BBC HD vs ITV1 HD

On Thu, 30 Apr 2009 23:17:49 +0100, "jamie powell"
wrote:


If so, that is what happens in bob deinterlacing.

Jeez! I don't believe this! Read my lips!

NO! NO! NO! NO! NO! NO! NO! NO! NO! NO! NO! NO! NO! NO!

Do you get the message now?

Each field consists of *alternate* lines, which are sent to the
display as a unit!

======================================

Please always reply to news group as the email address in
this post's header does not exist. Alternatively, use the
contact address at:
http://www.macfh.co.uk/JavaJive/JavaJive.html

On Thu, 30 Apr 2009 22:05:25 GMT, Dave Farrance
wrote:

Are you suggesting that in the cheaper flat-panels, each field is weaved
with the previous field like this:

frame1-odd/frame1-even
frame2-odd/frame1-even
frame2-odd/frame2-even
frame3-odd/frame2-even

Yes, that's exactly what appears to happen on my Panasonic LCDs, as
evidenced by the photo on my webpage. Further, as they were top of
the range or not far off at the time of purchase, and I've never seen
any substantial *evidence* (as opposed to mythological hearsay) to the
contrary, I am reasonably convinced that's how the majority of them do
it.

Of course, if someone were actually to come up with evidence to the
contrary, I'd have to, and would have no problem with, changing my
opinion (and my webpage), but it hasn't happened yet. All that ever
seems to happen is that people just repeat the same old myths without
thinking about them analytically.

It would still be subject to mice-teeth
artefacts and wouldn't mix well with the rescaling that's necessary to
display the image on the screen's native pixel resolution. I'd assume
that even the cheapest LCD telly would have selective-blending
deinterlacing as a minimum.

Well, these 'problems' don't appear to be problems on the Pannies, and
they don't appear to be doing any deinterlacing.

======================================

Please always reply to news group as the email address in
this post's header does not exist. Alternatively, use the
contact address at:
http://www.macfh.co.uk/JavaJive/JavaJive.html

I, and I suspect most others, would class that description as being
technically correct, and yours as being misleading.

On Thu, 30 Apr 2009 23:06:19 +0100, "jamie powell"
wrote:

No, you really can deinterlace to 50fps progressive without the need for
frame doubling I promise
Some video encoder packages (eg TMPGEnc Xpress) do misleadingly name this
option "deinterlace - double framerate", though.

======================================

Please always reply to news group as the email address in
this post's header does not exist. Alternatively, use the
contact address at:
http://www.macfh.co.uk/JavaJive/JavaJive.html

"Java Jive" wrote in message
...
On Thu, 30 Apr 2009 23:17:49 +0100, "jamie powell"
wrote:


If so, that is what happens in bob deinterlacing.

Jeez! I don't believe this! Read my lips!

NO! NO! NO! NO! NO! NO! NO! NO! NO! NO! NO! NO! NO! NO!

Do you get the message now?

Each field consists of *alternate* lines, which are sent to the
display as a unit!

And what is the display supposed to put in the space between these
"alternate" lines?.....

Choose the correct one from:
1) Nothing, it continues to display what was previously there,
just as a CRT does.
2) A white elephant
3) A flying pig

On Fri, 1 May 2009 00:03:01 +0100, "jamie powell"
wrote:

And what is the display supposed to put in the space between these
"alternate" lines?.....

======================================

Please always reply to news group as the email address in
this post's header does not exist. Alternatively, use the
contact address at:
http://www.macfh.co.uk/JavaJive/JavaJive.html

"Java Jive" wrote in message
...
On Thu, 30 Apr 2009 22:05:25 GMT, Dave Farrance
wrote:

Are you suggesting that in the cheaper flat-panels, each field is weaved
with the previous field like this:

frame1-odd/frame1-even
frame2-odd/frame1-even
frame2-odd/frame2-even
frame3-odd/frame2-even

Yes, that's exactly what appears to happen on my Panasonic LCDs, as
evidenced by the photo on my webpage. Further, as they were top of
the range or not far off at the time of purchase, and I've never seen
any substantial *evidence* (as opposed to mythological hearsay) to the
contrary, I am reasonably convinced that's how the majority of them do
it.

Of course, if someone were actually to come up with evidence to the
contrary, I'd have to, and would have no problem with, changing my
opinion (and my webpage), but it hasn't happened yet. All that ever
seems to happen is that people just repeat the same old myths without
thinking about them analytically.

