AVS Forum banner
Status
Not open for further replies.
1 - 20 of 33 Posts

·
Registered
Joined
·
16,627 Posts
Discussion Starter · #1 ·
A few questions for anyone in the know (or willing to hazard a guess:) about the new HD2+ chips and the new 7-segment color wheel.


If I've understood well from what I've read, the HD2+ is improved in two areas:


1/ "dimple fix" means that the mirror hinge is now more reflective so that the dark "dot" visible in the center of each pixel should now be less visible


(CAN ANYONE CONFIRM THIS - IS IT REDUCED OR ELIMINATED???)


2/ reduced screen door - I assume this means the dead space between pixels has been reduced but I have seen no explanation as to how/why


(CAN ANYONE CONFIRM THIS - ANY ESTIMATES ON HOW MUCH IT IS REDUCED RELATIVE TO HD2???)



Beyond this, there is now the new 7-segment color wheel. It appears that some manufacturers like DWIN are using the HD2+ with the old color wheel - can anyone confirm this? I assume the 7 segment wheel has been introduced either to reduce rainbows even further and/or to improve contrast ratio and/or apparent dark levels, especially for dark scenes. Can anyone confirm this? Does anyone have an opinion if the improvements related to this new color wheel are significant?


Reduced dimple and screendoor, are both nice improvements (I have an M20X, where both are noticable), but I am trying to understand if the new color wheel makes a significant further improvement, or whether a projector using the HD2+ with the older 6-segment color wheel (like the DWIN) is leaving anything significant on the table.



Thanks in advance for any answers or opinions...



-fafrd
 

·
Registered
Joined
·
1,166 Posts
Good questions, hopefully the gurus around here have a few answers.
 

·
Registered
Joined
·
1,016 Posts
Quote:
By Texas Instruments

The difference between HD2 and HD2+ is a new manufacturing process that enables the mirrors to be spaced more closely to each other and that enables the 'via' - the small 'hole' in the middle of each mirror - to be smaller. This provides a slightly larger reflective surface area to deliver more brightness, and reduces the opportunity for light to reflect from mirror edges, thus increasing contrast
 

·
Registered
Joined
·
1,016 Posts
Texas Instruments doesn't make color wheels for their customers (as far as I could know).

Look at manufacturers: I don't think there are two using exactly the same color wheel while using the same DMD.


But I assume Texas provides some advices/algorythms and maybe some specifically programable chips to help manufacturers in the implementation of the color wheel.
 

·
Premium Member
Joined
·
126 Posts
I spoke with a TI engineer at CEDIA, and here's what he said:


The HD2+ DMD partially fills the small dimple in the middle of the chip and also increases the surface area of the mirror. The increase in the surface area is what reduced the screen door effect. I saw several projectors with the new HD2+ chip and it definitely reduced the dimple almost to the point where it wasn't visible. The only place where I saw a hint of the dimple was on a full field white pattern, but even there it was very faint. As to the increase in surface area, I did not have a side by side comparison of an HD2 and HD2+ so really couldn't tell how much of a difference there was.


The new 7-segment color wheel includes a neutral density green segment that is used mainly during dark scenes to increase the bit depth, which reduces the need for dithering. I viewed some dark scenes of LOTR:TT on one HT2+ system and was impressed by the reduction in dithering.


Kevin
 

·
Registered
Joined
·
4,354 Posts
Ken

it seems the actual spacing between mirrors IS reduced also. so, a real screendoor reduction, closing in towards DILA screedoor.

the via filling indeed will also reduce significantly the parasiting luminosity reflect by it in the non HD2+ DMDs.

fascinating technology.
 

·
Registered
Joined
·
646 Posts
Hi,


I send a question to Sim-2 about the wheel they use in their HD2+ products. They mention it is an improved 6-segments wheel so I wanted to know the difference with the so-called 7 segments wheel.


Here is their reply:


"Regarding the 6-segment color wheel compared to the 7-segment color

wheel the comparison is very simple:


The 7-segment color wheel is said to increase the contrast and to do so it decreases the black level, however by doing this it also affects the

brightness.


