Warning: The extended set of examples may use large amounts of data.
This app visualizes the pitch and resonance of voices in recordings over time. Pitch is our perception of the frequency at which the vocal folds vibrate. Resonance is the effect exerted on a sound by the passage the air travels through. The program plots recordings in a 2-dimensional space, where the Y-axis represents pitch, and the X-axis represents resonance, with brighter sounds represented with a higher percentage.
Typically, voices perceived as female have brighter resonance and higher pitch (top right), while those perceived as male have darker resonance and lower pitch (bottom left). These features can be blended in different degrees to create a broad spectrum of perceptual and aesthetic qualities related to gender. There are no sharp cutoffs at which a voice is guaranteed to be gendered a specific way, just as the colors in the graph blend seamlessly into one another.
Press the play button in the top-left corner of the app to play the clip associated with the selected point on the graph. Examine some of the provided examples to get a sense of how positions on the graph correspond to your perceptions of gender. (Note: Example clips may take some time to show up on slow connections). You can visualize your own recordings by pressing the add clip button in the top-right and submitting a recording. It will take some time to process, so you may wish to examine the other clips while the recording is processed.
Your clips are sent to our server for processing, but they are not stored persistently. You can save the sound file from your recording using the download button in the "Details" tab. From the settings tab, you can also export all of your data in a single file which can be imported for viewing at another time. Otherwise, your data will be lost when you exit the page.
We hope that using this program will help you understand how voices are gendered and assess your own progress if you are trying to alter your voice to better match your gender identity or desired presentation. Good luck!
Most speech has a broad range of values in pitch and resonance, e.g. The highest and brightest phone spoken in a sample of typical male speech may well be similar in pitch in resonance to the average phone in typical female speech. Our perception of a voice is based on the overall tendencies within speech rather than the moment-to-moment features of individual phones.
When not playing, the markers are displayed at the median values across the whole clip. Because there's a wide range of values in most speech, this pushes the median towards the center. However, a small change in median values can have a big influence on how a voice is gendered
Not currently. If you add recording in another language, you might get an output, but it will probably not be very accurate.
Yes, so you may wish to upload recordings of other speakers with your accent or dialect to better compare with your own voice.
There are several ways. One is to raise (brightening) or lower (darkening) your larynx. Your larynx raises when you swallow and lowers when you yawn. You may be able to feel it move if you put your hand over the front of your neck while you make one of these movements. If you can learn to control your larynx position in normal speech, it can have a big effect on your resonance.
Your resonance is also affected by the space in your mouth. Holding your tongue higher in your mouth as you speak will brighten your resonance, and holding it lower will darken it.
There are many dimensions to voice other than pitch and resonance, so you might need to target other features. You might also be straining to reach a particular point, which can be audible to a listener. It might help to pick a less extreme target which you can reach comfortably.
Yes! You can donate through any of the following platforms:
This application is developed by Luna McNulty as part of a research project for her Sc. M. program at Brown University. Its content is available under CC BY-SA 4.0 and uses Creative-Commons-licensed sound files from Morgan and Christine Lemmer-Webber's podcast Foss and Crafts as well as public domain content audiobook content from Librivox.