On Aug 30, 1:40*pm, "Jake" <m...@privacy.com> wrote:
> "HEMI-Powered" <n...@none.gn> wrote in message
>
> news:Xns9B0A903621069ReplyScoreID@216.196.97.136.. .
>
>
>
>
>
>
>
> >>> I'm wondering about the contrast ratio spec of new LCD
> >>> monitors. *It varies between 1,000 and 3,000 in many cases,
> >>> but with the screens all lined up and displaying images in
> >>> the store I don't see much difference. Any opinions about how
> >>> much difference this makes in the real world, and how you can
> >>> actually assess the difference?
> >> If you were looking at a low key images side by side you may
> >> notice a slight difference between a monitor with a low
> >> contrast ratio and one with a high one, but to be honest
> >> 1000:1 isn't low at all.
> > Please see my reply to the OP first, but I would disagree that
> > 1000:1 isn't pretty low. But, unless one is into technical lab
> > testing to glean some scientific comparisons, it is vitally
> > important to view real world images and not just the canned
> > displays in the stores. In my shopping this spring I found the
> > 1000:1 displays were the lowest end monitors with the least sharp
> > display while the 5000:1 were barely better than the 3000:1 but
> > usually 2X-3X the price.
>
> > In the end, I think this is a highly subjective issue.
>
> I'm not talking about sharpness though, I'm talking about contrast
> graduations.
Keep in mind the human eye can only perceive a 2% contrast ratio
change.
Check out Sony's 1000000:1 OLED screen and you will see the difference.
Toby
"Victek" <Victek@invalid.invalid> wrote in message
news:g9bpvo$f01$1@registered.motzarella.org...
> I'm wondering about the contrast ratio spec of new LCD monitors. It
> varies between 1,000 and 3,000 in many cases, but with the screens all
> lined up and displaying images in the store I don't see much difference.
> Any opinions about how much difference this makes in the real world, and
> how you can actually assess the difference?
>>> I'm wondering about the contrast ratio spec of new LCD
>>> monitors. It varies between 1,000 and 3,000 in many cases,
>>> but with the screens all lined up and displaying images in
>>> the store I don't see much difference. Any opinions about how
>>> much difference this makes in the real world, and how you can
>>> actually assess the difference?
>>
>> If you were looking at a low key images side by side you may
>> notice a slight difference between a monitor with a low
>> contrast ratio and one with a high one, but to be honest
>> 1000:1 isn't low at all.
>>
> Please see my reply to the OP first, but I would disagree that
> 1000:1 isn't pretty low. But, unless one is into technical lab
> testing to glean some scientific comparisons, it is vitally
> important to view real world images and not just the canned
> displays in the stores. In my shopping this spring I found the
> 1000:1 displays were the lowest end monitors with the least sharp
> display while the 5000:1 were barely better than the 3000:1 but
> usually 2X-3X the price.
>
> In the end, I think this is a highly subjective issue.
>
It occurred to me to check out the specs of the 19" flat panel I'm currently
using. It's a couple of years old - contrast ratio is 500:1 and response
time is 16ms. These specs are poor by today's standards yet this monitor is
perfectly adequate for me. It's a good example of how specs can suggest
differences that don't matter in the real world (at least not to everyone
under all circumstances).
>You anticipated and answered my next question about DVI. My video card
>(Nvidia 6600GT) has both VGA and DVI connectors. I'm currently using the
>VGA connector, but the new screen I picked up has both so I will invest in a
>DVI cable. I don't know if the Nvidia card has the built-in monitor
>profiling you mentioned (?) I've read that using DVI eliminates screen
>centering issues - is that true?
The main point of a DVI connection is it delivers fully digital data
from the graphics card directly to the controller in the monitor,
eliminating several sources of noise and error.
With an analog connection, the digital RGB values stored in the graphics
card are converted to analog RGB values via a digital to analog
converter, with one pixel following another at a rate determined by an
internal oscillator in the graphics card. These signals, plus a small
amount of noise and hum picked up along the way, are fed to the analog
input of the monitor.
This circuitry has the job of guessing what resolution the graphics
card is working at by looking at the sync frequencies. Then it tries to
generate a local pixel clock oscillator which is as close as possible to
being in sync with the one in the graphics card. An analog to digital
converter converts the voltages in the input signal back into RGB
values, and those are stored in the monitor's local RAM which is used to
update the individual pixels in the display panel.
