The optics could be a limiting factor with digital projectors !

Alan,


I think what you are seeing and reacting to is very common with people viewing different resolution devices. I have heard over and over that the lower resolution devices look "sharper" and and in most cases they do, especially when you are at a distance where the pixel structure disappears.


But, what does sharper mean and is it always good? If you were displaying a high bandwidth 1080i source, the 800x600 device might look "sharper" but I can guarantee you that it isn't displaying as much "information". I've had the chance to compare live 1080i camera feeds on 1024x768, 1280x1024 and our 1365x1024 devices and there is a visible difference in the amount of fine detail visible in the image.


When you have very fine detail it and it can be displayed by your device, it begins to look more lifelike and had depth...the world we see is analog and we don't see a world of sharply defined edges of pixels.


Take a high resolution print...even though with a good 1440 dpi color inkjet printer you can't see the pixels unless viewed through a lupe, compare it with a high resolution printed brochure page...you cannot see the individual pixels in either one but one sure looks better to the eye.


I noticed the same thing when I viewed our 2048x1536 projector a while back...false sharpness defined by the edges of the pixels was replaced by "information" that allowed you to see deep into the picture. A case in point was a black and white clip from "Rumble Fish"...there was a scene looking down a sidewalk along a brick front building. With a 1365 or 1280 display device, the texture of the bricks disappeared a few feet down the wall, it became a blur with no detail...with the 1920 resolution, you could see the texture well down the sidewalk and you got a much better sense of depth and realism. In a clip from "The Truman Show", you could read the print on the papers on his desk with the 1920 device...just a blur with 1280 or 1365 devices...it was amazing what is missing with the resolution that some are trying to foist on us for Digital Cinema. Hopefully we can deliver the same resolution for Home Theater in the not to distant future...source materials may then be the big issue.


Back to your question about lenses though...lenses are one of the most expensive components in high quality projectors. Especially if you want to maintain high contrast ratio and good corner to corner focus in high resolution imaging devices. As you say, dispersion in the lens from either plastic lenses or poor quality glass and inexpensive or non existant coatings can wreak havoc on local area contrast ratio.


Low dispersion glass and high quality optical coatings are not cheap and in fact are, in the case of the best glass, in short supply.


Well, off to NAB in the morning...


Regards,




------------------

Tom Stites

Director, Business Development

Digital Systems Division

JVC Professional Products

"My opinions do not necessarily reflect..."

I must say that the optics on the Davis DL450 projectors are really good. When focused you can see the post of each mirror on the dlp chip in the projected image. Here is a pic of the DLP mirrors up close http://www.dlp.com/dlp/resources/images/bt_06.gif When you look at the projected image from the DL450 you can see the post in the middle of each mirror, its pretty cool!


------------------

lance

I'm looking for a job!

Resume online at http://resumes.dice.com/lkstitch


[This message has been edited by dlptheater (edited 04-20-2001).]

Interesting topic, however, I must agree with JoeFloyd regarding the quality of modern optics. I've been into amature astronomy for many years and there have been great strides in maintaining exceptional primary lens resolving power while reducing lens manufacturing costs. I would think that micro convergence, or rather, micro mis-convergence would have a more deleterious impact on image quality than the lens itself. As an owner of a single chip DLP and the fact that when I am up against the screen I am able to resolve the individual mirrors tells me that the lens is giving me all it can. It would be interesting if you could do the same comparisons between single chip DLP's and the Plazma display, both of which should realistically be SVGA devices.


This reminds me of the early public demonstrations of HDTV displays, on side by side comparisons many commented that the NTSC image looked sharper or the H-def image looked soft. They failed to pick up on the tremendous resolution difference. I agree with Tstites that you probably should be comparing devices with the same fixed resolution to avoid this extraneous effect.




------------------

Regards

GMan

I believe what Alan observed is not uncommon and is actually quite logical. It all comes down to the difference between resolution and the perception of sharpness, which are two quite distinctively separate things.


Resolution has the most straightforward definition: the smallest detail you are able to see within an image. For an ideal discretely adressable pixel panel, resolution is directly proportional to the number of pixels – the more pixels, the higher the resolution.


