How do people choose a new TV?

Steve Thackery wrote:

wrote:

There's no prediction required when creating
intermediate frames between the 25 that arrive every second.

You've got the start point, and you know what the end point is because
it's the next frame, sitting in your buffer. You "just" have to come up
with three intermediate frames. I imagine the process would involve
splitting the picture into "objects" and then performing a kind of
"tweening" process on them. Easy to do with vector graphics, of course;
not so easy with bitmap graphics.

But you have got the motion vectors for the macroblocks (unless it's an
I frame) as clues.

In message , Andy
Burns writes:
J. P. Gilliver (John) wrote:

" writes:

SD signals still often contain junk at
the edges that the creators assume will be lost in the overscan

What sort of junk (apart from DOGs)?

Most often, timecodes at the top (on news channels where they're showing
live feeds), also quite often, when a studio programme includes a
segment from "VT" it is faded up as an almost fullscreen overlay,
leaving a single row of "studio" pixels (or a green row) at the bottom
of the picture. And of course the odd boom mic that nobody thinks is
visible and vignetting in the extreme corners of the picture.

Thanks for all that. I still prefer accurate pixel matching, but I guess
I am unusual in that. (And I find all the things like timecodes
"interesting".)
--
J. P. Gilliver. UMRA: 1960/1985 MB++G.5AL-IS-P--Ch++(p)[email protected]+Sh0!:`)DNAf

Behind every successful organisation stands one person who knows the secret of
how to keep the managers away from anything truly important.

In message
,
"
writes:
[]
It remains on SD because SD pictures are still often imperfect at
their edges. Seeing varying amounts of blanking and crud at the edges
is undesirable.

I think only for old material or material from non-broadcast cameras
video surveillance, special ones used for wildlife etc.), for all of
which I don't mind a bit of a ragged edge - but, again, I'm probably
unusual in that.

As for HD - I don't know.

Cheers,
David.
--
J. P. Gilliver. UMRA: 1960/1985 MB++G.5AL-IS-P--Ch++(p)[email protected]+Sh0!:`)DNAf

Behind every successful organisation stands one person who knows the secret of
how to keep the managers away from anything truly important.

In message , Steve Thackery
writes:
wrote:

But there's still plenty of 50i, and that's more than
smooth enough (except on channel 4 where it's MPEG encoders seem to be
set to drop 50i to 25p whenever they feel like it!).

To be truthful, I haven't taken a note of the transmission format when
I've observed the jumpy motion. Panning, in particular, still causes
clearly visible jumping on a lot of TV content for me. I can only
report what I've seen, of course - I guess we are all different.
However, I seem to be sensitive to it.

As I said in an earlier post, I think it's not so much the display
technology, as fast shutter speeds having more or less killed off
"motion blur". I am aware of multiple images even on (and I _didn't_ say
this earlier) my twentysomething years old CRT telly, and I'm pretty
sure this has increased in recent years.

Also, a refresh rate of 50Hz is a real problem for me (I say "also"
because, of course, you can have a higher refresh rate than frame rate,
like a cinema picture).

I've done some experiments with a couple of CRT computer monitors, and
I find 50Hz and 60Hz refresh extremely unpleasant, 75Hz is quite usable
but I can still see a flicker, and 85Hz appears completely smooth.

You probably view a computer monitor from closer than you do a telly,
though.

As you'd expect, the bigger the screen the worse is the visible effect.

But as I can see very clearly a 50Hz refresh, I would think I can pick
up 50Hz panning jumps. Obviously I haven't done any lab tests on
myself.

Incidentally, I'm prone to migraines (with the visual disturbances),
and I wonder if that might be connected to my seeing flicker very
easily.

SteveT


--
J. P. Gilliver. UMRA: 1960/1985 MB++G.5AL-IS-P--Ch++(p)[email protected]+Sh0!:`)DNAf

Behind every successful organisation stands one person who knows the secret of
how to keep the managers away from anything truly important.

J. P. Gilliver (John) wrote:

" writes:

Seeing varying amounts of blanking and crud at the edges
is undesirable.

I think only for old material or material from non-broadcast cameras
video surveillance, special ones used for wildlife etc.), for all of
which I don't mind a bit of a ragged edge

Oh, you've reminded me of another thing visible without overscan, some
cameras seem to have no chroma on the pixels at the extreme left and
right of the picture, luma only ...

