Source Filter Theory

4 Key excerpts on "Source Filter Theory"

• 

  eBook - ePub

  Experimental Phonetics

  An Introduction

  • Katrina Hayward(Author)
  • 2014(Publication Date)
  • Routledge

    (Publisher)

  Chapter 4 The acoustics of speech production 4.1 Introduction In Chapters 2 and 3, our main concern was with the description of speech as sound. For phoneticians and linguists, however, this is not an end in itself. As explained in Chapter 1, we need to relate the sound of speech to speech production on the one hand and to speech perception on the other. In this chapter, we shall focus on the relationship between acoustics and production. The crucial link is provided by the source-filter theory, also often referred to as the acoustic theory of speech production. More specifically, the source-filter theory is concerned with the final stage of speech production, when sound is actually generated. It addresses the question of why particular articulations should give rise to particular sounds, but is not concerned with how the speaker produces the articulations in the first place. For example, the source-filter theory can explain why a vowel produced with a high front tongue position and spread lips ([i]) will have a spectrum like that of Figure 3.7, but it has nothing to say about how the speaker gets his or her tongue to the high front position. The basic idea behind the source-filter theory is quite simple. Sound production often involves more than one stage, and so it is necessary to distinguish between the original source of sound and later modifications (filtering) of that source. For example, in the case of a violin, the original source of sound is a vibrating string, but the vibration of the string is transmitted to a sort of box and it is the vibration of this second object which is responsible for the sound which is transmitted to our ears. At the risk of oversimplification, we may say that the string and the box make separate and independent contributions to the output sound

  Sign up to read

  Learn more about book
• 

  eBook - ePub

  Phonetics

  Transcription, Production, Acoustics, and Perception

  • Henning Reetz, Allard Jongman(Authors)
  • 2011(Publication Date)
  • Wiley-Blackwell

    (Publisher)

  9

  The Source-Filter Theory of Speech Production

  Chapters 3 and 4 described how speech sounds are produced with the speaking apparatus and how they can be transcribed. Chapters 5 and 6 presented the production of speech sounds in detail. Chapters 7 and 8 introduced the basic elements of acoustics and how sound can be analyzed. We now apply this knowledge to the acoustics of speech sound production. An important concept is the separation of speech sounds into one or more source signals and the modification of these signals by a filter .

  Sound sources for speech are rather limited. The sources are either the vocal folds for voiced speech sounds or a constriction that generates a noise somewhere along the vocal tract. Such a noise can be produced either continuously as in fricatives or only as a transition as for plosives. Another possible source is “jiggling” articulators, as in trills. For example, in a trilled [r], the vibrating tip of the tongue is a sound source. More than one sound source can exist at the same time, as, for example, in the case of a voiced fricative, which has a laryngeal source at the glottis and an oral source at the constriction at the place of articulation. Despite the small number of sound sources, a great range of speech sounds can be produced by filtering. This is what the vocal tract does to the sound source(s) at the larynx or in the vocal tract itself: it filters the sound sources and gives the speech sounds their characteristics.

  This situation is similar to a trumpet player who produces a louder or softer tone with lower or higher pitch but the sound always has a characteristic trumpet “sound.” To change the sound quality, a trumpet player can use a mute at the opening of the trumpet and move it during playing. This does not change the pitch of the trumpet sound, and varies the loudness only to some degree, but more importantly, it changes the quality of the sound. What actually happens is that the sound source, here the trumpet tone, is filtered

  Sign up to read

  Learn more about book
• 

  eBook - PDF

  Phonetics

  Transcription, Production, Acoustics, and Perception

  • Henning Reetz, Allard Jongman(Authors)
  • 2020(Publication Date)
  • Wiley-Blackwell

    (Publisher)

  Despite the small number of sound sources, a great range of speech sounds can be produced by filtering the sound sources. This is what the vocal tract does to the sound source(s) at the larynx or in the vocal tract itself: it filters the sound sources and gives the speech sounds their characteristics. This situation is similar to a trumpet player who produces a louder or softer tone with lower or higher pitch but the sound always has a characteristic trumpet “sound.” To change the sound quality, a trumpet player can use a mute at the The Source–Filter Theory of Speech Production 9 176 The Source–Filter Theory of Speech Production opening of the trumpet and move it during playing. This does not change the pitch of the trumpet sound, and varies the loudness only to some degree, but more importantly, it changes the quality of the sound. What actually happens is that the sound source, here the trumpet tone, is filtered by the mute. Something similar can be done with a stereo set by turning the “bass” and “treble” controls: this does not change pitch or rhythm, but it changes the sound quality by increasing or attenu-ating the low‐frequency or the high‐frequency regions. As the periodic and aperiodic source signals travel through the vocal tract, they are modified by the specific shape and diameter of the tract. Source and filter together determine the acoustic characteristics of speech sounds. Important parameters in describing filters are resonance and damping (the amplification and attenuation of frequencies by a filter), which can be computed from the size and shape of the vocal tract. Very important for sonorant sounds, especially vowels, are the main resonating frequencies of the vocal tract, which are called formants and which are the acoustic determinants of the vowel qualities (and can be associ-ated with the articulatorily‐based vowel quadrilateral).

  Sign up to read

  Learn more about book
• 

  eBook - PDF

  Proceedings of the seventh International Congress of Phonetic Sciences / Actes du Septième Congrès international des sciences phonétiques

  Held at the University of Montreal and McGill University, 22–28 August 1971 / Tenu á l’ Université de Montréal et á l’ Université McGill, 22–28 août 1971

  • André Rigault, René Charbonneau, André Rigault, René Charbonneau(Authors)
  • 2017(Publication Date)
  • De Gruyter Mouton

    (Publisher)

  The conventional view of sound generation in the vocal tract is, of course, that of a sound source which is filtered by the vocal-tract resonators to produce a sound output from the mouth or nose (Fant 1960). For most speech sounds, the sound source results either from vocal-cord vibrations or from turbulent airflow at a con- striction somewhere along the length of the vocal tract. The structures that give rise to the sound sources and are responsible for shaping the acoustic cavities which filter these sources can, for our purposes, be divided into four classes, as shown in Figure 1. These are (1) the entire respiratory system below the larynx, including the airways (trachea, bronchi, etc.) and the lungs; (2) the larynx; (3) the vocal tract between the larynx and the lips, including both the pharyngeal and oral portions; and (4) the nasal cavity. The dimensions and other properties of the subglottal respiratory system may vary greatly from one individual to another. The dimensions of the airways and the 208 KENNETH N. STEVENS elasticity and size of the lungs may be different, and these result in differences in vital capacity and in the resistance of the subglottal airways. Some of these aspects may also vary from day to day for a given individual, and will certainly change over a longer period of several years. These changes will have only a secondary effect on the properties of the speech sounds, however, since the subglottal respiratory system is only indirectly involved in sound production. The elasticity of lung tissue can have an indirect effect on speech, since it can in- fluence the subglottal pressure and the rate at which the subglottal pressure can be varied. Since the subglottal pressure changes are responsible for some of the variation in fundamental frequency during speech, modification of the lung characteristics can result in changes in the contour of fundamental frequency and in the range of fundamental frequency used by an individual when he talks.

  Sign up to read

  Learn more about book