[opendtv] Re: New DVDs already sparking copy-protection confusion
- From: "Kilroy Hughes" <Kilroy.Hughes@xxxxxxxxxxxxx>
- To: <opendtv@xxxxxxxxxxxxx>
- Date: Fri, 24 Feb 2006 08:47:40 -0800
[JS] So my question is: is doing only low-pass filtering a sufficient=20
implementation of this down-ressing, or is the actual sample-rate
conversion=20
with associated risk of aliasing and rather permanent loss of higher
frequencies mandatory ? What would be an acceptable low-pass filter, if
at all acceptable?
[KH] I don't recall the AACS license details, but I think filtering is
adequate. The resolution limit is 960x540, so that will look better
than 720x480 DVD-V scaled up to HD. (Especially because DVD video has
its vertical resolution reduced below 480 lines for interlaced display.)
[JS] I hope you mean 480p60, i.e. it has been de-interlaced to 480p24
using 3-2 pull-down reversal, and then up-converted to 480p60 again by
using frame=20
repetition in the same 3-2 pull-down sequence.
[KH] Yes, 480P60 on a settop player. Your process description is a
little optimistic though. =20
In a settop player, an unknown mix of 24P and 30i source, 30i
subpictures, menus, etc. is sent from the decoder to a deinterlacer chip
as 30i video, and the chip guesses whether it is seeing repeat fields
from pulldown or just a 30i scene without motion, etc., and does all
kinds of strange things (bob and scale, weave, temporal and spatial
prediction, motion estimation, etc.) to turn 60 fields into 60 frames.
(You know more than I about how that sausage is made.) Since everything
is usually decoded to 30i video, the deinterlacer is left guessing which
fields were created by capture and which were created in the decoder,
and it can't change the 60P output rate anyway.
The PC playback approach with known telecine content is to ignore the
repeat field flags and decode to 480P24, scale, then refresh at some
arbitrary rate; although people in the know use 72P. Field
"deinterlacing" is simple and perfect and no 3:2 motion judder. It is
unfortunate that actual vertical information is typically much less than
480 lines because it has been captured and/or pre-filtered for
interlaced display. A PC decoder typically has to play guessing games
based on the pattern of Top/Bottom repeat field flags to decide whether
it is seeing telecine 24P coded as 30i, or 30P coded as 30i, or format
converted 25i at 30i, or off-speed 30i, or regular 30i, etc. The actual
source format is well hidden by the MPEG 30i stream because video used
to be slaved to CRT refresh rate. The PC decoder has the advantage of
the repeat field flag metadata (lost once it is decoded to a 30i video
signal), and the flexibility to decouple the decoding rate from the
display rate (computer controls the refresh rate, not the monitor).
[JS] Why is that a bigger risk, if it is (slightly) more difficult than=20
stealing the HDMI+HDCP signal, and it adds analog noise and jitter too
?!
[KH] There are analog component HD realtime encoders relatively
available, and neither their inputs nor component outputs are
contractually controlled (by crypto IP, keys, license with dos and
don'ts, etc.). If someone tries to market an HDMI ripping device,
they'll be revoked, busted, etc.
Watermarks are useful for forensics, but not copy protection. They are
usually quickly breakable with easily distributed hacks, and it's a slow
process to replace all playback equipment with versions that screen for
watermarks and block unauthorized copies at playback. It's better if
the marks travel silently, useful if they indicate the general source
(e.g. theatrical, broadcast, disc, etc.), and possible in some cases to
indicate the player, showing, etc. to identify the traitor. But, that's
not copy protection; it's damage control for failed copy protection ...
but also a deterrent (your point).
Kilroy Hughes
Sr. Media Architect
Digital Media Interoperability Team
Microsoft Corporation
-----Original Message-----
From: opendtv-bounce@xxxxxxxxxxxxx [mailto:opendtv-bounce@xxxxxxxxxxxxx]
On Behalf Of Jeroen Stessen
Sent: Monday, February 20, 2006 21:12
To: opendtv@xxxxxxxxxxxxx
Subject: [opendtv] Re: New DVDs already sparking copy-protection
confusion
<Snip>
----------------------------------------------------------------------
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:
