3 Chip DLP vs 1 Chip DLP question. [Text View]

mlang46

02-24-09, 07:09 PM

Besides possible deeper color saturation, what are the "Significant" improvements that 3 chip DLP can provide over a very good/best 1 chip DLP? (example: marantz VP-11S2)

Not really looking for

Brighter... As I know 3 chip DLP's are usally brighter then 1 chip DLP's.

the only real advantage of a 3 chip DLP over a single chip is brightness

The higher grey scale depth due to the 3 chips having a 3 times longer integration time is not a real advantage because the RGB depth of blue ray discs is only 8 bits

Their are at least 3 manufactures who make single chip projectors which have both DI Variable Iris control they are Benq with the 20,000 and 5000 , Infocus IN83 and Planar. BenQ and Planar also include a variable intensity lamp.

There are some serious manufacturing problems with a 3 chip reflective DlP vs. a single chip.

1. Chip convergence Basically since there are 3 chips they can be out of alignment with each other both horizontally and vertically and with tip and tilt. Sim2 has had problems with its 3 chip projectors in this regard and it is very difficult to get them to replace a projector once you have bought it from them. The Lumis mentioned in this forum was out of alignment by .75 pixels which means at best it has the resolution of a 720P projector.

2. The combining prisms in a three chip projector require a much longer back focal length and induce spherical aberration which has to be corrected by the projector lens. This compromises sharpness. it has been my experience that single chip DLP projectors produce a much sharper image than three chip DLP projectors.

In conclusion: if you do not suffer from eyestrain or see rainbows and you do not need more than 1000 lumens at D65 , stay with a single chip projector when choosing between single chip DLP and 3 chip DLP.

Dennis Erskine

02-24-09, 07:49 PM

48 sounds better.
Wouldn't you want it to be three times the frame rate for optimal performance? I'm sure if I blow smoke rings for a while, I'll come up with the six or seven other issues that would interfere...just not coming to me off the top.

coldmachine

02-24-09, 07:55 PM

Wouldn't you want it to be three times the frame rate for optimal performance? I'm sure if I blow smoke rings for a while, I'll come up with the six or seven other issues that would interfere...just not coming to me off the top.

Thats a good point Dennis, and would be important for an actual implementation.

The number, and I defer to Darin's high 20s here, was not with reference to an implementation, its the actual number that was measured where the issue became irrelevant. I simply thought 48 was the measurement, I wasn't proposing that as an actual spec number.

Obviously if that were to be implemented, I assume we would go to the next higher multiple of frame rate that made engineering, optical and temporal sense.

darinp2

02-24-09, 08:04 PM

Their are at least 3 manufactures who make single chip projectors which have both DI Variable Iris control they are Benq with the 20,000 and 5000 , Infocus IN83 and Planar.The InFocus literature for the IN83 makes it look like they have a dynamic iris, but they don't. It is just a manual iris.

As far as DI's go, it looks to me like the Lumis likely has the best implementation so far if we look at native on/off CR, dynamic on/off CR, and how little it makes itself known, at least based on reports so far. That may or may not have much to do with being 3 chip vs 1 chip, but at this point could be said to be an example of where a specific 3 chipper has an advantage at this point. If SIM2 were to put basically the same system into a single chipper and do just as good a job then that may not apply anymore.
Wouldn't you want it to be three times the frame rate for optimal performance?I don't think this really has much to do with frame rate in that way. Playing back at 48Hz or 72Hz (or 96Hz or 120Hz) with 24Hz material makes sense, but the 48 or high 20s number discussed here was a subframe number (the x factor is the number of times all the primaries get updated every frame). Put another way, according to what was passed to me from a TI rep, the Air Force did some studies in the past and found that nobody would see color separations if the colors were updated at something like 28x (although I don't remember the exact number, but it was in the 20s). What that number means is that with 60Hz playback all of the colors would get updates every 1/28th of a frame, or every .6 msec (.2 msec of each if they all got equal time and no deadtime between them). Some LED DLP rear projectors have been reported by TI to run at an equivalent of about 48x, or the same as R, G, and B getting updated in a time that is only about .35 msec (or 3.5 ten-thousandths of a second), although they don't really run them sequentially as R, G, then B all the time from what I was told (they could switch between 2 colors for part of the frame or even have two going at once for short periods).

If we look at the 6x that the Marantz 11S2 can run at, they are about 20% of the way to the high 20s or 13% of the way to 48x, but I think the percentage of the population who actually saw color separations at 6x would be pretty low.

And there is the complication of what x means when playing back at 48Hz or 72Hz, but I won't do that math there. I think that as long as the updates are in the range of the high 20s from 60Hz the rainbows should basically be a thing of the past.

--Darin

mlang46

03-06-09, 03:53 PM

The InFocus literature for the IN83 makes it look like they have a dynamic iris, but they don't. It is just a manual iris.

As far as DI's go, it looks to me like the Lumis likely has the best implementation so far if we look at native on/off CR, dynamic on/off CR, and how little it makes itself known, at least based on reports so far. That may or may not have much to do with being 3 chip vs 1 chip, but at this point could be said to be an example of where a specific 3 chipper has an advantage at this point. If SIM2 were to put basically the same system into a single chipper and do just as good a job then that may not apply anymore.

--Darin

the planar and the BenQ do have a dynamic Iris.
the better numbers on one lumis are derived from measurements made by enthusiasts.
I do not know of any professional reviewer who has ever reviewed a 1080p sim2 projector

What do you think is going to give a sharper image a single chip DLP projector with a back focal length of 20 mm or a 3 chip dLP with a back focal length of 60 mm and 25 mm to 50 mm of glass inducing 3rd 5th and 7th order aberrations which have to be corrected for by the projector by adding more elements up front? How far does a DLP panel on a 3 chip have to be decentered with respect to the green panel before it degrades to the resolution of a 720p , 480p projector. How far in tip and tilt?

My point is that the variation in performance in 3 chip projectors from projector to projector is great.

Given the 8 bit limit depth on source material and using an osram lamp with a good DI implementation other than power output, do you agree or disagree that a single chip DlP should give a better delivered performance on average than a 3 chip dlp. This assumes that you are not effected by rainbows.

My prediction would be that the planar 8150 should give the best consistently delivered image performance of any single chip or 3 chip dlp on the market and would even have a higher probability of delivering this performance if it incorporated the dark ship 4 and the new insights TI has found for optimizing the optics.

Craig Peer

03-06-09, 04:52 PM

Pretty soon I'll be able to directly compare my dVision 1080p single chip,which is tack sharp, with my Lumis. I'll tell you this much - the Lumis at CES had an extremely sharp image.

Dizzman

03-06-09, 05:19 PM

What do you think is going to give a sharper image a single chip DLP projector with a back focal length of 20 mm or a 3 chip dLP with a back focal length of 60 mm and 25 mm to 50 mm of glass inducing 3rd 5th and 7th order aberrations which have to be corrected for by the projector by adding more elements up front? How far does a DLP panel on a 3 chip have to be decentered with respect to the green panel before it degrades to the resolution of a 720p , 480p projector. How far in tip and tilt?

I have to take some exception to this. by that logic, my portable camcorder will have better forcus et all than a broadcast sony with some big canon glass on it...

we have to factor in that the level of engineering, and the QUALITY of the glass going on that 3 chip is in most cases better. in some cases FAR better.

That seems far too simplistic a view. and it has been my experience that when you take (just about) any decent 1 chip, then A/B it on the same screen (or do a side by side) with (just about) any decent 3 chipper... and most everyone will prefer the 3 chipper. obviously we can find cases of an individual model of one typ ethat kicks the crap out of the other, but that is the exception and not the rule.

coldmachine

03-07-09, 08:27 AM

the better numbers on one lumis are derived from measurements made by enthusiasts.

Totally incorrect, on all points. Not all measurements were taken by enthusiasts at all. They were also not the result of testing one unit. The initial numbers seen here were actually derived from 4 units.

Some tests were of the most stringent nature, with SOTA equipment. Those were similar to Alan's and Wolfgang's numbers. ISCO in/out numbers, for brightness and ANSI, were actually validated by ISCO themselves.

You will also find that none of the tests results supplied out by the 3 people involved, totally independently, were of the fabricated nature we have seen elsewhere.

Alan Gouger

03-07-09, 09:48 AM

mlang46

03-07-09, 08:18 PM

I have found 1 chip DLP to show more banding due to its color bit depth. Add to this the color wheel is filtering high end information. 1 chip can look sharper but that is edge sharpness. Comparing to a well converged 3 chip DLP the 3 chip will always display more detail.

This may have been true before the pulsed lamp and brilliant color technology using the Osram or Vidi lamps were introduced and before a workable DI was introduced but since they have been introduced ,this is not true now because another 2 bits to 4 bits of depth has been added. Before these innovations the single chips had barely enough time to achieve 256 levels of grey. The least significant bit is close to at 22 usec the switching speed of dlps 15 usec and tricks like dithering the signal both spatially and temporarily were used to overcome this and in some cases caused artifacts

with the OSram pulsed lamp and brilliant color technology developed jointly by Osram and TI these artifacts are gone.

No doubt the 3 chip dlp can display a larger bit depth because it has 3 times the amount of time over a single chip to do so but that's a mute point because the highest RGB bit depth is only 8 bits

The qualifier "well converged" is the key to the problem. on a 3 chip to hold a half chip decenter you have to position and hold that the blue and red chip to the green chip to get half a pixel convergence to 5 microns and you still have tip tilt and rotation tolerances to consider.

during Cedia I compared the Sim2 Ht3000E which was the first projector to use the Osram lamp technology developed by Texas Instruments against their 3 chip 1080p. I went back 4 times and compared those two projectors and the Sim2 ht3000 was sharper , had richer colors and had better blacks.
The three chip was brighter. However I asked people around me whcih projector they preferred and to a man they all liked the image of the 3000e to the 3CX 1080p

It was noted in the Lumis Cheerleader forum after the ravings subsided that the projector they were raving about had a misconvergence of .75 pixels. that may not seem like much but it is enough to degrade the resolution in many scenes to that of a 720P projector.

Misconvergence is much worse than lateral color because it degrades the image at the center of the eye's vision where the eye has 6 times the resolution it does at its periphery.

there is no doubt in my mind that if you happen to be one of the lucky ones who happen to get a perfectly converged 3 chip Lumis from Sim2 with the dark chip 4 that you will be very happy with that projector but you also have a very good chance of getting one that is out of convergence "Quecumber" and when that happens you have paid an awful lot for a very bright low resolution projector and since their is no specification on convergence and Sim2 in the past has been extremely reluctant to replace projectors that work but under perform than you the owner will have a problem. They will just tell you it is in spec ,which of course, is true because there is no specification on convergence.

You make a single chip dlp projector with a dark chip 4, a good projector lens Ti's brilliant color with a good DI and an Osram lamp and except for power out no 3 chip projector will in production consistently be capable of competing with it.

Once they start making RGB leds with enough power the days of 3 chip projectors for use inside the home will be over as long as the bit depth on blue ray discs is only 8 bits.

darinp2

03-07-09, 09:06 PM

Given the 8 bit limit depth on source material and using an osram lamp with a good DI implementation other than power output, do you agree or disagree that a single chip DlP should give a better delivered performance on average than a 3 chip dlp.
...
My prediction would be that the planar 8150 should give the best consistently delivered image performance of any single chip or 3 chip dlp on the market and would even have a higher probability of delivering this performance if it incorporated the dark ship 4 and the new insights TI has found for optimizing the optics.I am buying a Planar 8150, but I suspect that I would like the images from the Lumis more overall. I should get to see a Lumis fairly soon, but unfortunately I probably won't get to see the 2 even close in time.
No doubt the 3 chip dlp can display a larger bit depth because it has 3 times the amount of time over a single chip to do so but that's a mute point because the highest RGB bit depth is only 8 bitsSources like Blu-ray are 8 bit component (basically YCbCr) and have a gamma component which basically needs to be reversed. I doubt that this can be converted to RGB and have the gamma in the range of 2.2 to 2.5 applied without extra errors if those only end up in 8 bits.
It was noted in the Lumis Cheerleader forum after the ravings subsided that the projector they were raving about had a misconvergence of .75 pixels.That claim looks pretty misleading to me given that the projector has digital shift functionality that should mean about .5 pixels max misconvergence in the center of the image. The corners or other areas could be off more of course, but where did your .75 pixel misconvergence claim come from?

--Darin

coldmachine

03-07-09, 10:14 PM

It was noted in the Lumis Cheerleader forum after the ravings subsided that the projector they were raving about had a misconvergence of .75 pixels. that may not seem like much but it is enough to degrade the resolution in many scenes to that of a 720P projector.

Again you grossly distort the facts, and demonstrate a rather weak understanding of the subject. Pointing out your continual errors and misrepresentations is becoming rather monotonous. I also see that every time you are proven wrong you simply fail to acknowledge the facts and clarifications that are given to you.

Firstly not all the PJs had that level of MC. I provided extreme close-up photos showing excellent convergence. I believe Alan's was also good.

Secondly, if the center MC was .75 pixel, then that is easily adjusted to 0.25. I would be happy with an overall center MC of 0.75 for this reason.

Your whole post, and many recent posts, simply smack of agenda and bitterness. I'm not the only one to have noticed and I'm not the only one to have publicly called you on it. It goes way back before the fabricated test results we saw posted.

joeycalda

03-07-09, 11:58 PM

Your whole post, and many recent posts, simply smack of agenda and bitterness

Mlang seems to really like the Sim 3000e so that statement isn't totally accurate, unless of course he has a bitterness and agenda towards 3 chip machines, but that seems unlikely IMO

brain sturgeon

03-08-09, 12:46 AM

Mlang seems to really like the Sim 3000e so that statement isn't totally accurate, unless of course he has a bitterness and agenda towards 3 chip machines, but that seems unlikely IMO

Hmmm....

It was noted in the Lumis Cheerleader forum after the ravings subsided...

... you also have a very good chance of getting one that is out of convergence "Quecumber" and when that happens you have paid an awful lot for a very bright low resolution projector and since their is no specification on convergence and Sim2 in the past has been extremely reluctant to replace projectors that work but under perform than you the owner will have a problem. They will just tell you it is in spec ,which of course, is true because there is no specification on convergence.

When was a Sim2 projector ever reviewed by a technical reviewer?

You want me to pay 6 times more for a projector, show me the specs.

You say the lens is superior in the Lumis show me the MTF curves that come with every lens that is purchased. I know they have them

perform a double blind tests on screen with lumens balanced to 16 ft lamberts and compare the lumis against the RS20, the Marantz 11S2 Benq 20000 and if the lumis is not considered dramatically better on each and every movie scene than , not just incrementally , than I say you got nothing

Show me baby.

And remember this "Just because you sleep and have sex with your overpriced 3 chip DLP projector does not mean you understand her"

the problem is that Sim2 is having problems producing this product and Sim2 does not stand behind their products. If you get a defective product or a poor performing product It is like trying to pull teeth to get them to replace it.

Once you have given them your money "Forget about it"

If it looks like a duck, quacks like a duck...

There are horror stories to be had with every projector company, many of which can be read here. But for most of the good companies, you can find many more instances where they did take care of their customers; and Sim2 seems to be one of them (I've not owned one yet, but I'll do a mea culpa if this is wrong in the future when I do).

Dude, you got some serious issues to get over with Sim2. We get it. You don't like them. You want their MTF data and they won't share! Time to move on... :rolleyes:

coldmachine

03-08-09, 10:02 AM

Mlang seems to really like the Sim 3000e so that statement isn't totally accurate, unless of course he has a bitterness and agenda towards 3 chip machines, but that seems unlikely IMO

Joey, the comment is deadly accurate. This has been going on for a considerable time, in different parts of AVS. "brain sturgeon" is pretty much on the money.

I am not the only one to have noticed it, or called it.

I did post yesterday explaining the issue, but the post was removed.

Art Sonneborn

03-08-09, 11:47 AM

Mlang seems to really like the Sim 3000e so that statement isn't totally accurate, unless of course he has a bitterness and agenda towards 3 chip machines, but that seems unlikely IMO

No doubt IMO ,a serious agenda that isn't even thinly veiled. Personally, I believe it is more the cost of 3 chip DLP that he dislikes despite using SIMs lack of published MTF data ,for example, as a way to hammer their products like the Lumis.

Art.

mlang46

03-08-09, 03:18 PM

I am buying a Planar 8150, but I suspect that I would like the images from the Lumis more overall. I should get to see a Lumis fairly soon, but unfortunately I probably won't get to see the 2 even close in time.
Sources like Blu-ray are 8 bit component (basically YCbCr) and have a gamma component which basically needs to be reversed. I doubt that this can be converted to RGB and have the gamma in the range of 2.2 to 2.5 applied without extra errors if those only end up in 8 bits.
That claim looks pretty misleading to me given that the projector has digital shift functionality that should mean about .5 pixels max misconvergence in the center of the image. The corners or other areas could be off more of course, but where did your .75 pixel misconvergence claim come from?

--Darin

the .75 pixel convergence came from a measurement by meyer reviewing the Lumis projector he had he said the pixel to pixel misconvergence was measured to be .75 pixels.

But lets go with your number .5 pixel misconvergence Again How far do you think the misconvergence has to be on a 1080p projector to degrade its MTF to that of a 720P projector. Reference for crt misconvergence can probably be applied: page 321 "Projector Design' Brehmholtz and Stuppe 2nd edition

I was not aware that the Lumis had a functional pixel adjustment equivalent to the Sony Vw200 if it does that would make a big difference in image quality

I agree with you that a strict 8 bit limit will have an effect on the gamma convergence but with the 2 bits of gray scale added from the Osram lamp and a DI system I doubt you will see any performance difference between the gray scale of a single chip to that of a 3 chip.

How many bits of grey scale do you need to have a gamma in the range of 2.2 to 2.5 with out producing errors?

I have not seen the Lumis either but I have a number of times directly compared 3 chip projectors from SIm2 to the single chip Osram lamp projectors they currently produce and every time I have viewed them I always come to the same conclusion. the Single chip Sim2 ht3000E and the Sim2 ht380 with the pulsed lamp technology produced sharper images with better blacks and as least as good if not fully saturated colors.

Given the fact that the 3 chip projector requires a much larger back focal length to accommodate the color combining prism which in themselves introduce aberrations in the converging beam and make the lens design of the projection lens significantly harder and in a 3 reflective chip projector you are always going to have some misconvergence, why would you think in production with projectors coming off the assembly line imperfect , especially if they are made by an Italian manufacture (think FIAT) that your odds of getting a 3 chip projector that produces as sharp an image as a single chip would be very good?

Your choice of the planar 8150 from a technological perspective is a good one but execution is everything.

GetGray

03-08-09, 03:36 PM

"especially if they are made by an Italian manufacture (think FIAT) that your odds of getting a [good product] be very good?" Man, now it's down to ethnic or cultural insults?? Actually, I believe the prism assemblies are handled in Japan at one of the worlds most repsected optic manufacturers, do orientals meet your expectations for mfgr QA? pfft. :rolleyes:

Stephan

03-08-09, 03:48 PM

especially if they are made by an Italian manufacture (think FIAT)

Yeah, let's all run out and buy something where GM is involved... oh wait... maybe not such a good idea after all. ;)

If all your comments are based on FIAT, they're really not worth anything. Why don't you try a Ferrari or Lamborghini? Now this must be a shock for you, but there are even some good products coming from China and Japan. It's hard to believe, I know but you have to get through this now.

coldmachine

03-08-09, 03:58 PM

I was not aware that the Lumis had a functional pixel adjustment equivalent to the Sony Vw200 if it does that would make a big difference in image quality

That is utterly typical of your recent outpourings. The adjustment feature has be mention on a great many occasions.

Perhaps your homework should be carried out with greater diligence. Having said that, when you fabricated your bogus C3X1080 measurements you totally failed to quote the correct specs. This being the fact that lead to uncovering the lie.......and the small matter of your CR "results" being around 4 times less than the genuine figure.

coldmachine

03-08-09, 04:07 PM

especially if they are made by an Italian manufacture (think FIAT) that your odds of getting a 3 chip projector that produces as sharp an image as a single chip would be very good?

Not the first time you have resorted to this sort of stuff.

I seem to remember you recently accusing one of the most respected members on all of AVS of being a dishonest dealer. Even after you were told this was not the case, you insisted in repeating the slur and then referring to him as "the German". That alone should tell anyone exactly what they need to know.

JlgLaw

03-08-09, 04:07 PM

mlang46

03-08-09, 04:23 PM

No doubt IMO ,a serious agenda that isn't even thinly veiled. Personally, I believe it is more the cost of 3 chip DLP that he dislikes despite using SIMs lack of published MTF data ,for example, as a way to hammer their products like the Lumis.

Art.

My agenda is simple to show that the emperor has no clothes

My agenda is simple in this particular forum and thats to show that other than power output and freedom from rainbows because blue ray disks only produce 256 gray levels , their is no inherent advantage of buying a 3 chip DlP over a single chip DlP and in fact there are many disadvantages in additon to the 3 to 8 times difference in cost.

I have bought two Sim2 dlp single chip DLP projectors and before buying them compared them to Sim2 C3X expecting to be blown away by C3X and every time I thought, what hell is everybody talking about . In almost cases I preferred the single chip and in no case was I blown away by the quality of the image in the 3 chip and for the money their asking I need to be blown away

If I then am a Little skeptical about the gushing claims by some enthusiasts about the Lumis image quality well EXcuuuuuse Me!

If you think the Lumis is the best projector than maybe one of you dealers can send this projector to Greg Rogers at Widescreen review and have him test it. His numbers I will believe.

As I have said before, to my knowledge, I have never seen a 1080P 3 chip DLP projector from any manufacturer technically reviewed by a professional reviewer and that ought to tell you something.

Until the, I have signed up for a demo of the LUMIS directly from SIM USA and when they can consistently produce one I will probably get to see one in action and if I think it is significantly better than any projector I have ever seen I will say so and if not well! you will hear about that too.

brain sturgeon

03-08-09, 04:50 PM

My agenda is simple to show that the emperor has no clothes

My agenda is simple in this particular forum and thats to show that other than power output and freedom from rainbows because blue ray disks only produce 256 gray levels , their is no inherent advantage of buying a 3 chip DlP over a single chip DlP and in fact there are many disadvantages in additon to the 3 to 8 times difference in cost.

I have bought two Sim2 dlp single chip DLP projectors and before buying them compared them to Sim2 C3X expecting to be blown away by C3X and every time I thought, what hell is everybody talking about . In almost cases I preferred the single chip and in no case was I blown away by the quality of the image in the 3 chip and for the money their asking I need to be blown away

If I then am a Little skeptical about the gushing claims by some enthusiasts about the Lumis image quality well EXcuuuuuse Me!

If you think the Lumis is the best projector than maybe one of you dealers can send this projector to Greg Rogers at Widescreen review and have him test it. His numbers I will believe.

As I have said before, to my knowledge, I have never seen a 1080P 3 chip DLP projector from any manufacturer technically reviewed by a professional reviewer and that ought to tell you something.

Until the, I have signed up for a demo of the LUMIS directly from SIM USA and when they can consistently produce one I will probably get to see one in action and if I think it is significantly better than any projector I have ever seen I will say so and if not well! you will hear about that too.

