or Connect
New Posts  All Forums:Forum Nav:

Optoma 8200 Review - Page 2

post #31 of 211
Quote:


No worries...I gave up on them a long time ago anyways. Too much work!

I am sure it's nice to know that Mark is taking the ball and running with it. You guys seem to have a great team over at AVS..might have to change the name to Audio Video Superheros
post #32 of 211
Can you advise your screen type and fabric used in your review please. thx
post #33 of 211
Great review. The one place that we differ on is the lumens. When I get to work I will post my calibration results but my lumens reading was allot higher. Your meter is better than mine though as I am using a i1Pro.
post #34 of 211
I am using a Stewart Studiotek 130G3. Projector is 15ft back and Ceiling Mounted. Disregard below 30 as the i1Pro does not read correctly under 30.

Mode - User
Contrast - 41
Brightness - 41
Color - 57
Tint - 47
Sharpness - 2
Noise Reduction - 0
Gamma - Standard *Curve type 4, offset -1*
Color Temp - Warm
Red Gain 5
Green Gain -3
Blue Gain -5

Red Bias 0
Green Bias 1
Blue Bias -1

Pure color Off
Pure Detail Off



post #35 of 211
Mark,

Thanks for a great review. Your observations match my brief demo of the 8200. I believe my HD81-LV to be a bit sharper edge to edge and brighter. It has a manual iris that I run pretty closed down(11) for my 100" wide Stewart Studiotek 2:35 screen.

It was great to read your comparison to the RS20. I have never found LCOS or LCD or SXRD as seductive as DLP, no matter the black and CR advantage.

I look forward to your comparing other DLP models, such as Planar 8150, InFocus IN83 and BenQ W20000 in the future. My concerns about Optoma QC and reliability make me hesitant in going that route again.
post #36 of 211
Thread Starter 
Quote:
Originally Posted by noah katz View Post

It's interesting how they're able get the greater brights along with darker blacks in dark scenes, given that the difference in lamp output between high and low lamp settings for pj's in general seems to be at most 25%.

Perhaps more (but many times more?) is possible on a short term basis.

I also wonder what effect it has on lamp life.

Yes, it's quite an achievement to deliver more brightness at 100 IRE while also reducing black levels by about 1/2. It does seem like it would drive the lamp very hard. I'd love to know more about this technology, if anyone has any links to white papers that discusses it in more detail, please let me know.

Quote:


On the Superwide, I was surprised looking at Optoma's picture at how little seemed to be lost on 2.35.

Turns out the picture is not accurate; I used calipers to measure and they show a 7% total loss in picture width where in fact it's 15% (2/2.35 = .85).

Regarding brightness, would you say that the 8200 is on par with the RS1?

Yes you're right. My original source for the Superwide numbers was TzungILin's post here (http://www.avsforum.com/avs-vb/showt...6#post15783316). After rereading it I can see that he said 7% on each side or ~14% total. Thanks for catching that and I'll update my post to reflect the change.
post #37 of 211
Thread Starter 
Quote:
Originally Posted by darinp2 View Post

That is probably part of it, but it also looks like they have higher intra-image CR in some of your tests than 6.5k would normally indicate. One of the main advantages of 15k:1 on/off CR over say 2.8k:1 on/off CR is up to around 5x or so higher intra-image CR in some scenes. 6.5k vs 2.8k would normally be like the 2.3x:1 range, but you got more than that in some cases.

Very good point. The intra-image contrast test shows much more contrast potential than is implied by the the on/off measurements alone. It does seem to have the potential to deliver the equivalent of 20k:1 although it's a mystery why the on/off measured relatively low.

Quote:


It sounds to me like that person from Optoma who told you that may not understand how the whole system works (including TI's part) or was referring to what happens after gamma is adjusted dynamically to account for the iris shutting down. Otherwise, how did you get the crushing from 80%stim to 100%stim on that one test pattern? Modulating the lamp shouldn't have done that.

It looks to me like their native or static on/off CR goes up significantly as the iris is closed. If that is the case then they could have also provided a higher native on/off CR with less lumens for white case with no DB, like some other DLPs.

This is another good point. When I first looked at the low APL greyscale test pattern, I assumed that I was seeing brightness compression very similar to what happens with dynamic gamma with a Sony DI implementation. But with dynamic gamma, it's the inability to boost luminance in the upper ranges that causes BC. If for example, the iris aperture reduces light output by 1/2, dynamic gamma can only make up for the light loss up until about 70 IRE or so (depending on gamma the 50% luminance point is around 60-70 IRE). Without an ability to raise the light output, dynamic gamma quickly runs out of headroom and the whites get crushed.

