TCX: Review buffer-sizes of Aq[][]-buffers

Basic info

• Fixed point:
  • Encoder (fixed): 187213dc
  • Decoder (fixed): 187213dc

Bug description

The buffer sizes of the Aq_fx[][]-buffers should be reviewed and possibly be reduced (also in floating-point code).

Example: In the function stereo_mdct_core_dec_fx() there following two buffers are allocated:

Word16 Aq_fx[CPE_CHANNELS][( NB_SUBFR16k + 1 ) * ( M + 1 )];
Word32 Aq_fx_32[CPE_CHANNELS][( NB_SUBFR16k + 1 ) * ( M + 1 )];

However, it seems that only a buffer size of [CPE_CHANNELS][NB_DIV * M] is actually needed. Not sure how easy it is to change this since, since typically a 2-dim array is passed to the lower-level functions.

Both, encoder and decoder should be checked.