http://common.ziffdavisinternet.com/util_get_image/17/0,1425,sz=1&i=177634,00.jpg

Dolby HDR promotional video (http://www.dolby.com/consumer/home_entertainment/hdr-video-demo.html)

http://www.dolby.com/uploadedImages/English_(US)/Professional/Video/Technical_Overview/dynamic-brightness-range-big.png

http://www.dolby.com/uploadedImages/English_(US)/Professional/Video/Technical_Overview/lcd-performance-comparison-big.png

http://www.dolby.com/uploadedImages/English_(US)/Professional/Video/Technical_Overview/dolby-hdr-comparison-big.png

It's interesting that Dolby is offering their local dimming technology in two flavors. I wonder what exactly distinguishes Dolby Contrast from Dolby Vision apart from higher brightness in Dolby Vision. Can't wait for more reports from CEATEC about this.

It's nice to see Dolby push advanced algorithms for LED LCDs. Most of the problems with 81 series could have been fixed with proper dimming algorithm. I hope that even 'Dolby Contrast' provides good greyscales and great colors. 'Dolby Vision' might offer more bit depth, dynamic range and ldr2hdr implementration.

It's nice to see Dolby push advanced algorithms for LED LCDs. Most of the problems with 81 series could have been fixed with proper dimming algorithm. I hope that even 'Dolby Contrast' provides good greyscales and great colors. 'Dolby Vision' might offer more bit depth, dynamic range and ldr2hdr implementration.It's the first time that Dolby shows this technology after BrightSide acquisition this year and it'll be interesting to see how Dolby Contrast which is probably the "affordable" local dimming flavor compares to what Samsung did with the "affordable" 81 Series. I really hope the reports will indicate that Dolby's implementation is superior to what is currently available which will prove that imperfections in 81 Series should be attributed to Samsung's implementation rather than to local dimming technology itself.

A set with Dolby Vision should certainly be capable of displaying HDR sources and, like you, I can't wait to sample Dolby's ldr2hdr algorithm. Sounds like things are moving in the right direction. Now, let's hope that display manufactures license these technologies ASAP.

http://translate.google.com/translate?u=http%3A%2F%2Fwww.watch.impress.co.jp%2Fav%2Fdocs %2F20071003%2Fceatec10.htm&langpair=ja%7Cen&hl=en&safe=off&ie=UTF-8&oe=UTF-8&prev=%2Flanguage_tools
http://www.watch.impress.co.jp/av/docs/20071003/cea13_04.jpg
http://www.watch.impress.co.jp/av/docs/20071003/cea13_07.jpg

http://www.news.com/8301-10784_3-9789531-7.html?tag=nefd.blgs
http://www.dolby.com/consumer/home_entertainment/video.aspx

http://www.computerpoweruser.com/editorial/article.asp?article=articles%2Farchive%2Fc0712%2F29c12%2F29c 12.asp&guid=

http://www.customretailer.net/story/story.bsp?sid=83181&var=story
The 37-inch prototype, featuring Dolby Contrast and Dolby Vision, was demonstrated side by side with a competitive LCD TV. According to Bharath Rajagopalan, Dolby’s business line director/image technologies, Dolby Vision technology improves both dynamic range and contrast ratio by controlling individual LEDs, while Dolby Contrast technology controls the LEDs in clusters, so that very high contrast ratios are possible, if a panel vendor so desires.

Giorgio Corazza, president of SIM2 USA, said his company was aiming at introducing a 46-inch panel as its initial offering and would likely show the technology at the 2008 CES, with an actual product introduction targeted for sometime after the Fall 2008 CEDIA Expo.

http://www.twice.com/article/CA6506581.html?desc=topstory
http://www.reuters.com/article/pressRelease/idUS70275+29-Nov-2007+BW20071129

So basically SIM2 is doing prototype Dolby Contrast&Vision displays in collabaration with Dolby. In addition SIM2 will provide Dolby manufacturing reference designs of these displays. This means that Dolby can license display manufacturers complete lcd designs that ascertain certain level of performance. :cool:

http://www.soundandvisionmag.com/features/2653/dynamic-range-rover.html

Gaven Wang, a senior product manager at Dolby, calls the [HDR] effect head-spinning. "It's almost real-life. Instead of looking at a flat picture, you feel like you’re watching something real through a window." (Yeah, we've heard the window analogy a million times before too, but still . . . cool!)

[John] Lowry is equally enthused. "The introduction of HDR sets is going to be a bigger step up from today's LCDs and plasmas than when we went from standard CRTs to HDTVs. The first time I saw that BrightSide screen — oh my heavens. It was startlingly good. The image just jumped out."

Should you begin saving your pennies? Probably not, unless you’re hell-bent on being among the earliest adopters (expect Dolby’s first HDR sets to be pricey, though the company declined to give us exact figures). Richard MacKellar, BrightSide's former CEO, believes that TV manufacturers will stick to the tried and true business model of introducing a marginally better product once or twice annually.