Sharpness is a more complicated parameter. What our visual system relies on for determining sharpness is the distance for a transition from a brighter area to a darker area as well as the brightness ratio of the areas. One could perhaps define sharpness as rate of brightness change. This means that the more well defined 'edges' the image holds, the sharper it looks. In the case of a pixel panel there are only discrete states so the distance is not a factor here, but rather the brightness ratio.


Consider an input signal that holds a relative brightness sequence of 1,2,4,2,1... The higher resolution device is able to accurately track the brightness variations and outputs the 1,2,4,2,1... sequence. Each pixel to pixel transition thus has a factor 2 relative brighness ratio. Now consider a device with half the resolution of the high rez device. The brightness steps are now twice as crude and this device outputs a 1,4,1... sequence. The pixel to pixel transitions now have a factor 4 relative brighness ratio which makes the lower rez device look sharper.


A CRT projector has continuosly varying brightness in the direction of the scan lines which is why a true resolution 960x1440 CRT will look softer than a 600x800 digital projector (the degree of overlapping lines/spot focus also has some influence).


Variations on this theme are also seen when edge enhancement is applied. By artificially increasing the brightness ratio for a specific frequency range of the signal (a frequency range translates to a size range in the reproduced image, or details of a certain scale), the image is perceived as having more sharpness.


Some viewers, as myself, do not readily accept the 'sharper' image and only perceive it as more 'electronic' or artificial looking, while the vast majority of viewers go for the sharper look.

The sony was zoomed to the same size as the plasma.


There was definetly some artifacts in the sonys picture due to the lens. Lens flare which appears as a color, mostly blue, around some bright edges

and also the picture was not as sharp around the outer edge of the picture.


The sony is a top performer on its own. Had I not done a side by side I would not have noticed this. I did pick up on the lens flare. This is also common in theaters when the theater is to cheap to buy a good quality lens. You can see this during the scrolling titles at the end of a movie and the white letters look like they are out of convergence with red or blue fringing.


Im not picking on the sony. This is very minor and the image poduced buy this projector is outstanding. Thats why I am using it!

But I do see a slight loss in picture quality due to the lens.


Stop and think about it.

The image from the plasma pass through nothing.

The image from a front projector pass through 5 or 6 eliments of glass!



------------------

Thanks Very Much!


Alan Gouger

AV Science
www.avscience.com


[This message has been edited by Alan Gouger (edited 04-21-2001).]

The resolving power of the lens is usually not a problem. Modern optics approach the theoretical performance limits, and even cheap lens systems in cheap projectors should not cause too many deleterious effects.


The optical path that leads up to the lens may be the cause of some of the problem. Trying to keep the projectors small and light while still delivering lots of light is sure to result in less than optimal light paths. I know that my projector has way too much internal light pollution that if better controlled would yield greater contrast which also add to the perceived image sharpness.


Even taking that into account, depending on the physical image size, the pixel structure, and the viewing distance the human visual system may respond differently than intuition would suggest. For example, if the sub pixels of the plasma are arranged in non linear offset pattern it's harder for the human visual system to pick out the individual pixels. This is especially true when viewing moving images. Color TVs and other video displays have used this trick for years. The older Sony 16:9 LCD projector also used this trick to reduce the screen door effect.


Since one of the features that the human visual system looks for are edges, a non liner offset pattern of pixels will tend to soften the edges as perceived by humans. Kind of like a full screen antialiasing but built into the pixel structure. Since the resolution has not changed the actual details of the image are still present, this appears to the human eye like the image is sharper.


There is good paper from Microsoft Research that goes into some of these issues in great and gory detail. If anyone wants me to dig it up, I can give it a shot.

Dan & Dan

No problem, in fact I'll post a couple of pictures aswell but you'll have to hold on for a couple of days.

The setup: PLC-XP21, LNS-T30K long throw lens, Draper Ambassador screen, AnalogWay SmartcutII scaler, complete Sonance directional in-ceiling eight speaker system (6.1 with two centers) and two Jamo THX 1015 active subs


Paul


------------------

Paul O'Hale
Your lies are like bananas,they're yellow and come in big bunches!

Leslie Neilson,Wrongfully accused
www.mcgahon-oneill.com