The stroboscope effect is pretty obvious even during motion at much higher frequencies. I just tried scanning with a pen in front of an LCD monitor. I can see distinct discontinuities despite a backlight frequency in the 200-300Hz range. I cannot see that in daylight. Try it.
PS I am human
PS I am human
The B/W sensors (rods) are much faster than the color sensors (cones). If you look directly at a stationary image, then you won't perceive much flicker above 15 Hz. This is our "shape" perception system that involves some pattern recognition neural net stuff in the retina, optic nerve, and brain.
But there are all sorts of things you can do to make other perception systems take precidence: Like rapid movement, which is a more primitive connection between the rods and the brain.
Even the part of your retina has an effect. There are a lot more rods around the outside and more cones in the center. So most people can detect 60 or 120 Hz flicker (ie. fluorescent lights) only in their peripheral vision.
This is co-evolved "monkey vision": We need to select ripe fruit by the color
, but we also need to leap when we detect a predator attacking from behind. 
But there are all sorts of things you can do to make other perception systems take precidence: Like rapid movement, which is a more primitive connection between the rods and the brain.
Even the part of your retina has an effect. There are a lot more rods around the outside and more cones in the center. So most people can detect 60 or 120 Hz flicker (ie. fluorescent lights) only in their peripheral vision.
This is co-evolved "monkey vision": We need to select ripe fruit by the color
, but we also need to leap when we detect a predator attacking from behind. I confirm that, concentrating less, stripping off shapes, helps flicker perception.
It seems the higher the contrast and speed involved, the easier that perception of stroboscopic motion. For the most part cinema movies and Tv can get away with it, also by lowering edge contrast with the so called motion blur. Thus fluid animation is possible at quite low frame rates, maybe even less than 15fps. Playing with a pocket laser on the wall I can see a thread forming sooner than that. With more power and a faster motion it should be possible to build an entire xy picture by modulating the spot, CRT style. Hmm, maybe a new DIY projector idea, with no objective lens!
On the other hand, the famous DLP rainbow effect seems to affect people differently, or is it that some are just more picky?
Many predators attack by night. I wonder why we haven't grown IR vision as well 🙂
It seems the higher the contrast and speed involved, the easier that perception of stroboscopic motion. For the most part cinema movies and Tv can get away with it, also by lowering edge contrast with the so called motion blur. Thus fluid animation is possible at quite low frame rates, maybe even less than 15fps. Playing with a pocket laser on the wall I can see a thread forming sooner than that. With more power and a faster motion it should be possible to build an entire xy picture by modulating the spot, CRT style. Hmm, maybe a new DIY projector idea, with no objective lens!
On the other hand, the famous DLP rainbow effect seems to affect people differently, or is it that some are just more picky?
Many predators attack by night. I wonder why we haven't grown IR vision as well 🙂
Polymer Optics Ltd may have the solution for a led based projector. In the brochure (.pdf file) they show a single sheet that has multiple hexagonal collimators that looks like it could provide a seamless source of collimated light from multiple leds. Maybe one of our UK members could contact the company and get more information.
individual collimator lenses
You might be able to do the same thing much cheaper with an array of small fresnel lenses. For example, from:
http://www.3dlens.com/shop/fresnellens.php
You can buy a pack of 100 credit card-sized 160 mm fl fresnels for $32 US, with much lower prices for higher quantities. Then you could cut them down to smaller squares or hexagons and assemble them into a fresnel array custom fit to your LED array. The geometry is pretty simple: If you put an LED 160 mm from the fresnel (smooth side facing the LED), then you would get a pretty parallel beam out the other side. You could adjust that distance slightly closer to make the individual lens beams overlap more or less with adjacent lens beams. Of course, the output angle of the individual LEDs should match the size of the individual fresnels at the 160 mm distance, or you would be losing light.