“They could leap to an HDR display today,” says MacKellar, "but they will make more profit by gradually improving the dynamic range by a few percent each year. In ten years the dynamic range will be vastly better than it is today, and I believe screens will be able to compete with sunlight." And all the while, prices will inch down.

http://www.engadgethd.com/2008/01/05/sim2-dolby-aiming-to-deliver-hdr-prototype-display-in-q1/

Video (http://www.youtube.com/watch?v=3nOigdmhc3o)
http://www.pocket-lint.co.uk/news/news.phtml/12151/13175/dolby-led-display-prototype-unveilled.phtml
http://www.trustedreviews.com/tvs/news/2008/01/09/CES-2008-Dolby-Brands-Brightside-Tech/p1
http://www.pcmag.com/article2/0,2704,2245940,00.asp

http://www.blogsmithmedia.com/www.engadgethd.com/media/2008/03/3-31-08-dolby_sim2_hdr.jpg
http://www.businesswire.com/portal/site/google/?ndmViewId=news_view&newsId=20080331006633&newsLang=en
NEW YORK--(BUSINESS WIRE)--Dolby Laboratories, Inc. and SIM2 Multimedia are proud to showcase a new prototype high-dynamic-range (HDR)–enabled LCD flat-screen display using Dolby’s new light emitting diodes (LED) local dimming technology.

Dolby’s HDR technologies utilize the capabilities of LED-based backlight units (BLUs) to provide outstanding contrast combined with crisp brightness to deliver picture quality that matches real-world visual perception of depth, detail, and color.

“This prototype is a stunning example of what Dolby’s HDR technologies can bring to the LCD market,” said Bharath Rajagopalan, Senior Marketing Manager, HDR, Dolby Laboratories. “We welcome the opportunity to showcase our innovative technology that delivers a dramatic improvement to LCD displays.”

As part of the collaboration effort, SIM2 designed and developed the BLU, which drives the electronics of the LCD display plus the BLU and BLU thermal management system. In addition to the prototype development, SIM2 will provide Dolby with manufacturing reference designs.

“This prototype exemplifies SIM2’s ability to incorporate breakthrough technology into its electronics and optical design, providing—once again—a stunning display that will be a key reference for next-generation LCD displays,” said Domenico Toffoli, Director, Professional Systems Business Unit, SIM2. “We are proud to be the first to support the development of a technology of this stature.”

The prototype is a 46-inch LCD with Dolby’s HDR LED backlight technology:

* 1,838 LEDs
* LCD panel providing a resolution of 1920 × 1080, fitting a native 16:9 aspect ratio
* Brightness: greater than 4,000 cd/m2
* Contrast ratio: Infinite! Full On/Full Off
* Step size: 16 bits of luminance
* Luminance uniformity: more than 95% through the LCD panel

Incorporated into the design is a Xilinx Virtex™ field programmable gate arrays (FPGAs) chipset, recently approved by Dolby for HDR innovations. The Xilinx FPGA chipset enables the implementation of Dolby’s complex high-dynamic-range algorithms in the SIM2 display.

This prototype sports a contemporary and stylish design, befitting the most diverse of home interiors; it is a new masterpiece by the award-winning designer Giorgio Revoldini.

Dolby and SIM2 are showcasing the prototype in New York City today and will be sharing its latest LCD display innovation in cities around the world.

Dang, you're a quick one vtms. I was just about to post that update here.

That design is quite ugly, but I'm still looking forward to the technology. I'm hoping we can get some critical views on how this prototype makes use of the tech.

Though I'm actually surprised they went all the way with this prototype, I know that Dolby Vision only requires about 2,500 cd/m2 - but I'm glad they went for the full 4,000.

http://displaydaily.com/2008/04/02/dolbysim2-show-off-46-hdr-lcd-tv/
http://www.hometheatermag.com/news/040108dolby/
http://eetimes.com/news/latest/showArticle.jhtml?articleID=207000950
http://www.bigpicturebigsound.com/dolby-sim2-hdr-lcd-hdtv-1485.shtml
http://www.twice.com/article/CA6546557.html
http://dvice.com/archives/2008/04/need_more_contr.php

Highlights:
- 1838 independently controlled zones (1 LED = 1 zone) [to put that into perspective, Samsung 81 has only 64 and upcoming LG75 has 128 zones]
- All white LEDs now (so only 92% NTSC), RGB LEDs later.
- 46" is probably the smallest size that will be sold.
- The design will change
- "In a demo reminiscent of the Pioneer Kuro concept at CES, SIM2's new LCD prototype produced inky blacks which made objects seem to float in space"
- Power consumption: 2000W (ouch!)

Dang, you're a quick one vtms. I was just about to post that update here.

That design is quite ugly, but I'm still looking forward to the technology. I'm hoping we can get some critical views on how this prototype makes use of the tech.

Though I'm actually surprised they went all the way with this prototype, I know that Dolby Vision only requires about 2,500 cd/m2 - but I'm glad they went for the full 4,000.
Not sure why they decided to go that high on the brightness. You're right, the bezel looks ugly but that will most likely change. I'm curious how much these are going to cost; hopefully much less than $50K for the previous 37" Brightside model. If it's under $10K, then maybe someone will buy it.

