Patent Grant
6928410
The present invention relates generally to modulating an audio stream with another audio stream and, more particularly, to a vocoding method where a speech signal is used to modulate a string of periodic tones.
The modulation of an audio stream indicative of speech data with another audio stream indicative of a periodic tone has been used to create synthetic music and certain sound effects. This modulation technique is usually referred to as vocoding, and the apparatus for vocoding speech is referred to as a vocoder or a phase vocoder. The term vocoding is derived from VOice CODING. Originally, the motivation for the development of the phase vocoder was to reduce the amount of data required for the transmission of speech over telephone lines or other speech signal transmission medium. For that purpose, vocoders extract pitch and voice information in order to time-compress the speech, and a phase vocoder may be considered as a series of bandpass filters, each having a center frequency. Through the bandpass filtering process, the speech signal is reduced to a series of signal segments carrying the center frequencies.
In an old-styled telephone set, the ringing tone that is used to signal an incoming telephone call is usually produced by a ringer repeatedly striking one or two bells. In a mobile phone, the ringing tone is produced by an electronic buzzer, which produces a pitch of a given frequency according to a value in a data stream representative of a series of musical tones. Likewise, in an electronic organizer or a personal digital assistant, such as a Palm Pilot, a beeping sound is used to remind the user of a scheduled event or the completion of a task requested by the user.
U.S. Pat. No. 5,452,354 (Kyronlahti et al.) discloses a ringing tone apparatus wherein subscriber identification information is used to generate the ringing tone. As disclosed in Kyronlahti el al., a ringing tone can be generated based on two or more binary digits of the subscriber identification number such as the mobile station identification number (MSIN), mobile identification number (MIN), etc. For example, if the lowest bits of the identification MSIN are described as a string of 11 binary digits, D10-D9-D8-D7-D6-D5-D4-D3-D2-D1-D0, these string of digits can be used to specify the parameters necessary for generating a ringing tone as follows: D1 and D0 are used to determine the duration of each ringing tone pulse; D3 and D2 are used to determine the frequency of the ringing tone pulses; D5 and D4 are used to determine the pulse number in one pulse sequence; D7 and D6 are used to determine the number of sequences to be repeated in the ringing tone; and D10, D9 and D8 are used to determine the silence period between pulse sequences. While this tone generation method is useful for producing different ringing tones for different subscribers, the ringing tones have no relevance to speech data, synthetic or natural. Japanese patent No. JP05346787 (Nakae Tetsukazu) discloses a method of extracting pitch data from a digital speech signal and generating a digital musical sound according to the pitch data. The digital speech signal and the digital musical sound are conveyed to a vocoder in order to generate a musical sound signal and a voice signal from which an envelope signal is produced. Finally, the sound signal is modulated with the envelope signal in order to add the nuance of a human voice to a musical sound. For most languages, the so-called musical sound, according to the pitch variation, is confined to one or two notes. For example, in a phrase like “I am Bond, James Bond”, there is not much in pitch variation and the resulting musical sound signal may sound like EEE_EE. U.S. Pat. No. 5,826,064 (Loring et al.) discloses a user-configurable earcon event engine, wherein auditory cues are provided responsive to command messages issued by tasks executed on a computer system. As disclosed, the command messages include an index to an earcon data file, which, in turn, includes a reference to an audio file and audio parameter data for manipulating the acoustic parameters of an audio wave. However, the audio wave does not have the content of speech.
It is advantageous and desirable to provide a method and apparatus for modifying a carrier stream indicative of musical tones with a speech signal, wherein a broad range of musical tones can be exploited, regardless of the pitch variation in the speech signal.
The first aspect of the present invention is a method for modification of a speech signal indicative of a stream of speech data having a plurality of syllables. The method comprises the steps of:
mapping the stream of speech data from the speech signal into a stream of tone data according to a predetermined rule regarding the syllables for providing a tone signal indicative of the stream of tone data;
forming a string of musical notes responsive to the tone signal for providing a carrier signal indicative of the string of musical notes;
modulating the carrier signal with the speech signal for providing a modulated signal; and
providing an audible signal representative of the speech signal, musically modified according to the predetermined rule.
Preferably, the predetermined rule includes a linguistic rule for assigning one, two or more notes to a syllable of the speech data based on a vowel of the syllable, a consonant of the syllable, or the intonation of the syllable.
