CurtPalme.com Home Theater Forum

[Spanky Ham]

It looks like we are having a wave of LED mania.

I know there has been some smoke and mirrors going on about the on/off cr. I believe Scott told me that they are around 3 to 5k to 1 at this time. The head Runco guy told Darin and me that there pj is around 5k to 1. Yes, you can turn the LEDs off, but you could also turn off a regular bulb. Obviously the difference is the time to turn back on, but it is still not like a CRT.

As for the local dimming, I think Brightside was doing something like this with their fourth panel pj. Supposedly they could achieve very high ANSI and on/off cr. It has been awhile since I read this over at AVS, so I can't be more specific. I went looking for the white paper about a year ago, but I can't find anything. Dolby bought Brightside and looks like they haven't done much with the company.

Oh, I agree that LED will probably be the future and I can see a lot of people migrating. I still think it needs another year or so. I will be interested in what comes out of Cedia this year.

[macgyver655]

[garyfritz]

You can turn the LEDs off, but not on a zone basis. LEDs off == totally pitch-black screen, like you unplugged the bulb. If you need ANY photons on the screen, the LEDs have to be on (and cycling). Which means you're basically back to a "bulb emits light all the time, DMD modulates it on the screen" mode.

[macgyver655]

Where are you guys seeing you can turn the LEDs off. When the projector is turned on the LEDs begin flashing in sequence and continue to flash until you turn the projector off. Have you guys read something different.

Scott and I discussed this in depth when I was at VDC along with other things.

[perisoft]

[macgyver655]

[perisoft]

[garyfritz]

Yes, cutting off the LEDS will provide perfect blacks. This feature is labeled "POWER" on the projector. Set it to the "OFF" position and you'll get perfect inky blacks.

Seriously. As far as I know, that's the only way you'll get "perfect" blacks and it's the only way the LEDs will turn off. With any displayed image, even a full-blacker-than-black image, those LEDs are on. It's just exactly like a DLP with a lamp and color wheel, except it's cycling LEDs.

[macgyver655]

[macgyver655]

[Spanky Ham]

[WanMan]

I would be curious to know what the measured minimal light output of these LED projectors are without the LED's actually being off. That would determine the black level capabilities. Of course, marketing has to lie in order to hype the otherwise mediocre they are shoveling.

This is not to say that digital projectors haven come a long way in their home application inception, but they are still being compared to CRTs. hahahahahahaha

[Kiev Savoie]

[cmjohnson]

I've spent a little time with my head stuck into the guts of a DLP engine and I can tell you that a LOT of things aren't usually being done that COULD be done to control stray light and keep it from reducing the contrast ratio.

Every DLP projector I've yet seen uses a black anodized aluminum target patch to absorb the "off" switched light from the DMDs. And eventually, they all bleach. The black anodized aluminum ends up being LESS than black. I've seen it bleached all the way to raw aluminum color.

DMDs are essentially a huge collection of minute mirrors. The light that falls on them from the lamp source is either directed TO the
screen, or AWAY from it. All the light that is directed away from the screen must be completely absorbed or diverted so that it does
not ever get to the screen. If it does, it will raise the black level and reduce the contrast level. In my opinion, the management
of this waste light is often done very poorly in many DLP engines. I've never YET seen one that handled it in a manner that I would
say I'm really impressed with.


A better way to do it would be to direct the light into a small tank filled with carbon ink that will NEVER bleach. That tank would have a
high quality glass entrance window which is treated with high grade anti-reflective coatings. And, every item in the head end of the
optical engine should be painted black or hidden behind black painted covers. Anything that is visible in the head end (after the DMDs)
will affect the CR to some extent.


I'd say that with careful application of paint and shielding, any DLP projector's contrast ratio could be improved by a worthwhile amount.

[Spanky Ham]

Interesting observations, CJ. I think your proposed solution would be expensive and only cost effective for the expensive pjs. A better alternative would be the new black coatings that have been developed. This might be cost effective for both the low end and the DIYer wanting to mod his pj.
http://www.rsc.org/chemistryworld/News/2007/March/01030701.asp

Doing a search for the above (which I don't believe has been developed commercially), I came across this company. I am not sure how expensive it would be, but at least their is something commercially available.
http://www.acktar.com/

[garyfritz]

[cmjohnson]

All that the LEDs do is provide illumination for the total DMD panel. They do not provide per-pixel illumination. They are a substitute for the lamp, and nothing more.

CJ

[Kiev Savoie]

Well, I probably am working from a skewed idea of how DLP works, but the fact that LED's are being used as a replacement for the lamp is the point i am getting at. Here we have a new technology being adapted to an old platform and at the the 1st gen level the capabilities of that new technology usually are just beginning to be explored. LEDs have many qualities (such as fast on/off times) that are probably not being utilized in the most effective way. Anyway, i'm excited to see how it all develops.

[garyfritz]

They *are* making use of the LED fast on/off. They're not just using white LEDs with a color wheel. They're using R/G/B LEDs and pulsing them for the 3 colors. No color wheel, no moving parts, much faster cycling = no rainbows (hopefully), no expensive bulb to degrade, colorshift, and burn out, and of course much less power and heat, therefore much better lifespan for the light engine, much less cooling issues, etc.

But it's still illuminating the entire DMD at the same time.

[cmjohnson]

At some point in the future, when manufacturing technology allows, we will eventually see LED matrix devices which will resemble a DMD in format, but each mirror will be an LED pixel in either a single color or an RGB triad. Given the brightness per unit area that LEDs are capable of achieving NOW, the technical obstacles to this accomplishment are all related to the consistency of output and color temperature
for each individual pixel in the display.

Right now, all LEDs that are produced for markets where color temperature and consistency of output is an important factor are sorted
and binned according to their actual performance. Manufacturing technology is not yet up to the task of cranking out millions of LEDs
at once that all are for all intents and purposes identical in terms of their color temperature and output curve. Only when that goal has
been achieved will unitized matrix LEDs become a viable product.

Light output from that device should not be a problem. An active area of just 12 square millimeters allows a Luminus Phatlight to
deliver over 2000 lumens (green) and they have a white LED package that delivers up to 6000 lumens out of a total emitter area of 36
square millimeters.

By way of comparison, a one inch class DMD in 16:9 format is ABOUT 400 square millimeters. With over 10 times the available surface area, it stands to reason that even with a less than perfect fill factor, it will be easy to build a complete self-luminous LED
imaging panel that is comparable to a DMD in overall size.

Thermal considerations will factor heavily into the design, but LEDs are getting more efficient all the time, meaning LESS heat generated
for any given light output level.

By the time a sufficiently consistent LED array can be fabricated on one piece of silicon, the efficiency will be high enough that the heat
elimination considerations will not present a monumental engineering challenge.


Be quite sure about it, this is the direction the industry will go as soon as it's a viable product.

The ultimate expression of it will be stacked transparent RGB pixels on the substrate. The red, green, and blue channels will be
vertically integrated with no horizontal offset, will not optically interfere with each other, and allow 3-chip performance with a single
chip engine.


I predict this will be seen for the first time in a technology demonstration within three to five years.


CJ