I've extracted the words "Mary" and "merry" from Richard Fontana's reading of "Mary dear, make me merry; say you'll marry me", and I've performed formant analysis on the first vowel of each of those words.

In each case, a segment has been extracted corresponding to the [E] sound only,  the intent being to try to avoid the suggestion of "m" at the beginning and "r" at the end.  You can listen to the segments for "Mary" and "merry".

The formant plots are shown in three ways: a plot of the "Mary" formants, a plot of the "merry" formants, and the two plots in different colors and superimposed.   The formants are referred to as "first", "second", "third", and "fourth" in ascending order from the bottom of a plot.

I've added the formant positions of the two vowels to an earlier vowel chart that showed formant positions of 14 vowels spoken by Richard.   In each case, the interval represented by the formant glide was taken between the first maximum and the minimum of the second formant. The same interval was used for the first formant.  The formant plots show that the formants practically coincide after the minimum is passed.

In consideration of the suggestion that there may be a difference in length between the vowels in the two words, I've plotted the time function of the "err" portion of each word.  You can listen to the "err" that I've extracted from "Mary" and "merry".

I've added formants for Richard's "a" vowel in "marry" to the vowel chart, but I haven't installed sound files or formant plots for that vowel.  I may or may not add them later.

My conclusions are:

• Although there is a slight difference between the two vowels, both of them are essentially the same at the beginning as the vowel in Richard's earlier pronunciation of "head".
• They both have a substantial glide toward a mid-central position.
• There's no significant difference in the lengths, and in fact the "merry" vowel seems to be slightly longer.  (I say "seems to be" because the extraction of the segments is a matter of judgment and so is subject to some error.)
• There's an interesting -- but slight -- difference between the third formants, but I can't say for sure what the significance of that is, except that it may be related to a difference in nasality.

Added conclusions, 16 May 2002:

• The start of the vowel in "marry" is quite close to Richard's earlier pronunciation of the vowel in "had".
• Like the other vowels, it glides toward a more central position.
• After further reflection, I think the glides shown in all three vowels are the effect of "r" coloring.  Although I tried to pick segments that avoided the "r", I suspect the effect of the "r" starts showing up in the formants almost immediately after onset of the vowel.

  -- Bob Cunningham