We can see from the relative white point graph that DB doesn't have this problem however. At 100 IRE, it's delivering 15% higher brightness at about 1/2 of the black level as the DB=off setting. Pretty impressive. So it seems like the technology is capable of a good greyscale without brightness compression. It could be that the technology just needs to be refined a bit more.

It'll be interesting to see where this technology is headed, especially given the interest in LED illumination systems which switch rapidly but tend to be dim. I can see where a manufacturer may be forced in the future to make a decision to tradeoff more rainbow artifacts for more intra-image contrast and vice versa.

It definitely makes for some interesting times in the DLP camp, which has been pretty quiet technology-wise the past few years.
post #38 of 211
Thread Starter 
Quote:
Originally Posted by Josh Z View Post

Mark, when PureMotion is turned off, will the 8200 display a 1080p24 input signal using 5:5 Pulldown, or does it automatically add 3:2 Pulldown and judder?

Due to lack of time I didn't spend a lot of time examining the video processing section and instead chose to focus most of my efforts on dynamic black. I would be surprised if they didn't use an even multiple of 24hz though. It's getting pretty rare for new designs to reinterlace and use 3:2 pulldown at 60hz. I wish I still had a unit to check this out though.

Quote:
Originally Posted by clrv View Post

clrv Great review. The one place that we differ on is the lumens. When I get to work I will post my calibration results but my lumens reading was allot higher. Your meter is better than mine though as I am using a i1Pro.

Throw plays a big role in lumens too and these were all measured at close to longest throw where there is usually less lumens.

Quote:
Originally Posted by blackdragon1 View Post

Can you advise your screen type and fabric used in your review please. thx

Stewart Studiotek 130. 110" 16x9. Completely light controlled room with black ceiling and black side walls transitioning to dark burgundy walls.

Quote:
Originally Posted by Jason Turk View Post

No worries...I gave up on them a long time ago anyways. Too much work!

Yeah the new guys always get stuck with all the crappy work (Just kidding of course) It was interesting to dig into dynamic black a little although it was a lot more work than I had expected.

Quote:
Originally Posted by Tom Huffman View Post

Mark, do you have post-calibration RGB and/or dE data for the grayscale?

Also, do you have post-calibration data for gamut and color decoding after you lowered the Color control 5 ticks?

I can dig that all up. I could have spent more time fine tuning the greyscale and getting it overall a little better, but there were some spikes that can't be addressed by gain and offset adjustments alone. An 11-step greysale adjustment like the new JVC's have would be nice. I will say though that the greyscale looked fantastic as far as a complete lack of banding or other artifacts.

Quote:
Originally Posted by Tom Huffman View Post

Finally, the white crush you saw in Cinema 1 and (to a lesser degree) in Cinema 2 sounds a lot like what Greg Rogers used to talk about with respect to the DI in Sony projectors. He called it "brightness compression." Since the Optoma uses a different technique for its DI than Sony, I'm wondering if you think this is essentially the same artifact?

Darin brought this up also and I talked about it a few posts back. I don't think it's the same artifact as BC, although visually they look similar. The term brightness compression, perfectly describes what is going on with a Sony DI because there is a brightness rolloff along the greyscale that is the inverse of the gamma curve shape that is needed to maintain brightness definition. So the whites do get compressed into each other with the Sony implementation.

From the luminance curve on the 8200 however, we can see that the 8200 doesn't rolloff in brightness like the Sony does. It's not a proper gamma shaped curve, but it is relatively linear which still should provide some delineation in the greyscale. Instead, it looks like whites are being crushed for some reason, so it's probably best to use a different term for it and white crush seems to fit.
post #39 of 211
Wing said you should tune blacks and whites with DB on to avoid the crushing.
post #40 of 211
Thread Starter 
Quote:
Originally Posted by guitarman View Post

"With DB enabled in Cinema1 mode, The whites from 80-100% were completely crushed and were not discernible from each other. As shown below:"

Most likely brilliant color is the culprit for the white crush. They must have added a level of brillaint color when using Cin1, probably a lesser amount of BC when using Cin2.

Was there anything in the menu's to disable brilliant color?

