view doc/AMR-study-utils @ 419:2a094af3d384

libtwamr: implement encoder top level
author Mychaela Falconia <falcon@freecalypso.org>
date Tue, 07 May 2024 04:51:04 +0000
parents f4f68c652e98
children e26b974f7ba3
line wrap: on
line source

As of this writing (2023-04), the main emphasis here at Themyscira Wireless is
on classic FR and EFR codecs, particularly the latter, rather than AMR.
However, given the close relation between GSM EFR and the highest MR122 mode of
AMR, we are starting to explore AMR a little too, focusing on the reference C
implementation and various published test sequences.  Working in this direction,
we have developed the following AMR study and exploration utilities:

amr-cod-parse	This utility reads the *.cod binary file format used by 3GPP
		reference code and test sequences for AMR-encoded speech, groks
		every frame and dumps all recorded parameters in human-readable
		form.

amr-ietf-parse	A similar AMR parse/dump utility to amr-cod-parse, but reading
		the more common IETF RFC 4867 *.amr binary file format instead.

amr-cod2ietf	These two utilities convert between 3GPP *.cod and IETF *.amr
amr-ietf2cod	formats.  amr-cod2ietf converts from *.cod to *.amr;
		amr-ietf2cod converts in the opposite direction.

Just like ETSI *.cod and *.dec formats for EFR, 3GPP *.cod format for AMR is
endian-dependent.  Our amr-cod-parse and amr-cod2ietf utilities expect LE byte
order by default, matching the official test sequences in
ts_126074v170001p0.zip; they also support BE byte order with -b option.
However, our amr-ietf2cod utility (rarely needed) emits its cod-format output
in the local machine's native byte order.