Not sure why they decided to go that high on the brightness. You're right, the bezel looks ugly but that will most likely change. I'm curious how much these are going to cost; hopefully much less than $50K for the previous 37" Brightside model. If it's under $10K, then maybe someone will buy it.
If it's over $15000 and still uses white LEDs then they're crazy.

It's a prototype so it doesn't have a price and it is designed to demonstrate the technology, not impress with a designer bezel. It makes sense to introduce the technology with white LEDs. If they added tri-color LEDs to the picture that might significantly increase the price and complexity. Once the technology has a foothold, tri-color LEDs or more LEDs will be the next logical step along with lowering power and cooling requirements.

Once the technology has a foothold, tri-color LEDs or more LEDs will be the next logical step along with lowering power and cooling requirements.
Well that's basically exactly what Dolby have said on the matter. The only fly in their ointment is that Sony is coming out with RGB LED based units this year, so they won't be able to claim complete technological superiority.

Dang, you're a quick one vtms. I was just about to post that update here.

That design is quite ugly, but I'm still looking forward to the technology. I'm hoping we can get some critical views on how this prototype makes use of the tech.

Though I'm actually surprised they went all the way with this prototype, I know that Dolby Vision only requires about 2,500 cd/m2 - but I'm glad they went for the full 4,000.
Having seen it, I think it was overkill (the brightness that is). There were people literally holding their hands up to shield their eyes at times.

And while the demo was "reminiscent" of the Kuro demo, this is not to say that it was as impressive as the Kuro concept plasma overall. The blacks were much better (than standard LCD) and uniformity was improved over traditional LCDs, but 92% NTSC gamut does leave you with under-saturated colors. And there was still some motion smear.

It's a huge step forward though, and here's hoping they're not the only ones to implement HDR.

Later,

-CB

http://www.electricpig.co.uk/2008/11/28/first-dolby-hdr-tv-scheduled-for-early-2009/
Dolby has been working with Italian projector company Sim2 on a prototype TV that uses individual LEDs behind the screen, rather than a uniform CCFL panel to give massive contrast ratios on screen at the same time (50,000:1), but Dolby’s director of sales for Europe and the Middle East, Andy Dowell, told us a consumer version will be out by “Q1 2009″ - before April, folks.

Dowell couldn’t go into any further details, other than to proudly boast that “it’ll be a very sexy product”. Sim2 specialises in high end stuff, so don’t expect an affordable price tag just yet, though Dowell did say he expected “more mainstream brands coming on soon.”

http://www.blogcdn.com/www.engadgethd.com/media/2009/02/solar_s_front1_600_020309.jpg

New SIM2 Solar Series With Dolby Vision LED Backlight Technology Scheduled to Be Available Q2 2009

AMSTERDAM, Netherlands, Feb 03, 2009 (BUSINESS WIRE) -- SIM2 Multimedia and Dolby Laboratories, Inc. are proud to unveil the latest in high-dynamic-range (HDR)-enabled LCD flat-screen display technology featuring Dolby(R) Vision during the 2009 Integrated Systems Europe (ISE) Show, February 3-5 in Amsterdam. SIM2 plans to make the new SIM2 Solar Series available in the second quarter of calendar 2009.

SIM2's Solar Series includes a 47-inch LCD display utilizing Dolby Vision technology. Dolby Vision features a proprietary algorithm that manages LEDs behind the liquid crystal panel. Each LED is controlled individually in concert with the image on display. By selectively turning off the backlight behind black areas in scenes, Dolby Vision allows those areas to become truly black. Dolby Vision also has the ability to selectively brighten the backlight behind bright areas, allowing them to truly radiate. SIM2's Solar Series delivers outstanding contrast combined with crisp brightness for a picture quality that matches real-world visual perception of depth, detail, and color.

http://investor.dolby.com/releasedetail.cfm?ReleaseID=363298

Nobody will be able to afford it, but at least it will give consumers an idea about how wide is a quality gap between current, half-baked LED LCDs with LD and the real deal IMLED HDR LCD.

Two comments:

1. Dial that beast down to 400 nits and it won't look noticeably better than the current Sony/Samsung in virtually any viewing condition.

2. Black bezel inside black bezel inside faux-silver bezel with visible(!) seams - that is one unattractive display!

3. The Engadget commentator had it about right - at full brightness that display will be like staring at a 200W light bulb. Hopefully they have good lawyers to write the warning labels. :)

Ok, I guess that was three comments.

I'm quietly curious how much it's going to cost. Unfortunately the small size and use of white LEDs are already a couple of black marks against it before it even gets out of the gate.

1. Dial that beast down to 400 nits and it won't look noticeably better than the current Sony/Samsung in virtually any viewing condition.

SIM2 will still have large static contrast advantage because of ~2000 dimming zones. S-IPS panel should also provide better viewing angles. SIM2 promises also almost perfect brightness uniformity.

