* get rid of globals in bitstream parsing
* get rid of statics in imdct.c
* avoid overflows, including reading the a52 stream !!!
* make dither code faster (generate dither table in advance ?)
* redo all bit allocation if previous frame had zero_snr_offsets
* make dynrng work in dual-channel streams

* fix and test mlib IMDCT

* finish a52_internal.h cleanup
	* use sensible data types
	* move some arrays into bit fields
* imdct.c
	* maybe use a single w[] array with only positive values
	* use split-radix algorithm
* cplfleak/cplsleak: move 2304-x in bit_allocate.c ?

* see if we can get rid of cplcoe in the phsflginu test

* update bitstream code


from aaron:

- Detect hrtime on solaris and enable it via config.h
- Add multilevel debug output
- change the divides/moduluos in coeff_get_mantissa to table
  lookups. Need to look at stats first though.
- rewrite imdct to use split 2/4 radix fft
- investigate strangeness in 256 point imdct (weird transients?)
	

DONE
- fix the bitstream interface to allow on aligned buffers - Feb 2000 -AH
- change dither.c, crc.c to use a lookup table - March 2000 -AH
- the output signal from the IMDCT sometimes is outside the rand
  of [-1.0,1.0]. This is why there is the normalization pass in
	output.c. Need to investigate and hopefully we can move the
	normalization. - Feb 2000 -AH
- Mute frame on error. Feb 2000 -AH
- Revisit convert_to_float and use int->float with scaling - Feb 2000 -AH
- Use proper frame sizes in auxdata parsing - July 1999 - AH
- fix clipping in output (normalize to max digital output) - Aug 99 -AH
- Fix coupling channel support - July 1999 - AH
- Add delta bit allocation support - July 1999 - AH
- Add downmixing support - Sept 1999 - AH, Yeqing Deng
