[gstreamer-bugs] [Bug 572854] New: rtph264pay failed to incorporate PPS for sprop-parameters-set

[GStreamer (bugzilla.gnome.org)]

If you have any questions why you received this email, please see the text at
the end of this email. Replies to this email are NOT read, please see the text
at the end of this email. You can add comments to this bug at:
  http://bugzilla.gnome.org/show_bug.cgi?id=572854

  GStreamer | gst-plugins-good | Ver: 0.10.11
           Summary: rtph264pay failed to incorporate PPS for sprop-
                    parameters-set
           Product: GStreamer
           Version: 0.10.11
          Platform: Other
        OS/Version: All
            Status: UNCONFIRMED
          Severity: minor
          Priority: Normal
         Component: gst-plugins-good
        AssignedTo: gstreamer-bugs at lists.sourceforge.net
        ReportedBy: waiming at ailuropoda.net
         QAContact: gstreamer-bugs at lists.sourceforge.net
     GNOME version: Unspecified
   GNOME milestone: Unspecified


Please describe the problem:
When a h264 PPS NAL is sent after a SPS NAL(both NALs are sent as individual
discret buffers), the payloader failed to incorporate the PPS into the caps'
sprop-parameters-set value. This is due to an extra loop inside
gst_rtp_h264_pay_decode_nal (Looping is not necessary because this is already
taken care of by handle_buffer()). The data being passed to
gst_rtp_h264_pay_decode_nal is already one complete NAL unit. All that is
necessary is to test the first byte for a NAL type 7 or 8 to identify SPS/PPS
NALs.

Steps to reproduce:
1. Send SPS, PPS and h264 coded slice NAL units as individual buffers
2. Observe GST_DEBUG traces. The PPS is never parsed.
3. Get the negotiated caps. the sprop-parameters-set does not contain the PPS
part (no comma (,) separating two base64 strings. Only one base64 string that
is the SPS).


Actual results:
Please refer to the GST_DEBUG traces emitted or do a
gst_pad_get_negotiated_caps() on the payloader pad after sending both SPS and
PPS.

Expected results:
The sprop-parameters-set value in the caps should contain 2 base64 strings,
separated by a comma (,)

Does this happen every time?
Yes

Other information:
The patch for the bug is to replace gst_rtp_h264_pay_decode_nal() completely by
the function below:

static void
gst_rtp_h264_pay_decode_nal (GstRtpH264Pay * payloader,
    guint8 * data, guint size, gboolean * updated)
{
  guint8 *sps = NULL, *pps = NULL;
  guint sps_len = 0, pps_len = 0;
  guint8 header, type;
  guint len;

  /* default is no update */
  *updated = FALSE;

  GST_DEBUG ("NAL payload len=%u", size);

  len = size;
  header = data[0];
  type = header & 0x1f;

  /* keep sps & pps separately so that we can update either one 
   * independently */
  if (SPS_TYPE_ID == type) {
    /* encode the entire SPS NAL in base64 */
    GST_DEBUG ("Found SPS %x %x %x Len=%u", (header >> 7),
               (header >> 5) & 3, type, len);

    sps = data;
    sps_len = len;
  } else if (PPS_TYPE_ID == type) {
    /* encoder the entire PPS NAL in base64 */
    GST_DEBUG ("Found PPS %x %x %x Len = %u",
               (header >> 7), (header >> 5) & 3, type, len);

    pps = data;
    pps_len = len;
  } else {
    GST_DEBUG ("NAL: %x %x %x Len = %u", (header >> 7),
               (header >> 5) & 3, type, len);
  }


  /* If we encountered an SPS and/or a PPS, check if it's the
   * same as the one we have. If not, update our version and
   * set *updated to TRUE
   */
  if (sps_len > 0) {
    if ((payloader->sps_len != sps_len)
        || !is_nal_equal (payloader->sps, sps, sps_len)) {
      payloader->profile = (sps[1] << 16) + (sps[2] << 8) + sps[3];

      GST_DEBUG ("Profile level IDC = %06x", payloader->profile);

      if (payloader->sps_len)
        g_free (payloader->sps);

      payloader->sps = sps_len ? g_new (guint8, sps_len) : NULL;
      memcpy (payloader->sps, sps, sps_len);
      payloader->sps_len = sps_len;
      *updated = TRUE;
    }
  }

  if (pps_len > 0) {
    if ((payloader->pps_len != pps_len)
        || !is_nal_equal (payloader->pps, pps, pps_len)) {
      if (payloader->pps_len)
        g_free (payloader->pps);

      payloader->pps = pps_len ? g_new (guint8, pps_len) : NULL;
      memcpy (payloader->pps, pps, pps_len);
      payloader->pps_len = pps_len;
      *updated = TRUE;
    }
  }
}


-- 
See http://bugzilla.gnome.org/page.cgi?id=email.html for more info about why you received
this email, why you can't respond via email, how to stop receiving
emails (or reduce the number you receive), and how to contact someone
if you are having problems with the system.

You can add comments to this bug at http://bugzilla.gnome.org/show_bug.cgi?id=572854.