[opendtv] Re: New DVDs already sparking copy-protection confusion
- From: "Kilroy Hughes" <Kilroy.Hughes@xxxxxxxxxxxxx>
- To: <opendtv@xxxxxxxxxxxxx>
- Date: Wed, 1 Mar 2006 21:44:09 -0800
Agree with everything you said Gary, except for terminology. People I
hang out with usually use the term "progressive encoding" to distinguish
between 24P frame encoding syntax (progressive sequence) and 30i field
encoding syntax (with Frame and Field pictures, repeat field flags for
telecine, or reel changes, or speed changes, or special effects, titles,
etc. laid down at 30i) as you described. If you go even further,
interlaced source Field pictures can use frame prediction on some or all
macroblocks and also be "progressive encoded", if you want to define it
that way. =20
Yes, a BD player can pull down 24P to 30i on output (such things aren't
defined or constrained by the disc format spec). =20
Yes, the BD format can play 30i encoded content, and it can use repeat
fields, so you could tunnel 24P source as 30i with 3:2 pull down. It is
likely to be treated as 30i content and not inverse telecined. If you
want 24P out, encode it as 24P.=20
I suppose you might use 30i with 3:2 for inconsistent or off speed 30i
masters of nominal 24P source where the encoder can't perform accurate
inverse telecine. Lots of "shot on film, edited on video" TV used to be
that way, and the original elements and edit decisions often aren't
available (or economical) for a new transfer and edit. Warner used to
have a visual comparator that would flash any frame cadence problems to
an operator who would manually enter timecoded corrections in an EDL
that would override the encoder on the next encoding pass until they had
solid ten field 30i sequences that could be decoded to 24P.
PS. How did Trip Hawkins slip in there and set progressive_sequence=3D
"3DO"?
Funny things happen on carriage returns on this reflector.
Kilroy Hughes
Sr. Media Architect
Digital Media Interoperability Team
Microsoft Corporation
-----Original Message-----
From: opendtv-bounce@xxxxxxxxxxxxx [mailto:opendtv-bounce@xxxxxxxxxxxxx]
On Behalf Of Gary Hughes
Sent: Wednesday, March 01, 2006 09:42
To: opendtv@xxxxxxxxxxxxx
Subject: [opendtv] Re: New DVDs already sparking copy-protection
confusion
At 04:04 PM 2/28/2006 -0800, Kilroy Hughes wrote:
>[KH] Progressive encoding is illegal on DVD-Videos.
That is not quite correct
>Progressive_sequence=3D3D0 is required, and only 30Hz or 25Hz frame =
rates
>allowed. =20
Progressive_sequence=3D0 does not prohibit progressive encoding. It only
says that interlaced encoding is permitted, but not required. The
picture
coding extension specifies how each picture is encoded. In the 30Hz
world, a correctly inverse telecined 24P source results in a sequence of
progressive encoded frames, along with the repeating pattern of
repeat_first_field
and top_field_first flags (also in the PCE). The sequence header will
indicate
a frame rate of 29.97, but each second of bitstream will contain 24
progressive
encoded frames. I think this meets the definition of "progressive
encoding".
>There has been a long battle to change production practices
>and tools to support consistent and invertible telecine so 24P could be
>tunneled through 30i transport streams. Accurate 24P cadence in 30i
>syntax was the exception for most movies for the first few years of
DVD,
>but is now the rule.
In my previous employment I designed the trick mode handling for a VOD
server
that makes use of the I frames as they exist in the original stream (to
avoid
the need for trick mode files). Film content that has not been through
IVTC
can exhibit terrible interlace flicker as you might imagine and I spent
much
of last year dealing with these issues.
Various people maintained that non-video data, such as closed
captioning, could
not be accommodated in IVTC content which is of course bogus.
Regrettably one of
the major content providers made a decision to not use IVTC in any of
their
content, largely because some of their production tools did not handle
the
stream timestamps correctly in IVTC content (if you repeat a field, that
has to
be reflected in the PTS and DTS values). In the end I designed a new
trick
mode sequence that forces the set top decoder to deinterlace by making
both
fields reference the same original field when in pause etc when playing
out
such content.
>BD format now allows 24P encoded Transport Streams. HD format doesn't,
>but it identifies 24P encoded as 30i to allow accurate decoding to 24P.
Which brings us back to the discussion in DVDlist :-)
If I understand correctly (all numbers modulo 1000/1001 as needed)
BD allows 24P content to be encoded at 24 progressive frame/sec with the
sequence header frame rate set to 24 and no use of the 3:2 pulldown
flags,
so the decoder could drive the display at 24 frame/sec.=20
HD DVD requires 24P content to be encoded as 24 progressive frames/sec
with
the sequence header frame rate set to 30 and the picture coding
extension
containing the necessary flags to reconstruct the 3:2 pull down, and
presumably
drives the display at 30 frame/sec.
Some questions...
Can a BD player can also perform the 3:2 pull down on its output if
needed?
Does the BD format also allow for the 24 frames plus rff/tff IVTC
approach,
or was that the gist of your original comment in DVDlist?
Sorry if this seems overly pedantic. MPEG does that to people (this is
your
brain, this is your brain on MPEG)
gary
Gary Hughes (ghughes@xxxxxxxxxxxxxxxxxx)
Stargate Video Systems
----------------------------------------------------------------------
You can UNSUBSCRIBE from the OpenDTV list in two ways:
- Using the UNSUBSCRIBE command in your user configuration settings at
FreeLists.org
- By sending a message to: opendtv-request@xxxxxxxxxxxxx with the word
unsubscribe in the subject line.
Other related posts:
