As LED technology is coming most likely within the year for FP. Please, post any news and advancements that have been made so that we can create a single thread with all the information that concerns LED as a light source for FP use.

I can't see it used for FP. Most projectors need a very intense point-source light to illuminate a small LCD panel, DMD etc. LEDs really need to be clustered to create lots of light. (and current hi-power LEDs need a heat sink).


They're probably more use on a flat-panel where they can be spread out (and their heat more effectively dissipated).

My first HDTV was a lamp based Samsung 50" DLP.(about two years ago)
I recently very shortly owned a LED Samsung 61" DLP.
I returned it for a couple reasons, but one was that is was too bright. At night, I would ramp the contrast and brightness down big time just so my eyes didn't feel like they were going to bleed. Which of course completely ruined the PQ.

I'm not sure how that transfers to LED PJ's though.
SSpivey could be correct.

If they are anything like what they have on Police vehicles these day's they should be bright enough. We have had some accidents on highways at night because of the intensity of the flashing lights on top of the police cars. I happened to drive by a work zone on the highway this week and the flashing lights on top of the car made me have dark spots in my vision. These new lights on police crusers are very bad. Way too intense.

Ron

There is alot of info going on about LED light sources. Are LEDs powerful enough for FP applications? Can they put out the same amount of lumens as a lamp? I know that when Samsung put LEDs into their RP DLP, people complained that the sets were not as bright as the lamp based models. So how are we gonna light up our 100"+ screens?

I have an automotive benchmark for LEDs. They may exist already, though I haven't seen them, but when LEDs make it into car headlights, then well be ready. In my opinion it isn't a matter of power efficiency or clustering (you can always put more) problems. I just don't think any one has been able to make them bright enough, where brightness is not a trade off...i.e. projectors, headlights, etc.

It's not just lumen output that matters--colour rendering index (CRI) is important too. A projector needs a lamp that emits light across the whole spectrum, otherwise colours will look awful. MH,HPS,Xe do this.

Those police cars just need to emit lots of light--they don't need to make anyone look good...

Yes, the colour accuracy is very important for the projectors. The 3 chip projectors separate 3 colours optically and recombines to produce the final picture.

"A projector needs a lamp that emits light across the whole spectrum, otherwise colours will look awful. "

The holy grail for pj light sources is the three correct R, G, and B wavelengths, i.e. lasers being the most famous example.

The trouble with UHP is that the colors don't have the optimum wavelength or distribution.

We shall see perhaps Cedia, perhaps a year later. remember LEDs don't dim with age so 650 lumens might be equivalent to a 1000 lumens bulb with 200 hours or so on it.

A cluster of a R,G, and B LEDs or multiple clusters might do the trick. something is coming though.

Alan thinks its a year or more off. I think we will see something in the $30K class at Cedia LED based. I dunno fo sure.

Is laser a powerful enough light source?

Ever see a James Bond movie? :)

Do you expect me to talk ? No mister Bond, I expect you to die.:D

Art

LED Front Projectors are currently being designed, built and tested by more than one manufacturer in Japan. Brightness is the issue right now. So far they've only been able to achieve 250 lumens with a projector the size of the old three gun CRT's.

Laser technology has not been a viable option for front projection so far.

If they are anything like what they have on Police vehicles these day's they should be bright enough. We have had some accidents on highways at night because of the intensity of the flashing lights on top of the police cars. I happened to drive by a work zone on the highway this week and the flashing lights on top of the car made me have dark spots in my vision. These new lights on police cruisers are very bad. Way too intense.

Ron

Its so idiots see them in their rear view mirror, and avoid hitting the actual police cruiser, ya know the idiots in the lane that should yield to an emergency vehicle pulled off the side of the road.

Perhaps its why they needed brighter leds, so that in the rain and heavy snowfall you can as a safe responsible driver see them to make proper judgment call.

Cheers

Bobby

Brightness is the issue right now. So far they've only been able to achieve 250 lumens with a projector the size of the old three gun CRT's.

awha? :eek:

That sounds like info from the early 2000s....

If you look at the Phlatlight technology (thanks Ohlson! :))
You'll see that they already have 16:9 (12mm^2 chip area) single LED chips blasting out 875 lumens for RED, 2100 lumens for GREEN and 400 lumens for BLUE.
Wavelengths are: 626nm RED, 528nm GREEN, 463nm BLUE.