It is also possible to assign one, two or more tones to a syllable of the speech data based on a combination of a vowel, a consonant and/or the intonation of the syllable. It is possible to assign a tone color (timbre), tempo, and/or a pitch range to the musical notes.
Preferably, the speech signal is provided in response to an incoming telephone call on a telephone, and the audible signal is indicative of the incoming telephone call.
Preferably, the speech signal is provided in response to a message on a telephone or a communicator, and the audible signal is indicative of the message.
Preferably, the speech signal is provided in response to a scheduled event in a personal digital assistance device, and the audible signal is indicative of the schedule.
Preferably, the speech signal is provided to indicate a user-interface event regarding an electronic device, wherein the user-interface event can be represented by an object positioned in the electronic device based on a hierarchy, and the predetermined rule is based on the position of the object in the hierarchy.
The second aspect of the present invention is an apparatus for modification of a speech signal indicative of a stream of speech data having a plurality of syllables. The apparatus comprises:
a mapping mechanism, responsive to the speech signal, for mapping the syllables into a stream of tone data based on a predetermined rule regarding the syllables, and for providing a tone signal indicative of the stream of tone data;
a forming mechanism, responsive to the tone signal, for providing a string of musical notes based on the stream of tone data, and for providing a carrier signal indicative of the string of musical notes;
a modulation mechanism, responsive to the carrier signal, for modulating the carrier signal with the speech signal, and for providing a modified speech signal indicative of the modulation; and
a sound production device, responsive to the modified speech signal, for providing an audible signal representative of the speech signal, musically modified according to the predetermined rule.
Preferably the modified speech signal is further combined with the unmodified speech signal in order to adjust the musical content in the audible signal.
Preferably, the modulation mechanism is a phase vocoder, and the modulation is according to the process of vocoding.
The present invention will become apparent upon reading the description taken in conjunction with
Instead of producing a ringing tone in a telephone that has no relevancy to the user of the called party, it is advantageous to provide a musically modified speech signal to signal an incoming telephone call or to reminder the user of a message left by a called party. For example, it is possible to provide a musically modified speech signal derived from the user's name, or the name of the called party of an incoming phone call. In certain languages, such as Italian, Spanish and Japanese, personal names such as Giacomo Puccini, Pablo Picasso, Akira Kurosawa can be represented by a string of syllables as GIA-CO-MO_PUC-CI-NI, PA-BLO_PI-CAS-SO, A-KI-RA_KU-RO-SA-WA. These strings of syllables can be made into a string of musically modified speech data according to a simple rule based on the vowel, the consonant or a combination of a vowel and a consonant in each syllable. In particular, Japanese words and syllables are made up of kana symbols. The kana symbols make it easy to assign a syllable to a musical note in order to generate a string of musical notes indicative of the syllables. For example, the vowels a, i, u, e, o can be mapped onto five musical notes, namely, C, D, E, G. A, as shown in TABLE I.
Thus, when a syllable includes a vowel ‘u’, as in ‘ku’, ‘tsu’, etc, is assigned the musical note E. Following this linguistic rule, we have
Fumiko Ichikawa (FU-MI-KO_I-CHI-KA-WA)=EDA_DDCC
Akira Kurosawa (A-KI-RA_KU-RO-SA-WA)=CDC_EACC
Yukio Mishima (YU-KI-O_MI-SHI-MA)=EDA_DDC.
The symbol ‘_’ signifies a pause the length of which can be made equal to or different from the musical notes. With a similar rule, a string of syllables such as “I-AM-BOND_JAMES-BOND” may be mapped into a string of musical notes as DCA_CA.
Similarly, a linguistic rule can be set up based on the consonant of the syllables. For example, the musical note C can be assigned to ‘ka’, ‘ki’, ‘ku’, ‘ke’, ‘ko’, and A can be assigned to ‘na’, ‘ne’, ‘nu’, ‘ne’, ‘no’, as shown in TABLE II.
