I notice there is a lot of quoting and not much new. I'd really like you to address this:



How so?

Also:



And while you're at it:



And maybe cite a source for:



And since we've already agreed (1) perfect alignment is unattainable and therefore it follows that some misalignment exists and (2) you assert it is not significant, can you support this by citing a source for or computing the actual amount of misalignment, or are you merely expressing your opinion it is not significant?

I'm OK with opinions (everyone has some), just not with opinions presented as fact.

Also, on a more constructive note, it occurred to me that one needn't restrict themselves to top->bottom scanning on some of the new systems. GLV for instance could be (it seems to me) arranged to scan 1080 from left->right (or right->left, whatever) and be an even smaller array.

Or even more interesting, if the scan was done this way and the array was a least common denominator of 480/720/1080 no external scaling would be needed - the projector could just gang a series of elements together.

For instance a GLV array of 4320 elements scanned from left->right would be able to directly display:

1080 - Gang 4 elements together.
720 - Gang 6 elements together.
480 - Gang 9 elements together.

Of course the same idea will work on a top->bottom scanned system, just with the need for a different sized array.

PS - Feel free to repeat this without citing the source, even though I believe it is an original idea.


[This message has been edited by charlie (edited November 26, 2002).]

"OK. So must the convergence error be greater than 0.1 to be visible, or are errors smaller than 0.01 visible? You really didn't add any useful info here."

Well, you edited my statement cutting off the 'useful' part. That's pretty cheap.

I said -Well, again... you 'suspect' wrong. Like I said... lets' stick with the 'real world'.
Ask around. There's millions of LCD projector owners out there. And lots of 3 chip DLP theaters across the country.

I guess you don't think the actual real world application of millions of 3 chip systems having no visible color convergence error is 'useful'??? Why don't you go look at ONE. You challenged me on my experience with them, and I told you my exp.

This is the same pathetic debate I had with you over my digital amp.

Where you would not just go LISTEN to one to hear the diff. yourself and tried to hide behind your sad 'missing data' defence. These weak 'data' questions just aren't needed to see/hear the 'real world' results in 'actual application'. And you should KNOW this.

"Supposing they probably have adjustments (what I said) and saying they must absolutely have adjustments are two quite different things. Can you see that?"

That's just NOT what you said. You just flat out wrongly guessed they had adjustments and I corrected you. Plain and simple as that.

What's your problem with understanding this fact?

"Again, I'm not being intentionally combative, I just want constructive dialog."

I don't mind if you're being combative or not, but you're certainly not being 'constructive' when you can't accept the fact that 3 chip designs don't need alignment. They are prefectly alligned as far as what the human eye sees at any normal viewing distance. All your 'data' measurments and guesses are meaninless when you can just LOOK at one of these displays and finally understand what I'm talking about. But you wouldn't want to be proved wrong so you ignore this.

If you want to debate something meaningless like the possibility of seeing a tiny hairline of misconvergence if you walk right up to a screen and stare at the pixels then it's a pointless debate.

Like I stated in my previous post... In the 3 chip DLP theater I've seen (the biggest screen display of a 3 chip system you're ever likely to see) I can see pixels once I walk so close that I'm not at any kind of 'normal viewing distance'... and I still don't see any color fringing error.

In the typical sub 12' size screens in home use your 'convergence' issue is even more pointless.

And you ignored my question about asking what you want in a display.
I think that'd be the best topic to follow here since it's the heart of what we're all looking to get -A better picture.

"GLV for instance could be (it seems to me) arranged to scan 1080 from left->right (or right->left, whatever) and be an even smaller array."

It DOES scan left to right. You should find out what it is and how it works before you say things like this.

"For instance a GLV array of 4320 elements scanned from left->right would be able to directly display:"

How do you define 'element' here? Whatever it is, it seems more complicated than what a 1080P GLV system is so I can't see why anyone would want to do that.

An array of 1080 vertical pixels can display 1080P and any lower resolution like the 480 and 720. The lower res. could be upconverted to 1080P internally and look better than 480P or 720P look natively.

HTPC users upconvert DVD's all the time to the native resolution of their projector improving the DVD's resolution beyond it's native 480P. The best native res. we can get it 1080i so a 1080P system can handle and improve every current common display resolution.

