Speech Analyzer 3.1 is speech-analyzing software used primarily by linguists.

It affords visualizing an
utterance through waveforms (Figure 1.1), which depict sentential- and word-level stress; pitch tracings
(Figure 1.2), which depict intonation; and, spectrograms (Figure 1.3), which depict both stress and intonation.













In addition, an overlay feature that is voice activated allows for the waveforms and pitch-tracings of a
new utterance, which appear in orange, to be visualized atop those of a pre-recorded target utterance, which
appear in black (Figure 2.1). There is a two-second processing delay between the utterance and visualization
thereof, but this ceases to be an issue after familiarization with the program.
Figure 1.2 pitch tracing
Figure 1.3 spectrogram
Figure 1.1 waveform


There are two affordances to the overlay feature (Figure 2.2), the ability to listen to the targeted
utterance, Original, which works well and is convenient, and the ability to hear that session’s recording, Play,
which breaks mid-utterance and is unsuitable for use; however, additional RAM may solve this issue.






Nevertheless, for recordation on a device with 4GB of RAM, one must exit the overlay screen and make a
general recording (Figure 3).

Figure 2.2 playback & original utterance
Figure 2.1 new utterance (orange) over the target utterance (black)













Additionally, because of timing issues pertaining to the novelty of spoken utterances, it is difficult to
match precisely the waveforms of the new utterance with those of the target utterance. I found that one word
works well (see Figure 2.1) and under 10 words are manageable; however, feedback becomes increasingly
more difficult to decipher with each additional word (see Figure 4).






It is important to note that a perfect visual reproduction of the targeted waveform is neither possible
nor desirable (in terms of intelligibly), nor is it necessary for the provision of adequate feedback. So long as
the utterance has been trimmed to a manageable length, improper alignment does not inhibit the ability to tell
which word bears the primary stress, if the stress lands in the proper position of the word, if pauses are used
Figure 3 to record
Figure 4 seven-word utterance ending with a timing discrepancy
correctly, and if the pitch is properly rising and falling. To illustrate the relative inconsequentiality of timing
discrepancies within shorter utterances, Figures 5 and 6 compare one that is target-like over one that is
targeted versus one that is non-target-like over the same. In figure 5, you will notice that despite the timing
being slightly off, the new utterance (orange) and existing utterance (black) have similar patterns, whereas in
figure 6 they do not.



Nevertheless, because shorter speech streams provide feedback that is easier to interpret, another
program such as Audacity is required for trimming (Figure 7). However, because of the drag and drop
capabilities and accessible toolbars within both programs, this is a very fast process. To select a file, drag it
into Audacity, take a clipping, export it in a .WAV format, and drag it into the Speech Analyzer took me just
under 30 seconds. The segmentation of the entire speech stream in Figure 7, for which I clipped 40
manageable files, took me just over 5 minutes.
Figure 6 non-target like stresses and pauses atop a targeted utterance
Figure 5 target-like stresses and pauses atop a targeted utterance with a timing discrepancy



In conclusion, Speech Analyzer is a way to provide visual feedback on the stress and pitch levels of an
utterance through comparisons with a targeted utterance. It could be used within the classroom to allow
students the ability to practice pronunciation, as opposed to waiting in silence, until the teacher can provide
individual attention. I recommend that those who are interested in prosody instruction at least allow their
students an opportunity to experiment with this program. Clearly, some will not like it, for they will find it
confusing and frustrating; conversely, others will figure it out intuitively. For the former group, I do not
recommend this program, but for the latter it has the potential to transform the ordinary skill and drill into a
wonderful challenge that could greatly assist them in their learning of English prosody.
Figure 7 Software such as Audacity can software used to edit longer speech streams.