I don't recall seeing any settings for brilliant color. There is a "PureColor" feature which I disabled. In fact for the contrast meaasurements I disabled everything that I could disable except for those things that I was directly measuring.
Quote:


Wing gave me some info, you should tune the whites and blacks with DB on to avoid the crushing if you're going to use DB. The machine just uses brilliant color for video choices other than Cinema which uses none. The bright choice will use the most brilliant color, video/TV use a very small amount.

He says you can definetly use DB and not have crushed B/W by using the gamma choices, or just bright/contrast and with the DB on.

This is good to know and it confirms what I posted about DB having the luminance headroom so that it can avoid brightness compression.

As far as tuning the whites go, as I'm sure you know, the norm for setting the contrast adjustment is to use very bright test patterns. With any typical contrast test pattern, dynamic black won't be engaged even if the user has selected the mode because the image is very bright. So I think the contrast adjustment setting will end up about the same independent of the mode chosen. Once the image gets dark and DB is engaged, there is a huge difference in whites with the different modes though.

Since most people never tune whites with a low APL greyscale test pattern, it would be asking a lot for people to tune using an overall very dark pattern. Even if they did, I think once the scene returned back to a bright scene, I don't think that the contrast setting will be correct. So what's really needed is for the user to be able to calibrate the contrast setting with a bright test pattern (DB disengaged) and then have it scale uniformly in dark scenes for the appropriate DB mode.
post #41 of 211
Thread Starter 
Quote:
Originally Posted by guitarman View Post

Wing said you should tune blacks and whites with DB on to avoid the crushing.

I was just writing a reply as you posted
post #42 of 211
"I can see that he said 7% on each side or ~14% total. Thanks for catching that and I'll update my post to reflect the change."

That's good, but Optoma is still on the hook for their misleading picture.
post #43 of 211
Quote:
Originally Posted by Mark Petersen View Post

Yes, it's quite an achievement to deliver more brightness at 100 IRE while also reducing black levels by about 1/2. It does seem like it would drive the lamp very hard. I'd love to know more about this technology, if anyone has any links to white papers that discusses it in more detail, please let me know.



Yes you're right. My original source for the Superwide numbers was TzungILin's post here (http://www.avsforum.com/avs-vb/showt...6#post15783316). After rereading it I can see that he said 7% on each side or ~14% total. Thanks for catching that and I'll update my post to reflect the change.

so is it 5% and 7% per side?

For a total of 10% of the Vertical and 14% of the Horizontal?
post #44 of 211
Quote:
Originally Posted by Mark Petersen View Post

I was just writing a reply as you posted


He told me crushed blacks or whites is a pet peave of his and the PJ wouldn't be designed with the problem. Also mentioned you could try different Cinema user choice - gamma's, I think there's five of them. But in the end simple contrast and brightness adjust ments should do the trick. Like when you engaged DB and saw the crushing in the high end you should just lower the contrast till the detail comes back.
post #45 of 211
Quote:
Originally Posted by Mark Petersen View Post

When I first looked at the low APL greyscale test pattern, I assumed that I was seeing brightness compression very similar to what happens with dynamic gamma with a Sony DI implementation.

I think you probably were, except that Optoma adds lamp modulation on top of it. Put another way, looks like the Optoma uses dynamic gamma to me (like the Planar does) and the suggestion to lower the Contrast setting would probably lower the native on/off CR, much like people could lower the Contrast setting on the Sony's to get rid of brightness compression, but at the expense of native/static and dynamic on/off CRs.
Quote:
Originally Posted by Mark Petersen View Post

But with dynamic gamma, it's the inability to boost luminance in the upper ranges that causes BC.

While the Optoma looks like it could use the lamp to avoid brightness compression, it still looks like it was doing dynamic gamma in your case. Maybe the TI part is doing the dynamic gamma and the Optoma part is doing the bulb, causing the compression. Either way I think your measurements show that gamma was changed dynamically, any way you look at it.
Quote:
Originally Posted by Mark Petersen View Post

We can see from the relative white point graph that DB doesn't have this problem however.

In this case it looks like the Optoma projector could have used the lamp to get more peak lumens for the same CR without using dynamic gamma, but TI's part may include dynamic gamma and need something to be done to avoid having it do that (like telling it different iris positions than you actually have because the lamp is compensating).
Quote:
Originally Posted by Mark Petersen View Post

At 100 IRE, it's delivering 15% higher brightness at about 1/2 of the black level as the DB=off setting. Pretty impressive.