Then you send your big parallel beam through the LCD and a big field fresnel to converge it into the projection lens. 3dlens also sells some field fresnels that match popular projection lens choices. (330 mm fl for a 320 mm fl OHP lens, or 550 mm fl for an 18" fl opaque projector lens)
You might be able to do the same thing much cheaper with an array of small fresnel lenses. For example, from:
http://www.3dlens.com/shop/fresnellens.php
You can buy a pack of 100 credit card-sized 160 mm fl fresnels for $32 US, with much lower prices for higher quantities. Then you could cut them down to smaller squares or hexagons and assemble them into a fresnel array custom fit to your LED array. The geometry is pretty simple: If you put an LED 160 mm from the fresnel (smooth side facing the LED), then you would get a pretty parallel beam out the other side. You could adjust that distance slightly closer to make the individual lens beams overlap more or less with adjacent lens beams. Of course, the output angle of the individual LEDs should match the size of the individual fresnels at the 160 mm distance, or you would be losing light.
Then you send your big parallel beam through the LCD and a big field fresnel to converge it into the projection lens. 3dlens also sells some field fresnels that match popular projection lens choices. (330 mm fl for a 320 mm fl OHP lens, or 550 mm fl for an 18" fl opaque projector lens)
High Definition LED Front Projector
Prism, Inc. is planning on unveiling a secret research and development project at CES 2007. A source inside the company says the project is a high definition (1080p) front projector that uses light emitting diodes (LEDs). In contrast to the LED pocket projectors currently on the market (example), this device will produce hundreds of lumens at cinema quality resolution and contrast. The best part, however, is the use of LEDs as the lamp because the lifetime of the projector is years of continuous use, it produces brilliant and lifelike colors, and it's whisper quiet. Check out Prism's projector if you're a hardcore gamer, a movie buff, or someone who wants the biggest screen on the block.
Prism, Inc. is planning on unveiling a secret research and development project at CES 2007. A source inside the company says the project is a high definition (1080p) front projector that uses light emitting diodes (LEDs). In contrast to the LED pocket projectors currently on the market (example), this device will produce hundreds of lumens at cinema quality resolution and contrast. The best part, however, is the use of LEDs as the lamp because the lifetime of the projector is years of continuous use, it produces brilliant and lifelike colors, and it's whisper quiet. Check out Prism's projector if you're a hardcore gamer, a movie buff, or someone who wants the biggest screen on the block.
6.2 million LEDs? (1920 x 1080 x 3)
I wonder how they manage to cool this array of LEDs on a chip, enough to get hundreds of lumens out of it. LEDs are not all that efficient, so they generate more heat than light.
They also don't work very well at elevated temperatures. When they get too hot, they put out even less visible light and more IR.
Doesn't mean it isn't possible. But it may have some limitations most of us miss at first glance. Like maybe an expensive Peltier cooling system.
I wonder how they manage to cool this array of LEDs on a chip, enough to get hundreds of lumens out of it. LEDs are not all that efficient, so they generate more heat than light.
They also don't work very well at elevated temperatures. When they get too hot, they put out even less visible light and more IR.
Doesn't mean it isn't possible. But it may have some limitations most of us miss at first glance. Like maybe an expensive Peltier cooling system.
Guy
Your idea using credit card fresnels..great creative thinking !
Do you think though you would see lines where each fresnel edge touches ?
Cheers
Your idea using credit card fresnels..great creative thinking !
Do you think though you would see lines where each fresnel edge touches ?
Cheers
I'm thinking on using 4.5mm so called "Ultra bright white" LEDs.
My rough calculation says that ca 400 to 700 LEDs are needed to get the same light behind the lcd's 16:9 format as 150W HQI bulb (13000lm) would give.
The plan:
* LED array format 16:9 (ca 2 more efficient than bulb)
* Bend each led so that they point to the same point, so that light goes through lcd directly to objective-len. And get rid of fresnel. (more light due to "no fresnel")
* LED array at least as big as lcd, (possibly bigger)
The projection of 1 LED beam should fit onto the objective-len, because of that 16deg LED must be as close as 36cm for 90mm len.
The only worry, i currently have, is the specter of the leds.
