[util] FR: stop at pre-proc 2 (.hlsyn): mismatch between raw outputs

I'm coming from the dectalk-3000b repo and most of the raw files do not match, i.e, the output from HLsyn is not the very same as Klatt's (which should be in fact).

I came across some problems using bash's od tool to convert binary data into short integers (16 bits per samples), but then I listened to the audios myself and at least one of them both was indeed buggy.

I can blame Python for calling Klatt and HLsyn binaries via subprocess.call() module, but I can't rely on that reason. All I know is that I generated the whole data again in another repo (dectalk-3000c), this time calling the HL and LL binaries via terminal itself over bash, added more sanity checks and everything seem to work fine.

Therefore, this feature request proposes changing the repo name to dectalk-datagen and step at the generation of the .hlsyn output file via DEC2HL script, leaving the later processes of calling HLsyn and Klatt binaries to generate the raw waveforms to another repo.