5

I've been looking all over the internet, but I can't seem to find out the default speaking rate for the in-built speech synthesis program on Mac OS.

say "What to say" -r <rate>

As shown above there is a way to specify the speaking rate (wpm) but there is no where that specifies the default rate?

Anyone have an idea of what it is?

4

The default speaking rates can actually be changed from System Preferences:

The slider can be set to a different position for different voices. When I tried dragging the sliders for Alex and Mikko to the middle, the WPMs shown in the output below were 175 and 180.

$ defaults read com.apple.speech.voice.prefs VoiceRateDataArray
(
        (
        1835364215,
        201,
        175
    ),
        (
        1919902066,
        546,
        180
    )
)

The outputs of say -r 175 and say without an -r option were identical for Agnes, Alex, and Bruce, and almost identical for some other voices. But for example the default speaking rate of the "Bahh" voice is much slower.

$ say -v ? | sed 's/  .*//' | while read l; do say -v "$l" -r 175 {1..9} -o "$l 1.aif"; say -v "$l" {1..9} -o "$l 2.aif"; done
$ stat -f '%z %N' *
120136 Agnes 1.aif
120136 Agnes 2.aif
131692 Albert 1.aif
180260 Albert 2.aif
122968 Alex 1.aif
122968 Alex 2.aif
231596 Bad News 1.aif
231820 Bad News 2.aif
131692 Bahh 1.aif
180264 Bahh 2.aif
318508 Bells 1.aif
318488 Bells 2.aif
131692 Boing 1.aif
164368 Boing 2.aif
116514 Bruce 1.aif
116514 Bruce 2.aif
131692 Bubbles 1.aif
138860 Bubbles 2.aif
196652 Cellos 1.aif
197088 Cellos 2.aif
131692 Deranged 1.aif
150720 Deranged 2.aif
131692 Fred 1.aif
138860 Fred 2.aif
331052 Good News 1.aif
331492 Good News 2.aif
301932 Hysterical 1.aif
301916 Hysterical 2.aif
131692 Junior 1.aif
144676 Junior 2.aif
131692 Kathy 1.aif
144676 Kathy 2.aif
118196 Kyoko 1.aif
114440 Kyoko 2.aif
420970 Mikko 1.aif
408432 Mikko 2.aif
228012 Pipe Organ 1.aif
228448 Pipe Organ 2.aif
131692 Princess 1.aif
144676 Princess 2.aif
131692 Ralph 1.aif
144676 Ralph 2.aif
128810 Sin-Ji 1.aif
124834 Sin-Ji 2.aif
131692 Trinoids 1.aif
138860 Trinoids 2.aif
119410 Vicki 1.aif
116118 Vicki 2.aif
122742 Victoria 1.aif
119444 Victoria 2.aif
131692 Whisper 1.aif
150724 Whisper 2.aif
131692 Zarvox 1.aif
138860 Zarvox 2.aif
0

Not a "nice" way of doing it, but you can do some samples, maybe starting at 120 wpm, and compare them with a control recording of the "default" speed (obviously, with all recordings using the same voice and text). For more precision, you could load the files into any basic audio editing application and view the waveforms to really compare the rate of one recording and another. I can't imagine it would take too long to narrow it down. I'd do it myself if I had the time right now.

Alternatively, you could get some text samples of known lengths (say 100 words, 200 words, etc.), record the voice speaking all of them at default settings, then use the durations to calculate the average wmp of all the recordings. With enough samples I'm sure you'll get close.

EDIT: Following the first procedure I described above, I have found the default speaking rate of Alex (the default voice) to be exactly 175 wpm.

At default and 175 wpm speeds, the two waveforms match perfectly.

0

720 seems max for the voice I am using

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .