone.
In modifications of the embodiments described above at least a second channel of processing is provided with one of more independent ambient noise microphone(s) 214, which can be placed on the printed circuit board 204. The independent ambient noise microphone(s) 214 will measure sound external to the air chamber 15. This provides two possibilities:
The transducer apparatus 200 will preferably retain in memory (104, 219) the master state of the processing and all parameters, e.g. a chosen training set. Thus the transducer apparatus 200 is programmed to update the process implemented thereby for all parameter changes. In many cases the changes will have been initiated by application software on the laptop, tablet or personal computer or smartphone, e.g. choice of training note. However, the transducer apparatus 200 will also generate changes to state locally, e.g. the pressure currently applied as noted by the pressure sensor 212 or the note currently most recently recognised.
The embodiments of the invention above could be modified by the addition of an accelerometer included in the apparatus. The signal from the accelerometer would indicate movement of the reed instrument and thereby provide the player with expression control and/or automatic power-up/power-down governed by instrument movement. This control could be implemented either in the electronics mounted to the reed instrument or in application software run on a laptop, tablet or personal computer or smartphone in communication with the device mounted on the reed instrument.
Whilst above an electronic processing unit (100, 217) is included in the device coupled to the reed instrument which provides both an excitation signal and outputs a synthesized musical note, a fast communication link between the instrument mounted device and a laptop, tablet or personal computer or smartphone would permit application software on the laptop, tablet or personal computer or smartphone to generate the excitation signal which is then relayed to the speaker mounted on the instrument and to receive the measurement signal from the microphone and detect therefrom the musical note played and to synthesize the musical note played e.g. by a speaker of the laptop, tablet or personal computer or smartphone or relayed to headphones worn by the player. A microphone built into the laptop, tablet or personal computer or smartphone could be used as the ambient noise microphone. The laptop, tablet or personal computer or smartphone would also receive signals from a pressure sensor and/or an accelerometer when they are used.
The synthesized musical notes sent e.g. to headphones worn by a player of the reed instrument could mimic the reed instrument played or could be musical notes arranged to mimic sounds of a completely different instrument. In this way an experienced player of a reed instrument could by way of the invention play his/her reed instrument and thereby generate the sound of a e.g. a played guitar. This sound could be heard by the player only by way of headphones or broadcast to an audience via loudspeakers. This can be particularly useful for the practice of certain reed instruments, e.g. bass reed instruments are very large and expensive, since being able to practice a piece of music on a Bflat clarinet fitted with the present invention will be far more convenient in many circumstances (e.g. when travelling) than practising on the bass instrument itself.
| Number | Date | Country | Kind |
|---|---|---|---|
| 1513036.2 | Jul 2015 | GB | national |
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/GB2016/052267 | 7/25/2016 | WO | 00 |
| Publishing Document | Publishing Date | Country | Kind |
|---|---|---|---|
| WO2017/013455 | 1/26/2017 | WO | A |
| Number | Name | Date | Kind |
|---|---|---|---|
| 2138500 | Miessner | Nov 1938 | A |
| 3558795 | Barcus | Jan 1971 | A |
| 3571480 | Techenor et al. | Mar 1971 | A |
| 4233877 | Okami | Nov 1980 | A |
| 5131310 | Kunimoto | Jul 1992 | A |
| 5245130 | Wheaton | Sep 1993 | A |
| 5668340 | Hashizume et al. | Sep 1997 | A |
| 5929361 | Tanaka | Jul 1999 | A |
| 7220903 | Bronen | May 2007 | B1 |
| 20060027074 | Masuda et al. | Feb 2006 | A1 |
| 20070068372 | Masuda | Mar 2007 | A1 |
| 20070144336 | Fujii | Jun 2007 | A1 |
| 20120230155 | Kawaguchi | Sep 2012 | A1 |
| 20140224100 | Vassilev | Aug 2014 | A1 |
| 20150101477 | Park et al. | Apr 2015 | A1 |
| 20180090120 | Kasuga | Mar 2018 | A1 |
| 20180218720 | Smith | Aug 2018 | A1 |
| Number | Date | Country |
|---|---|---|
| 104810012 | Jul 2015 | CN |
| 3839230 | May 1990 | DE |
| 1585107 | Oct 2005 | EP |
| 1760690 | Mar 2007 | EP |
| 1804236 | Jul 2007 | EP |
| 2650870 | Oct 2013 | EP |
| 2775823 | Sep 1999 | FR |
| 2537104 | Oct 2016 | GB |
| 2559135 | Aug 2018 | GB |
| 2559144 | Aug 2018 | GB |
| 2002278556 | Sep 2002 | JP |
| 2005122099 | May 2005 | JP |
| 2007065197 | Mar 2007 | JP |
| 2008076838 | Apr 2008 | JP |
| 2013164542 | Aug 2013 | JP |
| 2014232153 | Dec 2014 | JP |
| 2018138501 | Aug 2018 | WO |
| 2018138504 | Aug 2018 | WO |
| Entry |
|---|
| Gibiat et al., “Acoustical Impedance Measurements by the Two-Microphone-Three-Calibration (TMTC) Method”, The Journal of the Acoustical Society of America, American Institute of Physics for the Acoustical Society of America, New York, NY, US, vol. 88, No. 6, Dec. 1990, pp. 2533-2545. |
| Number | Date | Country | |
|---|---|---|---|
| 20180218720 A1 | Aug 2018 | US |