In order to get D6500k out of the above wavelegths we need the following mix:

43.1% RED, 33.5% GREEN, 23.4% BLUE


So, let's take the dimmest (BLUE) LED and work backwards:

400lm (BLUE) at 23.4% means that we'll need
737lm (RED) for 43.1% and
573lm (GREEN) at 33.5% to get D65

BUT!

If you look at the spec, we can safely overdrive BLUE to 600lm (50% overdrive) and still be in the safe/ comfortable region for both RED and GREEN.

So,

If we redo our calculations we have:

600lm (BLUE - OVERDRIVEN) at 23.4% means that we'll need
1,105lm (RED) for 43.1% and
860lm (GREEN) at 33.5% to get D65

Now, add to that the various advantages that LEDs bring:

Dynamic Iris Operation
80,000 hour halflife
Higher Efficiency (lm/W)
Lower thermal load - smaller heatsinks - simpler/smaller power supply
No colorwheel (for DLP)
Infinite contrast
Asynchronous operation (any color can be on or off)
Elimination of refresh artifacts



and we have a winner.

Now, back to the numbers, a single chip DLP system has a light conversion efficiency of about 15% (filters, colorwheels, light path losses, lamp inefficiencies etc. etc)

No doubt, an LED illuminated 1-DLP system would have a better efficiency than that but never better than the theoretical maximum of 33%.

Say, a good number would be 25% efficiency.
So, from the above numbers we get....

drumroll.....

a 640 lumen projector!

yay! :D

Seriously though, it is doable, but obviously, there are issues implementing it or we would have it already...

By the end of 2009, I'll bet we'll have a 1,000 lumen >100,000:1 CR projector.
Anybody willing to take me on? :D

My info is that 650 is doable now and that we might see it at Cedia, and definately we will see it by Cedia 2009.

My info is that 650 is doable now and that we might see it at Cedia, and definately we will see it by Cedia 2009.

Thanks Mark,

That would put it (650 stable D65 lumens) at the top 20th percentile of present 1080 projectors.

I say, that's OUTSTANDING!

Bring it on! so i can buy the 3rd generation for $2k :D

Wavelengths are: 626nm RED, 528nm GREEN, 463nm BLUE.If these each have very narrow bands then it should be possible to make a screen that will reflect them, but reject a lot of ambient light (by rejecting all wavelengths except for the 3 narrow bands). Might also make it easier/cheaper to make a meter for measuring color balance, where the meter is specifically for projectors putting out those wavelengths.
No doubt, an LED illuminated 1-DLP system would have a better efficiency than that but never better than the theoretical maximum of 33%.Why 33%? If that is from only have 1 of 3 colors at a time, it shouldn't apply to LEDs given how they can be driven harder while on if turned off for some of the time. That is a reason that LEDs should help single chip DLPs more than other projectors from what I've seen. The 2 technologies go together very well.

Another thing I think is an advantage of LEDs over current bulbs is the ability to put in some blank time between frames (to help motion perception) and then drive them harder when on. Blank time with bulbs means hurting on/off CR since the bulb is still on and also hurts lumens since the bulbs can't just be run harder the other time. Or at least not as much harder as I think LEDs can be. I'm not sure whether LEDs can be run enough harder to completely make up for the blank time, or just somewhat harder.

--Darin

There is a company that has a 1100 lumen LED FP. It may be shown at Cedia but in a shoe box case. They will try their best.

If these each have very narrow bands then it should be possible to make a screen that will reflect them, but reject a lot of ambient light (by rejecting all wavelengths except for the 3 narrow bands).

--Darin


Spectral BW at phi=50% is 19nm, 39nm, 25nm for RGB respectively.
You're right, it's tight, not LASER tight though...
Excellent idea Darin. Selective screens would help a lot in high ambient situations. Sports bars anybody? ;)
It may not be ideal for dedicated batcave HT due to other unforeseen issues.


Why 33%? If that is from only have 1 of 3 colors at a time, it shouldn't apply to LEDs given how they can be driven harder while on if turned off for some of the time. That is a reason that LEDs should help single chip DLPs more than other projectors from what I've seen. The 2 technologies go together very well.
--Darin

That was my thinking, yes. 1 color (of 3) at a time, 33% maximum theoretical.
I agree that 1-DLP and LED is an ideal marriage, but as far as I know, the LEDs can be overdriven 50% which wouldn't correspond to "50% brighter" since our vision is not linear. I may be wrong though. It's happened before... ;)




Another thing I think is an advantage of LEDs over current bulbs is the ability to put in some blank time between frames (to help motion perception) and then drive them harder when on. Blank time with bulbs means hurting on/off CR since the bulb is still on and also hurts lumens since the bulbs can't just be run harder the other time. Or at least not as much harder as I think LEDs can be. I'm not sure whether LEDs can be run enough harder to completely make up for the blank time, or just somewhat harder.

