Introduction
Welcome to psyphonics.
Here you can find a preview version of a web-based application for running forced-choice psychoacoustic experiments.
The software uses a text-based, human readable configuration file to define both the stimulus and the procedural details of the experiment.
The current version of the software supports stimuli comprising one or more (optionally modulated) sinusoidal waveforms, each of which can be associated with an independent temporal gate. The accompanying Stimulus Designer can be used to create and monitor stimuli and to generate the appropriate stimulus configuration.
Stimulus and Task Configuration
An example stimulus configuration file is shown below.
Example 1. Configuration for a 1-kHz tone with 5% AM and FM modulation
The sample configuration defines a 1-kHz pure tone modulated in both amplitude and frequency. The details of the carrier tone, the temporal envelope, the amplitude modulation, and the frequency modulation are specified separately in different sections of the configuration. At least one tone and one gate is required to define a valid tone, the modulation settings being optional. Note that the configuration uses JSON syntax (JavaScript Object Notation) a commonly used format for transmitting data over the internet.
Considering the configuration in more detail, the tones section defines one or more sinusoidal components, including the frequency, starting phase, attenuation relative to full scale (in decibels), the channel(s) the tone is presented to (0 = left, 1 = right), which gate to use (indexing is zero-based, so 0 is the first gate in the array), and which amplitude and frequency modulator to use, if any.
Each gate defines an ADSR (attack, decay, sustain, release) envelope, including the shape of the attack, decay, and release portions of the envelope (currently either Linear or Hann). For each of the two supported modulation types we can specify the frequency, starting phase, and modulation depth (in percent). All phases should be specified in radians and durations are given in seconds.
Example 2. Sample configuration for a pure-tone intensity discrimination task
The configuration for a forced-choice psychoacoustic task consists of a base stimulus, a signal function, and a set of key-value pairs specifying the constraints on the signal magnitude and the details of the adaptive staircase procedure.
The signal function takes the configuration of the base stimulus and the current signal magnitude as arguments and modifies the configuration for playback in the signal interval. The software also supports random perturbations of the stimuli, both between within and between trials. Such perturbations are commonly used in psychoacoustic tasks to restrict the available discrimination cues, for example, to eliminate within-channel loudness as a cue in spectral shape discrimination tasks.
Sample Threshold Estimation Tasks
The following examples demonstrate some of the capabilities of the Psyphonics Experimenter. Ideally, the listening tasks should be performed in a quiet environment using high-quality audio equipment. Headphones are required for binaural listening tasks (e.g. Example 3) and are strongly recommended for all examples.
Example 1. Pure-tone intensity discrimination
The perceived loudness of a sound is primarily determined by the intensity of the underlying vibration, with more intense vibrations appearing louder. Intensity discrimination thresholds therefore provide a measure of the change in intensity required to give a just noticeable change in loudness.
In this example, intensity discrimination is estimated for a 1-kHz pure tone using a two-interal, forced choice task, with the intensity of the signal adjusted dynamically using a 2-down, 1-up adaptive staircase procedure. Such a procedure targets the signal increment required for 71% correct discrimination (Levitt, 1971). Thresholds for such a task would be expected to be around 1 dB.
Example 2. Pure-tone frequency discrimination
The perceived pitch of a pure tone is primarily determined by the frequency of the vibration. This example estimates the frequency difference required for 71% correct discrimination. As in Example 1, the frequency of the base stimulus is 1-kHz. For this frequency, the expected threshold would be somewhere in the region of 5Hz (0.005%).
Example 3. Interaural intensity discrimination
This example, again using the same 1-kHz pure tone, measures sensitivity to a difference in intensity between the signals reaching the left and right ears. Such differences arise naturally for sounds that are spatially located to the left or right due to the "shadow" effect of the head. To ensure that discrimination is based on comparison of the signals reaching the left and right ears (and not on overall changes in intensity), the intensity difference in the signal interval is achieved by increasing the intensity in the right ear by half the signal magnitude, and decreasing by the same amount in the left ear. Additionally, the overall loudness of tones is randomized on each presentation. The expected threshold would be similar to that for monaural intensity discrimination, i.e., about 1 dB.
Example 4. Spectral profile discrimination
One of the physical properties of an auditory stimulus that determines its quality or timbre is the relative distribution of energy across frequency (spectral profile). This example measures sensitivity to a change in the intensity of a 1-kHz target tone relative to six flanking tones having frequencies of 200, 342, 585, 1710, 2920, and 5000 Hz. To ensure that discrimination is based on a comparison of the relative intensities of the target and flanking tones, the overall stimulus level is randomized on each presentation.