Two comments:

1. Dial that beast down to 400 nits and it won't look noticeably better than the current Sony/Samsung in virtually any viewing condition.



Dial it down to zero and it will be identical. Not a great argument :)

Dial it down to zero and it will be identical. Not a great argument :)
The black and white ends are irrelivant anyway, it's the stuff in between that makes the picture. Samsung have 6 bit (64 step) backlight control and the Sim2 has 16 bit (65,536 step) control. That, combined with the IMLEDs and superior video processor will make a difference.

But so will the small size, white LEDs and high price. If the Sim2 comes in at more than the 65XS1 does, then I really don't see the market embracing it with open arms.

The black and white ends are irrelivant anyway, it's the stuff in between that makes the picture. That's funny. I just read a paper by dolby labs that suggests this very thing. They want to redefine what dynamic range is. They want to change it from a max/min ratio to "Number of distinguishable gray levels" which is a count of visible luminence steps in between the min max.

SIM2 will still have large static contrast advantage because of ~2000 dimming zones. S-IPS panel should also provide better viewing angles. SIM2 promises also almost perfect brightness uniformity.The extra dimming zones sure do make a difference with regards to halo artifacts. However, what really makes the difference IMO is the use of pixel compensation and LED compensation. This should somewhat stabalize (but not perfect) the floating blacks/shifting contrast associated with local dimming and also greatly help with the black crush associated with local dimming.

That's funny. I just read a paper by dolby labs that suggests this very thing. They want to redefine what dynamic range is. They want to change it from a max/min ratio to "Number of distinguishable gray levels" which is a count of visible luminence steps in between the min max.
I think it would be a fantastic achievement if they did that. I got my first computer display with 8 bit video in 1991 and here we are in 2009 still with 8 bit video, it blows my mind. Then again, we still have 16 bit audio, too.

The range of possible modulation in the backlight is directly proportional to the peak brightness - the dimmer or more "standard" you set the peak, the less it can spatially modulate the backlight unless severe compromises in image quality are accepted (eg, extreme intra-frame floating whites). This is a function of the optics and independent of the number of individually controllable zones.

At "standard" brightnesses, the halos should be more or less the same as with the latest Sony/Samsung models - the increase in control density is offset by the increase in effective cluster size necessary to get light where it needs to go. At high brightness the halos won't matter much, of course, as you're more or less intentionally blinding your eye.

The display will almost certainly look good, as do the current Sony/Sammies, with the added kicker of high brightness. Two-thumbs-up of greatness on the brightness, as it should look terrific in ambient/daylight-like viewing conditions.

XRox - Sony, Samsung and Sharp all have published patents on per-pixel LCD compensation with backlights of this kind, and there is no reason to believe their current offerings don't implement the technology.

The range of possible modulation in the backlight is directly proportional to the peak brightness - the dimmer or more "standard" you set the peak, the less it can spatially modulate the backlight unless severe compromises in image quality are accepted (eg, extreme intra-frame floating whites). This is a function of the optics and independent of the number of individually controllable zones.
I don't think I follow you. The number of modulation steps is determined by the bit depth of the BLU controller, and the number of steps is the same irrespective of whether the peak brightness is 100cdm2 or ten trillion cdm2.

I don't see much point in going beyond 10bit video.

http://i41.tinypic.com/2886rlz.png

XRox - Sony, Samsung and Sharp all have published patents on per-pixel LCD compensation with backlights of this kind, and there is no reason to believe their current offerings don't implement the technology.I have Samsung's SID paper on this technique. There is no indication that they have implemented it into their products. In fact, the 9 series models still had the brutal black crush problem and the only fix that Samsung offered was to increase the overall black level in order to get light into the dark areas of the diffuser.

Even so, LCD compensation is simplistic and cannot improve static contrast uniformity unless LED compensation is also used. To my knowledge, only Dolby Vision employs both (Dolby Contrast is only LCD compensation). Of course the increased density of LEDs in Dolby vision is what makes it possible.

I don't think I follow you. The number of modulation steps is determined by the bit depth of the BLU controller, and the number of steps is the same irrespective of whether the peak brightness is 100cdm2 or ten trillion cdm2.

The number of steps available is determined by the controller; the number of steps that actually produce visibly perceptible differences is a function of both peak brightness and the BLU modulation depth allowed by the algorithm. And the modulation depth allowed by the algorithm is itself proportional to the peak brightness.

I have Samsung's SID paper on this technique. There is no indication that they have implemented it into their products.

I have 100% confidence the current Sony, Samsung, and Sharp TVs with spatial modulation of the BLU use the whole suite of compensation techniques. That problems persist with both black crush and intra-frame floating whites should IMO be taken as a sign of the intransigence of the problem rather than as an indication of an incomplete implementation.

They're all terrific TVs, and I'm sure the Sim2 will also look wonderful in many applications as it has the added benefit of very high brightness.

Onwards and upwards!

I have 100% confidence the current Sony, Samsung, and Sharp TVs with spatial modulation of the BLU use the whole suite of compensation techniques. That problems persist with both black crush and intra-frame floating whites should IMO be taken as a sign of the intransigence of the problem rather than as an indication of an incomplete implementation.