In a perfect world, the data in the monitor would be a perfect
reconstruction of what was in the graphics card, but in the real world
there are black level and gain errors, noise, and clock rate mismatches.
The DVI connection avoids all the analog stages, sending clock and
RGB intensity data in digital form across the cable.
>I wonder if someone who knows about densities and prints might care to
>comment on this. I'm too rusty! Is a 1000:1 range the same as a density
>range of 3.0?
Yes, it ought to be, since density is base-10 logarithm. On the other
hand, you need to check what you're comparing. Some monitors are
specifying a "dynamic contrast" rating that includes dimming the
backlight in dark scenes. What you want to know is the largest contrast
available *within one image at the same time*. Even then, this depends
on the test pattern.
>What is the best that prints can achieve? I would have
>expected less than 3.0, and if so presumably the blacks are compressed a
>little to fit the available dynamic range of the print? I remember
>sometimes having to do dodging and burning to produce the "best"
>black-and-white print.
If I remember correctly, the density range of good B&W paper developed
to have good blacks, and with a glossy surface, is about 100:1, maybe
200:1 in the very best cases. Colour prints are more like 30:1.
You can choose paper contrast during printing to make tradeoffs between
reproducing a relatively small scene brightness range (5-6 stops) with
normal contrast, or a wider scene range with lower than normal contrast.
Dodging and burning in provide local exposure changes to "push"
highlights and shadows towards midtone. You can even use multiple
exposures on vari-contrast paper in combination with
masking/dodging/burning to print different areas with different
contrast.
>I have been disappointed with the LINEAR dynamic range of typical (i.e.
>inexpensive) LCD monitors - there is very little display of the lower
>blacks so that, if you have an image of 0..255, all levels below 8 may
>appear as one black. To get the best visible greyscale reproduction, I
>have found that you typically need to decrease the contrast (gain)
>setting, and increased the brightness (offset) setting, so that the blacks
>are actually shown as dark greys. This then works well with colour
>images. Note that I print very little, so matching monitor to printer
>isn't an issue for me.
That might be deliberate on the part of the LCD makers. Pulling the
darkest tones to blacks causes the rest of the tonal scale to have
higher contrast and more saturated colours. Less accurate, but a lot of
people like the effect.
Dave Martindale wrote:
> "David J Taylor"
> <david-taylor@blueyonder.neither-this-bit.nor-this-part.co.uk>
> writes:
>
>> I wonder if someone who knows about densities and prints might care
>> to comment on this. I'm too rusty! Is a 1000:1 range the same as a
>> density range of 3.0?
>
> Yes, it ought to be, since density is base-10 logarithm. On the other
> hand, you need to check what you're comparing. Some monitors are
> specifying a "dynamic contrast" rating that includes dimming the
> backlight in dark scenes. What you want to know is the largest
> contrast
> available *within one image at the same time*. Even then, this
> depends
> on the test pattern.
Agreed.
> If I remember correctly, the density range of good B&W paper developed
> to have good blacks, and with a glossy surface, is about 100:1, maybe
> 200:1 in the very best cases. Colour prints are more like 30:1.
>
> You can choose paper contrast during printing to make tradeoffs
> between
> reproducing a relatively small scene brightness range (5-6 stops) with
> normal contrast, or a wider scene range with lower than normal
> contrast.
> Dodging and burning in provide local exposure changes to "push"
> highlights and shadows towards midtone. You can even use multiple
> exposures on vari-contrast paper in combination with
> masking/dodging/burning to print different areas with different
> contrast.
Thanks for the reminder!
>> I have been disappointed with the LINEAR dynamic range of typical
>> (i.e. inexpensive) LCD monitors - there is very little display of
>> the lower blacks so that, if you have an image of 0..255, all levels
>> below 8 may appear as one black. To get the best visible greyscale
>> reproduction, I have found that you typically need to decrease the
>> contrast (gain) setting, and increased the brightness (offset)
>> setting, so that the blacks are actually shown as dark greys. This
>> then works well with colour images. Note that I print very little,
>> so matching monitor to printer isn't an issue for me.
>
> That might be deliberate on the part of the LCD makers. Pulling the
> darkest tones to blacks causes the rest of the tonal scale to have
> higher contrast and more saturated colours. Less accurate, but a lot
> of
> people like the effect.
>
> Dave
Thanks for your reply, Dave. Perhaps there is also some problem about
reproducing dark very greys accurately, so the monitors are designed not
to reproduce them at all? I would much prefer accuracy to the garish
results sometimes seen. From what I hear, I may be in the market for the
OLED displays....
On Sep 1, 10:45*pm, da...@cs.ubc.ca (Dave Martindale) wrote:
> "David J Taylor" <david-tay...@blueyonder.neither-this-bit.nor-this-part.co.uk> writes:
>
> If I remember correctly, the density range of good B&W paper developed
> to have good blacks, and with a glossy surface, is about 100:1, maybe
> 200:1 in the very best cases. *Colour prints are more like 30:1.
>
> You can choose paper contrast during printing to make tradeoffs between
> reproducing a relatively small scene brightness range (5-6 stops) with
> normal contrast, or a wider scene range with lower than normal contrast.
> Dodging and burning in provide local exposure changes to "push"
> highlights and shadows towards midtone. *You can even use multiple
> exposures on vari-contrast paper in combination with
> masking/dodging/burning to print different areas with different
> contrast.
>
> * * * * Dave
The figures I had seen said 2% for the blackest black on fully exposed
printing paper. This corresponds well to about the same value for 3M
Black Velvet paint. I believe inks from printing press are a bit more
reflective.
I worked on a project developing blacks for sunshades on satellite
optical instruments. It was hard to find ANYTHING blacker than 1%.
The best blacks we had were between 1 and 2%.
Don Stauffer in Minnesota wrote:
[]
> The figures I had seen said 2% for the blackest black on fully exposed
> printing paper. This corresponds well to about the same value for 3M
> Black Velvet paint. I believe inks from printing press are a bit more
> reflective.
>
> I worked on a project developing blacks for sunshades on satellite
> optical instruments. It was hard to find ANYTHING blacker than 1%.
> The best blacks we had were between 1 and 2%.
Thanks, Don. I recall that for TV test targets they once used recessed
boxes (sort of a light sink) lined with black baize to try and get a very
low reflectivity value. However, all this doesn't explain why LCD
monitors, with claimed contrast ratios well in excess of 100:1, can
produce such lousy dark greyscales!
Don Stauffer in Minnesota <stauffer@usfamily.net> writes:
>The figures I had seen said 2% for the blackest black on fully exposed
>printing paper. This corresponds well to about the same value for 3M
>Black Velvet paint. I believe inks from printing press are a bit more
>reflective.
>I worked on a project developing blacks for sunshades on satellite
>optical instruments. It was hard to find ANYTHING blacker than 1%.
>The best blacks we had were between 1 and 2%.
In the case of the printing paper, I don't believe that the contrast
figures include specularly reflected light. After all, the paper
gelatin surface will reflect something like 4% of the incoming light.
But glossy paper, properly dried on a polished metal plate, gives a nice
specular reflection and you naturally hold the print *to direct that
reflection away from your eyes*. So the black you see is determined by
the *diffuse* reflectance of the paper with the specular reflection
removed.
On the other hand, a paint rated as 2% reflectance is likely talking
about total reflected light, including any specular component.
Exercise: if you want a small area of really black stuff, get a package
(or 10) of old-fashioned double-edge razor blades. Assemble all of the
blades into a single stack and clamp them together. Now look at the
two "faces" of the resulting structure that are a whole bunch of cutting
edges viewed edge-on - these faces will be very black. Even though the
cutting edges are actually shiny steel, they form a series of deep
V-grooves with such shallow angles that any light which reaches the face
gets bounced back and forth many times and eventually absorbed.
HEMI-Powered wrote:
> Victek added these comments in the current discussion du jour
> ...
>
<snip>
>>
> I'd make one observation here having purchased a 26" Samsung back
> in May. It seems that the popular trend in all LCD displays is to
> vastly over-saturate and over-brighten the display. As best I've
> been able to learn this is because the public thinks this makes
> the image jump off the screen better giving the illusion of
> greater dyamic range and greater sharpness. Well, maybe, but
Seconded!
Comparing LCD screen in the shop with their canned images I fell exactly
into that trap, only to find out the Samsung widescreen I bought
couldn't hold a candle to my old Philips in terms of smooth grayscale
rendering, no matter how I tweaked it. Swapped the Samsung for a cheaper
Philips widescreen which was tweakable...
<snip>
>
> Another tip: be sure to buy only from a store that will give you
> a 100% refund with no restocking fee. You see, the features and
yes!
And also: do not put much value in reviews from PC journals, they cater
to people with different needs, like gaming.
Re: Contrast ratio of LCD monitors 
Linear Mode
