# This unit test data file is a collection of TRAU-AMR-8k frames, to be # presented to osmo_amrt_decode_trau_frame() in sequence by way of trau2rtp_amr # test program. # The following frames have been emitted by Nokia TCSM2 TRAU under these # conditions: # # - PCM circuit type F; # - TRAU channel activated in AMR-8k mode with DTXd enabled; # - CMR set to 0 (MR475) in the UL frame stream fed to the TRAU; # - G.711 PCMA input was constant 0x54 (idle). # # Since the input was a form of silence (but not EHF) and DTXd was enabled, # the TRAU emitted 7 speech frames (encoding the silence) followed by AMR DTX # pattern of one Sid_Update every 8 frames. # # Please note that unlike FR/EFR, AMR TRAU frame format and semantics are # exactly the same in UL and DL directions. Therefore, frames emitted # by a historical hardware TRAU are just as good for unit test purposes # as those emitted by an E1 BTS - but much easier to produce in Themyscira lab. RESET 00fd8e53bb94908390889e8097c0f0c1f886dfff 00f6bc72e39e958f80889e81d7c0b192f886dfff 00f38d7ed1ee948380989e81d7c0b0b5f886dfff 00f3bd63e39e948380b89e81d7c0b0b5f886dfff 00ff8f71c1ee948380a09e81d7c0b0b5f886dfff 00f3bd63e39e948380b89e81d7c0b0b5f886dfff 00ff8f71c1ee948380a09e81d7c0b0b5f886dfff 0081d360bffffffffff883ffffffffffffffffff 0081c040bffffffffff883ffffffffffffffffff 0081d040bffffffffff883ffffffffffffffffff 0081c26089d8f89ba09883ffffffffffffffffff 0081d040bffffffffff883ffffffffffffffffff # The following frames have been emitted by the same Nokia TCSM2 TRAU # under these slightly different conditions: # # - G.711 PCMA input was constant 0xD5 (silence); # - CMR in the UL frame stream fed to the TRAU (itt-ater-16 test program) # was manually stepped through modes 0 through 4, the 5 possible modes # for TRAU-AMR-8k. # # G.711 A-law (but not mu-law) has this quirk whereby the silence code (0xD5) # decodes into 0x0008 in 16-bit linear PCM, rather than solid 0 - and all GSM # speech encoders, including AMR, interpret a frame of all 0x0008 samples as # an encoder homing frame (EHF). TCSM2 AMR speech encoder also behaves in this # manner, hence the DL output from the TRAU channel was nothing but endless # DHFs. This stream of DHFs exhibits mode switches on RIF=0 output frames # whenever CMR input to the TRAU was changed. # # The following extract exhibits just two frames of each mode, a pair of # RIF=0 and RIF=1 output. RESET 00ff8e49e9e3988394b09ef0d7d8e0c1fab6dfff 00ffbe49e9e3988394809ef0d7d8e0c1fab6dfff 00ff9f62bab8e680fbe89e81c5dfb7fbd7d1f7ff 00ffcf62bab8e680fbf89e81c5dfb7fbd7d1f7ff 00fcac78e2fdf480fbd09e80f3ec87e7eff083ef 00fcdc78e2fdf480fbe09e80f3ec87e7eff083ef 00fc93c6976fa00bc06f898fc2c1b8bc83d8e380 00fce3c6976fa00bc01f898fc2c1b8bc83d8e380 3f079145f7a006f0076208fc4c1ba6f01b003780 3f07f145f7a006f0136208fc4c1ba6f01b003780 # Note this quirk of Nokia TCSM2, visible in all of the DL outputs showcased # above: for both speech and DTXd No_Speech frames in 8k mode, it sets CMR in # its DL output to the same mode as its speech output mode (DL CMI, controlled # by UL CMR). See the corresponding comment in trau-amr-16k.in for further # explanation. # # When this quirk of TCSM2 is combined with CMR/CMI alternation in sequences # of speech frames and the stream of TRAU frames is converted to RTP, the # change in CMR will lag the change in FT by one frame. # Test decoder behaviour when a RIF=1 frame occurs first out of reset, # followed by RIF=0. Mode 2 DHFs were taken from TCSM2 output above, # but reordered for this test. RESET 00fcdc78e2fdf480fbe09e80f3ec87e7eff083ef 00fcac78e2fdf480fbd09e80f3ec87e7eff083ef 00fcdc78e2fdf480fbe09e80f3ec87e7eff083ef 00fcac78e2fdf480fbd09e80f3ec87e7eff083ef # Test decoding of No_Data frames beginning with RIF=1, including shifts in # RIF order (repeated RIF=0 or RIF=1) and UFE/DFE indication. RESET 0081d040bffffffffff883ffffffffffffffffff 0081c040bffffffffff883ffffffffffffffffff 0080d040bffffffffff883ffffffffffffffffff 0081c040bffffffffff883ffffffffffffffffff 0081c040bffffffffff883ffffffffffffffffff 0081d040bffffffffff883ffffffffffffffffff 0081d040bffffffffff883ffffffffffffffffff