I agree that it is impressive, but looks like another case of a DLP manufacturer using DB to deliver more lumens, instead of to deliver more CR. What I mean is it looks like they could have given just as much CR natively, but with lower lumens. That is useful for people looking for high lumens and still caring about CR, but for those who were fine with less lumens (like those using a high gain screen) it doesn't necessarily give them more on/off CR that they could have achieved just by leaving 2 irises closed down (like the Marantz 11S2 and Sharp 20k). And without the DB artifacts.
Quote:
Originally Posted by Mark Petersen View Post

So it seems like the technology is capable of a good greyscale without brightness compression.

If it is used for higher lumens when you have high CR and they start at a low static on/off CR iris position and end at a much higher static on/off CR iris position. If it was used for even higher CR at lower lumens that generally come with fairly closed down irises then I'm not sure it would still have the above advantage. It looks to me like they are using the fact that static on/off CR can go up significantly when going from a very open iris to closed down with DLP and that isn't likely to apply as much if they started with the irises closed down a fair amount for higher native or static on/off CR.

Just for the sake of trying to figure out how they achieve this, let's go with your 2000:1 without the DB on. If the lamp is turned down low (I'll say 60% of max for this example, although I'm not sure they could go that low) and they could shut 1 or 2 irises down in a way that would lower lumens by 30% and increase static on/off CR by 2.3x (to 4600:1) then they could have 1/2 the black level and 15% higher white level in a dark image with a little bit of white and the lamp at max with the iris(es) shut down.

In the above case what they largely would have achieved is higher lumens at around 5k:1 on/off as opposed to higher on/off CR.

Sorry if I missed it, but did you notice any difference in how loud the projector was between low lamp and high lamp? Is so, then how about between DB on with the lamp on low and DB off with the same? If they are pushing the lamp up to and maybe even beyond what it does for high lamp then some extra cooling might be necessary, especially for very dark movies where the iris might remain shut down a lot.

It would be interesting to see graphs like yours for the Lumis and see if they did things similarly.

--Darin
post #46 of 211
"the suggestion to lower the Contrast setting would probably lower the native on/off CR"

Darin no that's off because if they tune whites with DB on then blacks should be tuned with db on lowering the black level and keeping the CR intact.

I get the projector tomorrow can't wait to see what's going on.
post #47 of 211
I'll ask if when DB is set to off if the Iris is in wide open mode, maybe it's closed down a little for more CR.
post #48 of 211
Quote:
Originally Posted by guitarman View Post

Darin no that's off because if they tune whites with DB on then blacks should be tuned with db on lowering the black level and keeping the CR intact.

No, that is not the way it would work unless Optoma messed things up. If you set the Brightness so that the mirrors are just off for video 16 with the DB off they should still be just off when DB is enabled.

Put a different way, Optoma and TI's job with the DB was to lower the blacks appropriately as the iris is closed down. That is how a proper DI works. A user in a dark room shouldn't need to set the Brightness with the DI enabled. Your statement above seems to have an assumption that they aren't doing the right thing for video black.

You can try setting the Brightness with the DB enabled and set it with the DB disabled and see if you come up with different values for this Brightness setting. If they are different by more than one click then Optoma should probably fix their implementation.

--Darin
post #49 of 211
Mark the light pulsing is the same item Alan started a thread on a while back.

http://www.avsforum.com/avs-vb/showt...743&highlight=
post #50 of 211
Both should be tuned with DB on which is the first thing I'll do tomorrow to keep the CR at it's best. You wouldn't tune blacks with DB off and whites with DB on, that would lower the CR. We don't want to leave the machine crushing whites for sure.
post #51 of 211
Thread Starter 
Quote:
Originally Posted by HiHoStevo View Post

so is it 5% and 7% per side?

For a total of 10% of the Vertical and 14% of the Horizontal?

Yes, I believe that's right. The native 1920x1080 is masked top and bottom to 1920x960 to get to the 2.0:1 aspect ratio. So with 16x9 content it uses 1920x960 so that works out to about 11% total image being lost at the top and bottom.

For 2.35:1 content, it stretches the image vertically to the same 960 pixels to fill the 2.0:1 screen. So this would yield 2256x960 so then it crops the sides to 1920x960 to completely fill the 2.0:1 screen. So a little over 7% per side is lost.

