Is it possible to this project work if we use fixed-point quiet_sample_t instead of floating point quiet_sample_t?

[HengYongChao]

I try to redefine quiet_sample_t as int16_t in the encode_file example, It do output a file,but when I use decoder_file example to decode this file then it output nothing.

Is it possible to implement this routine using a fixed point number, since the chip I use reads the MIC sample as if it only supports a fixed point number.

Thanks a lots！