[M4IF Technotes] packet_header macroblock number

[Oztan Harmanci]

packet_header macroblock numberHello Ariel. I am not 100% sure but that is not supposed to happen for stream decoding. Check two things:
- are there any packet drops? ( I don't think this is the case because 5 MBs in a packet is pretty short . But still, depends on what you are doing.)
- are you counting the macroblocks in received packets correctly?
If there are not any packet losses you have a problem with your decoder. Potential problem points are "mcbpc"s, "cbpy"s and maybe "not_coded" fields. Make sure that they are used correctly.
Oztan
  ----- Original Message ----- 
  From: Richard Shields 
  To: technotes   lists.m4if.org 
  Sent: Thursday, January 30, 2024 12:10 PM
  Subject: RE: [M4IF Technotes] packet_header macroblock number
  Ariel,
  If you are receiving your packets through a protocol such as RTP then packet loss can occur.  So, yes, the situation you have run into can occur and the decoder must know how to handle it.  There are various methods of error concealment (handling the missing macroblocks) from just copying the collocated macroblocks from the reference frame or by using some type of interpolation from the surrounding macroblocks.  The handling of missing macroblocks is not covered by the specification.
  -Richard
  -----Original Message-----
  From: Ariel David [mailto:Ariel.David   ParthusCeva.com] 
  Sent: Thursday, January 30, 2024 7:48 AM
  To: technotes   lists.m4if.org
  Subject: [M4IF Technotes] packet_header macroblock number
  Hi,
  I've run across a problem where I decode a stream and get to a resync marker.
  I decode the packet_header and in it decode the macroblock number of the start of the next packet.
  The problem is that this macroblock number is not identical to my own macroblock count.
  This means that in my decoder I am supposed to be for example in macroblock 75 and the macroblock number in the packet_header shows 80.
  I now have 5 macroblocks that are unacounted for.
  I've looked through the standard and see no mention of such an even and how to resolve it.
  The main problem is that I don't know if this is a bitstream error or something that can possibly happen and I need to handle.
  Is this occurance at all possible?  Is this an error?  Should I decode these macroblocks as skipped?
  P.S  I hope I reached the correct address.  I am subscribed to technotes, but saw no address to send questions.
  Thanks,
  Ariel
-------------- next part --------------
An HTML attachment was scrubbed...
URL: /pipermail/mp4-tech/attachments/20030130/596ff5d3/attachment.html