So you've made it your mission in life to prove to everyone that your opinion regarding the image quality of single chip DLP versus 3 chip DLP must be right, despite the opinion of a multitude of other people here and elsewhere who have seen both and feel the opposite. And that only Greg Rogers' numbers/reports should be the basis of anyone ever ponying up any amount of money to buy a projector.

Get over yourself man! We all have our own visual systems, personal preferences, and pocketbooks to make decisions as to what projectors we choose to have; and none of them are perfect. But stop trying to convince people that they are committing some mortal sin by buying a three chip DLP as this is simply your opinion, and a suspect one at that. People buying triple chip DLP doesn't affect you at all (or does it? I'm starting to get suspicious!), just as people who buy single chip DLP, triple chip LCOS, or LCD panels don't affect you. Buy what you want, be happy with it, and stop with the "triple chip DLP sky is falling" BS that we are all tired of.

mlang46

03-08-09, 04:56 PM

[QUOTE=mlang46;15996052].......I was not aware that the Lumis had a functional pixel adjustment equivalent to the Sony Vw200 if it does that would make a big difference in image quality.

I can assure you it does. I just worked with it last Tuesday. It is found in the Service menu under "Convergence" and works as advertised.

Jim

When Sony first introduced that feature on the 200 no body appeared to take any notice and as an optical engineer who has personal experience with aligning optics to the micron level I thought " man that is huge " the problem is that even if you align and glue the panels to the beam combining prism if you are not careful they will move when you change altitude because of out gassing and even small shifts in the panels will degrade the resolution significantly. If you have a convergence control and especially if you have a sectional convergence control you can take out much of the errors including lateral color which is field dependent

If you do not have sectional convergence control but only macro you can take out most of the mechanical misconvergence but cannot correct for the misconvergence from the lateral color distortions of the lens. lateral color distortions increase as you move from the center to the edge of the screen but this distortion is offset by the fact that your eyes resolution drops off dramatically after the first 5 degrees down 1/6 th it was at the center.

You will see people on this forum who are engineers and have bought a Sony VW200 and a Jvc RS20 go for pages and pages about getting the color just right but never mention anything about using convergence control which makes a huge difference in sharpness on any 3 panel reflective projector.

Of course there can be no mechanical misconvergence in a single chip projector.

Art Sonneborn

03-08-09, 04:56 PM

GetGray

03-08-09, 05:07 PM

You should spend your talents explaining to Sim2, DPI, PD, and the others how wrong it is to build 3CDLP. Surely your retired talents would be welcomed by those dummies engineering all those 3C units. What are they thinking???

darinp2

03-08-09, 05:17 PM

the .75 pixel convergence came from a measurement by meyer reviewing the Lumis projector he had he said the pixel to pixel misconvergence was measured to be .75 pixels.In the corners. This could just be an honest mistake on your part, but there are certain clues that I look for to see if a person is trustworthy is far as the information they post here. You not only made it look like the information about the misconvergence came out later, but I think you made it look like he was getting .75 pixels of misconvergence even in the center. Did you mean to do that?
I was not aware that the Lumis had a functional pixel adjustment equivalent to the Sony Vw200 ...It doesn't (at least from what I've read). It has one basically equivalent to the JVCs (whole pixel shifts). The VW200 has sub-pixel adjustments using 2 adjacent pixels and can be done in zones, but most people probably don't use that feature (since it uses 2 pixels to accomplish this).
I have not seen the Lumis either but I have a number of times directly compared 3 chip projectors from SIm2 to the single chip Osram lamp projectors they currently produce and every time I have viewed them I always come to the same conclusion. the Single chip Sim2 ht3000E and the Sim2 ht380 with the pulsed lamp technology produced sharper images with better blacks and as least as good if not fully saturated colors.It could depend on what unit a person gets from either company, but from what I've seen with numbers here, a person would likely get better ANSI CR, native on/off CR, and dynamic on/off CR from a Lumis over the Planar 8150, unless they used the T1 lens with the Lumis, although the 8150 wouldn't be too far behind in any of those. Alan's unit definitely looks like it has better CR in all 3 of those categories than the 8150s I've heard of. Of course the actual black level will depend on multiple things like the screen chosen, but as far as this better blacks thing I think you are making a mistake if you assume that the 8150 would have better blacks than the Lumis because of some tests you did with different projectors than either of those.
... that your odds of getting a 3 chip projector that produces as sharp an image as a single chip would be very good?I didn't say it would be sharper, so not sure why you even asked me that (not even including your comment on Italian manufacturing). I may prefer a specific 1 chip DLP over a specific 3 chip DLP (especially if the 1 chip has much higher on/off CR), but I was talking about the 8150 vs Lumis. Either one can have certain problems, like bad uniformity of convergence with the Lumis and bad uniformity of focusing or poor chromatic aberration with the 8150. With either one I would of course want to get a good example.

--Darin

darinp2

03-08-09, 05:25 PM

When Sony first introduced that feature on the 200 no body appeared to take any notice and as an optical engineer who has personal experience with aligning optics to the micron level I thought " man that is huge " the problem is that even if you align and glue the panels to the beam combining prism if you are not careful they will move when you change altitude because of out gassing and even small shifts in the panels will degrade the resolution significantly. If you have a convergence control and especially if you have a sectional convergence control you can take out much of the errors including lateral color which is field dependent

If you do not have sectional convergence control but only macro you can take out most of the mechanical misconvergence but cannot correct for the misconvergence from the lateral color distortions of the lens. lateral color distortions increase as you move from the center to the edge of the screen but this distortion is offset by the fact that your eyes resolution drops off dramatically after the first 5 degrees down 1/6 th it was at the center.

You will see people on this forum who are engineers and have bought a Sony VW200 and a Jvc RS20 go for pages and pages about getting the color just right but never mention anything about using convergence control which makes a huge difference in sharpness on any 3 panel reflective projector.How can you go on and on about misconvergence taking a 1080p projector down to 720p and then be so excited about the VW200 having this zone thing? Do you understand how it works? What do you think Sony's controls do to the sharpness if actually used for sub-pixel adjustments?

--Darin

coldmachine

03-08-09, 05:50 PM

Alimentall

03-08-09, 06:05 PM

I think it's a valid argument that the lack of convergence issues with a 1-chip helps offset other gains made with a 3-chip. If you need the brightness for an >120" screen or just have tons of money, why not get a 3-chip and tweak the convergence as best you can? BUT, if you're a 'flipper' like me with a 110" screen, you buy a good single chip and upgrade every two years and get that nice little buzz each time. Because a single chip in 2 years will likely outperform today's 3-chippers except in brightness. And I believe that native 2.35:1 projectors will be out within 2 years (wouldn't be surprised to see some at next CES), and even 4K PJs and usable LED PJs will be available at recessionary prices within 2-3 years. Soooooo, spending $30K on a 3-chip is kinda silly unless you have the "gottahaveits"

Stephan

03-08-09, 06:53 PM

When Sony first introduced that feature on the 200 no body appeared to take any notice and as an optical engineer who has personal experience with aligning optics to the micron level I thought " man that is huge "

It is indeed huge, because if you don't apply the convergence adjustment by a full pixel across the whole panel and instead choose to use zone based convergence, you're causing far more damage than a 0.5 pixel misconvergence ever could. As an optical engineer, you should know this.

Unless of course you're speaking of displaying 2k content on a 4k or higher resolution machine with zone convergence. But you won't have much luck finding those.

Zone based convergence on a digital projector is not like zone based convergence on a CRT projector. What works for CRT doesn't work for digitals. Unless of course you want to physically deform the panels... you're welcome to try. ;)

There are several reasons to buy a 3-chip and there are also reasons to buy a 1-chip. The statement that a 3-chip is always better is not always true, just like the statement that a 1-chip is always better isn't always true either.

Personally I would not touch the Sim2 1-chip projectors because the rainbows are killing me. They have very solid 3-chip projectors if you happen to be lucky enough to get a unit that works as expected. It's really no secret that Sim2 has some major issues with QC right now. The Lumis is an excellent projector, but that's based on two very early units. I can understand that one might be disappointed if a rather poor performing unit is delivered. My advise would be to work something out with your dealer. Have him place an order, have a look at the unit and buy that specific unit if you're happy with it. If you're not happy, let your dealer return it and get another one.

Michael W.

03-08-09, 10:21 PM

Given the fact that the 3 chip projector requires a much larger back focal length to accommodate the color combining prism which in themselves introduce aberrations in the converging beam and make the lens design of the projection lens significantly harder and in a 3 reflective chip projector you are always going to have some misconvergence, why would you think in production with projectors coming off the assembly line imperfect , especially if they are made by an Italian manufacture (think FIAT) that your odds of getting a 3 chip projector that produces as sharp an image as a single chip would be very good?.

So, how do you explain that the best projector you saw at CEDIA was the JVC RS-20, which is a 3-chip projector? Especially with the RS20 likely having lower quality optics than 3-Chip DLP machines.

Art Sonneborn

03-09-09, 08:47 AM

Alimentall

03-09-09, 06:11 PM

I would be very surprised if we see native 2.35:1 projectors in that time line.

I like to say these things out loud as if it will make them happen :)

Of course, if you think about it, what better way for DLP to up their game than a 2.35:1 chip, especially with 4K on the horizon? Or for guys like JVC to leverage their advantage in the next generation (assuming you can align chips that wide)? The only thing I can think of that might prevent it would be 4K.

mark haflich

03-09-09, 09:36 PM

mark haflich

03-09-09, 10:49 PM

BTW. For my own viewing which is primarily sports and an occasional movie, I prefer the perfect convergence (there being nothing obviously to converge) of a singlr chip DLP. I do not see rainbows and prefer the edge sharpness associated with perfectly conrged image. If my viewing was predominately films, I would gowith a 3 chippper because of the better blacks and CR as well as the better color depth for lack of the correct term whatever it is. The best quality single chipper say costs less than 15K or 20K too whether it be the Marantz 11S2, the Samsung A900 or the Sim2 3000. That said when Sony replaces the VPL-vw70\80 this fall with a new 8K Buck 3 chip machine, I will probaably get one. Sony will make several significant improvements over the 80 with a revised optical engine, increased lumens, better blacks and higher on\off CR, I suspect a 240hz refresh rate and quite possibly a better SXRD chip. I`ll keep a 1 chipper for sports. Yes my CRT is gone. Sold.

Alimentall

03-09-09, 10:58 PM

From very reliable sources, JVC has no plans to do this.

I know! But I did plant the seed and spoke with one of the Japanese guys and one of the JVC America guys about it. I made a darned good pitch, futile though it may be.

Steve Bruzonsky

03-09-09, 11:07 PM

Yes my CRT is gone. Sold.

WOW! Really. I thought you'd never ditch your CRT. What digital are you using for now?

mark haflich

03-09-09, 11:27 PM

Right now a Marantz. An old 12S4 but the 720p native works great for all the 720p sports I watch, fed by my Radiance, and it costs only about 2K now. Things deteriorate when I convert 1080i to 720p. I am considering the new Samsung A900 and the Marantz 11S2. I use to have an 11S2 (the very one reviewed by Greg R). Thanks for you know what Steve.

mlang46

03-10-09, 04:31 PM

So, how do you explain that the best projector you saw at CEDIA was the JVC RS-20, which is a 3-chip projector? Especially with the RS20 likely having lower quality optics than 3-Chip DLP machines.

thats simple I liked the RS20 because of the high on off contrast ratio up to 50,000 1 but the sharpness of the projector is not as good as a good single chip dlp even so during Cedia I still thought it produced the best images at Cedia

when you design a lens for the highest resolution during the optimization process the software wants to shorten the back focal length to zero and if you fix that length the lens becomes much harder to design. The last surface of the lens wants to correct the aberrations by forming the last surface so that its radius of curvature is at the center of the image plane.

