When running the multi pipeline tests for capture, the test picks all the possible capture devices and if thats less than the max count allowed, the remaining ones are filled up with playback pipelines. That is a bit misleading. This becomes a different test than just multiple-pipeline-capture then. Not sure if multiple-pipeline-playback does something similar.
If the number of capture pipelines are less than max count, the test should process with just those.