--Darin

So, if I understand what you're saying:
For a 24Hz film, we refresh the LEDs at, say, 240Hz (for ease of calculation)

So, now we can show the same frame:
10 times - no blanks
9 times - 1 blank frame (11% brighter each of the 9 frames)
8 times - 2 blank frames (20% brighter each of the 8 frames)
.
.
5 times - 5 blanks (100% brighter each of the 5 frames)
.
.
.
1 time - 9 blanks (1000% brighter for the single frame)

to average the same frame brightness.

Of the 10 scenarios above, which is better?

Or, would it depend on the brightness of the frame itself?
More blanks for dark frames
less blanks for bright frames?

Am I in the ballpark at all?

There is a company that has a 1100 lumen LED FP. It may be shown at Cedia but in a shoe box case. They will try their best.

Is the shoebox made of brushed aluminum?
If so, sign me up! :D

Much better for Alan to blab it, than me.:) I think a fair inference is that an LED HT FP quality machine could be shown in some peotype form at Cedia 2008 but actually being able to order asnd take delivery will be some time off. i don't have a good feel for the tie frame. The beast will be expensive.

Alan. Did you ever try to pay a woman's shoe bills? :)

Its so idiots see them in their rear view mirror, and avoid hitting the actual police cruiser, ya know the idiots in the lane that should yield to an emergency vehicle pulled off the side of the road.

Perhaps its why they needed brighter leds, so that in the rain and heavy snowfall you can as a safe responsible driver see them to make proper judgment call.

Cheers

Bobby

I don't disagree with your comments about idiots on the road but these are way too bright. In the dark when they are flashing its like the sun flashing in your eyes. Maybe even worse. It causes black spots like a burn in your vision and you can not see the road after. They had a two car pile up on a highway near me because the cop had someone pulled over and the people driving on the highway lost vision due to the excessively bright flashing. I was surprised I couldn't see the road after I passed the cop on the side of the road. Thank God it was in a work zone and everyone was going slow.

Ron

Alan. Did you ever try to pay a woman's shoe bills? :)

Just the shoes, I thought it was a pair of shoes and a purse with every outfit :)

The thing I don't get about the wavelength-selective screens is how can they put 3 filters on top of each other w/o killing all reflection?

"That was my thinking, yes. 1 color (of 3) at a time, 33% maximum theoretical.
I agree that 1-DLP and LED is an ideal marriage, but as far as I know, the LEDs can be overdriven 50% which wouldn't correspond to "50% brighter" since our vision is not linear."

Brightness is measured in L, not our subjective response.

Anyway, what you called efficiency is actually duty cycle.

Which "efficiency" is actually "duty cycle"?
Are you talking about luminous efficiency of the light source (lm/W)?
If so, I don't think so.

I agree about the objective lumens measurement, I only wanted to convey the subjective feeling of 50% more lumens. It's not as much as people think.

Another thing I think is an advantage of LEDs over current bulbs is the ability to put in some blank time between frames (to help motion perception) and then drive them harder when on. Blank time with bulbs means hurting on/off CR since the bulb is still on and also hurts lumens since the bulbs can't just be run harder the other time. Or at least not as much harder as I think LEDs can be. I'm not sure whether LEDs can be run enough harder to completely make up for the blank time, or just somewhat harder.

--Darin

Would this blanking be in any way similar to the blanking that i see on almost every car that has LED tail lights to make them dimmer when the driver isn't stepping on the brakes?

Those led tail lights drive me crazy because when my eyes pan across the road, they leave annoying trails in my vision. This probably depends on the rate they are flashing. Some LED tail lights are more persistent than others. Vertical designs seem to be the most obvious though - like on Cadillacs. Come to think of it, vertical white lines on black backgrounds are the worst for RBE too, from my experience.

Anyway, I just hope they don't come up with another technology that only people who aren't sensitive to its defects can enjoy.

Some of us who have made the decision to go with large screens have painted ourselves into a corner if we don't see something closer to 2000 lumens.

Art

Some of us who have made the decision to go with large screens have painted ourselves into a corner if we don't see something closer to 2000 lumens.

Art

Patience, wise Yoda!
It will happen for you big boys too!

Just let the small ones play in the sandbox for now. ;)

"Which "efficiency" is actually "duty cycle"?"

