BrightSide DR37-P HDR Display

[Jeremy Charette]

<div class='os_post_top_link'><a href='http://www.bit-tech.net/preview/hardware/brightside_hdr_edr/1.html' target='_blank'>http://www.bit-tech.net/preview/hardware/brightside_hdr_edr/1.html</a><br /><br /></div><i>"In recent years, there has been a changing of the guard in display technology. LCD &amp; plasma screens have replaced the century-old Cathode Ray Tube (CRT), bringing us large-yet-thin displays used everywhere from lounge rooms to shopping malls; newsrooms and airports. Yet tucked away in an anonymous office building in Vancouver, Canada, there is an almost equally big revolution in display technology taking place. In the neat, open-plan offices of BrightSide Technologies, a team of researchers spun off from the University of British Columbia have developed the world's first true High Dynamic Range (HDR) display. In a recent visit, bit-tech was granted Access All Areas, and the experience was, if you can excuse the pun, enlightening."</i><br /><br /> <img src="http://www.digitalmediathoughts.com/images/03.jpg" /> <br /><br />No words can describe how incredible this technology is, or how it could revolutionize high-definition tv. Trust me, read this article, it's worth the time, and the pictures speak for themselves. I only have three words: I can't wait.

[Felix Torres]

Very interesting indeed.
But I do think these folks are underestimating the power of modern semiconductor manufacturing technology.
I see no reason why an addressable-pixel LED backlight like they use can't be manufactured as a 1920 by 1080 grid to start with. And any manufacturer of displays who looks at this tech will realize it is the obvious road to take.
For one thing, it'll simplify the system control electronics and firmware.

So, realistically; the configuration they describe will likely end up only being used on future *low-end* products! :lol:

Is it just me or isn't it looking like LCD light-valve tech in all its variations is starting to take over the HD display space?

[Jeremy Charette]

One huge advantage I see is that this will put LCDs on par with (or even above) Plasma displays in terms of color rendition and brightness/contrast. Plasma is going to be a very short-lived technology.

As far as using an individual LED behind every pixel on a 1080i/p display, well I can see one huge barrier: the number of LEDs required. BrightSide is using 1400 LEDs, whereas a 1920x1080 array would require over 2,000,000 LEDs. It would require an astronomical amount of time to place all of these LEDs on a PCB board, even with high-speed automated manufacturing. The heat generated would also go up exponentially. I can tell you that having worked with sub-pixel interpolation technology myself in the past, using just 1400 LEDs is a much more practical and reasonable approach.

If you read the complete article, it explains how BrightSide gets away with just 1400 LEDs, using a combination of pixel interpolation and by modifying the image displayed on the LCD panel. The resulting image has High Dynamic Range, even if each pixel's brightness isn't individually modulated.

Colored LEDs and sub-pixel interpolation are used widely in one particular place: Times Square. Almost all of the new Jumbotrons and scrolling banners use colored LEDs to modulate color and brightness, rather than backlit LCDs. The problem up 'til now has been color accuracy, which once solved should allow somebody to make a home theather sized LED tv (just LEDs, no LCD panel).

[Jeremy Charette]

One other thing: 1400 surface mount LEDs run about $42 at wholesale (3 cents a piece). If you wanted 2,000,000 of them, you'd be out about $60,000. :lol:

[Felix Torres]

One other thing: 1400 surface mount LEDs run about $42 at wholesale (3 cents a piece). If you wanted 2,000,000 of them, you'd be out about $60,000. :lol:

And what would they cost in 2010?

I don't doubt that there are good reasons to go with zone lighting today and since its a proof-of-concept thing its okay.
But these are semi-conductor devices, so price and size drops with time.
And I happen to think these savings will most likely be applied to increasing the granularity of backlight control.

We're looking at what, four LED tech generations down the road? maybe a 16-fold price cut? Five generations would cut your amusing price to less than $2k.

Since this is horizon tech and we're not going to see it much before 2010, its a safe bet semiconductor prices will come down by then. And flat panel display tech will have improved in resolution and cost. They might even pair these backlights with UHD LCD displays...

Finally, lets face it; the addressable-pixel backlight is really a monochome display in its own right, no?
So any kind of display tech could do the trick.
SED, or OLED monochrome panels...

Lots of roads to implement this tech; I just think one-to-one backlightling is going to be economically viable by the time they get to building these things. &lt;shrug> Besides, the first displays to use this tech will still be in the $5-10K range circa 2010, anyway.

[Jeremy Charette]

BrightSide is already selling their own display at $50,000.