This is a pretty slick feature that requires no adjustable masking, no anamorphic lenses and the only downsides are 1) scaling artifacts in the 2.35:1 vertical stretch and 2) cropping of the image. But as TzungiLin posted, there isn't a lot of area lost to cropping and directors try to maintan a safe zone for such things.
post #52 of 211
What a superb in depth review Mark! Aside from being so technically thorough, the review benefits from your being so in tune with what we AVS readers want to know about a display.

Thanks,

Rich
post #53 of 211
Thread Starter 
Quote:
Originally Posted by guitarman View Post

He told me crushed blacks or whites is a pet peave of his and the PJ wouldn't be designed with the problem. Also mentioned you could try different Cinema user choice - gamma's, I think there's five of them. But in the end simple contrast and brightness adjust ments should do the trick. Like when you engaged DB and saw the crushing in the high end you should just lower the contrast till the detail comes back.

It's good to know that readjusting contrast can bring it out of white crush, but as you and Darin alluded to, in a well behaved system the brightness and contrast settings should always apply as global settings. So whatever is chosen for gamma curves, or DB modes should still work properly with the previously selected brightness and contrast settings.

IIRC though, I believe that I cycled the DB modes when I was adjusting the contrast and it didn't seem to make a whole lot of difference. Once DB is engaged in a dark scene though there is a huge difference between modes as far as what happens to whites at least with that test pattern. Incidentally, with real video content I really didn't notice any crush. But it is hard to spot BC or crush, unless a person is very familiar with certain source content and knows what to look for (or is doing a side by side comparison).
post #54 of 211
Thread Starter 
Quote:
Originally Posted by darinp2 View Post

While the Optoma looks like it could use the lamp to avoid brightness compression, it still looks like it was doing dynamic gamma in your case. Maybe the TI part is doing the dynamic gamma and the Optoma part is doing the bulb, causing the compression. Either way I think your measurements show that gamma was changed dynamically, any way you look at it.

I agree that the light coming out of the lens is modulated in a dynamic way due the combined effects of the lamp and iris which inherently changes the shape of the gamma curve. But I'm not seeing how the data shows that dynamic gamma is being applied in the Sony sense of the technology. In other words, I'd expect a luminance curve that instead of being shaped like a proper power curve is instead shaped more like a log(x) function.

I do understand your point about reserving a lot of lamp headroom to ensure that the system doesn't run out of brightness as the iris aperture is reduced, but if it had enough headroom available to fully make up for the loss of light from the iris at 100 IRE (which it apparently does from the luminance data), then I don't think that there would be a need to use dynamic gamma.

Am I missing something?

Quote:
In this case it looks like the Optoma projector could have used the lamp to get more peak lumens for the same CR without using dynamic gamma, but TI's part may include dynamic gamma and need something to be done to avoid having it do that (like telling it different iris positions than you actually have because the lamp is compensating).

I agree that it is impressive, but looks like another case of a DLP manufacturer using DB to deliver more lumens, instead of to deliver more CR. What I mean is it looks like they could have given just as much CR natively, but with lower lumens. That is useful for people looking for high lumens and still caring about CR, but for those who were fine with less lumens (like those using a high gain screen)

That would be a nice tradeoff. Maybe in the future the technology will get refined enough where a person can use the lamp technology in a way that favors their own setup and tastes. It's an interesting technology that seems like it could be used in many places in front projection, but I'm not sure where it's headed with LED illumination on the horizon.

Quote:
Sorry if I missed it, but did you notice any difference in how loud the projector was between low lamp and high lamp? Is so, then how about between DB on with the lamp on low and DB off with the same? If they are pushing the lamp up to and maybe even beyond what it does for high lamp then some extra cooling might be necessary, especially for very dark movies where the iris might remain shut down a lot.

It's a couple of dB louder in high lamp than low lamp. IIRC, I do remember it getting louder at times with DB enabled. It's so quiet though it's easy to ignore

Quote:
It would be interesting to see graphs like yours for the Lumis and see if they did things similarly.

Yes I'd love to dig into the Lumis implementation and see how it compares. I think Alan mentioned that he measured 20k:1 on/off so there is a big difference right there.
post #55 of 211
Thread Starter 
Quote:
Originally Posted by guitarman View Post

Mark the light pulsing is the same item Alan started a thread on a while back.

http://www.avsforum.com/avs-vb/showt...743&highlight=

Thanks for digging up the reference. I'm surprised I missed it when it was first posted.

Quote:
Originally Posted by R.Harkness View Post

What a superb in depth review Mark! Aside from being so technically thorough, the review benefits from your being so in tune with what we AVS readers want to know about a display.