It should be noted that ‘n’ has been moved to the second row, and C2 denotes an octave higher than C. To use consonants as tone determinant, the tone range of two octaves is sufficient. Following the linguistic rule as set forth in TABLE II, we have:
Fumiko Ichikawa (FU-MI-KO_I-CHI-KA-WA)=C2D2D_CGDA2
Akira Kurosawa (A-KI-RA_KU-RO-SA-WA)=CDG2_DG2EA2
Yukio Mishima (YU-KI-O_MI-SHI-MA)=E2DA2_D2ED2
In many Western languages, however, there may be too many different consonants and multi-consonants, such as pr, pl, tr, chr and spl, in the syllables to be mapped into musical notes within two or three octaves. It is possible to use a linguistic rule similar to the rule as set forth in TABLE III. The linguistic rules, as set forth in TABLE I and TABLE II, are based on a monophonic implementation of the pentatonic scale. TABLE III illustrates a rule that is based on a polyphonic implementation of the major Western scale for consonants and pentatonic for vowels.
Following the linguistic rule as set forth in TABLE III, we have
Fumiko Ichikawa (FU-MI-KO_I-CHI-KA-WA)=C2D2D_CGDA2 E D A_DDCC
Akira Kurosawa (A-KI-RA_KU-RO-SA-WA)=CDG2_DG2EA2 CDC_EA C C
Yukio Mishima (YU-KI-O_MI-SHI-MA)=E2DA2_D2ED2 E D A_D D C
Furthermore, the voiced/unvoiced (nigori/maru) and compound kana characters can be mapped to the closest equivalent syllables in the system, or they can be designated their own musical notes. Moreover, when a string of musical notes derived from a name according to one rule (e.g., the vowel rule) sounds too monotonic, it is possible to substitute it with a string of musical notes using another rule (e.g, the consonant rule). The nigori symbols (ga, gi, gu, ge, go), (za, ji, zu, ze, zo), (da, ji, du, de, do) and (ba, bi, bu, be, bo), are derived from, respectively, the upper-case characters (ka, ki, ku, ke, ko), (sa, shi, su, se, so), (ta, chi, tzu, te, to) and (ha, hi, fu, he, ho). When they are combined with other words to become compounds, the characters from which they are derived become voiced. For example, hana (nose), plus chi (blood), combine to form hanaji, and the character, chi, becomes voiced. When treated as syllables in compounds, the nigori symbols can be mapped to the same musical notes as the characters from which they derived, if so desired. Similarly, the maru symbols (pa, pi, pu, pe, po) can be mapped to the same musical notes as the upper-case characters of (ta, chi, tzu, te, to) from which they are derived. As for the lower-case compound characters, kya, kyu, kyo, gya, gyu, gyo, cha, etc., they can be mapped to the closest equivalent syllables in the system, but they can have a different tempo or time-stretch. For example, ki and kya can be mapped to the same musical note with different durations or different tone colors. Another symbol, the lower-case tsu, doubles the next consonant when it is placed before that consonant. For example, by placing tsu before ka, ka is stretched out as kka. Accordingly, kka can be mapped to the same musical note as ka with a longer duration.
In languages such as Chinese and Vietnamese, a plurality of intonations are used to modify the pronunciation of single-syllable words. In Mandarin Chinese, four intonations are used to modify the pronunciation, and the intonations are denoted herein with subscripts 1, 2, 3 and 4. For example, the different intonations applied to ‘ba’ are:
ba1 (eight), ba2 (to pull out), ba3 (target), ba4 (dam)
It is thus possible to assign four different musical tones such as C, D, G, A to the intonations 1, 2, 3, 4 as further shown in TABLE IV:
Following this linguistic rule, the musical notes assigned to the Chinese pronunciation of the late Japanese writer Yukio Mishima would be:
san1 daO3
With rules as illustrated above, it is possible to assign a music note to a syllable in a speech signal in a variety of languages in accordance with the vowel, the consonant, or the intonation of the syllable.
It should be noted that, in a communication device such as a telephone, when using synthetic speech to make an announcement, the speech signal can simply be a stream of speech data having a plurality of syllables. From these syllables, it is possible to form a stream of musical notes based on a selected linguistic rule. The stream of musical notes can then be used as a carrier stream to musically modify, the stream of speech data. The musically modified speech data can be conveyed to a sound-producing device to make an audible signal. As such, the speech content is transformed into a musical form. Depending on the nature of the speech data, the musically modified speech data may or may not bear resemblance to the speech signal. Thus, it is possible to mix the musically modified speech data with the unmodified speech data. The mixing proportion can be adjusted so that the resulting sound will sound like speech having a certain mix of musical characteristics.
