headlvowels
: Distributions of vowels from Heald and Nusbaum (2015)Heald, S. L. M., & Nusbaum, H. C. (2015). Variability in vowel production within and between days. PLoS ONE, 10(9), e0136791–e0136791. doi:10.1371/journal.pone.0136791
devtools::install_github("kleinschmidt/healdvowels")
Because of human subjects protections, only aggregate summaries of vowel formants are included, in the form of means and covariance matrices for each vowel, conditioned on:
Time of day:
``` r healdvowels::by_time
```
Speaker
``` r healdvowels::by_speaker
```
Speaker and time of day joinly
``` r healdvowels::by_speaker_time
```
Additionally, the overall (marginal) distributions of each vowel are included:
healdvowels::marginal
#> # A tibble: 7 x 3
#> Vowel model Vowel_ascii
#> <chr> <list> <chr>
#> 1 æ <list [2]> AE
#> 2 ɑ <list [2]> AH
#> 3 i <list [2]> EE
#> 4 ɛ <list [2]> EH
#> 5 ɪ <list [2]> IH
#> 6 u <list [2]> OO
#> 7 ʌ <list [2]> UH
The model
column is a list column, where each entry is a list with mu
(mean
vector) and Sigma
(covariance matrix):
healdvowels::marginal$model[[1]]
#> $mu
#> F1 F2 F3
#> 844.9351 1794.3590 2608.0486
#>
#> $Sigma
#> F1 F2 F3
#> F1 11621.7654 312.2359 9679.805
#> F2 312.2359 46300.8625 28703.813
#> F3 9679.8054 28703.8127 71828.838
These models can be used with
phondisttools
in the same way
that models trained from raw F1×F2 values can be.
Finally, the models
tibble collects the four grouped model tibbls above into a
single tibble:
healdvowels::models
#> # A tibble: 4 x 2
#> grouping models
#> <chr> <list>
#> 1 Marginal <tibble [7 x 4]>
#> 2 Talker <tibble [56 x 4]>
#> 3 Time <tibble [21 x 4]>
#> 4 Time+Talker <tibble [168 x 4]>
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.