In message
,
"
writes:
[]
Do any current TVs allow you to display the source content "as is"?
i.e. one field flashed briefly once every 1/50th second?

It would have to be not the whole field (or frame) flashed briefly, but
each individual pixel (with perhaps some drag), in sequence.

Even with all fancy interpolation switched off, modern plasmas show
each field more than once in 1/50th second (causing double/triple
images with eye-tracked motion), and LCDs show it for almost the

I didn't know that - they (plasmas) multi-strobe, do they?

entirety of the 1/50th (causing smeared/blured images with eye-track
motion). Hence the frame creation technologies, which mitigate these
effects, but add their own artefacts.

Cheers,
David.

--
J. P. Gilliver. UMRA: 1960/1985 MB++G.5AL-IS-P--Ch++(p)[email protected]+Sh0!:`)DNAf

Behind every successful organisation stands one person who knows the secret of
how to keep the managers away from anything truly important.

Andy Burns wrote:
Steve Thackery wrote:

wrote:

There's no prediction required when creating
intermediate frames between the 25 that arrive every second.

You've got the start point, and you know what the end point is because
it's the next frame, sitting in your buffer. You "just" have to come up
with three intermediate frames. I imagine the process would involve
splitting the picture into "objects" and then performing a kind of
"tweening" process on them. Easy to do with vector graphics, of course;
not so easy with bitmap graphics.

But you have got the motion vectors for the macroblocks (unless it's an
I frame) as clues.

Not on hdmi inputs - my tv will double rate deinterlace or interpolate
when fed already decoded source from a computer.

I read some blurb for an ATI Xillion tv chip some time ago and it
claimed to use enhanced phase plane correlation for interpolation.

Java Jive wrote:

Digital displays such as LCDs and Plasmas are not subject to flicker

With respect, JJ, that's quite wrong. Wide area flicker is a
well-known bane of plasma screens.

It's discussed at length on Wikipedia (can't remember the reference,
now, but I'm sure you could find it).

SteveT

Java Jive wrote:



We've discussed here before that it's only the CRTs that might need
double (100Hz) or triple (150Hz) refresh to reduce flicker, rather as
with movie projectors. Both flicker by the very nature of their mode
of working, and this is more noticeable to some viewers than others.
Digital displays such as LCDs and Plasmas are not subject to flicker,
so don't need 2x or 3x refresh rates, so don't buy one that has this
feature, unless you can turn it off.



and if no-one had ever mentioned 100Hz scanning, the "rip-off" shops
would never be able to sell aerial cables "specially designed for 100Hz TVs"

Mike

Java Jive wrote:

"In principle, digital displays such as LCDs and Plasmas should not be
subject to flicker!"

I agree with you about LCD. I admit to being unclear exactly how the
LCD pixels are driven at the lowest (electronics) level, but they seem
to "latch" the picture, and then update it, rather than having the
picture redrawn each time.

A test I do regularly is to wave my finger vigorously from side to side
(like a heartfelt chastisement) in the viewing path between my eyes and
the screen. Most LCD screens produce no visible strobing whatsoever,
although with one of my PC monitors there is just the faintest
detectable strobing, but you'd hardly notice it.

With CRT screens the strobing is as clear as day, of course.

I haven't "wagged" at all that many plasma screens, but of those I've
tested, they all show a strong strobing.

I know even less about how plasma pixels are driven. Certainly, my
strobing experiments seem to suggest that they don't "latch" the
picture, and thus need redrawing at the refresh rate.

I must state clearly that I can only report my findings. I cannot
offer any intellectually useful content, such as an explanation. :-)

SteveT

PS: This is an old reference, but it does suggest that there is
something inherent in plasmas that makes them prone to flicker:

"This paper describes the occurrence of large area flicker in PDP's.
Flicker is mainly a problem in 50 Hz PDP's, although at high peak
brightness 60 Hz PDP's will also display flicker artifacts. The
description is based on existing theories on flicker that are applied
to the PDP's situation."

http://dx.doi.org/10.1889/1.1831750

I can truthfully say that my LCD monitors show no visible flicker even
at their 60Hz refresh rate, which strongly suggests that they aren't
actually being "refreshed" in the conventional sense, merely updated.
Plasmas do seem to behave more like a CRT.