My rough calculation says that ca 400 to 700 LEDs are needed to get the same light behind the lcd's 16:9 format as 150W HQI bulb (13000lm) would give.
The plan:
* LED array format 16:9 (ca 2 more efficient than bulb)
* Bend each led so that they point to the same point, so that light goes through lcd directly to objective-len. And get rid of fresnel. (more light due to "no fresnel")
* LED array at least as big as lcd, (possibly bigger)
The projection of 1 LED beam should fit onto the objective-len, because of that 16deg LED must be as close as 36cm for 90mm len.
The only worry, i currently have, is the specter of the leds.
I don't know what i meant with "using 4.5mm so" 🙂
"4.5lumens LEDs" is correct 🙂
Currently I have 250 of them = ca 1100LM and with that design i described, only restriction of LCD and the len must be subtracted to get the lumens on wall. In terms of quantity this should be fine, but i'm not sure about the quality.
I'm pretty sure that it isn't that simple task.
My 14.1" LCD's 16:9 region is 298x168mm=ca500cm2
When allocating the 250 LEDS evenly (2 cm2/LED), 1 led for every sqrt(2cm) = 1.4cm. It will be 21 x 12 LEDS. This is when the leds touch the lcd, but then it will be veeery uneven spotty light.
Since these are 12deg LEDS - they must be 10cm from the LCD for the beam to be 2cm wide (in case this 12deg is the 50% birghtness range?). Since the brightness doesnt vary linearly with the angle, the distance should be even bigger for spotless light.
But when moving leds farther from the LCD the beam would be too wide at the len. So lets keep them at 36cm from len.
It can be solved when moving LCD closer to len, lets say at 18cm. Then the LED array must be widened also twice, and there will be 1 led for every 28mm. And also the len should be accepting that wide angle... don't have enough time to write more clearly...
"4.5lumens LEDs" is correct 🙂
Currently I have 250 of them = ca 1100LM and with that design i described, only restriction of LCD and the len must be subtracted to get the lumens on wall. In terms of quantity this should be fine, but i'm not sure about the quality.
I'm pretty sure that it isn't that simple task.
My 14.1" LCD's 16:9 region is 298x168mm=ca500cm2
When allocating the 250 LEDS evenly (2 cm2/LED), 1 led for every sqrt(2cm) = 1.4cm. It will be 21 x 12 LEDS. This is when the leds touch the lcd, but then it will be veeery uneven spotty light.
Since these are 12deg LEDS - they must be 10cm from the LCD for the beam to be 2cm wide (in case this 12deg is the 50% birghtness range?). Since the brightness doesnt vary linearly with the angle, the distance should be even bigger for spotless light.
But when moving leds farther from the LCD the beam would be too wide at the len. So lets keep them at 36cm from len.
It can be solved when moving LCD closer to len, lets say at 18cm. Then the LED array must be widened also twice, and there will be 1 led for every 28mm. And also the len should be accepting that wide angle... don't have enough time to write more clearly...
The new led to get?
"P4 is making history in the lighting industry as the alternative to conventional fluorescent and incandescent lighting sources," said Jung Hoon Lee, CEO of Seoul Semiconductor. "Conventional LEDs have been known to emit more than 100 lumens with several dies. Seoul Semiconductor's P4 is the only LED product in the world with 240 lm via a single die."
http://displayblog.wordpress.com/tag/led/
Seems possible to use a single led instead of a whole panel. Especialy if they reach 145lm/W! in 2008.
I'd like to know if the commercial led projectors can be moded to use such leds. Anyone seen one inside? The cheapest one is close to 300$, but isn't very bright at 11lm.
"P4 is making history in the lighting industry as the alternative to conventional fluorescent and incandescent lighting sources," said Jung Hoon Lee, CEO of Seoul Semiconductor. "Conventional LEDs have been known to emit more than 100 lumens with several dies. Seoul Semiconductor's P4 is the only LED product in the world with 240 lm via a single die."
http://displayblog.wordpress.com/tag/led/
Seems possible to use a single led instead of a whole panel. Especialy if they reach 145lm/W! in 2008.
