i played around a little bit and found out how to listen to espeak without gstreamer.
one option ist to use mlayer to read from stdin:
unfortunatly, mplayer dont recognize end of stream or espeak does't send it, so must press Ctrl C to stop mplayer.

second option need no additional install as it works with the (i think) on board command play-sound:
based on the latter option, i wrote me a little python script to listen to txt file line by line. At least for "-v de" it is not easy to understand the text, even when adjusting the -s and -g options(speed and word gap) of espeak.
Sometimes it sounds very clear, sometimes not.
One of those meant to be easy things which is still a challenge to artificial intelligence