"PS - Feel free to repeat this without citing the source, even though I believe it is an original idea."

Yeah.. well we'll see if Sony can figure out what you're talking about then implement it. -heh

I've heard that images could be 'improved' but in practice I've never watched or heard of a scaler that was artifact free. By using a multiple of 480/720/1080 no scaling would be required. I like simpler where possible.

Also, I noticed that to modulate the intensity in a scheme like a GLV array there will be two basic ways, both with issues.

One could use a digital duty cycle modulation scheme (like DLP) and get very accurate results, but there would be in effect a binary stream of 1s and 0s (on spots and off spots) of sub-pixel size as the devices are switched on and off. Since they switch very fast it might not be noticed, as direct view color masks are not too noticable in most cases. Also, if more of one color is needed there will be places within the pixel where one 'pure primary' color shows unlike DLP, agian like direct view systems. This seems like it could reopen the door to moire problems if not treated with care. Also, if one used the concept of a larger than native (4320, etc.)array dithering could be used as needed.

The other option (according to you - i've not verified it) would be to use analog to modulate the GLV. This seems like a bad idea to me, but I'd need more info.

And back to our regularly scheduled programming:



Please cite a source to confirm 3 chip DLP theater systems have no convergence adjustment.

Here are my exact words:



As in (looking at a fast motorcycle) 'I bet that sucker can go over 160 mph easy'. It is a statement of expected or suspected fact, not an absolute statement of fact. I suspect anything not nailed down and thus not subject to piecework attack is bothersome to you and therefor understand that you feel some deep emotional need to nail it down and stomp it. Have fun, and then I suggest anger management courses before you piss someone off IRL that carries a gun or something. Beyond that, I have little interest in arguing over words. If you want that maybe former President Clinton can help you discuss what is is.



I'll make it simple for you. I said I suspect a convergence error of 0.1 pixel would be visible and 0.01 would not. You said I was wrong. So the possiblities are (1) error must be over 0.1 to be visible, (2) errors smaller than 0.01 are visible, or (3) you were wrong, or at least confused.

Please choose.

BTW I'm still waiting for a quantified specification of how much convergence error to expect (look up the big words if you have to) on a typical 3 chip DLP system, or a simple statement that your experience with 3 chip in general and a bit of 3 chip DLP in particular leads you to suspect (!} it is not a problem. In other words, if you have facts, spit 'em out, if not, be mature enough to say you have an opinion and strongly believe it to be correct. Presenting opinion as fact is weak.



I guess I missed the part where you told me how you knew DLP theater systems have no convergence adjustments, I'm sorry. Can you cite that source again? Thanks.

I'm curious how big the LCD projector devices are (optical part) since a bigger size imaging engine could make manufacture easier. Any idea? Any sources for this info?



I can almost agree on that statement.

Typical PCB materials display a CTE of ~17, and all materials display a CTE. A misalignment of 0.000055 (55 millionths) of an inch would cause a color registration error of 0.1 pixel on an HD2 device. This is well under the amount of movement the CTE of a PCB would typically be expeced to exhibit. Thus I am concerned about potential problems. I'm not asserting they cannot be engineered around. Engineers are resourceful people.

One last thing before you move on - can you explain in detail again the part about how color misconvergnece in single chip DLP systems causes the 'rainbow' effect? My understanding was it was due to time differences in the projection of the various colors and had nothing to do with the 'dots' failing to hit the same spot on the screen.

"I've heard that images could be 'improved' but in practice I've never watched or heard of a scaler that was artifact free."

Obviously you need hard to find 'data' to confirm things that people can just tell you or you can find out for yourself. I'm not a font of technical specs, but I don't have to be. Try AVS forum and you can talk to tons of HTPC owners who can tell you how upconverting DVD's can look better than the native 480P format. The artifact issue doesn't come up because you start with such a hi res format to begin with and the upconversion does far more good than any noticible harm, but YES there must be some form of damage.

Does this matter when the end result is a sharper picture that every HTPC user agrees looks better than the original DVD? No. Feel free to disagree without ever seeing this upconversion. It's your loss.

"By using a multiple of 480/720/1080 no scaling would be required. I like simpler where possible."

