What's dynamic contrast anyways?

by Ben Drawbaugh, posted Feb 14th 2008 at 7:50AM

In the world of video quality there aren't many specification more important than the contrast ratio -- if you enjoy movies or TV shows with dark scenes. Unfortunately, this also happens to be one of the most worthless specifications because every manufacture measures it differently. One of the latest tricks manufactures did was to introduce Dynamic Contrast. So instead of measuring the difference between the whitest white and the blackest black displayed on the screen at the same time, they display white and black on the screen one at a time and then measure the difference. Obviously this is next to useless, because this almost never occurs during normal programming -- of course there are exceptions to this, like alpine skiing or ice hockey. But it is the detail in the shadows and the ultimate appearance of black that is appealing in most programming. In the end the only way to compare apple to apples when looking for a new TV is by going to a trusted reviewer who always measure contrast the same way.

Filed under: Industry

Tags: Contrast, Dynamic Contrast, DynamicContrast

Reader Comments (Page 1 of 1)

Neutral

Jon Davis @ Feb 14th 2008 8:32AM

I must respectfully disagree. Many televisions have a tendency of increasing their brightness or gamma according to the total display brightness ratio. In other words, if you're looking at a skiing competition, the brightness or gamma is toned down. And in a scary, dark movie, the brightness or gamma is brought up.

The contrast ratio, measured correctly, will not allow this brightness factor to kick in. A fair measurement would indeed involve a screen that is 50% gray, with a black spot and a white spot, with two measurements, each with the white and black on switched sides. What is the difference between the two spots, is what I want to know. That in turn will measure the actual contrast ratio. Contrast ratio measured in this way ultimately MEANS IMAGE DETAIL.

Simply making the screen go black and measuring that, then making the screen go white and measuring that, will not be a fair contrast ratio measurement.

Now it sounds like you came across a manufacturer who was making a shift towards honesty. This should be cause for celebration, not contempt.

Highly Ranked

Ben @ Feb 14th 2008 8:32AM

What exactly do you disagree with? It seems we are in agreement. I'm confused.

Neutral

Darkest Daze @ Feb 14th 2008 9:00AM

Yea, that whole comment boils down to one word: oxymoron.

Neutral

Vidikron (FU) @ Feb 14th 2008 3:33PM

@Ben

The problem is in your wording. Read this sentence:

" So instead of measuring the difference between the whitest white and the blackest black displayed on the screen at the same time, they display white and black on the screen at one at a time and then measure the difference. "

Look closely at the second part. You have an extra "at" in there. I first read it as "they display white and black on the screen at one time and then measure the difference" the first time I read it too. I think my mind automatically took a screwed up sentence and made it make sense even though the outcome wasn't the intended one. He obviously read it the same way I did.

Neutral

Ben @ Feb 14th 2008 3:34PM

You're right, I didn't get it right the second time either. Thanks for the correction.

Neutral

Jon Davis @ Feb 14th 2008 8:34AM

Hm. Judging from your picture, I can see that they have a black screen with white spots. Nevermind; that is arguably not a fair measurement.

Highly Ranked

psp&dsFanboy @ Feb 14th 2008 9:05AM

The only time the dynamic contrast setting looks good on my LG TV is when I'm view a music channel always had a black background with mostly white text. Otherwise, LG's dynamic contrast sucks at displaying shadow details - which is why I leave that setting off.

Neutral

HD Guy @ Feb 14th 2008 9:33AM

The problem here is that Ben wrote "So instead of measuring the difference between the whitest white and the blackest black displayed on the screen at the same time, they display white and black on the screen at once and then measure the difference." That's tantamount to saying "so instead of doing X, they do X." What he meant to say was "So instead of measuring the difference between the whitest white and the blackest black displayed on the screen at the same time, they display a white screen and then a separate black screen and then measure the difference." Jon Davis is correct that the only way to properly measure contrast is for a monitor to display black, white and neutral on the screen at the same time. Of course, that the set also needs to be properly calibrated first goes without saying.

Neutral

Ben @ Feb 14th 2008 9:33AM

Duh, not sure how I manged to muck that up and not noticed it even though I re-read it so many times.

I've updated it, thanks for the correction.

Neutral

Charles @ Feb 14th 2008 9:34AM

So instead of measuring the difference between the whitest white and the blackest black displayed on the screen at the same time, they display white and black on the screen at once and then measure the difference.

This sounds like the same thing to me, i.e
at the same time vs at once.

Neutral

steedums @ Feb 14th 2008 9:51AM

dynamic contrast is annoying.

why would you want the screen getting brighter and then lighter. it drives me crazy

Highly Ranked

nathan @ Feb 14th 2008 10:09AM

It strikes me that there are at least three kinds of contrast, all of which tell one something useful. But, even taken together, those three don't tell the whole story.

1. Full on/full off, without dynamic contrast engaged. Often called "Native contrast", this is the difference between a set displaying an all black screen and all white screen.

2. Full on/full off, with dynamic contrast engaged (if applicable). This number is often 10x the native contrast. Is it gamed and unrealistic? Sometimes. My old Sanyo projector looked ridiculous with dynamic contrast engaged. It was like a guy was sitting there turning up and down the buld brightness as scenes changes. On the other hand, the Sony SXRD projectors use a similar technique, and it's at once imperceptible and improves the picture.

