From 758a454aa892657720ac7c779252ba9ac336ba32 Mon Sep 17 00:00:00 2001 From: "Matthias P. Braendli" Date: Fri, 7 Jun 2019 14:13:20 +0200 Subject: Prepare v2.4.0 --- ChangeLog | 12 ++++++++++++ configure.ac | 2 +- doc/TIMESTAMPS.rst | 30 ++++++++++++++++++++++++++++++ 3 files changed, 43 insertions(+), 1 deletion(-) create mode 100644 doc/TIMESTAMPS.rst diff --git a/ChangeLog b/ChangeLog index 452e353..9909265 100644 --- a/ChangeLog +++ b/ChangeLog @@ -1,6 +1,18 @@ This file contains information about the changes done to ODR-DabMux in this repository +2019-06-07: Matthias P. Braendli + (v2.4.0): + Add timestamp offset setting, and fix bug due to UTCO offset + not being properly applied. See doc/TIMESTAMPS.rst for more + details. + Implement FIG2 labels. + Fix bug for DAB MP2 in EEP_A. + Fix FIG0/24 for DAB+ services. + ODR-ZMQ2EDI: fix MNSC endianness swap bug. + Pull some RC changes from ODR-DabMod. + Implement EDI-over-TCP output. + 2019-01-07: Matthias P. Braendli (v2.3.1): Happy new year! diff --git a/configure.ac b/configure.ac index 21d7622..8c7c3b0 100644 --- a/configure.ac +++ b/configure.ac @@ -19,7 +19,7 @@ # along with ODR-DabMux. If not, see . AC_PREREQ(2.69) -AC_INIT([ODR-DabMux], [2.3.1], [matthias.braendli@mpb.li]) +AC_INIT([ODR-DabMux], [2.4.0], [matthias.braendli@mpb.li]) AC_CONFIG_AUX_DIR([build-aux]) AC_CONFIG_MACRO_DIR([m4]) AC_CANONICAL_SYSTEM diff --git a/doc/TIMESTAMPS.rst b/doc/TIMESTAMPS.rst new file mode 100644 index 0000000..458f462 --- /dev/null +++ b/doc/TIMESTAMPS.rst @@ -0,0 +1,30 @@ +Some knowledge accumulated about timestamping +============================================= + +The meaning of the timestamps changed between v2.3.1 and v3.0.0, this document gives some guidance about the interaction between different settings. + +The following table tries to summarise the differences. + ++-----------------------------+----------------------------------------------+-------------------------------------+-----------------------------------------------+ +| ODR-DabMux version | Meaning of timestamp inside EDI | ODR-ZMQ2EDI wait time w | Offset that should be set in the mod | ++=============================+==============================================+=====================================+===============================================+ +| Up to and including v2.3.1 | t_frame = t_mux (No offset in mux available) | positive, meaning delay after t_mux | Something larger than w + mod processing time | +| Later than v2.3.1 | t_frame = t_tx = t_mux + tist_offset | negative, meaning delay before t_tx | Something larger than mod processing time | ++-----------------------------+----------------------------------------------+-------------------------------------+-----------------------------------------------+ + +For historical reasons, ODR-DabMod decodes absolute timestamp from MNSC, not from “EDI seconds”. +The edilib tool decodes both EDI timestamp and MNSC, and can be used to verify both are identical. + +Issues in ODR-DabMux v2.3.1 +--------------------------- + +Running ODR-DabMux against the absolute timestamp firmware has uncovered a few issues: + + * At startup, the UTCO was not properly applied to the EDI seconds. This offset was 5 seconds (TAI-UTC offset - 32s, see EDI spec); + * odr-zmq2edi did not compensate for UTCO, hiding the above issue; + * ODR-DabMux needs a configurable offset; + * (minor) MNSC and EDI timestamps did not use the same internal representation, making it difficult to prove that they encode the same value; + * (minor) odr-zmq2edi swapped endianness when regenerating EDI from ETI (minor because only ODR-DabMod considers MNSC, and usually isn't used with EDI); + +**Important** Do not combine odr-zmq2edi with odr-dabmux of a different version! + -- cgit v1.2.3