Thanks,

Rich

Thanks Rich, Tryg, Craig and everyone else who have posted such nice comments. As you can tell I really enjoy this hobby and it's always a pleasure discussing it with everyone and exchanging ideas and viewpoints. AVS is pretty unique in this respect.
post #56 of 211
Quote:
Originally Posted by guitarman View Post

Both should be tuned with DB on which is the first thing I'll do tomorrow to keep the CR at it's best. You wouldn't tune blacks with DB off and whites with DB on, that would lower the CR.

It seems like you are sticking with your previous claim that adjusting the Brightness while DB is on will give you a different black level than adjusting the Brightness with the DB off and then turning the DB on. I don't mean this to offend, but it doesn't seem like you understand what the Brightness does with a DLP projector. If you think I am wrong, please explain why you think you would end up with a different Brightness setting if you calibrated it with the DB off than with the DB on.
Quote:
Originally Posted by guitarman View Post

We don't want to leave the machine crushing whites for sure.

I don't agree. Some people may feel that way, but then why even using the DB? That is a serious question. What benefit do you or Wing think the DB brings if there is never any crushing? And I wonder if it is Wing's position that this projector doesn't do any dynamic gamma.

A DB brings some positive, but there is some negative that comes along with it if you use it to its full or very large advantage. If you want to make a rule that there will be no crushing near white ever or no brightness compression near white then that is also going to be very limiting as far as how much benefit can be applied to the on/off CR or total CR range. Put another way, there are benefits that most would like from a DI that they won't get if there is never any compression or crushing, even with that extreme image that Mark used (almost all black with a small gray step area).

I would like to see some actual numbers, like what native and dynamic on/off CRs you get after setting the white levels so that there is never any crushing near white, even and extreme image that is almost all black with just a little bit of stuff near white. We could then compare those to what people would get if they just turned the DB stuff off and set the white point for that way.

From what I am reading my doubts about Optoma knowing what they are doing with respect to DIs are increasing. This is the same company that finished dead last in the DI rankings by the Cine4home guys with the HD80. The Planar projector hadn't been reviewed by them at that point, but when they did review it they liked the DI very much from what I recall. And I think Bob Williams is one of the most respected engineers around here. I hope he won't mind me saying that I know his recommendation is to set Contrast and Brightness with the DB off and then turn the DB on with the Planar. And he has spent a ton of time optimizing their DI. I am sure that he would disagree with claims that you want to set things up to avoid any white crushing under any circumstances. That would be very limiting. Making the DI work the best possible includes some tradeoff when getting more benefit.

BTW: Just to add one example with respect to wanting no white crushing, what do you (or Wing) think the projector should do if there is an extended fairly dark scene where there is all of a sudden a very small glint off of something that then goes away? Do you think the projector should make the whole image brighter when that glint is there to avoid any crushing, not give much CR for the extended fairly dark scene because there might be a small glint of bright stuff coming up, go ahead and give that dark scene high CR and then deal with the small glint by doing some crushing or compression to keep the rest of the image from showing a bad pumping effect, or something else? I'm thinking of a scene in Johnny English like that and I think there is one in Donnie Darko that is somewhat like that, although it doesn't start as dark. There is also one near the beginning of There Will Be Blood as Daniel Day-Lewis takes a swing at the wall and sparks fly, but they might be bright enough that pumping would be noticeable.

--Darin
post #57 of 211
Quote:
Originally Posted by Mark Petersen View Post

I agree that the light coming out of the lens is modulated in a dynamic way due the combined effects of the lamp and iris which inherently changes the shape of the gamma curve. But I'm not seeing how the data shows that dynamic gamma is being applied in the Sony sense of the technology.

How would they get crushing like you showed in that one image if there wasn't a component that was raising 80%stim up to be equal (or close to equal) to 100%stim? The lamp wouldn't do that and the iris wouldn't do that. They may use dynamic gamma much less than Sony and others, but I don't see what else would cause 80% and 100% to be at the same levels in that one fairly extreme image.
Quote:
Originally Posted by Mark Petersen View Post

It's an interesting technology that seems like it could be used in many places in front projection, but I'm not sure where it's headed with LED illumination on the horizon.

