Is there interest to optimize size of speech recordings stored in your system?
For example such codecs as Speex or AMR give a possibility to choose quality
factor when compressing voice files. The trick would be to choose the
optimal one that gives best compression and lowest audio size. This can
easily be done using AQuA as a brute force (or more intellectual) routine as
described here, but for speech:
We can provide this both as a service or as product deployment with
correspondent methodology, which is though quite simple.
Thanks for your feedback and attention!