Where you said efficiency was 33% because they were on for 1/3 of the time.

"Which "efficiency" is actually "duty cycle"?"

Where you said efficiency was 33% because they were on for 1/3 of the time.

I meant "Illumination System Efficiency" defined as follows:

D65 Lumens out
---------------------- = 33% (max) <---- for 1-DLP
Light Source Lumens In

I hope it's clearer now.
I'm not talking about LED duty cycle.

from CES:

"PhlatLight Home Theater Projector
Luminus will demonstrate a home theater front projector prototype powered by PhlatLight LEDs. The latest PhlatLight LEDs have achieved a dramatic increase in brightness, producing over 700 ANSI lumens, and the front projection system is on track to break the 1,000 lumen barrier in 2008. This will enable home cinema front projector manufacturers to provide the benefits of PhlatLight LED technology, which include instant on operation, ultra high contrast, uniform brightness and wide color gamut performance with the elimination of the need to replace lamps during the life of the product. "
http://www.luminus.com/content1245

http://www.luminus.com/

Having just bought a LED DLP set I found this write-up interesting... it's a long article with a few highlights quoted below.

http://www.soundandvisionmag.com/innerworkings/2601/inside-an-led-based-light-engine.html

There are many advantages to using LEDs instead of conventional lamps in DLP rear-pro sets. For one, LEDs are more reliable and last longer. Where the bulbs in conventional lamps cost several hundred dollars and need to be replaced every 3,000 to 5,000 hours, LEDs last about 120,000 hours — which means you can expect them to last for the life of the TV. LEDs also don't wind up in landfills like spent conventional bulbs, and they don't contain the hazardous substances (such as mercury, xenon, or metal halide) found in arc lamps.

Also, while bulbs require a warm-up period, LEDs turn on instantly. This not only allows the picture to appear immediately with full color and brightness, but (when properly synced with the DLP chip) can also help improve contrast and picture quality. And the simple optical design of PhlatLight LEDs, which use a single emitting area per color instead of an array of small LEDs, allows for shallower cabinet depths.

But maybe the biggest advantages of LEDs are in color reproduction and brightness. Because individual LEDs are used to create the primary colors, they can produce a wider, richer range of colors than sets using a white-light bulb and filters or color wheels to split the light into separate colors. For example, the PhlatLight PT120 chip set used in the Samsung RPTVs can produce about 40% more color than traditional lamp-based sets, according to Luminus. They can also help reduce or eliminate the visual artifact known as "the rainbow effect."

The PhlatLight PT120 chip set used in Samsung's RPTVs was developed and optimized for Texas Instruments' 1080p DLP chips. Each chip set includes three LED panels — one each for red, green, and blue — that have an emitting surface with a 16:9 aspect ratio that matches the TV display, eliminating the need for tapered light tunnels to reshape the projected image to the correct aspect ratio. The PT120's design allows the chip set's LEDs to produce enough lumens to illuminate a 61-inch rear-pro HDTV. The next generation of PhlatLight chip sets will likely be bright enough to light even larger screens. And the PhlatLight's low-profile design, which allows the primary optical elements to be placed very close to the device's emitting area, helps manufacturers shave inches off the depth of their rear-pro cabinets. One of Samsung's current Slim LED DLP sets, for example, is just 13 inches deep.

So, if I understand what you're saying:
For a 24Hz film, we refresh the LEDs at, say, 240Hz (for ease of calculation)

So, now we can show the same frame:
10 times - no blanks
9 times - 1 blank frame (11% brighter each of the 9 frames)
8 times - 2 blank frames (20% brighter each of the 8 frames)
.
.
5 times - 5 blanks (100% brighter each of the 5 frames)
.
.
.
1 time - 9 blanks (1000% brighter for the single frame)

to average the same frame brightness.

Of the 10 scenarios above, which is better?The above is the general idea, but I don't think it needs to be whole frames. They could just blank the last 20%, 10%, 5%, or whatever at the end of each frame, no matter the frequency. I really don't know which value would be best and figure it would probably depend on the individual. Those most sensitive to rainbows may want low or no blank time, while others may want a fair amount. This is one reason I would like them to make it a user option with multiple choices.
Or, would it depend on the brightness of the frame itself?
More blanks for dark frames
less blanks for bright frames?I hadn't really thought about that part. My first guess is that the motion issue matters more on certain frames than others, but might be brighter frames where having the blank time is more important. I know that with a pan across stars on a black background I can often see 2 of each star and I believe I've seen this effect even with film. It would be interesting to play with different blank times and see if there is an amount where I would just see one of each star (in motion).

