I'm also not sure about this:
    if (mpegts_packetizer_has_packets (base->packetizer)) {
      while (1) {
        /* Eat up all packets */
        pret = mpegts_packetizer_process_next_packet (base->packetizer);
        if (pret == PACKET_NEED_MORE)
          break;
        if (pret != PACKET_BAD &&
            mpegts_packetizer_get_seen_pcr (base->packetizer) >
            initial_pcr_seen) {
          GST_DEBUG ("Got last PCR");
          done = TRUE;
>>        break;
        }
      }
    }

Why is the break there? This will find the first PCR in the block, but that is not necessarily the last PCR in file.

I pushed a refactoring of pcr/offset handling. Please try again, the results are quite different now.

The reason why it breaks is to limit amount of data to scan (and get results faster). With the new pcr/offset handling it ends up giving a proper duration.

Created attachment 280211 [details] [review]
mpegtbase: Improve last PCR detection

When dealing with random-access content (such as files), we initially
search for the last PCR in order to figure out duration and to handle
other position estimation such as those used in seeking.

Previously, the code looking for that last PCR would search in the last
640kB of the file going forward, and stop at the first PCR encountered.
The problem with that was two-fold:
* It wouldn't really be the last PCR (it would be the first one within
  those last 640kB. In case of VBR files, this would put off duration
  and seek code slightly.
* It would fail on files with bitrates higher than 52Mbit/s (not common)

Instead this patch modifies that code by:
* Scanning over the last 2048kB (allows to cope with streams up to 160Mbit/s)
* Starts by the end of the file, going over chunks of 300 MPEG-TS packets
* Doesn't stop at the first PCR detected in a chunk, but instead records all
  of them, and only stop searching if there was "at least" one PCR within
  that chunk

This should improve duration reporting and seeking operations on VBR files

commit 56ae254dd31f6c1ca468aa43bb3262d2fc84422e
Author: Edward Hervey <bilboed@bilboed.com>
Date:   Wed Jul 9 10:11:40 2014 +0200

    mpegtbase: Improve last PCR detection
    
    When dealing with random-access content (such as files), we initially
    search for the last PCR in order to figure out duration and to handle
    other position estimation such as those used in seeking.
    
    Previously, the code looking for that last PCR would search in the last
    640kB of the file going forward, and stop at the first PCR encountered.
    The problem with that was two-fold:
    * It wouldn't really be the last PCR (it would be the first one within
      those last 640kB. In case of VBR files, this would put off duration
      and seek code slightly.
    * It would fail on files with bitrates higher than 52Mbit/s (not common)
    
    Instead this patch modifies that code by:
    * Scanning over the last 2048kB (allows to cope with streams up to 160Mbit/s)
    * Starts by the end of the file, going over chunks of 300 MPEG-TS packets
    * Doesn't stop at the first PCR detected in a chunk, but instead records all
      of them, and only stop searching if there was "at least" one PCR within
      that chunk
    
    This should improve duration reporting and seeking operations on VBR files
    
    https://bugzilla.gnome.org/show_bug.cgi?id=708532