I'd like to know if the commercial led projectors can be moded to use such leds. Anyone seen one inside? The cheapest one is close to 300$, but isn't very bright at 11lm.
need help
hi,
i am trying to build LED based projector.
initially i tried with 100 LEDs with 50000mcd each ( angle was not mentioned on the packets) for a 6" LCD screen. got image , was sharp but not bright enough. have packed them on a 13cm x 13cm panel. runs on 12v 3 mA adapter. cant add more LEds as the size of the panle will be much bigger than the LCD itself.
to push up brightness, i am planning to add 2 CFLs 30 watt each giving additional 3500 lumens. yet to take tests with them.
if they work, nothing like it.
has anyone tried such dual light source?
hi,
i am trying to build LED based projector.
initially i tried with 100 LEDs with 50000mcd each ( angle was not mentioned on the packets) for a 6" LCD screen. got image , was sharp but not bright enough. have packed them on a 13cm x 13cm panel. runs on 12v 3 mA adapter. cant add more LEds as the size of the panle will be much bigger than the LCD itself.
to push up brightness, i am planning to add 2 CFLs 30 watt each giving additional 3500 lumens. yet to take tests with them.
if they work, nothing like it.
has anyone tried such dual light source?
I have tried an array of 25 CFL with a 15" LCD, 550 mm fl field fresnel, and a 22" fl opaque projector lens. (I got them at a utility company subsidised price of $0.50 US for each 25 Watt lamp.)
The results were dissappointing. Most of the light from the CFLs goes off in directions that do not end up in the projection lens, so the image is very dim. I could also see the ill effect of using multiple light sources: Each LCD pixel sends rays to every part of the projection lens, so the resulting screen pixel suffers from all the distortion possible. The result is blurry pixels.
It should be very cheap and easy to try, so you should do it. But I don't think it will work very well.
When you use a point-source like a MH lamp. it is easy to get most of the light reflected and refracted into the projection lens. When my projector is adjusted properly, I can stand at the screen and look back into the lens (wearing dark glasses of course!). I see a fairly small spot of bright light that comes from about 1/10 th of the lens area. As I move my head across or up and down the screen, that spot moves within the lens area. This demontrates that the light is well-colimated so that each screen pixel's rays are only coming through a small part of the projection lens. This minimizes the effects of lens aberrations and distortions. I think it is analagous to the image-sharpening effect you see in photography when you change to a smaller lens aperature.
The results were dissappointing. Most of the light from the CFLs goes off in directions that do not end up in the projection lens, so the image is very dim. I could also see the ill effect of using multiple light sources: Each LCD pixel sends rays to every part of the projection lens, so the resulting screen pixel suffers from all the distortion possible. The result is blurry pixels.
It should be very cheap and easy to try, so you should do it. But I don't think it will work very well.
When you use a point-source like a MH lamp. it is easy to get most of the light reflected and refracted into the projection lens. When my projector is adjusted properly, I can stand at the screen and look back into the lens (wearing dark glasses of course!). I see a fairly small spot of bright light that comes from about 1/10 th of the lens area. As I move my head across or up and down the screen, that spot moves within the lens area. This demontrates that the light is well-colimated so that each screen pixel's rays are only coming through a small part of the projection lens. This minimizes the effects of lens aberrations and distortions. I think it is analagous to the image-sharpening effect you see in photography when you change to a smaller lens aperature.
Yes, but you may still get lateral colour distortion. If one pixel gets rays only from the perifery of the lens it may be even worse than the average from the whole area.
It doesn't pay to use multiple light sources, because of the etendue issues. Not to mention this zillion leds approach is a patient nightmare. I think single efficient led dies, and in the near future laser leds are the next steps. Can't beat simplicity, and order or collimation come together.
It doesn't pay to use multiple light sources, because of the etendue issues. Not to mention this zillion leds approach is a patient nightmare. I think single efficient led dies, and in the near future laser leds are the next steps. Can't beat simplicity, and order or collimation come together.