They're all terrific TVs, and I'm sure the Sim2 will also look wonderful in many applications as it has the added benefit of very high brightness.

Onwards and upwards!Well, you said there is no reason to believe that these companies have not yet implemented compensation. I gave you two good reasons. The most important being they have not stated so. The second being the main benefit of LCD compensation (improvement of black crush) is not seen in with the current implementation.(I'm speaking solely of Samsung here since I've not demoed the Sony or Sharp)

If you have evidence they are actually in a product rather than just a patent please post as I'd love to learn more.

As for LED compensation (Dolby Vision), the main benefit is to somewhat stabalize the static contrast across the screen and limit crosstalk. Logically speaking, I don't see how this can be accomplished with so few zones in the current implementations. And even with thousands of zones, the inherent weakness of a backlight and diffuser combination will never be ideal compared to self emitting pixels.

Please visit this thread for more info on the subject.
http://www.avsforum.com/avs-vb/showthread.php?t=947924&highlight=local+dimming

number of steps that actually produce visibly perceptible differences is a function of both peak brightness and the BLU modulation depth.
Or in simpler terms; the granularity and dynamic range. The capability of the human vision system in both respects is regularly underrated, which is why you still get people coming out of the woodwork claiming that 8 bit video is "good enough" when it isn't.


Well, you said there is no reason to believe that these companies have not yet implemented compensation. I gave you two good reasons. The most important being they have not stated so. The second being the main benefit of LCD compensation (improvement of black crush) is not seen in with the current implementation.(I'm speaking solely of Samsung here since I've not demoed the Sony or Sharp)

If you have evidence they are actually in a product rather than just a patent please post as I'd love to learn more.

As for LED compensation (Dolby Vision), the main benefit is to somewhat stabalize the static contrast across the screen and limit crosstalk. Logically speaking, I don't see how this can be accomplished with so few zones in the current implementations. And even with thousands of zones, the inherent weakness of a backlight and diffuser combination will never be ideal compared to self emitting pixels.

Please visit this thread for more info on the subject.
http://www.avsforum.com/avs-vb/showthread.php?t=947924&highlight=local+dimming
On top of the "do they, don't they?" question, there is also the matter of how good these algorithms are. Just like the dreadful quality of the motion compensation algorithms used in Sony and Samsung TVs, it's quite possible for them to have a pixel compensation algorithm that is completely ineffectual if their accountants decided that doing it properly would demand too much cost in silicon. Just because a company has a patent for a really good implementation of a technology doesn't mean that commercial products will do it that way.

The black and white ends are irrelivant anyway, it's the stuff in between that makes the picture. Samsung have 6 bit (64 step) backlight control and the Sim2 has 16 bit (65,536 step) control. That, combined with the IMLEDs and superior video processor will make a difference. Absolutely. Display's ability to show all the shade gradations is the most important determinant of PQ. I'd take a 16-bit set with poor blacks over 8-bit set with zero blacks any time.

But so will the small size, white LEDs and high price. If the Sim2 comes in at more than the 65XS1 does, then I really don't see the market embracing it with open arms.There will be a professional version of this set available in Q2 and consumer version in September. I don't see any way either set will cost less than $15K. But you know what? The HDR picture will probably compel a lot of rich folks to buy the set. After all, it's been said that going from HD to HDR is like going from SD to HD and there will probably be a lot of people who'll want to be the first to own first ever HDR tv.

Absolutely. Display's ability to show all the shade gradations is the most important determinant of PQ. I'd take a 16-bit set with poor blacks over 8-bit set with zero blacks any time.I personally don't believe you would. Or at least I believe that your definition of poor black level is actually not that poor.

In a low APL scene or even worse, a low APL entire movie, a high black level can destroy PQ irrespective of distinguishable gray levels. Think about it........

After all, it's been said that going from HD to HDR is like going from SD to HD
Yes, with HDR content.

And that's quite a big caveat right now.

Yes, with HDR content. No, actually with LDR HD content shown on HDR set. Brightside created algorithms that extend dynamic range of LDR content. A year or two ago I posted links to demonstration video of the technology and research papers done on people looking at HDR picture (on a Brightside set) that seem to support my SD->HD like HD->HDR analogy.

I personally don't believe you would. Or at least I believe that your definition of poor black level is actually not that poor.

In a low APL scene or even worse, a low APL entire movie, a high black level can destroy PQ irrespective of distinguishable gray levels. Think about it........Yes, of course, there's some black level above which PQ would be unacceptable, especially during dark scenes. I guess current plasma black levels would be ok. Unfortunately, most people focus just on black levels thinking it's the key to great PQ while it's the ability to display all the shades that they should be after. Lowest luminance level is important only to the extent that it gives a set potential or access to display even more shades on the low end. A tv can have lowest luminance level of 0cd/m2 and still fail to generate better picture than a tv with higher black levels, yet better shade gradation control.