Alimentall

03-10-09, 04:57 PM

I agree with this ^ Single chip DLPs are more precise in their ability to render a precise pixel. Though the JVC has an incredibly smooth, coherent picture for the price. Easy install too!

mlang46

03-10-09, 05:56 PM

How can you go on and on about misconvergence taking a 1080p projector down to 720p and then be so excited about the VW200 having this zone thing? Do you understand how it works? What do you think Sony's controls do to the sharpness if actually used for sub-pixel adjustments?

--Darin

I am assuming not having talked to Sony engineers that they are shifting the timing between panels to shift in space by temporal shifting the readout and should increase the perceived sharpness. From the comments of JG law and cold machine this appears to work well even when it is just macro-shift

when someone says misconvergence I am assuming if they have not said it is from lateral color that it is from a physical misalignment of the panels with respect to one another which would of course show up as misalignment at the center of field which is really really bad because that is where your vision has at least 6 times the visual acuity it has even at 10 degrees.

Why was I excited by an electronic fix? Thats simple because three panels are almost impossible to align in tip tilt rotation and decenter physically to a high enough precision to not degrade the MTF or spatial frequency response of the image and you can calculate these changes in closed form. You don't even have to ray trace or do a scatter analysis of the image. Once thay move far enough they will degrade the image to a point it is equivalent to a 720p projector. Its about yield and opto-mechanical tolerance analysis . it an't just about the software

If on the other hand you can show me that these electronic fixes to mechanical misalignment really don't improve the sharpness of the image and that the sony engineers don't know what their doing than you have strengthened my case for single chip projectors.

but better yet since you live in Seattle go to Sim2 dealership take neutral density filters to balance brightness and compare a Sim2 C3X 1080p against Pulsed lamp based Osram SIm2Ht3000E which I believe has a Dark chip IV and compare a number of images and use your eyes to make a judgment

I did at least 3 times and every time I preferred the image of the Sim2 Ht3000E to the Sim2 C3X 1080p I though it was sharper and I thought and this really surprised me that the colors were richer (had as great or greater bit depth) than the 3 chip. As soon as the Sim general manager explained what they had done I knew immediately why the 3000E put out a better image than any single chip projector, I had ever seen. Before the introduction of the Osram modulated lamp technology this was not the case , the richness and depth of the color of 3 chips and the lack of dithering artifacts made the 3 chip DLP projectors significantly superior to single chip. Thats not the case anymore. especially with the bit depth limitation of blue ray source material.

Now I have seen the planar 8150 once at CEDIA and I was no where near as impressed with it as I was with the Sim2 Ht3000E even though it has DI. it could have been the screen, the vidi vs, the Osram lamp better optic or better software.

Now the Lumis with the introduction of a DI and and taking advantage of the pulsed lamp probably puts out a gorgeous image but if you apply that same technology to a single chip DLP projector you will get as good ,not as bright but as good a projector with a much better yield.

by the way I have developed products and worked with both Japanese and Italian engineers on very complex electro-optical products with multimillion dollar budgets. A Fiat is more fun to drive but a Toyota lasts longer and is more reliable.

mlang46

03-10-09, 07:59 PM

From very reliable sources, JVC has no plans to do this.

LEDs at current light output levels are not viable now or in the very near future for normal size and gain screens.

The Sony subpixel multizone convergence fix`is a joke. It should not be used because it lowers resolution and destroys sharpness among other things. Full pixel shifts work well on all types of three chip projectors and have no adverse effects other than at the extreme vertical edges of the screen and slight overscan to hide a one or two vertical line loss will be essentially unnoticeable.

Now let us talk about Sim2. the light engine in all their three chip machines is made by Delta. On all machines the 3cx`s the engines come in from Delta and the chip alignment is not adjustable. I believe alignment issues are inherent in the delta manufacturing process and one does indeed roll the dices. perhaps Delta could do a better job but I don`t know. Clearly Sim2 could reject many of them but costs would substantially I suspect delta is required to meet a certain spec but what that spec is I haven`t a clue. In the HT5000 line up the chips are adjustable by Delta in its quality control process. The chips are adjustable by a jig in the xy directions only and are locked in place by screws rather than permanentl glued or welded into place in the c3x engine. The who insides including engine and internal light path in the HT5000 line make the insides of the c3x`s look like a toy. YOU GET WHAT YOU PAY FOR and I suspect the HT5000 will be revised to bring it inline with the CR advances wrought by the Lumis. Of course some CR and black level advantage may go the Lumis because of its lower lumens.

While not convergence in the strictest some optical appearance of misconvergence will be less in the HT5000 line because of its superior lens and much greater sweet spot allowing for greater lens shift without aberrations.
I frankly look at distain the animosity between cetain posters here. I learn much from the various posters here. SOME ARE HOWEVER CLEARLY WRONG IN SOME OF THE OBSERVATIONS AND STATEMENTS BUT EVEN THOSE HAVE MUCH TO OFFER IN THEIR VARIOUS SUBJECT MATTER EXPERTISE. GET it.

Frankly I don`t understand MLANG`S refusal to acknowledge that misconverce of 3\4 pixel can be full pixel shifter to 1\4 pixel which is a fine spec. Much ,ore objectionable would be .5 pixel misconvergence which can not be shifted to better it. WHETHER ONE SHOULD EXPECT BETTER IN A ERR CHEAP 36K MACHINE IS ANOTHER THING. Certianly for 60K I would expect no worse than .25 center.

Good point Mark on pixelshift , I did not know at the time you could shift pixels on the lumis. A good lens will help but not eliminate mechanical misconvergence. the lens MTF is just multiplied by the decreased MTF of the misconvergence. Its almost like you made the pixels on the chip larger but since the lens is better you get a better final contrast.

My point is that the optical system of a 3 chip projector is so much more complex than the single chip and the mechanical alignment tolerances on the reflective 1080p chips are so tight that tip tilt decenter and rotation that the variation in performance from projector to projector will be very large with some people being very happy and others not so much. ALL you have to do is scan this forum to find people who were extremely unhappy with their msconverged three chips. Until the Lumis I don't believe that their was any pixel shifting control on three chip DLP projectors. the worse shift you could get is half a pixel. I guess it takes out tip tilt and vertical and horizontal decenters but rotation im not sure about. anyway macro pixel shifting could really help

The coatings on the dichroic filters in a compact color combining system with
non telecentric optics have to be layered across the prism so you don't get color shifts from the non convergent non-telecentric imaging system otherwise you get color shifts from different field rays as you go across the prism. manufacturing that prism has to be nightmare and is probably not very repeatable. Sometimes you get a good 36.000 dollar projector and some times you do not. Again a 3 chip is very complicated and an order of magnitude as hard to reproduce

Varyiing the gamma is important to have a good functioning DI system but you can't change the gamma without screwing up the color because in changing the gamma you change how the colors are mixed in proportion. if you have a wide dynamic range in the image DI will not help. simpler is better and a projector with a natively high contrast ratio is going to produce a better image than dI system all other things being equal. Unfortunately you really do need a bat cave to appreciate a high contrast projector.

Now most of the DLP 3 chip projector manufacturers are boutique small companies with limited resources and do not have the production engineers to really engineer and set up the proper tooling to insure that each projector leaving the assembly line meets the same specifications and again I think a cursory look at AVS form bears this out. Most of these small companies are just what we used to call system integrators where most of added value is contributed by OEMS. They don't manufacture the chip or the optics or even come up with the drivers for the chip or the wheel or the light source to how the light source is driven. for the most part they are packaging not just other peoples components but their complete systems.

On the other hand a company like Sony can invest in the Optical tooling developed from years of producing other products and have 1000s of engineers who can be enlisted to solve problems and they produce their own chips.

I am not offended but I am amused by the animosity I generated and I am surprised. These guys really take their toys seriously. I mean if I went over their house for dinner and afterwords I said the dinner was just okay but your wife is ugly and stupid they would reply " would you like another glass of champagne" but if instead you said dinner was perfect , your wife is beautiful and charming but your 36.000 dollar 3 chip DLP is not as good my single chip and does not have as good an MTF as my single chip DLP, he would threaten to shoot me and throw me out of the house. Of course I would not know if he threw me out because I insulted his projector or because he thought I was cursing at his wife

FrantzM

03-10-09, 08:28 PM

Hi

I will retreat in my lurking spot after this post, this discussion is over my head but it is looking to me as if that single chip have the potential to be great. No convergence issues, simpler optical path seems to be their prime advantage.. The other side of the coin is the need to have at least RGB to make a complete color picture and there, they have to do all kind of tricks, most of these are caught by the eye.
It seems to remain that with a technology that would eliminate the wheel (or find a way to make it spin much faster or) , things such as LED, we have the POTENTIAL to have one-chipper surpassing than 3-chippers... LEDs are not yet that bright from what I think I know, so we are frm there yet...
So in the Here and now 3-chip PJs are superior to One-chip in term of general PQ... It will take a few more years before their general performance is reached by single panel PJs.. Am I correct in this simplistic assessment?

mark haflich

03-10-09, 08:32 PM

These last two Mlang`s posts contain a lot of good and valuable information and is one of the reasons we should cherish Mlang`s participation here. I am friends with many of the participants here and it pains me when we question and attack participants posting motives when I know such motives are beyond reproach. We indeed all have bad posts and make bad plays you know who. Forgive, live and honor the good plays they far outweigh the bads.

Steve Bruzonsky

03-10-09, 09:46 PM

I am not offended but I am amused by the animosity I generated and I am surprised. These guys really take their toys seriously. I mean if I went over their house for dinner and afterwords I said the dinner was just okay but your wife is ugly and stupid they would reply " would you like another glass of champagne" but if instead you said dinner was perfect , your wife is beautiful and charming but your 36.000 dollar 3 chip DLP is not as good my single chip and does not have as good an MTF as my single chip DLP, he would threaten to shoot me and throw me out of the house. Of course I would not know if he threw me out because I insulted his projector or because he thought I was cursing at his wife

If given a choice of the projector or his wife, he would undoubtedly keep the projector and let you have and pay for the wife!!!@@@

I would only shoot you if you tried to date my young daughters.

Steve Bruzonsky

03-10-09, 09:47 PM

Alimentall

03-10-09, 10:02 PM

People forget this is just audio gear. Not fighting over how to fix the economy. And the people who get hypertensive and, well, bitchy, over this stuff try to shift blame to those that don't.

darinp2

03-10-09, 10:03 PM

I am assuming not having talked to Sony engineers that they are shifting the timing between panels to shift in space by temporal shifting the readout and should increase the perceived sharpness.Were you thinking this was for adjusting convergence during movement in the images? I'm trying to figure out what you thought they were doing. They still have the pixels where the pixels are. They just use adjacent pixels to try to simulate good convergence. Their system works on static patterns. At full pixel shifts it is like the JVC, but partial pixel shifts are basically simulated by decreasing the brightness of one pixel and increasing the brightness of the pixel next to it in the direction the person is trying to move that primary color. There are 10 steps total from pixel to pixel, so I think halfway between would have two pixels of thickness at the same level (but not at 100%) for a test pattern with a single pixel wide line.
when someone says misconvergence I am assuming if they have not said it is from lateral color that it is from a physical misalignment of the panels with respect to one another which would of course show up as misalignment at the center of field which is really really bad because that is where your vision has at least 6 times the visual acuity it has even at 10 degrees.W. Mayer posted .75 off for the corners, but less than that for the center and then you used the .75 as the misconvergence of the projector. If the center is what mattered then you shouldn't have claimed that his projector had .75 pixel misconvergence.

Also, people are likely to look more toward the center of the screen, but since they generally aren't tied down with devices on their eyes like in A Clockwork Orange the rules about what our vision does at the center apply to wherever people are looking, not just the center of the screen. If people's eyes were locked to the center of the screen then multiple things would be easier as far as projector design.
If on the other hand you can show me that these electronic fixes to mechanical misalignment really don't improve the sharpness of the image and that the sony engineers don't know what their doing ...I don't think it is a case of the sony engineers not knowing what they are doing. I'm pretty sure they know what they are doing. It sounds like you thought it would do something it doesn't, but doesn't seem like that is their fault. I am still curious about what you thought the system did. Did you think they were actually doing sub-pixel shifts where a pixel would physically move where it shows up on the screen by less than a full pixel?
but better yet since you live in Seattle go to Sim2 dealership take neutral density filters to balance brightness and compare a Sim2 C3X 1080p against Pulsed lamp based Osram SIm2Ht3000E which I believe has a Dark chip IV and compare a number of images and use your eyes to make a judgmentWhy do you suggest I go backward to the previous 3 chipper from SIM2 when your original prediction was against any 3 chipper on the market, the Lumis is on the market, and we were discussing the Lumis and the Planar? I already made it clear that I was discussing those 2 projectors. You said before that you preferred the blacks on the single chipper in what you saw, but I think that would be much less likely to be the case against the Lumis. And I would love to see SIM2 put the major advancements of the Lumis into a single chip DLP, but that isn't a choice right now. I would be happy to compare SIM2's latest single chipper to the Lumis, but I wouldn't want to take up a dealer's time doing that.

--Darin

mlang46

03-10-09, 10:11 PM

It is indeed huge, because if you don't apply the convergence adjustment by a full pixel across the whole panel and instead choose to use zone based convergence, you're causing far more damage than a 0.5 pixel misconvergence ever could. As an optical engineer, you should know this.

Unless of course you're speaking of displaying 2k content on a 4k or higher resolution machine with zone convergence. But you won't have much luck finding those.

Zone based convergence on a digital projector is not like zone based convergence on a CRT projector. What works for CRT doesn't work for digitals. Unless of course you want to physically deform the panels... you're welcome to try. ;)

There are several reasons to buy a 3-chip and there are also reasons to buy a 1-chip. The statement that a 3-chip is always better is not always true, just like the statement that a 1-chip is always better isn't always true either.

Personally I would not touch the Sim2 1-chip projectors because the rainbows are killing me. They have very solid 3-chip projectors if you happen to be lucky enough to get a unit that works as expected. It's really no secret that Sim2 has some major issues with QC right now. The Lumis is an excellent projector, but that's based on two very early units. I can understand that one might be disappointed if a rather poor performing unit is delivered. My advise would be to work something out with your dealer. Have him place an order, have a look at the unit and buy that specific unit if you're happy with it. If you're not happy, let your dealer return it and get another one.

there are two contributers to misconvergence mechanical tip tilt decenter and rotation of the individual panels and misconvergence from the lens from lateral color. by far mechanical misconvergence is the most problematic because it effects the center of the field where the optical acuity is the highest. that can addressed by applying a constant convergence correction across the whole panel

lateral color on the other hand is a change in magnification as function of color. since your eye drops in resolution rapidly as the field increases so at the center of the panel you may have 1 minute of resolution at the 10 degree field point you would have only 6 minutes of resolution , correcting it is far less important than correcting mechanical panel misalignment but as an optical engineer it is not obvious to me as it was not obvious to the many optical engineers who work for Sony why zonal correction is bad.

you can not correct lateral color with a macro shift because lateral color pixel shift increase as a function of the field angle. at the center of the field I have only mechanical misconvergence say one pixel If I am .7 field have one pixel of convergence from mechanical and 2 pixels from lateral and I can control the pixel shift by zone independently than I should be able control lateral by introducing a small amount of unnoticeable lateral distortion. The question in my mind is how do the individual zones stack up using fixed pixels as opposed to a CRT beam. that might be tricky but the change from zone to zone will be continuous and predictable. it also depends how big the zones are.

the important correction is macro to correct the mechanical misalignment and hopefully the lens will be constructed with little lateral color distortion. before once those 3 panels were glued to a very expensive dichroic prism you were stuck.

I agree with you that a three chip is not always better than a one chip but I will extend that argument to say that the yield on 3 chips is going to be much worse because the optical system is much more complicated and considering the amount of money and current QC problems of Sim you are rolling the dice.

Once you put your money down it is hard to almost impossible to get it back especially since there is no convergence spec. Its not like you can say this projector is out of spec because their is nospec but an uncorrected misconverged projector can really hurt the resolution of the image

if rainbows are a problem or tension headaches you have to go with a 3 chip projector. No question about it.

darinp2

03-10-09, 10:26 PM

... but as an optical engineer it is not obvious to me as it was not obvious to the many optical engineers who work for Sony why zonal correction is bad.Before going into zones, do you understand how doing a half pixel shift over the whole screen with Sony's system has a negative side effect to go with the positive? It isn't all bad, but using 2 pixels to simulate correct convergence is going to bring something bad also. A picture might be worth a thousand words here, but I don't have one.

If you understand why simulating shifting the whole panel by sub-pixels by using 2 adjacent pixels brings a negative factor, then it would seem to follow that using zones brings a negative factor, since it is going to need to simulate sub-pixel changes by using 2 adjacent pixels, or have whole pixel jumps at some points on the screen.

--Darin

HoustonHoyaFan

03-10-09, 10:57 PM

...The Sony subpixel multizone convergence fix`is a joke. It should not be used because it lowers resolution and destroys sharpness among other things. ...People who have compared the results including myself may not agree with you.

Here are some user data which makes the opposite case
http://www.avsforum.com/avs-vb/showthread.php?p=15940508#post15940508

Well firstly, another shot of that test pattern from the centre of the image. No adjustments, 1px adjustments, 0.1px adjustments:
http://sr-388.net/images/ht/projector/convergence/_MG_1587.jpghttp://sr-388.net/images/ht/projector/convergence/_MG_1589.jpghttp://sr-388.net/images/ht/projector/convergence/_MG_1588.jpg

You get that "haze" around things with the 0.1px adjustments, but it does not seem to affect detail—the number 3 is made up of single pixel-wide lines and is still fully resolved.

Having that uniform "haze" seems to be much less noticeable at a distance than having one edge with a stronger colour. That said, it doesn't seem to show up much on video content rather than test patterns.

I used some of the B&W scenes from Memento on Blu-Ray for "real-world" comparisons, as I figured B&W would show off misconvergence easier (as there shouldn't be any colour at all) and I seemed to remember them looking pretty sharp/detailed when I last watched. (actually not quite as good as I thought now that I'm viewing at 120")

Unfortunately there's a little bit of moiré in these photos, but these are the ones that turned out best.

First scene. (to show how tight a crop it is)
http://sr-388.net/images/ht/projector/convergence/_MG_1597.jpg

Crop of his collar, without, 1px, 0.1px:
http://sr-388.net/images/ht/projector/convergence/_MG_1599.jpghttp://sr-388.net/images/ht/projector/convergence/_MG_1598.jpghttp://sr-388.net/images/ht/projector/convergence/_MG_1597-2.jpg

Note the magenta/green fringing without adjustment, the predominantly green fringe on the left with 1px adjustments, and effectively none with 0.1px adjustments. (I should have really moved the camera closer/zoomed in to show it off better)

Second:
http://sr-388.net/images/ht/projector/convergence/_MG_1601.jpg

This time I did move the camera in closer. Again, no adjustment, 1px, 0.1px.
http://sr-388.net/images/ht/projector/convergence/_MG_1606.jpghttp://sr-388.net/images/ht/projector/convergence/_MG_1605.jpghttp://sr-388.net/images/ht/projector/convergence/_MG_1604.jpg

I've not done the best job capturing these shots as the difference between 1px and 0.1px adjustments is more noticeable by eye, with the reduction in colour fringing using 0.1px adjustments being much more apparent.

Nothing I've seen from "real-world" content has been able to produce the discolouration shown further up, and I haven't noticed any detail loss.

So while logic dictates that you should use 1px adjustments rather than 0.1px ones, and test-patterns agree. Real-world video/game content shows the most improvement with the 0.1px adjustments, and I have yet to notice any real downside to using it.

I would still prefer to have a projector that is either much better converged physically, or whole pixels out rather than fractions of a pixel, but I'm pretty happy with the results of the convergence adjustment now that I've done more testing with it.

amirm

03-10-09, 11:37 PM

People who have compared the results including myself may not agree with you.

Here are some user data which makes the opposite case
IndianaGeorge in that thread has the right explanation which agrees more or less with both of you :).

To repeat what he said there, the algorithm is just plain interpolation. And like any filter, it will roll off the high frequencies some (the "haze") and/or cause ringing.

So you are trading convergence error (which also reduces resolution) for filtering (which reduces resolution by definition). Based on your quoted post, the feature seems worthwhile for extreme misconvergence and lateral CA. For other cases, it can be subjective although I think it is more useful than not for movie viewing.

At the risk of stating the obvious, if you have convergence errors, you have already dropped below 1080p resolution whether you use this function or not.

mlang46

03-11-09, 12:04 AM

Were you thinking this was for adjusting convergence during movement in the images? I'm trying to figure out what you thought they were doing. They still have the pixels where the pixels are. They just use adjacent pixels to try to simulate good convergence. Their system works on static patterns. At full pixel shifts it is like the JVC, but partial pixel shifts are basically simulated by decreasing the brightness of one pixel and increasing the brightness of the pixel next to it in the direction the person is trying to move that primary color. There are 10 steps total from pixel to pixel, so I think halfway between would have two pixels of thickness at the same level (but not at 100%) for a test pattern with a single pixel wide line.
W. Mayer posted .75 off for the corners, but less than that for the center and then you used the .75 as the misconvergence of the projector. If the center is what mattered then you shouldn't have claimed that his projector had .75 pixel misconvergence.

Also, people are likely to look more toward the center of the screen, but since they generally aren't tied down with devices on their eyes like in A Clockwork Orange the rules about what our vision does at the center apply to wherever people are looking, not just the center of the screen. If people's eyes were locked to the center of the screen then multiple things would be easier as far as projector design.
I don't think it is a case of the sony engineers not knowing what they are doing. I'm pretty sure they know what they are doing. It sounds like you thought it would do something it doesn't, but doesn't seem like that is their fault. I am still curious about what you thought the system did. Did you think they were actually doing sub-pixel shifts where a pixel would physically move where it shows up on the screen by less than a full pixel?
Why do you suggest I go backward to the previous 3 chipper from SIM2 when your original prediction was against any 3 chipper on the market, the Lumis is on the market, and we were discussing the Lumis and the Planar? I already made it clear that I was discussing those 2 projectors. You said before that you preferred the blacks on the single chipper in what you saw, but I think that would be much less likely to be the case against the Lumis. And I would love to see SIM2 put the major advancements of the Lumis into a single chip DLP, but that isn't a choice right now. I would be happy to compare SIM2's latest single chipper to the Lumis, but I wouldn't want to take up a dealer's time doing that.

--Darin

the image at he screen does not care what number the physical pixel has as long as the same optically conjugate position as a contributing pixel otherwise time shifting pixels would make no sense If the optically conjugate position for a green pixel for a particular scene is pixel 101 and the red is 102 for that same conjugate and mechanically 101 because of misalignment that you increase the resolution of the projector by doing a pixel shift making the green 101 pixel optically conjugate to the 102 pixel. its a timing change full pixel by pixel. as far as static images their all static even the ones that move

Why is that a big thing? because if you can electronically even on a non zonal level shift a misaligned panel to where instead of say being off by 3 pixels you only off by .5 pixel you have improved the crap out of that resolution. remember those panels are glued to that prism thats why I am excited about electronic alignment of 3 panel projectors.

The title of this particular forum was single chip DLP vs. 3 chip DLP

my arguments in favor of single chip were

1. Single chip DlP produce a sharper image because
a. there is zero panel misalignment
b. the projector lens with a shorter back focal length and no color combining prism produces a sharper image than even a much more expensive single chip projector with a much more expensive lens

2. The increased integration time of the 3 chip with its 100 percent almost color efficiency has no real advantage over a single chip with a 33 percent color conversion efficiency other than lumen output power because
a. the source material has only an 8 bit depth per color
b. the osram pulsed lamp increases the grey level by at least 2 bits and balances the color of the UHP by boosting the output of the lamp when the red section and the blue section of the color wheel goes buy. It also reduces
temporal digital artifacts

Why do I want you to look at a Sim2 single chip before buying a planar 8150

a. Because I have seen the Planar 8150 maybe under poor circumstances and I think the Sim2 ht380 3000 are much better projectors even with out the DI as is probably the D80E. I also think they have implemented the pulse lamp technology better than anyone else.

b. why compare a c3x1080p projector to a single chip sim2 as oppose to looking at the LUMIS
your strongest argument for 3 chips not strong but your strongest is that 3 chips have deeper gray scale and thus produce a richer image

that argument can be settled by observation and the best way would be to compare a single ship from the same generation from the same manufacturer with its 3 chip from that generation and then you can really judge the difference. Compare a 3000E host against a CX1080p

as far as pixle misconvergence of Mayer .75 pixels I was writing from memory and since the Lumis has convergence correction, something I did not know at the time, it would be better to be off 3 whole pixels than .5 pixels at the center

goodbye

as far as wasting a projector dealer's time in this economy he will be glad to have the company

darinp2

03-11-09, 01:19 AM

the image at he screen does not care what number the physical pixel has as long as the same optically conjugate position as a contributing pixel otherwise time shifting pixels would make no sense If the optically conjugate position for a green pixel for a particular scene is pixel 101 and the red is 102 for that same conjugate and mechanically 101 because of misalignment that you increase the resolution of the projector by doing a pixel shift making the green 101 pixel optically conjugate to the 102 pixel. its a timing change full pixel by pixel.It seemed that the sub-pixel stuff is what you were excited about. Was it the full-pixel stuff like JVC and SIM2 do that you were really excited about? I'm still not sure what you thought Sony was doing that is different than they are actually doing.
Why is that a big thing? because if you can electronically even on a non zonal level shift a misaligned panel to where instead of say being off by 3 pixels you only off by .5 pixel you have improved the crap out of that resolution.Why would you shift 3 pixels to .5 pixel? Why not shift it whole amounts if it is off by whole pixels?

And if .5 a pixel off is okay with you, then why get excited about feature doing basically a simulation of sub-pixels?
remember those panels are glued to that prism thats why I am excited about electronic alignment of 3 panel projectors.Getting excited about full pixel shifts makes sense to me. And for some people getting excited about something like the Sony with a kind of sub-pixel thing using 2 adjacent pixels, but for somebody like you who kept going on about misconvergence causing a loss of resolution from 1080p to 720p, I don't understand you getting excited about a feature that basically hurts resolution to try to overcome misconvergence of something other than whole pixels.
b. why compare a c3x1080p projector to a single chip sim2 as oppose to looking at the LUMIS
your strongest argument for 3 chips not strong but your strongest is that 3 chips have deeper gray scale and thus produce a richer imageWho is "your" in your above statement?

You made a statement:
My prediction would be that the planar 8150 should give the best consistently delivered image performance of any single chip or 3 chip dlp on the market and would even have a higher probability of delivering this performance if it incorporated the dark ship 4 and the new insights TI has found for optimizing the optics.and I addressed it with respect to what I would consider the best 3 chipper based on reports from people I trust and now you want me to compare to a lesser model instead.

If you want to counter some argument that somebody else made then that is fine, but as far as your statement about "any single chip or 3 chip dlp on the market" I would expect that to apply to the Lumis and not have a need to go back to last year's model.

--Darin

coldmachine

03-11-09, 03:58 AM

that argument can be settled by observation and the best way would be to compare a single ship from the same generation from the same manufacturer with its 3 chip from that generation and then you can really judge the difference. Compare a 3000E host against a CX1080p

Perhaps you should ask someone who owns or owned both.....That would be me. I ran the 3000e and the C3X1080 for some time. The 1080 doesn't offer any greater light output than the 3000e, but does offer a better image. I also owned the HT5000 at that time as well. The HT5k was, unquestionably, a step up again from the 1080.

At one point, as has been reported here many times, I had a friend use my main HT for part of a research project and actually had all 3, along with a number of other units, in the same room.

Anyone who has used those machines properly will confirm this.

coldmachine

03-11-09, 04:07 AM

People who have compared the results including myself may not agree with you.

I have used it and found it to have a very negative impact on the image. See post #36. I posted exactly the same when I first tested the Sony PJ.

mark haflich

03-11-09, 09:50 AM

Re the bad effects of sub pixel adjustment, there is even a caution someone stuck in the Sony manuals (here some poorly translated Japanese) cautioninf that use of the subpixel system (zonal and if I remember correctly global) may cause loss of resolution and sharpness. Obviously if a system varies the brightness between 2 adjacent pixes to cause say a grid line on a chip to appear to move, remember the line of pixels on the chip is fixed) two pixels are now being used where only one was being used before. Regarding any subpixel shift (a subpixel shift being any shift not in a whole non fractional number (1, 2, 3 etc), the spacing between the pixels will further reduce sharpness and increase line width.

I agree that one may prefer a global shift of less than a full pixel to see fringing in an image. Say a 0.1 or 0.2 global subpixel shift. But its still a trade off, one loses resolution and sharpness Even the nearly blind (I am exagerating) can see the losses associated with 0.5 amnd 0.4 shifts. Ugh. There is no free lunch here sort of paying more for your beer but getting some snacks included for the price.

I FINALLY GET TO ENGAGE DARIN. Seriously, why should one stae a misconvergence spec as the misconvergence in the center. If there must be only one number for a misconvergence spec, it I thing needs to be the maximum anywhere on the screen. Better would be a two part spec, one for center and one for corner. Absent that, perhaps a weighted spec but a two part spec would be better and easily field verifiable. Sim2 does in fact give a 2 part spec for its HT5000 line or is it really a Delta spec.

Now lets talk about how one measures misconvergence. Does it have to be measured after the lens. And if the lens is involved then what actually is being measured at the extremes.

And what should a manufacturer strive for, minimizing global misconvergence or maximizing center convergence.

When you start messing with the multitude of zones, one must deal with vary degrees or err correction in adjacent areas. The effects are a mess.

mark haflich

03-11-09, 10:08 AM

I really agree with CM if for no reason that he agreed with me and is so much larger than me in physical size, the big lug.

The Sony adjustments are for shutting customers up about misconvergence and for marketing. If you like small subpixel globally, by all means enjoy your perceived improvements. If you want to adjust zonally, do what you want. Its all about what you see.

Sorting out the correct from the incorrect here on the forum, one can become very educated. One sometimes learns about and then unfortunately begins to sees real adverse effects not noticed before.

One can feed their egoes through posting on a forum. I saw it first. Trying to score style points. Wouldn`t it be better to say I see it, please look and see if you see it too. Then lets discuss why we see it.

Let`s say someone makes an argument and makes factual or rationality errors. Rather than attack, I gotcha you moron, far better to state I think you argument breaks down at this point because or do I misunderstand what you said.

There is no class leader here who has all the correct answers and explanations. It is a group which can increase the knowledge of all the participants and knowledge supplied to the group could lead to insights and knowledge not previously possessed by anyone in the group.

Its not about the egoes of one or two here, its about the net sum. So much better to pretend we are at someones house all together and having a group discussion between friends. It sounds too sweet but the net results would be so much more if we all did this and besides it is the right thing to do.

howdydoody

03-11-09, 12:00 PM

its not about the egoes of one or two here, its about the net sum. So much better to pretend we are at someones house all together and having a group discussion between friends. Its sounds too sweet but the net results would be so much better if we all did this and it besides is the right thing to do.

+1

FrantzM

03-11-09, 12:22 PM

<snip>......

Sorting out the correct from the incorrect here on the forum, one can become very educated and unfortunately see adverse effects not noticed before. One can indeed let their ego be fed, I saw it first. Its all style points. I see it, please look and see if you see it too. Then lets discuss why we see it.
Let`s say someone makes an argument and makes factual or rationality errors. Rather than attack, I gotcha you moron, far better to state I think you argument breaks down at this point because or do I misunderstand what you said.

There is no class leader here who has the correct answers and explanations. It is a group which can increase the knowledge of all the participants and knowledge supplied to another could lead to insights and knowledge not previously possessed by anyone in the group.

Its not about the egoes of one or two here, its about the net sum. So much better to pretend we are at someones house all together and having a group discussion between friends. ITS SOUNDS TOO SWEET BUT THE NET RESULTS WOULD BE SO MUCH BETTER IF WE ALL DID THIS AND IT BESIDES IS THE RIGHT THING TO DO. Damn cap key, sorry.

Mark.

One of the best posts I have read since I have been in this forum.. You will allow me to quote it in other parts of the forum where others and I have let our ego get the best of a good discussion... Thanks for a wise, level and intelligent post...

darinp2

03-11-09, 01:31 PM

I FINALLY GET TO ENGAGE DARIN. Seriously, why should one stae a misconvergence spec as the misconvergence in the center.I don't think that it should. What I took exception to was:
It was noted in the Lumis Cheerleader forum after the ravings subsided that the projector they were raving about had a misconvergence of .75 pixels. that may not seem like much but it is enough to degrade the resolution in many scenes to that of a 720P projector.both because it made it look like that misconvergence came out after the fact when W.Mayer was up front about the misconvergence he saw on the first one he got and then the 2nd one, and because W.Mayer reported the misconvergence for the center and at least one corner (which for the 2nd one at .75 was worse than the center) and the above made it look like that projector had that misconvergence even at the center. If somebody said that the first Lumis W.Mayer got had close to perfect convergence without mentioning that this was only for the center of the screen I think that would be misleading too.

Mlang has also taken a position about the center of the field from the projector being really important here:
... which would of course show up as misalignment at the center of field which is really really bad because that is where your vision has at least 6 times the visual acuity it has even at 10 degrees.In that case it would be center misconvergence that would matter most and that is true to a large degree, but as I basically pointed out, the center of the image from the projector and the center of a person's vision are not the same thing. A person can look wherever they want. Most people will look toward the center of the image more often than the edges, but will look around some while watching content.

I think mlang has some good stuff to contribute and I'm sure he thinks he is doing good, but IMO the 2nd part of this statement:
My agenda is simple to show that the emperor has no clotheshas been appropriate to portions of his posts too often in his attempts to disparage the SIM2 3 chippers.

--Darin

noah katz

03-11-09, 01:38 PM

sierraalphahotel

03-11-09, 02:09 PM

Would the best mis-convergence solution be to allow for physical adjustment such as that found on the HT5000, assuming it were financially viable?

I assume that other than potentially cost, that the reason the C3X family has fixed panels is due to chassis size constraints?

Sean

amirm

03-11-09, 04:24 PM

"People who have compared the results including myself may not agree with you."

That b/w movie is probably from a lower res source that has been upscaled, so there's not much to lose and a lot to gain from the Sony fix.
He used a blu-ray title (Momento):

" used some of the B&W scenes from Memento on Blu-Ray for "real-world" comparisons, "

I haven't watched but reviews say it is better than SD version.

As to core of this topic, better subpixel interpolation can be designed than the one Sony has. We use such correction in digital photography (e.g. photoshop) to great effect to remove CA. So to the extent we are talking about $30K projectors, I think they should offer it and let people decide which one bothers them more, convergence errors or some loss of resolution due to interpolation. They could also make it source specific so that it is not always on.

amirm

03-11-09, 04:31 PM

Expanding on above, the lens (and misalignment in the panels) could be profiled and reverse correction be done using signal processing in the projector. That way, the correction can be applied correctly across all pixels and avoid the edge problem in Sony design.

Of course, this will only work for 2-D distortion. I would think 3-D distortion (tilt, etc.) can be helped with micro alignment of the panels using mechanical means.

mark haflich

03-11-09, 08:14 PM

Art Sonneborn

03-11-09, 09:19 PM

People forget this is just audio gear. Not fighting over how to fix the economy. And the people who get hypertensive and, well, bitchy, over this stuff try to shift blame to those that don't.

But bitching about the gear allows us to forget about the economy for a few minutes. This seems to be an important point lost.

Art

Art Sonneborn

03-11-09, 09:21 PM

Would the best mis-convergence solution be to allow for physical adjustment such as that found on the HT5000, assuming it were financially viable?

Sean

It does seem that it would be a better solution.

Art

mark haflich

03-12-09, 12:13 AM

Mechanical adjustment of two chips in the xy direction is possible in the HT5000. The alignment is done to bring those chips into better alignment with the third chip. However this is done at the light engine OEMer, Delta. It is not done at the Sim2 factory in Italy or at the US facility in Florida. From what I imagine, it requires very delicate hand adjustment using very expensive jigs and measurement devices. It is not dependant on the lens chosen by the user\installer and is purchased separately. The lenses are not sourced from Delta. The whole alignment process is very expensive. Production of the light engines used in the c3x line is done I believe by a different alignment procedure costing far less and accomplished much quicker obviously lowering labor costs. Tighter tolerances on any part raises the costs.

I spent a day at the US SIM2 facility a little over a year ago serially viewing the picture thrown by each SIM2 projector then in the line. I also examined the insides of many of the projectors especially focusing on the c3x 1080 and the HT5000. The quality of the picture substantially improved as we went from the cheapest to the most expensive. There was a hugh jump in PQ going from the 3000 to the c3x 1080. The c3x 1090 made the 3000 unwatchable. Going to the HT5000 was another hugh improvement. Looking at the light emgine, the entire inside optical path and the various lenses made the c3x 1080 look like a toy compared to the HT5000. I can hardly wait until the DI and lamp technolgy used in the new Lumis are employed in the HT5000. Unfortunately I will never be able to afford one. Hell for my 54 x 96 1.3 gain screen, the Lumis let alone the HT5000 is too bright.

The source here was DVD movies.

Michael W.

03-12-09, 01:59 AM

My point is that the optical system of a 3 chip projector is so much more complex than the single chip and the mechanical alignment tolerances on the reflective 1080p chips are so tight that tip tilt decenter and rotation that the variation in performance from projector to projector will be very large with some people being very happy and others not so much. ALL you have to do is scan this forum to find people who were extremely unhappy with their msconverged three chips. Until the Lumis I don't believe that their was any pixel shifting control on three chip DLP projectors. the worse shift you could get is half a pixel. I guess it takes out tip tilt and vertical and horizontal decenters but rotation im not sure about. anyway macro pixel shifting could really help

There was. The C3X1080 has the exact same pixel shift mechanism as the Lumis. So does the 720p C3X-E.

Craig Peer

03-12-09, 08:08 PM

I can hardly wait until the DI and lamp technolgy used in the new Lumis are employed in the HT5000. Unfortunately I will never be able to afford one.
I can't wait to see what SIM comes up with for a new HT5000 too.

Hell for my 54 x 96 1.3 gain screen, the Lumis let alone the HT5000 is too bright.

I don't think the Lumis will be too bright on my 104" wide screen, but if it is I'm willing to use a ND filter until the lamp dims somewhat. The picture is just too good!!

mark haflich

03-13-09, 09:20 AM

Fine for your width. I found it too bright on my screen with the lamp at 200 watts. The answer would have been to switch to a lower or negative gain screen material and one that would have improved the black level (I am talking the c3x 1080, not the Lumis).

mlang46

03-15-09, 03:31 PM

It seemed that the sub-pixel stuff is what you were excited about. Was it the full-pixel stuff like JVC and SIM2 do that you were really excited about? I'm still not sure what you thought Sony was doing that is different than they are actually doing.
Why would you shift 3 pixels to .5 pixel? Why not shift it whole amounts if it is off by whole pixels?

And if .5 a pixel off is okay with you, then why get excited about feature doing basically a simulation of sub-pixels?
Getting excited about full pixel shifts makes sense to me. And for some people getting excited about something like the Sony with a kind of sub-pixel thing using 2 adjacent pixels, but for somebody like you who kept going on about misconvergence causing a loss of resolution from 1080p to 720p, I don't understand you getting excited about a feature that basically hurts resolution to try to overcome misconvergence of something other than whole pixels.
Who is "your" in your above statement?

You made a statement:
and I addressed it with respect to what I would consider the best 3 chipper based on reports from people I trust and now you want me to compare to a lesser model instead.

If you want to counter some argument that somebody else made then that is fine, but as far as your statement about "any single chip or 3 chip dlp on the market" I would expect that to apply to the Lumis and not have a need to go back to last year's model.

--Darin

The thing I am excited about is the full pixel shift , shifting less than one pixel will not improve the image and frankly is impossible if you dither the image all you will do is blur the image. Zonal shifting does not mean sub pixel as far as I know.

from a manufacturing point of view once those panels are glued they are physically fixed. They can easily be off by a couple of pixels. If you don't have electronic convergence you will have a significant degradation in MTF which can easily be calculated. if you have full pixel convergence this limits your optical misconvergence to less than .5 pixels. of course even .5 pixels will under certain circumstances degrade the image resolution significantly because the optical transfer function is a convolution process.

It is not unusual in a projector system to have a lens with as much as 3 -5 percent lateral color aberration. that is 5 to 10 pixels. I am assuming when Sony is talking about Zonal convergence they are not talking about pixel averaging but they have figured out how to a zonally shift a variable number of full pixels as a function of field so as to correct for lateral color and our scaling the image to make room for these shifts. I don't know exactly what they are doing but having worked with a lot of Japaneses engineers, I know Unlike Microsoft they do a lot of testing and prototyping before they send it out to the public so what ever they are doing I am sure it improves the image.

But in any case electronic full pixel shifting is a big deal for 3 chip projectors and is now used by Sony, Sim2 and JVC.

its true that the Sim2 Ht380 and Sim2 Ht3000E do not have an adjustable Iris but having seen Planar 8150 I still think they are better projectors although I will take another look at the 8150 again

again should we forget the title this forum was single chip vs. 3 chip dlp and my position is that there is no significant technical advantage and there are significant price and production disadvantages to a 3 chip and any of advantages of the Lumis ,what they are in theory and what they are in practice will soon be incorporated into the single chip projectors

But theory is wonderful and practice is something else.

Have you actually seen a 3 chip DLP projector and compared it to a single chip projector from the same manufacturer?

in fact have you ever demoed a 3 chip DLP projector?

mlang46

03-15-09, 03:56 PM

I really am getting a bit confused from you last post Mlang. Maybe you can edit it a bit. Especially the one chip with a particular lens vs a one chip with a more expensive lens. Did you mean a 3 chipper with a more expensive lens.

I just don`t get how using two pixels and the space between them to do what one pixel can do can help the image. I do get how resolution and sharpness would be decreased. Not a good thing.

I do get the benefits of electronic full pixel shifts. Almost all good and little little bad. 3 pixel error fine, I can shift it 3 full pixels. 3.5 pixels not too bad. Tthe best I can end up with is .5 error, and that would I think would be the center spec for most three chip machines with most ending up with some error that was smaller than .5 using the electronic pixel shift.

Noah. I don`t know about most here, but I would be few watch very much black and white DVDs and I would bet that most watch mainly Bluray and HD TV. I certainly am buying anymore DVDs, just Bluray and I watch NTSC low def only when I have no alternative.

In the big boy forum, I suspect there is not much low def watching anymore.

Mark

What I was trying to say is that in a DLP projector with the latest dark chip IV the pixel to pixel contrast is according to the TI papers is 1100:1. Now there is no projector lens on the planet that can deliver that contrast to the screen without degrading the image so the image limitation is primarily coming from the optical system. So the fastest way to improve the image is to improve the optics because the limitations in the image resolution is not coming from the chip. The dynamic range , the black level is coming from the reflections off the back of the chip but the resolution is limited by the lens.

I am glad you were able to visit the factory and there is no doubt in my mind that the optics in the C3X 1080P are more complicated and expensive than the optics in the Ht3000E but every time I have compared their projected images and I did so on 3 separate occasions I liked the the 3000E better every time.

The only thing more I will say about electronic Pixel shift is that only full pixel shifts will improve the image and when 3 panels are out of alignment by more than .5 pixels it can improve the image.

dramatically. I was surprised to hear they actually have mechanical adjustments for the panels in the 5000. that has to be very expensive and will not be as stable as just optically gluing the panels to the color combining prism but it does make it repairable. What are they doing with the C3X systems?

darinp2

03-15-09, 04:06 PM

The thing I am excited about is the full pixel shift , shifting less than one pixel will not improve the image and frankly is impossible if you dither the image all you will do is blur the image. Zonal shifting does not mean sub pixel as far as I know.
...
I am assuming when Sony is talking about Zonal convergence they are not talking about pixel averaging but they have figured out how to a zonally shift a variable number of full pixels as a function of field so as to correct for lateral color and our scaling the image to make room for these shifts.How do you think a full pixel jump at different points would look in order to do full pixel shifts of different amounts in different zones? Do you really expect that they wouldn't use the sub-pixel thing done with 2 pixels (as I've explained) to smooth things between zones that had been adjusted to different convergence settings? In case it isn't clear, lets use a case where the red has been shifted one to the right digitally at the center of the screen and 3 to the right on the far right side of the screen. That is 2 pixels worth of change that have to be done somewhere.
again should we forget the title this forum was single chip vs. 3 chip dlp and my position is ...We shouldn't forget anything and I have think I have made it extremely clear what position you took that I commented on. You shouldn't need to switch to discussing a different position you took to address that, as if I had argued against your other position. For instance, there is no need for me to go backwards to a previous generation 3 chipper with regard to your prediction about how the 8150 would do against the best currently available 3 chipper. The previous SIM2 stuff could leave me preferring the single chipper and it wouldn't change my comments at all. Especially if the single chipper had the better on/off CR there, but doesn't between the Lumis and the 8150 (in general).
Have you actually seen a 3 chip DLP projector and compared it to a single chip projector from the same manufacturer?I don't believe I have ever compared them in the same room other than at shows like CEDIA, which I largely take with a grain of salt. I did own a SIM2 3 chip DLP (one of the older 720ps). Have you ever owned a 3 chip DLP? Not that it really matters that much since I think I've made it about as clear as possible that my comment was about the 8150 against the Lumis and there isn't much need to waste both our times by trying to deflect away from that and argue for stuff I didn't comment on, as if I had argued against it.

How about addressing the 8150 vs the Lumis knowing that the Lumis has full pixel shift? This is what you said that I commented on:
My prediction would be that the planar 8150 should give the best consistently delivered image performance of any single chip or 3 chip dlp on the market and would even have a higher probability of delivering this performance if it incorporated the dark ship 4 and the new insights TI has found for optimizing the optics.--Darin

Art Sonneborn

03-15-09, 04:21 PM

from a manufacturing point of view once those panels are glued they are physically fixed. They can easily be off by a couple of pixels.

Of the dozen or so three chip DLPs I've seen I've never experienced panel alignment that far off. In my own case my HT 5000 has over 70% of the image being perfect while the corners range from 1/6 to 1/3 of a pixel.

Lower right corner
http://i127.photobucket.com/albums/p146/CarlosMeat/Focus08a.jpg

It is not unusual in a projector system to have a lens with as much as 3 -5 percent lateral color aberration. that is 5 to 10 pixels.

Indeed that is unusual not usual. Even with my ISCO III in place I have about one pixel CA at one corner .

If your assumed numbers were correct I'd run from three chip DLP as well but they are far from accurate.