3. Mixed-image contrast -- sometimes called ANSI contrast -- typically measured by having black-and-white checkerboard boxes on the screen and measuring the difference between the heart of a black square and the heart of a white square. This test most closely approximates real world video content, but doesn't tell the whole story. Old CRT's often had poorer ansi contrast than current PDP displays, but still had more "pop" in their picture.

I think one needs to know all three, and as Ben notes, they need to be measured in the same way on different sets, to get an idea of image quality. But without even getting into the real of color balance, image processing, resolution, etc, there is another closely related "spec" that really does require a trusted reviewer: Shadow detail. Sets can measure very well in all the above contrast measures, but still lack good detail in the shadows.

Neutral

kcjones @ Feb 14th 2008 10:24AM

What Ben described was a "Full On/ Full Off" contrast ratio test, where a small window box of white is displayed and measured. For black the TV is unplugged and the non-powered screen is measured. "Dynamic Contrast" is a spec typically used with LCDs that underpower the backlight during dark scenes, unfortunately this excessively darkens the small areas of a screen that are happen to be bright. Both techniques are completely worthless only results in consumer confusion. A typical cinema only gives a 500:1 contrast ratio when measured using the ANSI technique. No HDTV on the market today comes close to that of your local Cinema.

Neutral

Bob Mc @ Feb 14th 2008 10:48AM

Was the brightness all the way up on the old Kuro?

Neutral

jrepetti @ Feb 14th 2008 11:09AM

My thoughts exactly. I have a 6010 and I've never seen a picture that bad looking before.

Neutral

Ben @ Feb 14th 2008 11:10AM

jrepetti,
I also have a PDP-6010FD and I actually saw the display and I'm telling you it is how it looks.

It was absolutely amazing, you have to see it to believe it really.

Neutral

Ben @ Feb 14th 2008 10:53AM

Guys I have a kuro and yes, that is what mine looks like on the left.

If you want to reproduce it on your TV at home do this. Wait till it's good and dark outside, turn off all the lights in your house and let you eyes adjust to lack of lighting in the room. Now put on some content with lots of dark scenes and look at your TV, you will see it glows just like the picture on the left. The new kuro concept truly is amazing and really is that much better than the old one.

Neutral

Bob Mc @ Feb 14th 2008 11:54AM

Thanks, Ben, for clearing that up. I've seen that TV many times and I really want one (unfortunately, my wife doesn't and she's in charge.) I've just never seen a Kuro (or any recent plasma) with blacks that poor. I just figured they adjusted it to make the new one look better.

Neutral

Kevin @ Feb 14th 2008 5:34PM

Dynamic Contrast is probably more comfortable in ways, like to your eye, you don't get the backgrond lightning glare of the screen and probably reduces headaches... :)! That what I guess... :)!

Neutral

Allen @ Feb 14th 2008 6:05PM

The best way to measure contrast, I think, is to measure how dark any given black can be next to the brightest white a TV can reproduce. If it can get super white with the brightness on max and everything set up, great. I bet it will reproduce dark black with the brightness and contrast all the way down to.

The best way to actually get GOOD COLOR, though, is to run a THX DVD through your screen after 100hours of the screen constantly being on. The video optimizer will do wonders. Of course, the absolute best way is to set the TV's color to 6500K and then run an ISF optimizing sequence. But not a lot of us have 4 hours and $5000 to spend on making our color as great as a TV can reproduce, so meh.

Neutral

Gary @ Apr 29th 2008 9:55PM

I sometimes wonder if it wouldn't be a good idea if someone were to use a spectrometer and photomultiplier to actually MEASURE the light output from any given television versus the actual video drive voltage above black level (in the analog domain, of course). A television with really good ooverall video contrast and low harmonic distortion should produce nearly a straight line relationship between video voltage and light output and that linear response should be at the all of the emission wavelengths of the given display.

The point at which linearity of this response begins to compress is in fact the contrast dynamic range on the hign end (whiter than white) and when noise begins to dominate the measurement on the black end, you have reached the black level. The ratio of the two voltage readings between the two linearly related drive settings is the contrast ratio, expressed as a ratio of voltages, at any given color (or summation thereof).

Each color has its own relationship between light output and video drive voltage. If these slopes are not the same for each primary, the set cannot possibly track in gray scale. If gray scale does not track, it is not possible for colors to track at differing levels.

Liverpool wants to build 'games academy' Rob Thomas knows digital marketing, wants in on Rock Band Bethesda's classic The Elder Scrolls II: Daggerfall goes free	Twitter statistics that make you go "hmmmm" Hot new Mozilla browser released for ... Mac OS 9? Tumblr introduces Submissions for community-powered blogs
First Look: Make cool slide shows with Pix Remix Lifted tutorial code winds up in $0.99 App Store tennis game Loss of Wi-Fi connectivity plagues the iPhone 3G as well	Options Update: Arcelor Mittal volatility low at 69 on banker leverage discussions Before the bell: Futures point to lower start ahead of consumer confidence data Warren Buffett, tells us more!
X3F TV -- XBLA in Brief: Battlefield 1943 Ghostbusters demo busts out on XBLM, PSN Ion Assault will assault XBLA this summer	AutoblogGreen for 07.10.09 VIDEO: 2010 Jaguar XJ in motion, interior and interviews + expanded gallery and pricing! Spy Shots: Mini Cooper Crossover spotted in California
get.game vel.shock() Video reveals first gameplay footage from MechWarrior Arkane Studios working on parts of BioShock 2	Ellis: Ware Hid From Coaches to Help Me Coaches, Teammates Remember McNair Talk About Overpaid: Mike Brown Paid More Than Dan Rooney