We have reached an optimal compromise between brightness and contrast on our products by maintaining a 6-segment color wheel which also allows

for a better color reproduction.


We feel that with this new improved 6-segment color wheel (smaller and more silent) we are able to offer top quality colors and contrast/brightness

ratio."


Cheers,

Michel
 

·
Registered
Joined
·
16,627 Posts
Discussion Starter · #9 ·
Thanks to all for the replies. So if I understand correctly, the improvements from reduced via and reduced screendoor are real and apparent, whereas the improvements from the 7 segment wheel are still being debated between the manufacturers. In any case, the fact that the new technology supports 10-bit resolution(over what in the HD2 generation? - 8 bit?) must be behind the new concepts in color wheels.


I suppose that 10-bit versus 8-bit signalling depth should mean that the "dead zone" between colors can be reduced. This can be optimized in two different ways:


1/ maintain the same intensity using less area of the wheel (and so freeing up some additional area of the wheel for a possible additional color possibly for better perfromance in one corner of the performance space, such as dark scene resolution/dithering - the 7 segment wheel)


2/ use the additional "space" to get more brightness out of the same lamp (by using more of the wheel for projecting)


Sounds like Sharp is opting for the first approach while Seleco appears to be opting for the second (assuming they are using new drivers)


Because of their "upgrade policy" I am concerned that DWIN is using the same basic drivers with both HD2 and HD2+ technology, which ought to mean that they are not exploiting the additional bit depth in any manner. Anyone have a comment on this? Has anyone checked with DWIN if the drivers are changed of not, if the HD2+ projector is exploiting full 10-bit image depth, if DWIN is claiming any additional performance numbers either in terms of effective brightness or dark scene performance from the new technology, or if they are merely taking advantage of dimple-fiz and screendoor improvements but leaving contrast/brightness/darkscene perfromance as they were???



-fafrd
 

·
Registered
Joined
·
32,172 Posts
I'm all-but certain that the enabling technology for the new color wheel comes from TI. In fact, I'd guess that de facto, TI produces / sources the new wheel.


Because implementation of the DMD and the associated chipset is up to the projector manufacturer, it is possible that SIM simply chooses not to turn on that implementation.


As for reduced screendoor, I respectfully disagree that anything meaingful has changed vis a vis the interpixel. The smaller "via" or dimple certainly results in a higher reflective surface area. That certainly would increase the true fill ratio. But the inter-pixel grid? That only improves if the mirrors are actually larger (the DMD size is the same between HD and HD2+).


If anything has improved vis a vis screendoor, there wasn't anyone at CEDIA who was even remotely hinting that was a benefit of the HD2+ -- at least no one at any of the HD2+ projector demos.


And, fwiw, I believe most of the CR increases from the HD2+ projectors come from irises and/or new color wheels, while the chip contributes something somewhat less.


Mark
 

·
Registered
Joined
·
4,735 Posts
Quote:
Originally posted by fafrd
In any case, the fact that the new technology supports 10-bit resolution(over what in the HD2 generation? - 8 bit?) must be behind the new concepts in color wheels.
My understanding is that the 10-bit resolution is only for the green channel, the red and blue stay unchanged with 8-bit processing. It is this 10-bit resolution and the dark segment that allows them to use the green channel to get dark gray resolution and reduce dithering in the darks. I wouldn't expect extra bit resolution to help make anything brighter. I guess I just don't see how the bit resolution would effect the no-man's-land between pixels or help the overall brightness.


I know I've heard multiple times that they made the mirrors bigger, but I don't remember where it was. It seems like it might have been during the Sharp 10k demo. I meant to go look at one up close to see how the gap compared, but didn't get around to it.


--Darin
 

·
Registered
Joined
·
3,211 Posts
I never see rainbows, but I get very bad headaches when viewing DLP projectors. Is it possible that the source of the headaches is not rainbows, but dithering? If it is, then perhaps the new HD2+ DLP projectors with the 7 segment colour wheels designed to reduce dithering would eliminate or at least reduce headaches in those who are susceptible.
 

