FreeCalypso > hg > gsm-codec-lib
view miscutil/amrts-pcm8-compact.c @ 605:63f774192906
gsmhr_decoder_twts002_in(): set BFI=1 SID=1 for invalid SID
When a received TW-TS-002 RTP payload indicates invalid SID,
which of the 3 possible BFI/SID combinations should we pass to
our internal ETSI-based speech decoder or TFO engine?
Our original code passed BFI=0 SID=1, but upon further reflection,
BFI=1 SID=1 is a better choice. In the corner case where received
invalid SID is fed to a full decoder in homed state, setting BFI=1
allows that decoder to emit zeros on PCM and stay homed, instead of
launching into full decoding.
| author | Mychaela Falconia <falcon@freecalypso.org> |
|---|---|
| date | Thu, 04 Dec 2025 21:01:46 +0000 |
| parents | 7c50864deaff |
| children |
line wrap: on
line source
/* * The set of AMR test sequences shipped by 3GPP as TS 26.074 includes * not only linear PCM (13-bit left-justified) and AMR-encoded files, * but also 8-bit PCM sequences in both A-law and mu-law. However, * those PCM8 sequences are shipped in a stupid and inconvenient format: * each 8-bit PCM sample is expanded to a 16-bit word, written in LE * byte order. This utility converts a PCM8 test sequence file * from this weird format into sane PCM8 format with one byte per sample. * For this conversion, it does not matter whether the PCM8 test sequence * in question is A-law or mu-law. */ #include <stdio.h> #include <stdlib.h> main(argc, argv) char **argv; { FILE *inf, *outf; int cdat, cpad; if (argc != 3) { fprintf(stderr, "usage: %s in-file out-file\n", argv[0]); exit(1); } inf = fopen(argv[1], "r"); if (!inf) { perror(argv[1]); exit(1); } outf = fopen(argv[2], "w"); if (!outf) { perror(argv[2]); exit(1); } for (;;) { cdat = getc(inf); if (cdat < 0) break; cpad = getc(inf); if (cpad < 0) { fprintf(stderr, "error: %s has odd length\n", argv[1]); exit(1); } if (cpad != 0) { fprintf(stderr, "error: presumed padding byte in %s is not 0\n", argv[1]); exit(1); } putc(cdat, outf); } fclose(outf); exit(0); }