The linguistic rules, as described above, can also be used in an electronic device to provide auditory cues indicating a user-interface (UI) event. Typically, the UI events on an electronic device, such as a computer, are represented by objects or icons. According to the present invention, the UI objects or icons are further represented by auditory icons so that the user of the electronic device can be notified of the UI events using the auditory cues. For example, an auditory icon for arriving e-mail could be represented by musically modified syllables of “mes-sa-ges”. The musical notes can be assigned to these syllables according to the vowel, the consonant or the syllabic intonation. Similarly, the UI event of “reply to a message” could be represented by musically modified syllables of “re-ply-to-mes-sage”. It should be noted that the objects in device UI can be categorized in a hierarchical manner. For example, the hierarchy of a UI event indicates whether the event is related to a folder, a file, or the file's place in the file list. The division and the placement of objects in device UI can be further indicated by timbre, tempo and a pitch range. Timbre is a tone color of a sound, imitating the sound of a piano, an English horn, a flute and so forth. Tempo is a measure of time, or the duration, of each musically modified syllable. TABLE V lists a few examples of the auditory cues representing UI events, wherein the musical notes are assigned to the syllable according to syllabic intonation.
Accordingly, the vocoded end result is as follows:
Messages (MES-SA-GES)=G2E2C2
Calendar (CAL-END-AR)=A2D2F#2
Inbox {Messages} (MES-SA-GES_IN-BOX)=E3C3G2_C3C3
View day notes {Calendar} (VIEW_DAY_NOTES)=F#3_D3_A2
Delete the note (DEL-ETE_THE_NOTE)=B3A3_F#3_D3
In the examples shown above, the musical form for each UI event is designed such that there are as many musical notes as the spoken content has syllables. It should be noted that, while the mapping of musical notes to a string of syllables is predetermined by a linguistic rule, the assignment of the pitch range, timbre, and tempo to the objects of device UI is more or less arbitrary. It is more of a question of design.
The method 1 for musically modifying a speech signal, according to the present invention, is summarized in FIG. 1. As shown, the speech signal is organized into a string of syllables at step 2. Using a selected linguistic rule, the string of syllables is mapped into a string of tone data at step 4. The string of tone data is transformed into a carrier stream of musical notes at step 6. Optionally, the carrier stream of musical notes is modified to include timbre representing the sound of a musical instrument, at step 8. The carrier stream is modulated with the speech signal to produce a musically modified speech signal at step 10. Optionally, the musically modified speech signal is combined with the unmodified speech signal at step 12 so as to adjust the amount of musical content in the speech signal. It is understood that the resulting signal can be a completely musically modified speech signal or completely unmodified speech or anything in between. The resulting signal is conveyed to a sound-producing device to produce an audible signal at step 14.
How much the audible signal 122 sounds like speech depends on a variety of factors. It may depend on the language itself, or on the linguistic rule (TABLE I to TABLE V, or the like). Thus, it is also preferred that the amount of musical modification be adjusted so that the audible signal 122 can be more speech-like than music-like.
The audible signal 122 can be used in many different ways.
The vocoded signal, or the audible signal 122, as shown in
Thus, although the invention has been described with respect to the preferred embodiments thereof, it will be understood by those skilled in the art that the foregoing and various other changes, omissions and deviations in the form and detail thereof may be made without departing from the spirit and scope of this invention.
| Number | Name | Date | Kind |
|---|---|---|---|
| 4731847 | Lybrook et al. | Mar 1988 | A |
| 4856055 | Schwartz | Aug 1989 | A |
| 5452354 | Kyronlahti et al. | Sep 1995 | A |
| 5481594 | Shen et al. | Jan 1996 | A |
| 5747715 | Ohta et al. | May 1998 | A |
| 5826064 | Loring et al. | Oct 1998 | A |
| 5930755 | Cecys | Jul 1999 | A |
| 6308086 | Yoshino | Oct 2001 | B1 |
| 6385581 | Stephenson | May 2002 | B1 |
| 6459913 | Cloutier | Oct 2002 | B2 |
| 6697796 | Kermani | Feb 2004 | B2 |
| 20020085700 | Metcalf | Jul 2002 | A1 |
| Number | Date | Country |
|---|---|---|
| 5346787 | Dec 1993 | JP |