Art

coldmachine

03-15-09, 04:45 PM

Of the dozen or so three chip DLPs I've seen I've never experienced panel alignment that far off. In my own case my HT 5000 has over 70% of the image being perfect while the corners range from 1/6 to 1/3 of a pixel.

Lower right corner
http://i127.photobucket.com/albums/p146/CarlosMeat/Focus08a.jpg

Indeed that is unusual not usual. Even with my ISCO III in place I have about one pixel CA at one corner .

If your assumed numbers were correct I'd run from three chip DLP as well but they are far from accurate.

Art

Pretty much my own experience, and that of many others. I cant believe some of the fecal outpourings I'm reading here.

Its bad enough when some ass flange is flapping their lip everywhere, but when they're also all over the place, like a mad woman's sh1t, it gets a bit much.

noah katz

03-15-09, 04:51 PM

"the pixel to pixel contrast is according to the TI papers is 1100:1. Now there is no projector lens on the planet that can deliver that contrast to the screen without degrading the image so the image limitation is primarily coming from the optical system. So the fastest way to improve the image is to improve the optics because the limitations in the image resolution is not coming from the chip. The dynamic range , the black level is coming from the reflections off the back of the chip but the resolution is limited by the lens."

Something isn't adding up here.

How much pixel-pixel contrast is required to be considered as having achieved a given resolution?

It's got to be way less than 1100.

With ANSI CR of ~250:1, my RS1's adjacent-pixel CR must be a fraction of that.

Yet every pixel is clearly resolved on the screen and no one has claimed it's not 1920x1080 res because of that.

mlang46

03-15-09, 04:53 PM

Mlang has also taken a position about the center of the field from the projector being really important here:
In that case it would be center misconvergence that would matter most and that is true to a large degree, but as I basically pointed out, the center of the image from the projector and the center of a person's vision are not the same thing. A person can look wherever they want. Most people will look toward the center of the image more often than the edges, but will look around some while watching content.

I think mlang has some good stuff to contribute and I'm sure he thinks he is doing good, but IMO the 2nd part of this statement:
has been appropriate to portions of his posts too often in his attempts to disparage the SIM2 3 chippers.

Actually for all practical purposes they are unless of course you are watching a tennis match. All optical engineers and lens designers ,including myself, take advantage of the fact that a movie goer's attention is directed at the center of the image. People for the most part do not turn there heads to watch the edge of the screen they can but they don't which is why we let the lenses resolution drop as the field increases.

Disparage My arguments have all been technical. I do not understand why other than power output why a 3 chip dlp projector will produce a better image than a single chip projector

1. Do you disagree with me that a 3 chip projector requires a longer back focal length and that that longer back focal length makes that lens much harder to manufacture than a lens with a shorter back focal length

2. do you disagree that placing more glass in a converging beam increase the aberrations in the system

3. Are their any blue ray discs which encode more than 8 bits of RGB

4. You don't think you can get another 2 bits of gray level out of a single chip dlp projector using a pulsed lamp made by Osram

5. How much do you think the MTF is degraded in a projector if the panels are misaligned by only .5 pixels

6. You don't think its extremely hard to produce a multicolor beam combining prism with out producing color shifts in a compact projector like the C3x 1080p or the Lumis or do you think its easy. If you think it is easy why do you think their are complaints on this forum about color misconvergnece in 3 chip Dlps

7. have you ever done an actual A/B comparison of a 1080P 3 chip projector from the same company with a single chip projector

8. finally who do you think is more likely to have a hidden agenda ,an optical engineer who actually designed these systems or a dealer who makes 3 to 10 times the money selling you a 3 chip DLP projector than if he sold you a single chip DLP projector?

--Darin[/QUOTE]

amirm

03-15-09, 04:56 PM

The thing I am excited about is the full pixel shift , shifting less than one pixel will not improve the image and frankly is impossible if you dither the image all you will do is blur the image.
The technology is not "dither." It is interpolation. You know, the same thing used to upscale DVDs. Or even produce 4:2:2 out of BD's 4:2:0.

It is not unusual in a projector system to have a lens with as much as 3 -5 percent lateral color aberration. that is 5 to 10 pixels. I am assuming when Sony is talking about Zonal convergence they are not talking about pixel averaging but they have figured out how to a zonally shift a variable number of full pixels as a function of field so as to correct for lateral color and our scaling the image to make room for these shifts.
There is no magic about what Sony is doing as a couple of us explained earlier. It is a simple interpolation limitted to adjacent pixels and variable based on each zone. BTW, "pixel averaging" is a form of interpolation. Not a good one but a form of one nevertheless.

I don't know exactly what they are doing but having worked with a lot of Japaneses engineers, I know Unlike Microsoft they do a lot of testing and prototyping before they send it out to the public so what ever they are doing I am sure it improves the image.
Say what? What does testing and Microsoft have to do with this? The algorithm is textbook signal processing 101. It trades off one kind of distortion (convergence) for another (softness of the interpolating filter). It works as well as the technology lets it. But not because Sony tested it more or did some fundamental research here...

coldmachine

03-15-09, 05:05 PM

have you ever done an actual A/B comparison of a 1080P 3 chip projector from the same company with a single chip projector

Many times, as an owner of a number of both. Ive had them on the same screen. Others here have direct experience of both.

What do you want to know?

finally who do you think is more likely to have a hidden agenda ,an optical engineer who actually designed these systems or a dealer who makes 3 to 10 times the money selling you a 3 chip DLP projector than if he sold you a single chip DLP projector?

Which dealer are you referring to there?

On deciding who is operation to an agenda, I find applying Occam's razor to the evidence to be fairly reliable. It certainly seems so in this case.

mlang46

03-15-09, 07:42 PM

[QUOTE=darinp2;16048576]How do you think a full pixel jump at different points would look in order to do full pixel shifts of different amounts in different zones? Do you really expect that they wouldn't use the sub-pixel thing done with 2 pixels (as I've explained) to smooth things between zones that had been adjusted to different convergence settings? In case it isn't clear, lets use a case where the red has been shifted one to the right digitally at the center of the screen and 3 to the right on the far right side of the screen. That is 2 pixels worth of change that have to be done somewhere.

So I take it as a NO that you have never compared a 3 chip 1080P DLP to a single chip 1080p DLP and NO I have never bought a 3 chip DLP projector because when I compared them at CEDIA and the Dealers showroom and in my home I preferred the single ship.

As far as Sony is concerned as I said before I do not know how they are doing it but if they are not shifting full pixel shifts they are degrading the resolution at that zone. One way they could compensate for the variation in color misconvergence with field is to introduce geometrical distortion. of course unlike most people who watch movies you actually ignore the center and concentrate on the edge, so actually you may notice the geometric distortion while most people won't.

the Planar 8150 may be a very good projector in principle but alas principle is only a dim desire when it comes to production

noah katz

03-15-09, 09:47 PM

mlang,

"1. Do you disagree with me that a 3 chip projector requires a longer back focal length and that that longer back focal length makes that lens much harder to manufacture than a lens with a shorter back focal length

2. do you disagree that placing more glass in a converging beam increase the aberrations in the system

6. You don't think its extremely hard to produce a multicolor beam combining prism with out producing color shifts in a compact projector like the C3x 1080p or the Lumis or do you think its easy. If you think it is easy why do you think their are complaints on this forum about color misconvergnece in 3 chip Dlps"

Those may all be true but it doesn't automatically follow that 1-chips outperform 3-chips in those areas.

It does follow that the same performance will cost more to achieve in a 3-chip.

darinp2

03-16-09, 01:32 AM

Actually for all practical purposes they are unless of course you are watching a tennis match.So why were you so excited about the zone thing with the Sony? Why is that necessary if people are only looking at the center of the screen?
1. Do you disagree with me that a 3 chip projector requires a longer back focal length and that that longer back focal length makes that lens much harder to manufacture than a lens with a shorter back focal length

2. do you disagree that placing more glass in a converging beam increase the aberrations in the system

3. Are their any blue ray discs which encode more than 8 bits of RGB

4. You don't think you can get another 2 bits of gray level out of a single chip dlp projector using a pulsed lamp made by Osram

5. How much do you think the MTF is degraded in a projector if the panels are misaligned by only .5 pixels

6. You don't think its extremely hard to produce a multicolor beam combining prism with out producing color shifts in a compact projector like the C3x 1080p or the Lumis or do you think its easy. If you think it is easy why do you think their are complaints on this forum about color misconvergnece in 3 chip Dlps

7. have you ever done an actual A/B comparison of a 1080P 3 chip projector from the same company with a single chip projector

8. finally who do you think is more likely to have a hidden agenda ,an optical engineer who actually designed these systems or a dealer who makes 3 to 10 times the money selling you a 3 chip DLP projector than if he sold you a single chip DLP projector?I thought I had been pretty clear when I said:You shouldn't need to switch to discussing a different position you took to address that, as if I had argued against your other position.
...and there isn't much need to waste both our times by trying to deflect away from that and argue for stuff I didn't comment on, as if I had argued against it.

How about addressing the 8150 vs the Lumis knowing that the Lumis has full pixel shift?and there you go again refusing to address your prediction about the 8150 against "any single chip or 3 chip dlp on the market" with respect to the Lumis (which is obviously a 3 chip DLP on the market) and waste my time with other questions as if I had disagreed with those. If you can't address your prediction with respect to the 8150 vs the Lumis then just say so. Of course it is harder and more expensive to make a good 3 chipper in some ways, but there is room for some stuff like that between $8k and close to $40k MSRP.

Do you stand by your position about the 8150 vs the Lumis or don't you? I don't know how I can make it clear enough that needing to move to other projectors is wasting time for both of us.

As far as agendas, I don't think yours is hidden at all. And I already know you have been an optical engineer and I think you have some good stuff to add from that. But I also think you step into areas that you don't know and then think you must be right because of your background. For instance, if I should trust you because you have been an optical engineer then please explain why you were excited about Sony doing a zone system and what you think it should do if it can't do anything at the sub-pixel level, especially if only the convergence in the center of the image matters. You made a comment about maybe the Sony engineers not understanding this stuff, but yet it is you who looks like you didn't get it. Again, what do you think jumping a whole line for one color at one point in order to have different convergence adjustments in different zones would do?

I could have trusted you before based on your background, but if I had I would have thought that the C3X1080 was rated at 5k:1 on/off CR like you claimed or that it came out later that the projector people were raving about had .75 pixel misconvergence, when it didn't come out later and if the center convergence is what matters, it definitely was not .75 like your claim about what the misconvergence was. The first projector he got was close to perfect for convergence in the center and the 2nd one was better than .75 pixels in the center according to him. I realize the above were likely honest mistakes, but I think they fit with the agenda you had and you probably remembered things the way you did because you were looking for ways to disparage the Lumis.
of course unlike most people who watch movies you actually ignore the center and concentrate on the edge, so actually you may notice the geometric distortion while most people won't.Seriously, why the stupidity? It is funny how you were so upset about misconvergence from tip and tilt and now you have stooped to this level. Why don't you tell people again, if the center is off by less than .5 pixels in all cases, then why are you concerned about misconvergence from tip or tilt outside there? What I said is true, the center of the image and the center of a person's vision are not the same thing. As an optical engineer you should have been able to figure that out even though people are likely to look more toward the center of the screen, like I said. If they were the same thing then our current video systems could have been designed quite a bit differently than they are.

Does convergence 5 degrees away from the center of the image with respect to the viewer matter much? How about 10 degrees? If not, why were you so concerned about tip and tilt and why were you so excited about the Sony having a zone system?
the Planar 8150 may be a very good projector in principle but alas principle is only a dim desire when it comes to productionSo, why not address your comment about the 8150 vs the Lumis? No need to run from it with questions about different things. Just address the 8150 vs the Lumis. Is this a position you agree with at this point knowing what you know:
the planar 8150 should give the better consistently delivered image performance than the SIM2 Lumis--Darin