It's not simpler when you have an array that's made up of 4 times as many pixels as the original 1080P GLV design. But to clarify, this is still a lot less pixels than DLP needs in a 2D array. I mentioned before about how it's very easy for the GLV design to drastically increase it's resolution.

Let's just see if Sony even gets this sucker into production. and then into our homes and then increase the resolution to the range you suggest. I think I won't hold my breath, but I'd sure welcome it as much as you would.

"Also, I noticed that to modulate the intensity in a scheme like a GLV array there will be two basic ways, both with issues.-" "-but there would be in effect a binary stream of 1s and 0s (on
spots and off spots) of sub-pixel size as the devices are switched on and off."

You'll have to clarify your point here.

-"as direct view color masks are not too noticable in most cases." Yeah, mine isn't noticble at all.
Have you looked at current CRT HD RPTV's. from any reasonable distance they're rock solid.

"Also, if more of one color is needed there will be places within the pixel where one 'pure primary' color shows unlike DLP, agian like direct view systems. This seems like it could reopen the door to moire problems if not treated with care."

I'm not sure about the details of this train of thought.

"The other option (according to you - i've not verified it) would be to use analog to modulate the GLV. This seems like a bad idea to me, but I'd need more info."

Uh huh. Bad. Well, you do need more info.

Personally I'm only going by the reports I read years ago from people who actually saw the 1080P prototype FP in action. At the time the reporter from www.twice.com called it the best looking picture he's ever seen. There's a pic of the display in action on the GLV's web site not that you can tell anything for that pic other than the system does work and does exist (and has since 1998 or 99).

"Please cite a source to confirm 3 chip DLP theater systems have no convergence adjustment."

I saw pictures of the actual entire light engine on TI's web site, and read where they explain how the three chips are mounted so that they're alligned and don't need convergence.

I can't find it there anymore, but it might still be there. There's lots of DLP info still there.

I don't care if you don't believe this, there's already so much that you don't believe and won't even consider possible... let's just add it to your list of things you refuse to understand.

I'm not going to address your remarks on convergence again. That's beyond enough for me. If you don't get it by now you probably never will.

You may not have seen my remarks in a previous post. I changed computers during the post and had to go back and edit it to post what I meant to. By then you had already posted several times.

Go back and read my comments on my actual viewings of a 3 chip DLP system and LCD projectors.

"BTW I'm still waiting for a quantified specification of how much convergence error
to expect (look up the big words if you have to)-"

No need to insult. I answered this question in an above post.

"Presenting opinion as fact is weak."

Yeah... I present actual 'real world' application results from countless 3 chip system users. The 'fact' element is that they all (including myself) see no color convergence error at any normal viewing distance. I guess seeing no color error isn't good enough for you? Your loss.


"I'm curious how big the LCD projector devices are (optical part) since a bigger size imaging engine could make manufacture easier. Any idea?"

Yes, but why waste my time for you?

"-Any sources for this info?"

Why don't you go find your own info? Why ask me? You don't believe anything I tell you anyway. That info isn't hard to find. I know this info and othe related subjects like LCoS light engines and more about the GLV system than you. Go find out for yourself.

"One last thing before you move on - can you explain in detail again the part about how color misconvergnece in single chip DLP systems causes the 'rainbow' effect?"

I explained what I meant in my edited post above.

"My understanding was it was due to time differences in the projection of the various colors and had nothing to do with the 'dots' failing to hit the same spot on the screen."

Yes, since that's not possible with a one chip system. That's not what I meant though. Read what I wrote.

And you YET AGAIN ignore me asking 'what you want in a display' which should be the point of this thread. If that's of no interest to you then I won't answer anymore of your questions.

Actually, why AM I answering them anyway? You'll never 'get it'. Like you never tried the digital amps I told you about to HEAR the difference for yourself, you'll probably never SEE a 3 chip LCD or DLP system to see the actual color alligment in real life for your own eyes.

But hey... what would be the point of THAT right? Actually seeing ang hearing the real world results of theses designs... pointless right!?

I give up on you. Go hide behind your 'data crunching' as you refuse to live in the real world and see/hear these things you only wish to debate in theory.

Yeah - now that you edited it has more content. I read the original and thought, huh, that's all quote!?!???!!

You're usually much more verbose than that!



I was actually looking for a number. As in a quantity. Anything else is mere opinion, although it may be correct. And note that I'm not saying the above is wrong (or right either), just that it's not useful to me.

