[Mp4-tech] Microsoft MPEG-4 Reference Decoder with packet loss

[Steve Wright]

Hi,
MPEG-4 provides the tools for error resilience but the designer has to
implement/use the tools. Maybe the MS Encoder/Decoder Reference Software
does not include error resilience functionality?
Here at IndigoVision our hardware MPEG-4 Codec IP implements Video packets
with HEC and synchronization markers in the encoder hardware. In the decoder
we have error resync and concealment in hardware and data partitioning and
RVLC's in software. This means our MPEG-4 Codec IP can operate over lossy
networks where packet loss is encountered with minimum degradation of video
quality to the user.
The start of a frame in MPEG-4 is marked by the VOPStartCode which has a 32
bit value of 0001b6.
Hope that helps.
Steve Wright
ASIC Manager
IndigoVision Limited
The Edinburgh Technopole
Bush Loan, Edinburgh
Scotland, UK
EH26 0PJ
+44 (0)131 475 7346
+44 (0) 7780 633101 (mobile)
+44 (0)131 475 7201 (Fax)
http://www.indigovision.com/site/sections/products/mainstream.asp
-----Original Message-----
From: mp4-tech-bounces lists.mpegif.org
[mailto:mp4-tech-bounces lists.mpegif.org]On Behalf Of Stefan A. Goor
Sent: 06 January 2024 18:31
To: mp4-tech lists.mpegif.org
Subject: [Mp4-tech] Microsoft MPEG-4 Reference Decoder with packet loss
Hey All,
Hopefully someone can help me with this.
I have used the MPEG4IP tools to create hinted mp4 files that contain mpeg4
bitstreams created with the MS Encoder Reference Software and I have placed
these files on the Darwin Streaming Server to allow RTSP access.
When I write the payload of the packets received by the RTSP client to a
file, I can decode it with the MS Reference Decoder.  However if I drop any
packet, the decoder stops with a dialog box on the screen that says "Debug
Error!... ...abnormal program termination..." and the decoder prints the
following message in the dos window:
	Assertion failed: iMbType >= 0 && iMbType <= 3, file
d:\ms_om\sys\decoder\mbheaddec.cpp, line 476
>From what I understand, the MPEG-4 decoder should be able to handle losses
and corruption (isn't one of the main goals of MPEG-4 to be highly error
resiliant?).
Could anyone explain what the problem may be and how I may get around it?
I would like to be able to decode content has been sent over a network and
that may have approximately 0.5% packet loss.
Also, I assume that each packet or at least an integer number of packets
corresponds to a single frame, is this true?
>From RFC 1889 for RTP, I see that there is a marker bit that generally
indicates frame boundaries.  In MPEG-1 and 2 there was a specific code that
determined the start of a frame, is there something similar for mpeg4 video
stream?
Any help with these problems and questions would be greatly apreciated,
Many thanks,
Stefan
_______________________________________________
Mp4-tech mailing list
Mp4-tech lists.mpegif.org
http://lists.mpegif.org/mailman/listinfo/mp4-tech
Note: Conduct on the mailing list is subject to the Antitrust guidelines
found at
http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php