250 Lumen LEDs
And a Ushio 250 Watt S250-DD MH lamp puts out 20000 lumens, but everybody who has built a projector using that lamp is much happier when they change to the S400-DD with its 33000 lumen output.
Even if we say an LED-based projector can be more efficient because the LED output is in a much smaller cone, I think we would still need 10000 lumens. So that would require 40 of those high-powered LEDs. And the result would still be very dissapointing because of the etendue effects, not to mention the inconvenient fact that LEDs are less efficient heat-wise than MH lamps. So you would need MORE heat removal, not less.
When somebody builds a 10000 lumen point-source LED, let me know. Of course, it would be somewhat difficult to build a 3 volt 67 amp power supply to run it!
And a Ushio 250 Watt S250-DD MH lamp puts out 20000 lumens, but everybody who has built a projector using that lamp is much happier when they change to the S400-DD with its 33000 lumen output.
Even if we say an LED-based projector can be more efficient because the LED output is in a much smaller cone, I think we would still need 10000 lumens. So that would require 40 of those high-powered LEDs. And the result would still be very dissapointing because of the etendue effects, not to mention the inconvenient fact that LEDs are less efficient heat-wise than MH lamps. So you would need MORE heat removal, not less.
When somebody builds a 10000 lumen point-source LED, let me know. Of course, it would be somewhat difficult to build a 3 volt 67 amp power supply to run it!
First of all, I don't see any problem with making a ~150w power supply. 
The point is, utilizing LED's as a projector light source is completely feasable, I never said it was better than a halide bulb. Also consider this, alot of people dont need an 18 foot wide picture, I would be content with a 4 or 5 foot image, as I don't have a gymnasium in my home to house an enormous screen.
I've noticed that alot of people have tried 5mm led's, which is obviously a complete waste of time and money. I was just making sure you guys were aware that an entirely new generation of high flux LED's exist now.
The point is, utilizing LED's as a projector light source is completely feasable, I never said it was better than a halide bulb. Also consider this, alot of people dont need an 18 foot wide picture, I would be content with a 4 or 5 foot image, as I don't have a gymnasium in my home to house an enormous screen.
I've noticed that alot of people have tried 5mm led's, which is obviously a complete waste of time and money. I was just making sure you guys were aware that an entirely new generation of high flux LED's exist now.
Sorry if I seemed abrupt. I really did mean it when I wrote "Let me know..." because I would probably try building an LED-based projector if the right LED existed. And it probably will, sooner or later. Lots of work is being done to make higher and higher power LEDs for architectural lighting. The high end of that has belonged to specialty MH lamps for quite some time, but LEDs have some attractive features.
It would be interesting to build a current-limiting power supply to drive a single LED with 67 Amps: I have done design work on marine voltage regulators and auxilliary battery charging circuits, so I can see a couple of ways to do it. But just delivering 67 Amps to anything presents its own set of problems. Like winding a custom transformer or choke with 8 gauge wire in the secondary... 😀
BTW: My experience with a "too dim" picture was on a 95" diagonal screen, not some behemoth. My projector-to-screen distance is just over 12 feet. The LED-based projectors I have read about can make an acceptably bright picture around 32" diagonal, but with a picture that small you might as well just buy an LCD TV right off the shelf.
It would be interesting to build a current-limiting power supply to drive a single LED with 67 Amps: I have done design work on marine voltage regulators and auxilliary battery charging circuits, so I can see a couple of ways to do it. But just delivering 67 Amps to anything presents its own set of problems. Like winding a custom transformer or choke with 8 gauge wire in the secondary... 😀
BTW: My experience with a "too dim" picture was on a 95" diagonal screen, not some behemoth. My projector-to-screen distance is just over 12 feet. The LED-based projectors I have read about can make an acceptably bright picture around 32" diagonal, but with a picture that small you might as well just buy an LCD TV right off the shelf.
- Status
- Not open for further replies.