I have very good uncorrected vision, and for instance I can see the lines 'between' pixels on the Zenith 9K projector I watched (the same one I noted scaler artifacts on) so I'm concerned with any technology that has discrete pixels and multiple color sources. I don't think concern is unreasonable given the dearth of actual products to test and fiddle with.



Why should I believe functionality from LCD maps to DLP? And do you have a service manual for the LCD projector? If not, how do you know there's no adjustment?



How was that clear?



I've never looked at a color wheel in person, so I can't say, but I do know that most high quality filters are glass, and that dyes and pigments (even in plastic) vary in resistance to fading. So I still don't have an answer. Saying something is so doesn't prove it to be so, no matter how much you say it. Got any accelerated aging figures, or just guessing?



Actually it is much simpler. Switching 4/6/9 devices on at the same time to widen the effective scan line, then dumping the raw data into the device is trivial, unlike doing artifact free scaling which is impossible. Scaling can be very very good, but there will always (as long as computing power is finite) be some loss. I'd welcome at least the option to view without a scaler. And as you noted, 4K devices is still a pretty low count compared to other technologies.



Basically - wow, this is hard without a drawing. Imagine the scan 'bar' as it crosses the screen. As it enters the next 'pixel' the elements switch on. As it scans across the with of the pixel, at some point, depending on the intensity of the image in that pixel, it will switch off. Thus the 'brightness' of any discrete point on the screen will be 100% or 0%, a 1 or 0, although the average brightness for the area of the pixel would be correct.

Now imagine 3 colors, a non gray pixel, and you can see that not all the red will always be covered with blue, etc. Might be OK (esp since the resolution is so high) but they may need to dither to prevent moire.



Um, yeah, that would be because they have no color masks. Only color direct view sets use color masking to allow a single tube to generate multiple colors. Look really close at a direct view tube and you'll see the masking. Just one of the many reasons I actually like RPTV images better than direct view.



Yeah - I saw it somewhere too. It also didn't show the electrical connections, fasteners and other construction details, but I bet (as in I suspect) it has those too. A service manual or physical inspection is probably the only way to be sure, although the word of an experienced tech would be close to definitive.



I appreciate your input. I believe you saw no issues. Good for you. The fact you didn't see it doesn't mean it cannot exist. Do you understand that? It proves nothing except you cannot see any problem. There may or may not be a problem, and the fact that you see nothing encourages me to like the 3 chip concept but proves nothing else.

Perhaps as an artist you are more in touch with your inner child or something, but as an engineer I need actual proof. If that frustrates you, join the club. But there are guys who think like me and approach problems like I do that are right now getting the next great piece of gear ready for you to enjoy, so a rational approach is not without benefit.



It seems you have done some research on LCD and LCoS and I have not, so I was hoping that, as a courtesy, you would share any knowledge and even better, sources, with me and by extension all who might through some twisted sense of curiousity still be reading this thread.



Well now that you have added more to the original post it makes sense, but the original post was senseless to me.



You're not really very much, so whatever. I'm asking for facts and getting anecdotal evidence. They are distinctly different things.

And:

What I want in a display? For a 2D system I want a 1 mm thick flexible film direct view system that can display at 1200 dpi and 48 bits of color, 100 fps and should cost $5 per square foot. I'll use it for wallpaper.

I'll need some time to think about 3D, but I doubt there's any hurry.

You asked.

I'll be adding to the digital amp thread soon, as I'm finishing some measurements on my home system. It measures pretty well, but I'll post results there after the measurements are confirmed.


Have a good one.

Bringing some info to the table here to bat around. I ran into some items doing some quick surfing, as we will probably be purchasing another TV for a gift so I’m looking (again at the moment) into the current products. The debate in this thread centered on which technology will become the most prevalent in future projectors and Direct View had me digging for specifics on the topic. Azyran you had some excellent points that sent me off to dig for LCOS based products but I ran into some discouraging information. Regardless of the merit of which technology should earn the backing dollars of large manufactories. I wonder. First the manufactories have to get the products to us…..to see which the masses will jump to and on….and some of what I found does not look good for Locos at THIS moment. I ran into multiple names just putting on the market their various versions of LCD and DLP including but not limited to: Samsung, Sharp, Philips, Mitsubishi. Yet I ran into this announcement concerning the withdrawal of the
LCOS based Scenium L50000 dated July 2002.
http://www.insightmedia.info/news/ThomsonSTORY.htm

On the opposite side is this Co in CA promoting their LCOS components for manufactories.
http://www.elcos.com/ and I also ran into a company (can’t remember whom) who just sold a large double order for their equipment to 2 undisclosed company’s for inspecting LCOS based panel production.

