• use Exact-width integer types (eg uint16_t)

  • Add libresample support to libtwolame (twolame_set_in_samplerate etc.)

  • parameter checking in twolame.c using assert

  • Create a new twolame.spec (be sure to include twolame.pc)

  • quite a lot of duplicated code between twolame_encode_buffer_interleaved and twolame_encode_buffer

  • sort out changing parameter for twolame_set_VBR_q from FLOAT to int (like LAME)

  • better use of verbosity settings

  • full options sanity checking/verification in twolame_init_params

  • don’t allow twolame_set_* after twolame_init_params() has been called

  • Fix broken DAB support

  • with this VBR mode, we know the bits aren’t going to run out, so we can just assign them "greedily".

  • VBR_a_bit_allocation() is yet to be written :)

  • Add a layer 2 decoder ? mpglib ?

  • use 32-bit floats internally where possible