No, actually with LDR HD content shown on HDR set. Brightside created algorithms that extend dynamic range of LDR content. A year or two ago I posted links to demonstration video of the technology and research papers done on people looking at HDR picture (on a Brightside set) that seem to support my SD->HD like HD->HDR analogy.
You can't pull dynamic range out of thin air any more than you can pull resolution out of thin air. The best that an HDR display can do with normal content is to present 100% of the dynamic range in the source material.

As far as I'm aware, Brightside presented their displays with HDR content. They also displayed their screen far too bright for the sake of fooling people's eyes.

A tv can have lowest luminance level of 0cd/m2 and still fail to generate better picture than a tv with higher black levels, yet better shade gradation control.
Or a TV could have better picture quality than a display with both lower blacks and more shades of grey. There are so many variables and tradeoffs involved in video displays that claiming one particular aspect of picture quality is The One will only lead you down a dead end path.

You can't pull dynamic range out of thin air any more than you can pull resolution out of thin air. No, resolution is different. One can't fake resolution but one can fake dynamic range because we know how light scatters (and can't know what pixels are missing). Or, at least, one can fake higher dynamic range more convincingly than fake higher resolution.

There are so many variables and tradeoffs involved in video displays that claiming one particular aspect of picture quality is The One will only lead you down a dead end path.I said "most important", not "the only one that's important".

No, resolution is different. One can't fake resolution but one can fake dynamic range because we know how light scatters (and can't know what pixels are missing). Or, at least, one can fake higher dynamic range more convincingly than fake higher resolution.
I don't agree. The algorithms may be different, but you're still talking about interpolation of dummy data based on clues in the source material.

I said "most important", not "the only one that's important".
Even so, I regard a holistic approach as the better one. It's very easy for R&D dollars that are spent in a too targetted manor to end up offering little in the way of returns.

You can't pull dynamic range out of thin air any more than you can pull resolution out of thin air. The best that an HDR display can do with normal content is to present 100% of the dynamic range in the source material.


I don't agree. The algorithms may be different, but you're still talking about interpolation of dummy data based on clues in the source material.


Of course you can pull dynamic range out of thin air. There is no interpolation of dummy data, to stretch a low range image into high range. There are just quantization effects. However quantization effects in light levels are much less noticable than q effects in color range.

Example: Take an underexposed digipic (indoors, lower light, no flash). Bring up in any editor. Look at histogram. All bunched up on the left side. Using maybe 20% of the range available. Then have the editor do a spread (usually known as level adjust, or Auto light enhance, etc.). Takes the brightest of the original image and assigns it to the max avail. Spreads the rest of the light levels evenly along the full dynamic range. Of course there are unused light levels in this new full range (i.e. the histogram is a series of spikes, instead of a smooth curve, that a well light picture would have). However, the picture now has a larger dynamic range, and is quite viewable.

Of course there are other, more intricate algorithims, then just a linear spread, but you get the idea.

Of course you can pull dynamic range out of thin air. There is no interpolation of dummy data, to stretch a low range image into high range. There are just quantization effects. However quantization effects in light levels are much less noticable than q effects in color range.

Example: Take an underexposed digipic (indoors, lower light, no flash). Bring up in any editor. Look at histogram. All bunched up on the left side. Using maybe 20% of the range available. Then have the editor do a spread (usually known as level adjust, or Auto light enhance, etc.). Takes the brightest of the original image and assigns it to the max avail. Spreads the rest of the light levels evenly along the full dynamic range. Of course there are unused light levels in this new full range (i.e. the histogram is a series of spikes, instead of a smooth curve, that a well light picture would have). However, the picture now has a larger dynamic range, and is quite viewable.

Of course there are other, more intricate algorithims, then just a linear spread, but you get the idea.
How does video scaling work? I think you'll find it quite close to what you just described.

http://www.cs.ubc.ca/~heidrich/Projects/Ldr2Hdr/

don't confuse taking image date and blowing it up to a higher "resolution" with dynamic range expansion it s totally different.

If you look at a regular tv the brightest image it can display is nowhere near the brightest you eyes can see but it looks really bright if the surronding ambient light is low.

Imagine trying to watch a Front Projector outside on a sunny day. you would think the picture was very bright compaired to the light from the sun would you?

The luminance of a scene works the same way. If you can make the darkest parts seem dark but the bright parts brighter the "contrast" is enhanched like how the dynamic contrast works on displays now.

The issue is todays displays cannot make a image as many orders of magnitude brighter than its darkest part as what our eyes can see. But the idea of the Dolby Vision tech is that the whites are pretty much able to be much brighter than the darks and contain all the details that exist in between.

So if a image is comprised of many shades of gray but displayed on a 8bit or 10bit display you are limited numerically on how many levels of brightness of each shade can be produced 256 times brighter for 8bit and 4096 times brighter for 10bit.

With 8bit to 16bit expansion the availability for the levels of brightness to now fill a much greater range of values is so much larger that it makes the image look better even without being captured in 16bit.

You aren't "inventing" data like with scaling resolution, you are taking the values for black and moving them much farther away form the maximum values for whites than 8bit can numarically represent.

