Appendix

We show in the following list all the available commandline options for the H.263 encoder, which we used to generate the distorted video sequences.

TMN (H.263) coder version 3.0, University of British Columbia, CANADA,
based on Telenor's coder version 2.0, Copyright (C) 1995, 1996 Telenor
R&D, Norway

Encoding format: QCIF (176x144)

Usage:  mn [options] -i <filename> [more options]
Options:
   -i <filename> original sequence [required parameter]
   -o <filename> reconstructed frames [./out.raw]
   -B <filename> filename for bitstream [./stream.263]
   -a <n> image to start at [0]
   -b <n> image to stop at [0]
   -x <n> (<pels> <lines>) coding format [2]
      n=1: SQCIF n=2: QCIF n=3: CIF n=4: 4CIF n=5: 16CIF
      n=6: Custom (12:11 PAR)
          128x96   176x144    352x288   704x576   1408x1152
          pels x lines
   -s <n> (0..15) integer pel search window [15]
   -q <n> (1..31) quantization parameter QP [13]
   -A <n> (1..31) QP for first frame [13]
   -r <n> target bitrate in bits/s, default is variable bitrate
   -ren <n> target bitrate in bits/s, in the enhancement layer
   -rfi RATE_FILE target bitrates contained in the file RATE_FILE
   -C <n> Rate control method [3]
   -k <n> frames to skip between each encoded frame [2]
   -Z <n> reference frame rate (25 or 30 fps) [30.0]
   -l <n> frames skipped in original compared to reference
      frame rate [0]
   -e <n> original sequence has n bytes header [0]
   -g <n> insert sync after each n GOB (slice) [0]
      zero above means no extra syncs inserted
   -w write difference image to file "./diff.raw" [OFF]
   -m write repeated reconstructed frames to disk [OFF]
   -t write trace to tracefile trace.intra/trace [OFF]
   -j <n> force an Intra MB refresh rate every <n> macroblocks [0]
   -D <n> use unrestricted motion vector mode (annex D) [OFF]
      n=1: H.263 n=2: H.263+ n=3: H.263+ unlimited range
   -E use syntax-based arithmetic coding (annex E) [OFF]
   -F use advanced prediction mode (annex F) [OFF]
   -G use PB-frames (annex G) [OFF]
   -U <n> (0..3) BQUANT parameter [2]
   -M use improved PB-frames (annex M) [OFF]
   -I use advanced intra coding mode (annex I) [OFF]
   -J use deblocking filter (annex J) [OFF]
   -c <n> frames to select number of true B pictures
      between P pictures (annex O) [0]
   -d <n> to set QP for true B pictures (annex O) [13]
   -i <filename> enhancement layer sequence
   -u <n> to select SNR or spatial scalability mode (annex O) [OFF]
      n=1: SNR n=3: SPATIAL(horiz) n=5: SPATIAL(vert) n=7: SPATIAL(both)
   -v <n> to set QP for enhancement layer (annex O) [13]
   -S use alternative inter vlc mode (annex S) [OFF]
   -T use modified quantization mode (annex T) [OFF]
   -p <n> a I picture every n frames
   -CR <n> Conditionnal Replenishment With treshold: n
   -MS Mode and QP Selection (base layer)
   -MSUI Mode and QP Selection (base layer): only INTRA&UNCODED Modes
   -MSEP Mode and QP Selection (enhancement layer)
   -FOR FORWARD Mode forced (with -MSEP)
   -FUP UPWARD Mode forced (with -MSEP)
   -PG parameter P of the Gilbert Model
   -QG parameter Q of the Gilbert Model
   -ARQ <filename>  Prise en compte des infos de Feedback
   -PL <filename>  Sortie des limites de paquets dans
      un fichier ASCII
   -PLEL <filename>  Sortie des limites de paquets SNR_EL
         dans un fichier ASCII
   -RF <filename>  Sortie des debits instantannes utilisés
   -EL <filename>  Sortie de la layer de raffinement SNR
       dans le fichier filename
   -vi Utilisation de l'algo de Viterbi pour l'optimisation R-D
   -L <nb_loops> Bouclage sur la sequence par nb_loops aller-retours
   -sk [filename] Write the number of the skipped frames in the
       file filename (default : skip.txt)
   -aa <filename> Write the results of intra MB of each packet
       in a file [Mos_file]
   -ab n the default packet size in byte <536>
   -h Prints help

    Default filenames and other options in square brackets
    are chosen in config.h

The commandline options of the H.263 decoder, which we used during the subjective video quality test and in the video demonstration application:

Usage:   tmndecode options bitstream outputfilename

Options:
   -vn  verbose output (n: level)
   -on  output format
      n=0 : YUV
      n=1 : SIF
      n=2 : TGA
      n=3 : PPM
      n=4 : X11 Display
      n=5 : YUV concatenated
      You have to choose one output format!
   -q disable warnings to stderr
   -r use double precision reference IDCT
   -t enable low level tracing
   -s <filename>  output reconstructed frame to filename
      (YUV concatenated)
   -p enable tmn-8 post filter
   -c enable error concealment
   -x interpolate pictures to double size before display
   -fn  frame rate
      n=0 : as fast as possible
      n=99: read frame rate from bitstream (default)
   -l loop sequence
   -TR enable Absolute Temporal References trace
   -TREP enable Absolute Temporal References trace
   -E enhancement_file: decode SNR-ENH layer in file
      enhancement_file
   -EF enh_recon_sequence: reconstruct SNR-ENH layer in file
       enh_recon_sequence
   -PL base.lim: specifie le fichier de limites de paquets de
       la couche de base
   -PLEL enh.lim: fichier de limites de paquets de la couche
         de raffinement
   -A arq.txt: fichier de numeros de paquets perdus
      (couche de base)
   -AE enh_arq.txt: fichier de numeros de paquets perdus
       (couche de raffienement)
  -AM <Mos_file> is a must file name that contains the
      #intra and inter MBs in each packet
  -Znn % loss rate <00>
  -Ynn no. of consecutively lost packets <1>