I was really surprised to find ‘LCD’ related products popping up all over the web. When I purchased my Sony it was hard to find. Now I note that over at AVS forum the GW11’s seem to be sparking a lot of interest with at least 3 purchasers Choosing the Sony over the latest Samsung DLP displays.

So my question for this honorable board of debaters is: Do you think that the sheer momentum of more manufactories seeming to head to LCD for many new model releases currently in both projection and Direct View, Will sink LCOS ship before it sails? Here is a VERY interesting Industry overview on some of the problems in the industry LCOS might have to overcome.
http://www.insightmedia.info/news/LCOS%20Leadership%20Needed.htm

Here is one reviewer not convinced about Philips promoting LCOS single panel.
http://www.emedialive.com/r13/2002/next8_02.html
The first reviewer was just roaming the show. The Second reviewer to balance the opinion had an appointment with Philips
http://www.whatvideotv.com/articles/frame.html?http://www.whatvideotv.com/articles/interviews/200201_DougStanton.shtml
Reading those two reviews alone, I think I conclude that Philips thinks it is positioning to able to produce the LCOS technology by cutting the past costs to target the in between plasma and CRT crowd. But several other LCOS sets I saw reviewed which never made it to market exhibited problems and were dropped from lineups in recent past years.
So I wonder can Philip (or anyone) put out something that looks so good at a VERY desirable price point which will have the rest backing away from the seemingly current LCD trend to shift towards LCOS?
Philips is BIG maybe they will provide the ‘Leadership’ mentioned as lacking in the In Sight Media link.

Remember I how I stated I found the real room effect of my Sony’s poor black level, to not be an issue for me personaly, and that I felt it was an incredibly good display as regards dealing with ambient light issues?
Ran into this…which has thoughts about this subject as compared to plasma.

“Clarity points out that simply looking at the brightness and contrast spec of a panel is insufficient for understanding how well it will look in a normal ambient environment (40 foot-candles). For example, the phosphors in plasma screens reflect quite a bit of ambient light, about 15%. But LCDs' reflectivity is much lower, around 2.5%. The result is that a plasma panel with 3000:1 dark room contrast will have an actual contrast of about 38:1, whereas an LCD panel with 600:1 and lower reflectivity produces an actual contrast of 109:1.

Clarity takes this "viewability" analysis one step further. It defines a Quality Viewing Metric (QVM) as the product of the panel's peak brightness times its contrast in ambient light. This makes good sense to us as a more logical way to evaluate the actual image quality of a display.

Accordingly, the 131 Ft-Lamberts of brightness for the LCD panel yields a QVM of 14,370. The higher peak brightness of the plasma panel-- 228 Ft-Lamberts-- times a much lower ambient contrast produces a QVM of 8,990. This analysis says that the LCD panel looks better in normal lighting conditions.”
Clarity, Ed Kiyoi, 503-570-0700, http://www.clarityvisual.com/

This is a Company biased report however and they are starting to promote commercial message boards in LCD.


Enough links for one post!!!. What do you guys think?

I am! I'm finding the discussion very intriguing. I've been interested in these different technologies but haven't had the time (ok... haven't made the time) to investigate them. I feel like I'm getting the crash course introduction here! Thanks guys! Oh, SLL, thanks for the links. I'll be investigating them soon!

SLL:

I tend to be cautious, so take this with a grain or two of salt, but IMO it's way too early to call a winner reliably. It's pretty clear that CRT is near its' peak and due to be overtaken, but which one (or ones) of the newer technologies will jump forward is still up for grabs I think.

I would tend to favor the MEMS based devices because (1) the technology is young and promising [always a good sign] and (2) there is tremendous R&D investment for a huge variety of MEMS applications outside imaging.

In the short term - CRT still rules the roost overall.


Just, as always, IMO and YMMV