--Darin

"I meant "Illumination System Efficiency" defined as follows:

D65 Lumens out
---------------------- = 33% (max) <---- for 1-DLP
Light Source Lumens In

I hope it's clearer now.
I'm not talking about LED duty cycle."

33% brightness because the LED is on 1/3 of the time (= 33% duty cycle) is not the same thing as losing 2/3 of the light by filtering.

33% brightness because the LED is on 1/3 of the time (= 33% duty cycle) is not the same thing as losing 2/3 of the light by filtering.

I understand.
What, you think, might be the (theoretical maximum) efficiency of a LED 1-DLP system then?

I've decided to rename the Thread title as it has been getting alot of attention. I think we need this thread as more and more people will be interested in LED as a light source. Hopefully we can keep it up to date. THanks

I just read at Luminus Devices website that the company will be showing a front projector illuminated by its proprietary LED ligth source Phlatlight at the SID conference taking place in LA on May 20th/22nd; Anyone planning on attending to see whether or not the company comes true ?
Luminus Devices released a statement before CES 2008 specifying that it would also be displaying a front pj at the show illuminated by Phlatlight LEDs; Did it come to pass or was it just hype ?

"What, you think, might be the (theoretical maximum) efficiency of a LED 1-DLP system then?"

Do you mean optical efficiency = [light out of lens]/[light emitted by LED's]?

The only losses I can think of are the reflectivity of the DMD, the iris, and the lens absorption.

An optics person could make a more educated guess than me, but my guess fort he first and third would 15 - 25% loss.

Someone like Darin could make a better assessment on the iris, but I'd guess between 10 and 50% depending on whether it's set for high brightness vs high contrast.

I just read at Luminus Devices website that the company will be showing a front projector illuminated by its proprietary LED ligth source Phlatlight at the SID conference taking place in LA on May 20th/22nd; Anyone planning on attending to see whether or not the company comes true ?
Luminus Devices released a statement before CES 2008 specifying that it would also be displaying a front pj at the show illuminated by Phlatlight LEDs; Did it come to pass or was it just hype ?

I didn't go to CES, but I did see a picture of the projector from one of the reports from the convention. I have no idea if the PQ was any good- the picture was of the display showed a lit room with a floor lamp next to the screen. Did anybody go and see it?

So, will these projectors pass the "hand puppet" test?

That is to say, in a movie scene that is faded completely to black, will one be able to place one's hand in front of the screen and NOT see a shadow?

That would be the "holy grail" of black levels, as far as projectors are concerned. The only technology I can think of that would exceed it would be a laser front projector. Technically, though, a laser front projector would still be a LED projector since the laser source would likely be a laser LED.

He has never been called a wise yoda if I understand the meaning of yoda. Now he is wise but given his keeping in shape, he just doesn't look like a yoda. Now Tryg sort of looks like an unshaven yoda but wise?

Now Tryg sort of looks like an unshaven yoda but wise?

Isn't Tryg a bit big for Yoda? Maybe if he grew a little more hair, I could see him as Chewie.

regarding police cars;
the alternators on the cars were not beefy enough for the old lights. the car would stall out with a dead battery. , with labtops,lights and so on. They went to the more efficient emergency lights.

This is funny to me since I discussed this with my works city garage Machanic
today.

Larry

I've been following this topic and this thread with interest for a while.
Have any of you seen the latest press release by Luminus?
I can't post links here, so just go to the Luminus website (luminus.com), and you'll see it on the top right side.

It seems like we’re not too far from seeing LED DLP home theater front projectors on sale. According to the press release they are displaying another prototype with TI at the InfoComm in Las Vegas right now. Apparently we will see them coming to market before the end of the year.

I wonder whether this prototype uses the PT120 chipset, like the one at SID, and which brightness it will be able to achieve. Also, considering that previous LED DLP projectors based on the PT54 chipset (e.g. LG HS101) are on sale for less than $1000, what are the chances of this next generation to be sold for less than $2000? That would even justify moving this thread to another category from "$3000 and up" :)

Also, considering that previous LED DLP projectors based on the PT54 chipset (e.g. LG HS101) are on sale for less than $1000, what are the chances of this next generation to be sold for less than $2000? If they are 1080p, close to zero chance of that out of the gate IMO. If somebody does a 480p or 720p model, then a lower price like that is a lot more likely.

--Darin