The present invention relates to the field of synthesizing of speech or music, and more particularly without limitation, to the field of text-to-speech synthesis.
The function of a text-to-speech (TTS) synthesis system is to synthesize speech from a generic text in a given language. Nowadays, TTS systems have been put into practical operation for many applications, such as access to databases through the telephone network or aid to handicapped people. One method to synthesize speech is by concatenating elements of a recorded set of subunits of speech such as demisyllables or polyphones. The majority of successful commercial systems employ the concatenation of polyphones. The polyphones comprise groups of two (diphones), three (triphones) or more phones and may be determined from nonsense words, by segmenting the desired grouping of phones at stable spectral regions. In a concatenation based synthesis, the conversation of the transition between two adjacent phones is crucial to assure the quality of the synthesized speech. With the choice of polyphones as the basic subunits, the transition between two adjacent phones is preserved in the recorded subunits, and the concatenation is carried out between similar phones.
Before the synthesis, however, the phones must have their duration and pitch modified in order to fulfil the prosodic constraints of the new words containing those phones. This processing is necessary to avoid the production of a monotonous sounding synthesized speech. In a TTS system, this function is performed by a prosodic module. To allow the duration and pitch modifications in the recorded subunits, many concatenation based TTS systems employ the time-domain pitch-synchronous overlap-add (TD-PSOLA) (E. Moulines and F. Charpentier, “Pitch synchronous waveform processing techniques for text-to-speech synthesis using diphones,” Speech Commun., vol. 9, pp. 453-467, 1990) model of synthesis.
In the TD-PSOLA model, the speech signal is first submitted to a pitch marking algorithm. This algorithm assigns marks at the peaks of the signal in the voiced segments and assigns marks 10 ms apart in the unvoiced segments. The synthesis is made by a superposition of Hanning windowed segments centered at the pitch marks and extending from the previous pitch mark to the next one. The duration modification is provided by deleting or replicating some of the windowed segments. The pitch period modification, on the other hand, is provided by increasing or decreasing the superposition between windowed segments.
Despite the success achieved in many commercial TTS systems, the synthetic speech produced by using the TD-PSOLA model of synthesis can present some drawbacks, mainly under large prosodic variations.
EP-0363233, U.S. Pat. No. 5,479,564, EP-0706170 disclose PSOLA methods. A specific example is also the MBR-PSOLA method as published by T. Dutoit and H. Leich, in Speech Communication, Elsevier Publisher, November 1993, vol. 13, N.degree. 3-4, 1993. The method described in document U.S. Pat. No. 5,479,564 suggests a means of modifying the frequency by overlap-adding short-term signals extracted from this signal. The length of the weighting windows used to obtain the short-term signals is approximately equal to two times the period of the audio signal and their position within the period can be set to any value (provided the time shift between successive windows is equal to the period of the audio signal). Document U.S. Pat. No. 5,479,564 also describes a means of interpolating waveforms between segments to concatenate, so as to smooth out discontinuities. When a noisy signal is to be synthesized by means of a known PSOLA method, the signal is repeated periodically. This way an unintended periodicity is introduced into the frequency spectrum. This is perceived as a metallic sound. This problem occurs for all noisy signals which do not have a fundamental frequency, such as unvoiced speech parts or music. An unvoiced speech part, like the “s” sound, has no pitch. The vocal chords are not moving as they do for a voiced sound. Instead, a noisy hiss-sound is produced by pushing air through a small opening between the vocal chords. Whisper is an example of speech containing only unvoiced parts. Where there is no pitch, there is no need to change it. However, it can be desirable to change the duration of an unvoiced speech part.
The present invention therefore aims to provide a method of synthesizing a signal which enables to modify the duration of unvoiced speech parts or music without introducing an unintended periodicity in the signal.
The present invention provides for a method of synthesizing a signal, in particular a noisy signal, based on an original signal. Further the present invention provides for a computer program product for performing such a synthesis, as well as for a corresponding computer system including a processor configured to perform the signal synthesis method, in particular, a text-to-speech system for outputting the synthesized signal as a speech signal from a speaker.
In accordance with the invention the required pitch bell locations of the signal to be synthesized are determined. This is done based on, for example, an assumed frequency of for example 100 Hz. This chosen frequency corresponds to a pitch period. The required pitch bell locations of the signal to synthesized are spaced apart on the time axis by intervals having the length of the pitch period. The required pitch bell locations are mapped onto the original signal to provide pitch bell locations in the domain of the original signal. The pitch bell locations in the domain of the original signal are randomly shifted. Preferably the randomization is performed by shifting the pitch bell locations in the original signal domain within +/− the pitch period.
In accordance with an embodiment of the invention the windowing is performed by means of a sine-window. The advantage of a sine-window is that it helps to reduce any residual periodicity. In particular using a sine-window is advantageous in that it ensures that the signal envelope in the power domain remains constant. Unlike a periodic signal, when two noise samples are added, the total sum can be smaller than the absolute value of any one of the two samples. This is because the signals are (mostly) not in-phase. The sine-window adjusts for this effect and removes the envelope-modulation.
In the following, preferred embodiments of the invention are described in greater detail by making reference to the drawings in which:
The flow chart of
Preferably the randomization of the pitch bell locations i is performed in accordance with the following formula:
i′=i+(R×p)
Where i denotes the original pitch bell location on the time axis 202, i′ is the new pitch bell location after the randomization, R is a random number between −1 and 1 and p is the pitch period. The result of the windowing of the original signal is a pitch bell. This pitch bell is placed at the first required pitch bell location within the domain of the signal to be synthesized on time axis 200 as illustrated in
Number | Date | Country | Kind |
---|---|---|---|
02078853 | Sep 2002 | EP | regional |
Filing Document | Filing Date | Country | Kind | 371c Date |
---|---|---|---|---|
PCT/IB03/03544 | 8/8/2003 | WO | 00 | 3/14/2005 |
Publishing Document | Publishing Date | Country | Kind |
---|---|---|---|
WO2004/027754 | 4/1/2004 | WO | A |
Number | Name | Date | Kind |
---|---|---|---|
5479564 | Vogten et al. | Dec 1995 | A |
5581652 | Abe et al. | Dec 1996 | A |
5664051 | Hardwick et al. | Sep 1997 | A |
5890118 | Kagoshima et al. | Mar 1999 | A |
6208960 | Gigi | Mar 2001 | B1 |
6801898 | Koezuka | Oct 2004 | B1 |
6963833 | Singhal | Nov 2005 | B1 |
Number | Date | Country |
---|---|---|
0706170 | Apr 1996 | EP |
0706170 | Apr 1996 | EP |
0706170 | Nov 1997 | EP |
61292700 | Dec 1986 | JP |
63199399 | Aug 1988 | JP |
10214098 | Aug 1998 | JP |
2001513225 | Aug 2001 | JP |
9933050 | Jul 1999 | WO |
Number | Date | Country | |
---|---|---|---|
20060053017 A1 | Mar 2006 | US |