·
Registered
Joined
·
1,204 Posts
I attended a roundtable discussion sponsored by TI at the show, and DLP was the focus. On the topic of bit depth, the TI engineer who was moderating said that the HD-2 chip was capable of 10 bit resolution all along, as is the HD-2+, but whether the system is actually 10 bits with either one depends on each manufacturer's implementation. He stated that there are some 10-bit and some 8-bit out there, and declined to be more specific. The increase in contrast ratio provided by the improved dimple is mainly due to a reduction in light spray or uncontrolled diffraction in the dimple itself, since it is smaller. Therefore, less stray light leaks through the lens when a mirror is "off". The dark substrate of the chip is not substantially different from the HD-2. There are improvements to be made, since they said that the current Dark Metal DLP Cinema chips are superior in that regard. They mentioned inter-pixel spacing as something they have improved over the life of DLP technology, but I don't recall that being a specific improvement from HD-2 to HD-2+, although I might have missed that point.
 

·
Registered
Joined
·
1,204 Posts
Oh, and I also got the impression we shouldn't hold our collective breath for 1920 x 1080 DLP. I specifically asked him about it, and mentioned that I had seen one at their facility that was produced in 1994. He acknowledged the chip, but said that TI has already demonstrated the "actual resolution" advantage of 1280 x 720 DLP over some devices with "higher rated resolution" and implied that what we have today is good enough. He did admit that if the market demanded it, they could build it quickly, but were in no hurry otherwise.
 

·
Registered
Joined
·
710 Posts
Quote:
He acknowledged the chip, but said that TI has already demonstrated the "actual resolution" advantage of 1280 x 720 DLP over some devices with "higher rated resolution" and implied that what we have today is good enough. He did admit that if the market demanded it, they could build it quickly, but were in no hurry otherwise.
That engineer must have received marketing training; it reminds me of that Michael Fox TV show "Spin City".


Don
 

·
Registered
Joined
·
16,627 Posts
Discussion Starter · #16 ·
Quote:
I guess I just don't see how the bit resolution would effect the no-man's-land between pixels or help the overall brightness.
In principle, 10-bit versus 8-bit could allow the "cutoff" to be spatially/temporally located closer to the boundary between colors. I assume that today there is a "no-mans land" where the mirrors are shut off during the transition. With 8-bit resolution this gap would have to be at least 1/256 of the full segment. With 10-bit resolution it could in principle be 1/4 this size (1/1024). This all assumes that the mechanical accuracy of the color wheels themselves is within these limits.


Quote:
If anything has improved vis a vis screendoor, there wasn't anyone at CEDIA who was even remotely hinting that was a benefit of the HD2+ -- at least no one at any of the HD2+ projector demos.
I believe there was someone who took some close up screenshots (of the Sharp 12000, I believe) and also claimed that he was told reduced screendoor was a benefit of the 12000 over the 10000. This would of course require larger mirrors and a reduced space between the mirrors. There were alot of detailed posts way back when quantifying screendoor between different technology - I would think it is just a matter of someone getting ahold of a unit and remaking some of those similar measures. The ultimate test, of course, would be to hold a side-by-side between two like projectors with HD2 and HD2+ (Such as Sharp 10k and 12k, for example)


Quote:
It is this 10-bit resolution and the dark segment that allows them to use the green channel to get dark gray resolution and reduce dithering in the darks.
I get it - so the "dark green" segment is specifically to throw very low levels of gray in a kind of "dark B&W mode" (I assume for this to do anything useful, no light is thrown through any of the other segments when the dark green segment is used [for a given pixel in a given frame])


I am pretty familiar with screendoor and the dimple, having owned a Sharp M20X. Dithering I am less familiar with, but I assume it refers to the dim "jumping around" of the image in very dark area noticeable from close up. I guess they do this to compensate for limited grayscale by turning pixels on with the minimum gray level for one frame and completely off the next frame to try to create the appearance of an intermediate shade of "1/2"


Thanks to anyone to let me know if I've got this right.