As far as pricing, I was way off base (I was thinking of a different type of LED). White SMT LEDs run $1.80-2.20 at wholesale in quantities of 10 or more. Even assuming a 50% discount for massive quantities, that would be at least $180,000 (today's dollars) for a per-pixel display. Even with a 16-fold decrease in price (which I don't see happening in 4 years), that's still over $10,000 just for the LEDs alone. Then there's the question of if you can even place 2,000,000 LEDs on a PCB in a reasonable amount of time. Assuming you can place 20 LEDs per second (which is incredibly high), it would take over 27 hours to assemble a complete backlight panel. You'd get less than 300 panels per year from a piece of equipment that costs over $250,000. The laws of physics take over, and the costs become prohibitive.

Of course, I'll be the first to admit, this is a totally pointless debate for us to have. Neither of us can really say what will come down the road in 5 years. I certainly didn't see this coming 5 years ago!

My question is this: if 1400 LEDs give the desired effect, and more LEDs don't make the display look any better to the naked eye, why bother using more? There has to be a point of diminishing returns, and I'm sure BrightSide and their licensees will figure out where that point is. I just want one of these in my living room right now. :D

[Felix Torres]

My question is this: if 1400 LEDs give the desired effect, and more LEDs don't make the display look any better to the naked eye, why bother using more? There has to be a point of diminishing returns, and I'm sure BrightSide and their licensees will figure out where that point is. I just want one of these in my living room right now. :D

The point of diminishing returns starts at one-to-one. :-)

See, Brightlight is (apparently) using software, their "secret sauce" to modulate the chroma data that drives the LCD light-valve layer to compensate for the lack of granularity in their chroma layer. This is essentially the same thing all LCD displays do today since all the luminance has to be regulated through chroma variations in the LCD because the backlight is static. Once you move into dynamic backlighting, you start to move luminance control into the backlight. Well, the obvious end-point of this process it to move all lumina data to the backlight and leave the lcd light-valve to work off a pure chroma signal.

To put it another way, if 1400 is good, 2 million is better. :wink:
Way better.

And, remember; like I said, a one-to-one backlight is essentially a monochrome display. You don't have to use LEDs to make your dynamic backlight and if you use LEDs you don't have to use discrete LEDs. An LED is a semiconductor laser, after all. The same technology that makes the driver electronics of LCDs could make LED-matric panels of the required resolution if there were a demand for the panels. You could also use a cheap plasma panel in monochrome mode driving a uniform white phosphor as your back-panel if you wanted to. (Those can be had today for a couple hundred bucks, oem.)

One-to-one is doable, honest! ;-)

It is the logical end-point *if* this technology takes off.
*If* it does, we'll see it.

Me, I hope it does because it gives us the best of both the digital and analog display worlds in one flat display and it opens the door to very high quality/very high resolutions at affordable prices in the mid-term (5-10 years) that we otherwise wouldn't be seeing.

If nothing else, medical imaging is about to experience a big leap forward.

[Jeremy Charette]

What if you backed an HD (1920x1080) LCD panel with an HD monochrome CRT (1080i or 1080p), and used the CRT to drive luminance only? Wouldn't you get the same effect, without having to place and drive 2 million LEDs? (Kind of like your Plasma/LCD hybrid, but cheaper.)

8O

*runs down to the patent office as fast as possible*

[Jason Dunn]

&lt;swooosh>

That's the sound of most of this thread going right over my head. :lol:

[Felix Torres]

What if you backed an HD (1920x1080) LCD panel with an HD monochrome CRT (1080i or 1080p), and used the CRT to drive luminance only? Wouldn't you get the same effect, without having to place and drive 2 million LEDs? (Kind of like your Plasma/LCD hybrid, but cheaper.)

8O

*runs down to the patent office as fast as possible*

You could.
Bulky and heavy but doable.
Only thing is, CRTs *already* have separate chroma and lumina control built-in so you're better off engineering a CRT that can natively display 1080 than trying to mate it with an LCD light valve.
(Remember, CRTs are analog display devices by nature.)

Still, building a proper high-def shadow mask has proven hard so, while baroque, a hybrid LCD/CRT might have some commercial merit.

*I* wouldn't invest on it but somebody might...

Maybe Thomson would go for it; it would give the french a TV system nobody else uses... :twisted:

[Felix Torres]

&lt;swooosh>

That's the sound of most of this thread going right over my head. :lol:

...he said, tongue firmly in cheek... :wink:

[Jonathon Watkins]

&lt;swooosh>

That's the sound of most of this thread going right over my head. :lol:

...he said, tongue firmly in cheek... :wink:

Well, it's certainly for real with me. :lol: (I'll leave you boys to on with your discussion). :wink: