Graphing Vocal Vibrato

    Most Western vocalists have a vibrato that far exceeds the bounds of one semitone, so the display on the lower graph becomes quite useless. Any particular note changes its semitone identity rapidly, during a time we wish to think of it as one particular semitone.

                            Vibrato in normal mode

    The Lower Graph is useless.

    The solution is to expand the pitch range covered by the lower graph; this results in diminished magnification, but keeps the vibrato within the range of the graph. To effect this, select a
View | Lower Graph Scaling that is greater than the default of 10 cents per division. If, for instance, the vibrato oscillates two semitones top to bottom, then it would just fit in the 20 cents per division scaling, if it were exactly in tune. Since it won't be, 30 cents per division is the appropriate choice.
    Here is a short trace to explain the concept. As before, the orange boxes show the area being mapped by the lower graph. Note the labels - the scaling has changed, and each dashed line represents a tuning differential of 30 cents. Since the lower graph now extends into the surrounding semitones, the centers of those semitones are also colored in on the graph. [to make that clear, the semitone lines C, B, and A# are labeled for the note B, the third box in the trace].

                      Lower Graph Scaling 30

    Parsing the beginning and ending of individual notes is quite imperfect in this mode. You will almost always encounter incorrect assignment of the semitone at the beginning of a note; it is unavoidable. Nevertheless, you will have the precise pitch feedback for the main part of the note - this is certainly enough information to help you refine your intonation. Furthermore, the upper graph will be correct, it doesn't depend on separating the tune into this or that semitone. So it can always serve to show you how your notes begin. The end of a note is usually correct in the lower graph, just that it may continue in that color while the next note begins, until enough time has elapsed for the new note to become clear to TuningMeister.
    In the above example, it takes a while for TM to lock in on the E, which is the first note. 
Then the pitch of E, which starts out well-centered, rises to about 30 cents or more sharp, as can be read from the black averaging trace in the lower graph. (Once TM settles on E as a solid note, the averaging graph can't begin until one cycle of the vibrato has occurred.) The D lies a bit sharp from the beginning, and gets sharper and sharper at the end until finally it breaks into the region of Eb. Then it drifts down preparatory to the consonant that begins the B - the upper graph shows some breaks where the consonant is pronounced, then the lower graph locks onto the B. Some wobble in the average pitch is quite common in vocal production. Then another consonant begins the A; again, you can see the averaging graph doesn't begin until after the first cycle of the pitch. This A starts 30 cents sharp and improves slightly towards the end.
    Those deviations from the center pitch look sort of small in the graph, but remember the scale is much more compressed. At the end of the box depicting the D, the pitch is now halfway to Eb, right between the semitones. It is in fact quite out of tune. The other notes aren't very accurate either, but this seems to be something we have stopped listening to.
    If your vibrato is even wider, you will need to select
View | Lower Graph Scaling to be 40 or 50.

Back to the Index