Another way to think of it is like how 0ire is supposed to be the lowest brightness a display can produce and the highest being 100ire. If 100ire can be brighter than 256 steps and brighter than 4096 steps it isn't a matter if the color info only has 256 shades of difference, the luminance can be "expanded" to fill a range.

the difference isn't invented, it just has to be calculated and is the amount of values avalible to represent the calculations is vastly bigger with 16bit versus 8bit the image can look better by a lot.

Scaling video has to do with the data to represent each pixel not being there so the video processor has to "guess" what the info is. It cannot simply display the data without chosing the values of each pixel to be displayed first.

The luminance data only has to be assined a value from a range available to be displayed. If you make it dark were it is dark and bright were it is bright the picture show is proper no matter if the dark parts are way way darker now or vise versa.

Resolution is either there or it isn't, luminance is relative to the lack of light as far as how we see it. There is no "detail" in it. It is simply brighter or darker so if you ca represent those values more on a much larger scale they can be that much brighter or darker than if constrained by a smaller scale.

the difference isn't invented, it just has to be calculated
See, I fail to see the semantic difference there. Interpolation is interpolation. You're taking a range of scalar data points and remapping them to a new one. Remapping 8 or 10 bit video to 16 bit can no more create additional granularity that wasn't there in the source material than scaling 480 line video to 1080 line can.

If you don't care about greyscale resolution then sure, you can just make the image brighter, just as you could buy a bigger TV of the same resolution, but that's not technological process. TVs on the market today can be made to pump out 80-90ft-L.

light has no "resolution" to interpolate. the same way a tv form a few years back has maybe 600:1 contrast ratio but now most have 1000:1 or more native contrast.

This is not because they interpolated the data, the video data is the same but the display either has better blacks or white or both.

But after a point you become numericaly constrained by the bit depth of the panel and control electronics. With 8bit there can only be 256 levels of data for any given pixel which might be seen as a lot but 16.7 million possible colors is limited as compaired to human vision. It does not encompass the full range of possible shades of color a human can see.

Or more accurately the to end of the brightness range on a 8bit panel combined with standard backlights is notable to reporduce the top range of luminance the someone can see.

Resolution is finite in so much as if I have a video clip that is 480p, the resolution can never be more or less. It can be interpolated using math and scaled to take up more pixels on a screen, but the resolution doesn't change.

With luminance data the bottom and to value for any source are limited by the bit depth they are captured at but not the bit depth of your display.

A cellphone camera has a small lens and a small sensor to capture with right? So when you take a picture if there is not enough light the details captured suffer. Now if you take a picute of the same scene with a capable dSLR or high quality camera with a bigger lens and bigger sensor more light is captured and you can often see details in a picture the would be crushed on a lesser camera. Yes part of that IS resolution but if they both were the exact same in megapixel count haveing a lens and sensor capable of capturing more light you will see more.

If the lowest possible amount of light output of a display is zero you can have a very good contrast ratio with a few nits of light because the ratio is a difference between two values. you can have a very dim display process the data very well and get good images as long as the ambient light was low enough and so was the black level.

So if you have a display that has both very low black level and super high white levels you would have good contrast. The issue being unless you are able to reproduce the levels of brightness in between you get crushed details.

I can have 100ire red defined by a 8 bit panel at say 30% the brightness of white. We get 76.8 out of 256 right on a 12 bit panel that is 1228.8 out of 4096.

See how I can represent the exact same percent or red compaired to white but using a larger number. I'm not making up the number, I'm not adding pixels to something to make it take up more of the screen.

It isn't a arbatrary number it is 30% red of the 100% white value. Only the precision with which I can define that percent is much larger in scale. In 16bit that is 19660.5 out of 65535.

This is were I can take the luminance values for any color in a 8bit video and multiply them to get their 16bit equilant. Like if I am 29 years old I'm 348 months or 10592.0238 days or 254208.57 hours old. The data is the exact same but by having a larger scale I can represnt it in much more detail.

I cannot take 480 lines of resolution and show them in any other way, they are not part of a range of values possible they are 480 lines. Luminance is a value in a range possible limited by the rate that it was sampled/recorded in.

I wish someone that knows a better way to explain how luminance is represented might chime in becaue I'm sure Im not the best to try to break it down but is is not inventing data for thin air it is changing the scale that the data is measured on and expanding that data to a larger one.

Like if I am 29 years old I'm 348 months or 10592.0238 days or 254208.57 hours old.

The problem is your "more detail" values are only correct if you are *exactly* 29 years old. And that won't actually be true very often. Put another way, for most of the time you can legitimately call yourself 29 years old, your "more detail" numbers will hold incorrect values.

No, actually with LDR HD content shown on HDR set. Brightside created algorithms that extend dynamic range of LDR content.

There was an interesting paper out of one of the German research houses showing viewers do not exhibit a meaningful preference for "HDRed" LDR content over simply cranking the brightness on LDR content with a constant backlight, across a wide range of viewing conditions. I believe they actually used a Brightside display for the tests, but I would have to check as I could be wrong on that.

