quote: Marcel Birgelen

---

I don't understand how any real contrast improvement can come only from a laser source.

---

Exactly! X projector is going to be stuck at what ever X contrast ratio capability the prism/lens combinations is capable of, irregardless of the light source used. It can not improve it and Imax dispensed with the prism completely because of this. When you have light paths crossing each other that actually lowers the effective contrast ratio the system is capable of. This is why lenses some times utilize light baffles to eliminate back reflections inside the lens which would lower the effective MTF of the lens. D-Cinema lenses utilize some very ingenious series of light baffles so they maintain the highest MTF possible!

Mark
 

---

Posted by Steve Guttag (Member # 268) on 11-07-2017, 01:28 PM:
 
Actually, doing a bit of research, Laser Phosphor CAN improve its contrast ratio a little bit due to the lower etendue. Barco is claiming around 20%. I have not verified their implementation to prove/disprove it. But the people I trust that know about such things back up the potential of Barco's claims.
 

---

Posted by Marcel Birgelen (Member # 6801) on 11-07-2017, 02:06 PM:
 
Well, I tend follow Mark on this, unless I'm missing something here:

A light source with a smaller etendue should increase the efficiency of the light source, since there's less light getting lost in reflectors, lenses and other optics trying to focus the light on the DMD/LCoS imager. Yet, a more efficient light source alone doesn't imply a higher contrast in the image.

Besides the quality of both the lenses and the prisms, the contrast is primarily going to be defined by the quality of the imagers, the DMD or LCoS type devices. More precisely by the difference between the amount of light they're able to reflect towards the screen and deflect away from hitting the screen. A 100% perfect system should be able to control this per pixel and should reflect 100% of all the light in an on-status versus 0% of all the light directed at it in an off-status.

One thing where I think laser could help is, because the laser light in itself is more coherent, there will be less "non-directional" or "stray" light hitting the imager at an odd angle and causing leakage into the image at large. Although I'm pretty sure this effect exists to some extend, I have no real idea how much it affects current designs. There might also be other solutions for this than lasers, like polarizers. Obviously those will not enhance the efficiency of the light sources.

Another thing could be the lenses and prisms themselves, but this only applies for 3P/6P laser systems. The monochromatic aspect of the primaries will probably make it easier to correct for chromatic aberration, since you only have to account for three very discrete wavelenghts. This however, should increase the sharpness of the image, not so much the contrast.
 

---

Posted by Mark Gulbrandsen (Member # 72) on 11-07-2017, 07:43 PM:
 
Someone needs to step up to the plate and get rid of all the glass, meaning the Prism itself. It can't be THAT difficult, except getting around the Imax patents. Hey, others have done THAT over the years. LCD projectors have never used big prisms in them that I have ever encountered and they do their tricks with just high temperature plastic frames. A precision machined frame supporting high end front surface and partial mirrors is long over due in regular cinema. Then the laser light does not have to be combined before hand. Pump in each section of light to it's appropriate DMD and back out and combine it going into the projection lens itself. Not unlike the reverse of a three CCD TV camera.
 

---

Posted by Adam Martin (Member # 641) on 11-07-2017, 10:13 PM:
 
You seem to know a lot about this, so why don't YOU do it? Then we can all sit around bitching and moaning about your device and tell stories of how the equipment WE use has never had a problem.
 

---

Posted by Carsten Kurz (Member # 5396) on 11-08-2017, 06:36 AM:
 
LCD projectors do use prisms as well.

You can always increase contrast easily by sacrificing light. But who wants to do that, other than for screening rooms/post production or simulators?

There are other ways to recombine light, but prisms do not only have disadvantages, they also have a great advantage - stability, once built, there is no drift in registration. The Laser IMAX solution is very expensive in order to reduce that issue. And the Kodak/IMAX patents may prevent other companies from doing it commercially.

There are some references that the maximum achievable on/off contrast for current DLP imagers is around 10.000 to 12:000:1. While Laser IMAX comes close, other manufacturers slowly seem to achieve something in the 3000-5000:1 ballpark with their LIPs. Sony quotes a very solid 8000:1 for their UHP SRX-R 5xx series. I guess it wasn't that complicated to improve this towards the 10.000:1 they now quote for their phosphor laser - even their (non cinema) SRX-T615 UHP counterpart is quoted with 12.000:1. Dolby Vision/Cinema is a bit different as it not only cares about sequential contrast improvement with it's (expensive) 'local dimming' approach.


With exit/emergency lights, auditorium design and patrons clothes and faces bouncing light back to the screen, I guess it doesn't make sense to achieve much more than 5000-8000:1 on the projectors part. How that translates to in-picture/ANSI contrast for the different technologies is an interesting question. Usually, more work is spent on analyzing high end home cinema projectors than in evaluating DCI machines.
The Barco LHC machines seem to target these realistic numbers. Barco started to quote sequential and ANSI contrast figures for their RGB laser machines. From these numbers, the LHC machines look very promising. I hope they also improve on the speckle issue.

At some point, self illuminating screens will take care of that issue as well, and also improve in-picture/ANSI contrast greatly.

- Carsten
 

---

Posted by Tom Bert (Member # 5719) on 11-08-2017, 08:31 AM:
 

quote: Marcel Birgelen

---

A light source with a smaller etendue should increase the efficiency of the light source, since there's less light getting lost in reflectors, lenses and other optics trying to focus the light on the DMD/LCoS imager. Yet, a more efficient light source alone doesn't imply a higher contrast in the image.

---

quote: Marcel Birgelen

---

One thing where I think laser could help is, because the laser light in itself is more coherent, there will be less "non-directional" or "stray" light hitting the imager at an odd angle and causing leakage into the image at large. Although I'm pretty sure this effect exists to some extend, I have no real idea how much it affects current designs.

---

quote: Mark Gulbrandsen

---

Someone needs to step up to the plate and get rid of all the glass, meaning the Prism itself. It can't be THAT difficult, except getting around the Imax patents. Hey, others have done THAT over the years. LCD projectors have never used big prisms in them that I have ever encountered and they do their tricks with just high temperature plastic frames.

---

quote: Tom Bert

---

Using a prism reduces size, complexity and cost: there are good reasons to do so.

---

quote: Tom Bert

---

You're mixing up things here:
- "coherence" in the context of laser is something else, it's not spatial coherence
- what you call coherence here is actually the same thing as étendue: a low étendue/high-F-number system helps to boost contrast.

---

I do know that "coherence" isn't the right word here, hence the quotes.

I do understand why a "High-F-number" (I'll remember that one ) light source will be more efficient in terms of energy you put in v.s. the amount of light you get out of it (for the purpose of projection), but maybe you can explain why a "High-F-number" system eventually will improve the contrast of the image.

Is it because a "high-f-number" light source will cause less unwanted reflections inside the light path at large or are there other reasons for this?

quote: Tom Bert

---

Many LCD projectors are single chip: different world than 3-chip, which is the only way to achieve image quality that cinema expects/deserves.

---

quote: Mark Gulbrandsen

---

All 3-LCD projetors I have repaired, mostly Panasonic, all have integral light engines in a single plastic frame. The dichro filters, the LCD's themselves, and extra heat and trimming mirrors are all in that frame. Not that I am advocating using plastic in a DLP... But all I am saying is they never had a prism to start out with.

---

Why not go old-school and have three separate light paths like the CRT projectors did? They were so bulky, nobody even tried to combine the output of the three tubes back to one beam. Also, I somehow miss the old days of converging and edge-blending those things together for the better part of the day.
 

---

Posted by Dave Macaulay (Member # 813) on 11-08-2017, 04:31 PM:
 
The Kodak/Imax system is rather bulky, the light paths are complex and long inside the light engine case. And maintaining optical alignment and convergence through thermal changes is tricky. The prism system won out because it is smaller, simpler, and less prone to misalignment. Like Betamax vs VHS, both have advantages. Contrast is better with the Kodak/Imax system but no projectors except Imax Laser use it now... this is not just because of patents.
 