Sorry but the photo on your webpage isn't conclusive "evidence" of anything,
so don't accuse others of spouting mythological hearsay when your own
"findings" - which you've even published on the web, as if there were some
authority behind them - amount to nothing more.

Most deinterlacers operate on a 3-field principle.
Where frames are represented by numbers and fields by the letters A and B:

Information from 1A,1B,2A is used to create progressive-frame 1.
Information from 1B,2A,2B is used to create progressive-frame 2.
Information from 2A,2B,3A is used to create progressive-frame 3.
etcetera.

The central field of the three (eg. 1B for progressive frame 1) is used as
the "master field" for the newly-created progressive frame, because they are
both temporally-aligned. Information from the two adjacent fields is used to
provide the extra information required to create a full-height (576 lines
for UK SD) progressive frame.


It would still be subject to mice-teeth
artefacts and wouldn't mix well with the rescaling that's necessary to
display the image on the screen's native pixel resolution. I'd assume
that even the cheapest LCD telly would have selective-blending
deinterlacing as a minimum.

Well, these 'problems' don't appear to be problems on the Pannies, and
they don't appear to be doing any deinterlacing.

If mice teeth aren't visible on interlaced material, then some form of
deinterlacing is *definitely* taking place. It's as simple as that!

"Mike Henry" wrote in message
...

My original statement - an appropriately-simplified description for the
OP -
was not "TECHNICALLY" incorrect.

Yes it was technically incorrect, because you said "the" image. See above
and all the posts by Java Jive elsewhere.

I clarified what I meant by "the image" immediately after typing it.
The discussion between me and Java Jive has no relevance to this issue.

"Java Jive" wrote in message
...

Nothing, it continues to display what was previously there,
just as a CRT does.


This is a key issue in your understanding of how this stuff works:
A conventional CRT *doesn't* continue to display what was previously there -
there is some minimal residual phosphor lag, but field A has all for all
intents and purposes disappeared from a given area of the screen before
field B reaches the area.
That's why you *don't* see so-called "mice teeth" artefacts on a CRT -
adjacent fields are never displayed on-screen at the same time.

Java Jive wrote:
Hard information is hard to come by, but it's debatable whether LCDs
*need* to deinterlace at all, really they only *need* to buffer, and
therefore that's much more likely what the majority do:
http://tinyurl.com/daw2gz
... standing in for ...
http://www.macfh.co.uk/JavaJive/Audi....html#CriticTV

You say in that article that the LCD TV has a vertical resolution the
source so I wouldn't expect you to be able to see interlacing artifacts.

On Fri, 1 May 2009 00:58:05 +0100, "jamie powell"
wrote:


"Java Jive" wrote in message
...

This is a key issue in your understanding of how this stuff works:

.... but even more so yours ...

A conventional CRT *doesn't* continue to display what was previously there -
there is some minimal residual phosphor lag, but field A has all for all
intents and purposes disappeared from a given area of the screen before
field B reaches the area.

No it hasn't.

The term persistence when applied to phosphors has a specific,
mathematical meaning: it's the time taken for the luminescence to
decay to 10% of its starting value. It is NOT the time taken to fade
away to nothing (that would be infinite) NOR even the time taken for
luminescence to become invisible to the eye or any other detector. I
get the impression that many, perhaps including yourself, do not
understand this fundamental point.

Each phosphor dot as it is refreshed momentarily reaches a brightness
many orders of magnitude greater than its brightness during the
greater part of its exponential decay. This extreme brightness if
maintained would probably blind you (and may be related to stories of
eyesight deterioration using CRTs, as in my experience, though I am
not aware of independent evidence proving exactly what the mechanism
of deterioration is). Thus, by the time the next field is being
drawn, the previous one can still be picked up by cameras:

http://en.wikipedia.org/wiki/File:Refresh_scan.jpg

The previous frame is still visible to this camera at only 1/3000s
f/1.6 exposure. Though darker, it is clearly discernible as two
people facing the camera over the top of a counter or table, the top
of the right hand person's head having just been refreshed.

As the human eye is much more sensitive than most cameras, it beggars
belief that this would not also be visible to the naked eye.

======================================

Please always reply to news group as the email address in
this post's header does not exist. Alternatively, use the
contact address at:
http://www.macfh.co.uk/JavaJive/JavaJive.html