
HARMONIC ANALYSIS OF CERTAIN VOWEL SOUNDS. 763 
TABLE X. (FouRTH PARTIALS.) 


Pitch of Third Partial. |e” Ag bh” a” oe g”’ FR" ve el! eb” q!’ of cl b! bh! a oR 9%, 



Amplitudein Voice 1,/ 10 | 14) 8) 22)...| 6) ...| 4) 7]... | 16] .., | 83] 31 |.47 | 29) ... | 40 
- Peo lea koh Zl ek Oe ce eSuluesaliee | EO a|. oa lb bSt#26.10 36 | 39 
x era 3. 140] 0% | 8 6 1S; 5 *25 | 15 | 21 










A very remarkable instance of an abrupt change in the constitution of a 
vowel sound is afforded by the letter @, on which we have made a large number 
of observations which will now be described. Like 0, @ was well spoken by 
the phonograph. Figs. 7 and 8 give sets of curves for uw, sung by voices 5 
and 1 respectively. In voice 5 the curves given for the sound a above 
@ are very approximately simple harmonic curves; almost the whole of the 
sound consists of the prime tones. This result agrees so far with that 
obtained with the phonautograph by DonprErs. But at @ a sudden and 
remarkable change takes place. The period split up into two halves, each 
very nearly simply harmonic, and the analysis of this form shows that it 
consists of a feeble prime tone with an excessively strong second partial. 
This form is given on g, f, e¢, and d. One, d, and B the third partial 
appears, but we cannot say that we have got any really good w’s from any 
voice or notes below ¢; at these low pitches the vowel sound of this letter 
become very poor, the marks on the tinfoil were feeble, and the reproduc- 
tion by the phonograph was little more than an inarticulate groan. For 
the sake of brevity, we will call the approximately simple harmonic form 
obtained for « above a the simple u, and the form shown for the notes ef, gf, 
and @, the duplex u. At the first glance it is difficult to distinguish the duplex 
curve for w on ¢ from the simple one for w on @. 
Table XI. gives the analyses of the #s in fig. 7. It will be observed that 
the note a is wanting in the examples given. It has been omitted because we 
had difficulty in obtaining a good « back from the phonograph when the voice 
was sung on this note. The voice in question (No. 5) was very apt to give an 
6 quality to the sound, both as originally sung and as reproduced by the phono- 
graph, and in that case the form of the curve obtained was at once recognisable 
as resembling 6. The attempts of this voice to sing # at this pitch generally 
resulted in a vowel sound of very variable quality, and a correspondingly 
irregular form in the phonographic records. The voice seemed to have a diffi- 
culty in keeping the second partial sufficiently weak to give the simple form, 
or in making the second partial sufficiently strong to give the duplex form. 
But although no good specimen of voice 5 has been transcribed on this note, 
