audiotestsrc

AudioTestSrc can be used to generate basic audio signals. It support several different waveforms and allows to set the base frequency and volume. Some waveforms might use additional properties.

Waveform specific notes:

Gaussian white noise

This waveform produces white (zero mean) Gaussian noise. Volume sets the standard deviation of the noise in units of the range of values of the sample type, e.g. volume=0.1 produces noise with a standard deviation of 0.132767=3277 with 16-bit integer samples, or 0.11.0=0.1 with floating-point samples.

Ticks

This waveform is special in that it does not produce one continuous signal. Instead, it produces finite-length sine wave pulses (the "ticks"). It is useful for detecting time shifts between audio signal, for example between RTSP audio clients that shall play synchronized. It is also useful for generating a signal that feeds the trigger of an oscilloscope.

To further help with oscilloscope triggering and time offset detection, the waveform can apply a different volume to every Nth tick (this is then called the "marker tick"). For instance, one could generate a tick every 100ms, and make every 20th tick a marker tick (meaning that every 2 seconds there is a marker tick). This is useful for detecting large time offsets while still frequently triggering an oscilloscope.

Also, a "ramp" can be applied to the begin & end of ticks. The sudden start of the sine tick is a discontinuity, even if the sine wave starts at 0. The* resulting artifacts can often make it more difficult to use the ticks for an oscilloscope's trigger. To that end, an initial "ramp" can be applied. The first few samples are modulated by a cubic function to reduce the impact of the discontinuity, resulting in smaller artifacts. The number of samples equals floor(samplerate / sine-wave-frequency). Example: with a sample rate of 48 kHz and a sine wave frequency of 10 kHz, the first 4 samples are modulated by the cubic function.

Example launch line

 gst-launch-1.0 audiotestsrc ! audioconvert ! autoaudiosink

This pipeline produces a sine with default frequency, 440 Hz, and the default volume, 0.8 (relative to a maximum 1.0).

 gst-launch-1.0 audiotestsrc wave=2 freq=200 ! tee name=t ! queue ! audioconvert ! \
     autoaudiosink t. ! queue ! audioconvert ! libvisual_lv_scope ! videoconvert ! autovideosink

In this example a saw wave is generated. The wave is shown using a scope visualizer from libvisual, allowing you to visually verify that the saw wave is correct.

 gst-launch-1.0 audiotestsrc wave=ticks apply-tick-ramp=true tick-interval=100000000 \
     freq=10000 volume=0.4 marker-tick-period=10 sine-periods-per-tick=20 ! autoaudiosink

This pipeline produces a series of 10 kHz sine wave ticks. Each tick is 20 sine wave periods long, ticks occur every 100 ms and have a volume of 0.4. Every 10th tick is a marker tick and has the default marker tick volume of 1.0. The beginning and end of the ticks are modulated with the ramp.