The result seems to make sense. In most viewing conditions the achievable black levels aren't going to be determined by the backlight, they are going to determined by the ambient conditions. Especially with a super-bright display that is going to self-illuminate its own screen to a very high degree. Since both the brightness and blackness of the two approaches are so similar, the only real difference will be the contrast curve between the two end points. And I suspect anybody spending much time at AVS knows just how subjective a contract curve preference can be! :)

light has no "resolution" to interpolate.
Neither does space...until you quantize it.

You guys are hijacking this thread. :)

I just read an article on LD artifacts (static and moving halos). Here is data on the halo artifact. The article does not discuss the inherent black crush and dynamic black level issue of LD though. Even so, interesting data IMO.

Reference : Information display 2009 Vol-25 No-1 pg22
http://i222.photobucket.com/albums/dd126/xrox/LED.jpg

Also demonstrated at the show was the breathtaking Solar Series LCD screen. The technologies implemented in this £20,000 display deliver an astounding 65,536 shades per colour, smashing the 1,024 shades typical of current LCD displays. The Solar Series offers a theoretically infinite contrast ratio, with measurements of well over 1,000,000:1. This is the world’s first production High-Dynamic-Range (HDR)–enabled LCD display using Dolby Vision technology.

http://uk.cinenow.com/articles/9293-sim2-showcase-c3x-lumis-host-at-bristol

http://i.cinenow.net/ymages/5/5374.jpg

http://www.consumerelectronicsnet.com/articles/viewarticle.jsp?id=848202

Meanwhile, the company formally presented its 47-inch “Solar47HT” high dynamic range LED-backlit LCD monitor with local dimming technology by Dolby. The product, which is expected to ship in the fourth quarter, will carry a $24,995 suggested retail price, the company said.

Dolby’s HDR system, aka “local dimming technology,” manages 2,206 individually controlled LEDs behind the liquid crystal panel. By selectively turning off the backlight behind black areas in scenes, Dolby’s HDR algorithm allows those areas to become truly black.

The HDR system in tandem with Sim2’s proprietary flat-panel 16-bit color processing delivers more than 65,000 shades per color.

The HDR-based system uses an array of individually addressed LEDs behind the liquid crystal panel. Each LED is individually controlled in concert with the image signal so that less light will be produced for darker small portions of the screen while more light is produced for brighter small portions of the screen, Sim2 said.

Sim2 said the LCD panel’s 2,206 high-power LEDs produce 4,000 cd/m2 of peak brightness, and an “over 1,000,000:1 contrast ratio.” The 16-bit processing produces 65,536 shades per color.

Do any of the Main St. local dimmers - quite an extensive list now: Samsung, Sony, LG, Toshiba, Vizio - have the "Dolby Vision" or "Dolby Contrast" licensing badges? I've seen I think 3 of these on the shop floors, but didn't notice anything.

Do any of the Main St. local dimmers - quite an extensive list now: Samsung, Sony, LG, Toshiba, Vizio - have the "Dolby Vision" or "Dolby Contrast" licensing badges? I've seen I think 3 of these on the shop floors, but didn't notice anything.No. All these "mainstream" LED tvs with local dimming have a much more primitive version of the technology compared to Dolby Vision. Judging from the price Sim2 wants for their Solaris set - $25K- I guess IMLED technology remains extremely expensive. Too bad.

No. All these "mainstream" LED tvs with local dimming have a much more primitive version of the technology compared to Dolby Vision.

Technology in all of them is very similar, including the Sim2. Only real difference is brightness; those non-Dolby TVs are all doing the "pixel compensation" etc tricks. The electronics/firmware to control clusters is not appreciably less complex than that to control at very high granularity, however the price differential between Main St and Sim2 is very consistent with the difference in required LED pricing and mechanical/thermal implications of generating thousands of nits of peak brightness.

Technology in all of them is very similar, including the Sim2. Only real difference is brightness; those non-Dolby TVs are all doing the "pixel compensation" etc tricks. The electronics/firmware to control clusters is not appreciably less complex than that to control at very high granularity, however the price differential between Main St and Sim2 is very consistent with the difference in required LED pricing and mechanical/thermal implications of generating thousands of nits of peak brightness.Mainstream sets have pretty much the same number of LEDs as Sim2, so other than the cooling requirements, the difference in cost seems to be explained by higher cost of controlling individual LEDs vs. controlling clusters of LEDs.

Mainstream sets have pretty much the same number of LEDs as Sim2...

There are nearly two orders of magnitude price difference between LEDs suitable for "few hundred" nit vs "few thousand" nit output. This is the single biggest cost differentiator, by far.

...so other than the cooling requirements....

This is also a significant expense, not only for the material/design, but because of the size, as it requires retooling for the entire display.

...the difference in cost seems to be explained by higher cost of controlling individual LEDs vs. controlling clusters of LEDs.

This is the easiest part to price out, as we know exactly how many more control lines, drivers, etc are required, which is why it is known with a high degree of certainty there isn't a meaningful cost difference due to this.