The limit to brightness

R. Kennedy McEwen wrote:

need for highlights. Gamma attempts to even out the quantisation, so
approximately the same precision is used to quantify discernable
differences in shadows as highlights.

Gamma 2.2 significantly over-corrects and results in shadows having
limited perceptual resolution. The gamma used by the Mac is closer to
the human eye's response.

True sRGB actually goes linear for the darkest shadows.

Gamma 2.2 is dictated by the empirical behaviour of CRTs and was
presumably chosen to make TV receiver hardware cheap (and similar for
IBM PC hardware).

Graham Murray wrote:

from black to white with a contrast between each section. I wonder how
many people now take the trouble to adjust a new TV set - first with the

Probably only slightly fewer than did so for analogue, i.e. not many at all!

"Steve Thackery" wrote in message
...
Mortimer wrote:

For me, the big problem with modern TVs (and this includes CRTs) is that
they cannot handle highlights well, and suffer horrendous crushing of
anything over a certain threshold. At best, when all three colours are
maxed-out, you just get featureless white.

Yep, good point, and there seem to be two potential causes. Firstly, the
screen itself may not have a linear response to brightness, such that
luminance values higher than a certain level tend to "crush".

But I think the more likely scenario is that the picture data itself too
often "tops out" at 256, 256, 256 RGB (so to speak - I don't know the
actual dynamic range of a TV picture).

If the latter is the case the screen itself is probably behaving correctly
and displaying "fully white" when it is told to, and the problem is
earlier in the chain.

I wonder whereabouts in the chain this "topping out" occurs. My suspicion
is at the broadcast end. But it is also possible that the bit-depth of
the video circuitry in the TV, or of the LCD screen itself, is less than
the bit-depth transmitted, in which case you'd get that same visible
problem.

We need someone in the industry to tell us which it is.

Of course, if the problem is at the broadcast end of the chain (as I
suspect) we would see exactly the same effect on a CRT TV. Perhaps the
problem isn't to do with LCD vs. CRTs, but really about analogue vs.
digital transmissions.

The fact that this "crushing" problem seems to occur only occasionally
makes me think it's due to crap cameras or broadcast equipment.

It's more noticeable on documentaries such as house makeover or
house-hunting where they are probably using low-budget (and physically
smaller) cameras to keep production costs down and make the cameras less
intrusive, but I've also seen it on some dramas with high production values
(properly graded pictures, etc).

The fact that it is more apparent with modern TVs than older ones, for the
same broadcast, makes me think that it's partly due to the electronics and
screen in the TV.

In article , David Woolley
writes
R. Kennedy McEwen wrote:

need for highlights. Gamma attempts to even out the quantisation, so
approximately the same precision is used to quantify discernable
differences in shadows as highlights.

Gamma 2.2 significantly over-corrects and results in shadows having
limited perceptual resolution. The gamma used by the Mac is closer to
the human eye's response.

The human eye response is only one factor which determines the ideal
gamma, and the 1.8 gamma of the Mac is no better, given the other
factors, than 2.2.

True sRGB actually goes linear for the darkest shadows.

Correct, which is why I said "attempts" and "approximately". However,
the reason for the linear segment is to reduce the quantisation in
linear space to practical limits.

Gamma 2.2 is dictated by the empirical behaviour of CRTs and was
presumably chosen to make TV receiver hardware cheap (and similar for
IBM PC hardware).
Which made the choice of 1.8 for the original CRT based Macs even more
complex, requiring another (undocumented!) layer of LUT to achieve
linear response.

However, as I said, this is merely an approximation and colour
management/calibration adds another layer of correction to the whole
process.
--
Kennedy

Graham Murray wrote:
Paul writes:

On Thu, 30 Aug 2012 19:19:14 +0100, Steve wrote:

Contrast is another issue, though. Cheap LCD TVs still seem to suffer
from glowing blacks, greatly reducing the apparent contrast ratio.

But the manufacturers tell you they've got a million to one contrast
ratio on their screens, so you must be wrong and it's your eyes that
are at fault.

Or they could just be making it up as they go along.

Maybe that is the reason that the TV stations no longer broadcast the
testcard - so that people will not discover that it is impossible to
correctly adjust the set so that the greyscale bars correctly progress
from black to white with a contrast between each section. I wonder how
many people now take the trouble to adjust a new TV set - first with the
saturation right down (ie monochrome display) and iteratively adjust
brightness and contrast to get black black and white white with the
correct greyscale progression, and only what that is correct turn up the
saturation to make the colour bars and the skin tone look correct.

That is indeed the correct technique - for analogue CRTs :-)

BugBear

Steve Thackery wrote:

good stuff

Thank you for a detailed and informative post.

BugBear

Room lighting?
I turn lighting in the room down, my wife turns room lighting up to full.
They dim the lights at cinemas and theatres.
Even at a modern church we go to.
Regards
David

R. Kennedy McEwen wrote:

Fascinating post - thanks.

If it was at the broadcast end then all TV's, including CRTs, would
suffer from it at the same time, on the same images. They don't, so it
isn't at broadcast.

That's the only bit I might disagree with. I've certainly noticed the
occasional "topping out" effect on various TVs, including older CRTs.
I've never, though, had the chance to study them in a row to see if the
effect occurs (or doesn't) at the same time on all of them. Have you
done that? Can you confirm that you've seen it occur on some but not
all screens showing the same picture at the same time?

And it is often not associated with full brightness - probably the most
noticeable (because of how we work psychologically) is the familiar
bright orange face, where details are lost.

My TV shows faces beautifully most of the time, but sometimes they cut
to some old material or maybe some kind of Skype link and the reporter
has a topped-out orange face. Surely this is from the broadcast end of
the chain.

Mind you, going back to my first paragraph - we need to be careful
because of course by maladjusting the brightness/contrast/colour
controls you can force your picture to top out, or bottom out, even
with good content. So clearly the effect *can* be more of a problem on
some TVs than others, depending on how they are set up.

In summary, though, I have to say that crushing or topping out is not a
consistent problem on my TV (as judged by my eye). If it were, then
clearly I would agree that the TV was causing it. The fact that it
only occurs sometimes strongly suggests that it has more to do with the
picture *source* in those cases.

But I also agree that you can force it to happen at the TV end by
deliberately maladjusting the brightness/contrast/colour controls.
This shows that the cause may be complex.

--
SteveT

On Thu, 30 Aug 2012 21:27:42 +0100, Graham Murray
wrote:

Maybe that is the reason that the TV stations no longer broadcast the
testcard - so that people will not discover that it is impossible to
correctly adjust the set so that the greyscale bars correctly progress
from black to white with a contrast between each section.

Is that the case? A few years back there was some key sequence to use on BBC
digital text on Freeview to get a test card on the screen.


--

On 31/08/2012 09:41, The Other Mike wrote:

Is that the case? A few years back there was some key sequence to use on BBC
digital text on Freeview to get a test card on the screen.

The MHEG test card is still available on the DTT platform: go to LCN105
(BBC red button channel) and press yellow as soon as the MHEG welcome
page appears. Then go to any other channel, and back to 105. This time
press green at the welcome page and you should see a 'status page'.
Press green for the test card and a rendering of TCW will appear after a
short delay.

As noted here before though, this test card isn't terribly useful for
setting up displays. I find, with a CRT display, that for pictures to
look right the TC has to to be set with the blacks somewhat crushed.

--
Andy