---

Posted by Carsten Kurz (Member # 5396) on 11-08-2017, 05:12 PM:
 
https://www.barco.com/en/News/Post/2013/6/26/Why-laser-is-just-another-part-of-the-image-quality-puzzle---part-two

---
In this installment, we’ll take a look at how étendue is impacting image quality. In addition, we’ll see how laser technology has a direct impact on the way light is transmitted, thus creating better contrast and uniformity.


Étendue

Étendue describes how big or compact a light source or bundle of light is. It is a combination of the surface and the angle at which light is emitted. The most specific thing about étendue (also the biggest challenge for optical designers) is that étendue can only grow. Applying this to our case above where we were thinking of combining multiple lamps to achieve a higher light input, we are confronted with the –relatively huge étendue of lamps: the arc where the light is formed is several mm long; furthermore the light is emitted in all possible angles. Combining more lamps only makes the system étendue even bigger. So if you want to make use of all this light, you will have to dimension your projector accordingly: a big light path, bigger chips, bigger angles, ... all too expensive. The nice thing about a laser light source is that the étendue is tiny: the surface of one beam is micrometer scale and thanks to the collimation the angular distribution is almost zero. This makes it easier to stack and scale lasers and accumulate a huge (but still low étendue) light source. Net result: the possibility of higher brightness, and better image quality!

Stray light

Lasers are not only about more raw power and pumping more light into the projector light engine; the higher collimation (i.e. a lower étendue, our word of the day) also has a direct positive effect on efficiently transporting light and an indirect contribution to better contrast and uniformity. In any projection system you want to control the light flowing through the system as tightly as possible: light bouncing off surfaces and going places where you don't want it, never has a positive impact on the image you are trying to project. Looking at digital cinema, the image-generating component (the chip) always has a certain light acceptance angle: any light in the system outside that cone is stray light. Stray light is out of your control and exits the lens on top of to the actual image you are trying to create. Added to a very bright image, the impact will be small. Added to dark scenes, it will ruin your contrast ratio. Existing cinema set-ups operate at a contrast of around 2000:1. The geometry of the lamp light distribution plays an important limiting role here. When taken into account in the system design, laser light sources and their more collimated nature can boost contrast ratio. Do note that other components (such as the chip) are still impacting and an infinite contrast ratio will never be possible.

The power of uniformity

Another important part of the image quality puzzle is uniformity. Whereas brightness and contrast ratio are local parameters (measured in one location on the screen, sometimes averaged out); uniformity is a global parameter. It quantifies how well the image quality is maintained across the entire screen, from center to edges. A good uniformity is achieved (among other things) by illuminating the surface of the imaging chip with a nicely filled and rectangular-shaped bundle of light. As mentioned above, the angles of this light cone affect the contrast ratio; the shape and fill of the cone impact uniformity. In the case of lamp-illuminated projectors, at the very start of the light emission, we start from an ‛arc-shaped’ source of light: between the anode and cathode of the lamp, a small 'bridge of light' is where the actual light generation happens. To achieve perfect uniformity, the optical components between the lamp and imaging chip (reflector, light rod, ...) have to reshape this arc (long, thin) into the rectangular shape of the chip. If you are not willing to compromise the size (and cost) of the set-up, this is a very hard thing to do. That is why the current cinema DCI spec for uniformity takes realistic values around 80%. This is where the laser comes in. Light generation no longer starts from an arc between two electrodes, but as a stable collimated bundle out of an aperture. This makes the life of the optical designer easier: by opening up the bundle (uncollimating) or adding up several bundles (here our good friend étendue again) it is possible to fill that rectangular aperture that the imaging chip is built for. The net result is not only a better uniformity and better image quality; but ‒ once again ‒ a more efficient light use since you don't have to start applying compensation tricks for inherent lamp non-uniformity. Do note that ‒ just like in the case of contrast ratio ‒ other parameters, like the lens, also impact the on-screen uniformity. Even parameters outside of the projector’s control, like screen gain, impact the perceived uniformity. So yes, laser can positively impact uniformity; but it will not magically lead to 100% perfect uniformity. This is true of any laser-illuminated projector.
---

