Subsampling in an audio pipeline

[wavlew]

I am using a gst spectrum element in a python environment in this pipeline:

pipeline = Gst.parse_launch(autoaudiosrc ! spectrum interval=100000000 bands=256 threshold=-70 multi-channel=true ! fakesink)

Everything works fine but the samplerate is with 96kHz very high. For a spectrum analyzer, the quarter of it would be enough. so I could save FFT processing time.

How can I force subsampling (to lets say 20kHz)?

[Baby Octopus]

autoaudiosrc ! audioconvert ! audio/x-raw,rate=20000 ! spectureum ....

[wavlew]

... perfect, thank you !