What has me confused is that going from 8-bit to 10-bit resolution should mean that there are now "true" intensity levels of "1/2" and "1/4" (on a relative scale keeping 256 as the max) without needing to use any artificial techniques such as dithering and also without the need to use any new "dark gray" segment on the color wheel. Only going towards intensity levels of "1/8" and less would dithering still be needed (or a new color wheel segment to give the equivalent gray-scale intensity without the use of dithering).


Anyone who can shed any light on this would be greatly appreciated. In particular, what I am curious to know from anyone who can do the side-by-side is how do the following projectors compare in very dark scenes:


Sharp 12k vs 10k

Infocus 7205 vs. 7200

DWIN TV3 HD2+ vs HD2

SIM2 HD300 XTRA vs HT300


To me it seems that it is probably much more important to know who is using true 10 bit drivers than who is using the 7 segment wheel. If DWIN and SIM2 are just using old 8-bit drivers, there should be little apparent change in dark scene/dithering performance, while if they have moved from 8-bit resolution to true 10 bit resolution , the improvement should be very evident and probably more significant than the additional improvement that is possible with a "dark gray" section on the wheel.


Any insights appreciated.


-fafrd
 

·
Registered
Joined
·
4,735 Posts
Quote:
Originally posted by fafrd
In principle, 10-bit versus 8-bit could allow the "cutoff" to be spatially/temporally located closer to the boundary between colors. I assume that today there is a "no-mans land" where the mirrors are shut off during the transition. With 8-bit resolution this gap would have to be at least 1/256 of the full segment. With 10-bit resolution it could in principle be 1/4 this size (1/1024). This all assumes that the mechanical accuracy of the color wheels themselves is within these limits.
My impression is that they turn all mirrors off when a transition point on the wheel reaches the first mirrors and don't turn them back on until it passes across the other side. I don't think they follow the line across, but I could be wrong.


If true, I think the effect of 8-bit vs 10-bit would not effect this transition time much at all. I don't know how big this no-man's-land is, but I bet it is more than 5%. I know that Optoma has claimed a big brightness boost by having a mode that leaves the mirrors on during the transitions. I think the Sharp 10k might be doing the same thing, since they have a white-peaking mode.


--Darin
 

·
Registered
Joined
·
3,465 Posts
In the Secrets report, the reviewer saw no dithering while standing at the screen while viewing the IF 7205, one of the HD2+/7 segment wheel projs. He found a "noticeable improvement in CR and black level [in HD2+ projs] over HD2 projs."


On the Dwin, it is my understanding that the only change being made to the TV3 is going from the HD2 to HD2+ chip; everything else is the same. It is interesting to note that w/only this change Dwin claims little to no difference in picture from the TV3 w/HD2 to the TV3 w/HD2+.


On the new SIM2s HT300s using the HD2+ chip, SIM2 is using a 6 segment color wheel. SIM2's press release claims that this is an improved 6 segment wheel.


Of the four HD2+/7 segment wheel projs. released, Marantz and Sharp both claim 10 bit processing w/their 7 segment wheels. InFocus/Toshiba press releases do not address this issue, but it appears from the above posts that the 7 segment wheel yields 10 bit processing on green.
 

·
Registered
Joined
·
32,172 Posts
Again, fafrd, I just don't think interpixel spacing / screendoor is improved. And jogging my brain, I don't remember anyone else saying it was either.


Mark
 

·
Registered
Joined
·
9,224 Posts
Quote:
Originally posted by darinp
My understanding is that the 10-bit resolution is only for the green channel, the red and blue stay unchanged with 8-bit processing....

--Darin
This is an option of the driver electronics design. Remember it's a single chip DLP. If it's capable of 10bit operation in Green, there is no reason it can't do the same for red and blue. It's the same chip, just at a different point in the mux timeline for red and blue.
 
1 - 20 of 33 Posts
Status
Not open for further replies.
Top