- Carsten
 

---

Posted by Tom Bert (Member # 5719) on 11-09-2017, 01:19 PM:
 

quote: Marcel Birgelen

---

maybe you can explain why a "High-F-number" system eventually will improve the contrast of the image.

Is it because a "high-f-number" light source will cause less unwanted reflections inside the light path at large or are there other reasons for this?

---

quote:

---

(I believe I wrote it :-))

---

Doesn't sound very scientific.

Well, thanks for essentially confirming my "theory". Still, with all the reflectors and collimator lenses in place in an ordinary Xenon light path, I didn't expect there would actually be a 20% increase in REAL contrast due to less stray light of a "high-F-number" light source.

The highest contrast in DCI I've seen until now is the Dolby Cinema / Dolby Vision setup. We know they're using a dual Christie projector with 3P lasers. Yet, their setup seems to achieve a higher contrast than any native 3P Christie Laser setup. They claim this is due to the Dolby Vision specific modifications they made to the light enigine. Unfortunately, the information available on the kind of modifications is sketchy. Does anybody know what kind of modifications they did?
 

---

Posted by Tom Bert (Member # 5719) on 11-13-2017, 07:53 AM:
 

quote: Marcel Birgelen

---

The highest contrast in DCI I've seen until now is the Dolby Cinema / Dolby Vision setup. We know they're using a dual Christie projector with 3P lasers. Yet, their setup seems to achieve a higher contrast than any native 3P Christie Laser setup. They claim this is due to the Dolby Vision specific modifications they made to the light enigine. Unfortunately, the information available on the kind of modifications is sketchy. Does anybody know what kind of modifications they did?

---

They use a dual-engine system. Think of it as two 2000:1 projectors in sequence. 2000x2000 = 4M ... in theory (only)

The big con is light efficiency (lm/W): e.g. a 5% system in series with a 5% system yields 0,25% net efficiency. I've heard stories of the Dolby Cinema system consuming >20kW electrical power
 

---

Posted by Carsten Kurz (Member # 5396) on 11-13-2017, 12:12 PM:
 
http://www.patentsencyclopedia.com/app/20130321475

---
Abstract:

A zonal illumination system for use in a projector, comprising a source of illumination; a first modulation stage for dividing light from the source of illumination into a plurality of zones having respective light intensities based on zonal lighting intensity of an image to be projected; and a second modulation stage for modulating light from the first modulation stage to generate the image to be projected.
---

- Carsten
 

---

Posted by Marcel Birgelen (Member # 6801) on 11-13-2017, 03:46 PM:
 
I've heard them using a "stepped light engine" approach. I didn't read the whole patent and I've encountered it before. Someone referred to it as a possible scenario how the "Dolby Vision" HDR stuff works on the AVS Forum. But it describes a zonal system using white light with a "zonal modulator" that regulates the luminosity per zone. But I guess the Dolby Vision system doesn't combine the color primaries first, only to split them again afterward?

So, in this case, they're using a modulator per color, so that would either be three extra DMDs or LCoS like devices.

I was wondering if you could not develop a more elegant zonal system using laser sources. Since you've got a whole bunch of "high-F-number" sources at your disposal, isn't it possible to bundle them in an orderly fashion where every module represents a zone? It's one of the fancy aspects of lasers that you can modulate them at GHz speeds if you require to do so.

quote: Tom Bert

---

The big con is light efficiency (lm/W): e.g. a 5% system in series with a 5% system yields 0,25% net efficiency. I've heard stories of the Dolby Cinema system consuming >20kW electrical power

---