Double precision arithmetic in combined formats

There are used double precision mathematical functions in the MDCT core:

• \lib_enc\ivas_omasa_enc.c(130): set_f( hOMasa->prev_object_dm_gains[i], (float) sqrt( 0.5 ), MASA_MAX_TRANSPORT_CHANNELS ); -> sqrtf()
• \lib_enc\ivas_osba_enc.c(108): set_f( hOSba->prev_object_dm_gains[i], (float) sqrt( 0.5 ), MAX_INPUT_CHANNELS ); -> sqrtf()
• lib_rend\ivas_omasa_ana.c(153): set_f( hOMasa->prev_object_dm_gains[i], (float) sqrt( 0.5 ), MASA_MAX_TRANSPORT_CHANNELS ); -> sqrtf()