It will be interesting to see. One nice thing about using LEDs like a DI is that they could still have manual irises like the Marantz 11S2 and Sharp 20k so a person could choose to start with less lumens, but higher native on/off CR. And I would expect that it would be easy to know how much the images had been dimmed when the LEDs were reduced, where a system with something like one manual iris and one dynamic iris with a DLP could have problems with the projector not knowing how much the light had been reduced by the dynamic iris (although this likely depends on the particular design).
Quote:
Originally Posted by Mark Petersen View Post

Yes I'd love to dig into the Lumis implementation and see how it compares. I think Alan mentioned that he measured 20k:1 on/off so there is a big difference right there.

I wonder if they have a similar thing to what this Optoma does, where some of the intra-image CRs are higher than the measured dynamic on/off CR would normally indicate, even for a projector with as much native on/off CR as that.

I may have missed it, but did you measure full screen white with the DB enabled? I saw that you got around 15% more for a small amount of white on black with DB on than DB off, but was wondering if full screen white was that much brighter or even more than that.

--Darin
post #58 of 211
Mark,

I just noticed something that seems a little odd. I noticed that you said the on/off CR was about 2000:1 with the DB off. But your chart of intra-image CRs as gray is swept from 0 to 100%stim shows less than 1000:1. I had been thinking that you were using just a little bit of non-black in that image, so that with a high ANSI CR compared to native on/off CR projector like this you would get close to the native on/off CR for your measurements in that image when you used 100%stim. Do the images for that one contain just a little bit of non-black? Do you do your measurements for black near those non-black parts? I'm trying to understand how you got so much less intra-image CR than on/off CR there without DB on and then what that means for the almost 2k:1 intra-image CR you got with DB on.

--Darin
post #59 of 211
Quote:
Originally Posted by Mark Petersen View Post

It's good to know that readjusting contrast can bring it out of white crush,

The problem, though, is that lowering the contrast to eliminate crush reduces picture quality on brighter scenes where DB had no crush in the first place. Its the achilles heel of these DI algorigthms IMO.

Quote:


Incidentally, with real video content I really didn't notice any crush.

Any scene that has near peak bright areas but is overall dark so that the DI clamps down will show it I would imagine...at least thats how it worked on my W5000. On Wall-e you can spot it during the 'space flight' scene when the AXIOM first comes into view. The bright edge of the ship is totally crushed. Its also easy to spot in side by sides with the JVC PJs...you can spot it more frequently if you look for it.

Once you spot it, you can use those scenes as a reference and tweak the settings to see the difference.
post #60 of 211
Thread Starter 
Quote:
Originally Posted by darinp2 View Post

How would they get crushing like you showed in that one image if there wasn't a component that was raising 80%stim up to be equal (or close to equal) to 100%stim? The lamp wouldn't do that and the iris wouldn't do that. They may use dynamic gamma much less than Sony and others, but I don't see what else would cause 80% and 100% to be at the same levels in that one fairly extreme image.

Yes I see your point now. I guess it comes down to whether there is another mechanism that can cause white crush other than using dynamic gamma. It does look like video levels have changed and if the video levels have changed then it implies something like dynamic gamma is being used. I think it's Tom and Wing's position though that the white crush is caused by the white (contrast) setting simply needing to be adjusted better for that specific DB mode. I don't know that I buy that argument because there is such a huge difference between what happens with higher IRE whites between those 3 modes. If I get access to another 8200 I'll take a look at this though, or Tom can verify it when he gets his unit. The low APL greyscale is in the AVS Contrast thread.

Incidentally, getting the contrast and brightness settings correct is critical for taking dynamic contrast measurements and it's something that I spend a lot of effort getting right and I check it with several different test patterns. It's absolutely essential not to elevate black levels for example by having to high of a brightness setting, but setting it to low can make it look like it's crushing black contrast at the 0-10% range. IIRC, I didn't think that the DB mode affected the contrast adjustment, but I could be wrong. In general though, I had a lot of difficulty getting the contrast setting dialed in correctly with a HTPC and why I wish that it had an explicit video level setting. There were also all of the other issues that I mentioned that can affect static test patterns like the lamp flickering and shutting down every minute or so and the lamp and iris sometimes working inconsistently with the same test pattern. All of these issues combined are reasons why I mentioned that I would consider this to be only preliminary data. I'd like to retake this data once the software issues were fixed.

Quote:


It will be interesting to see. One nice thing about using LEDs like a DI is that they could still have manual irises like the Marantz 11S2 and Sharp 20k so a person could choose to start with less lumens, but higher native on/off CR. And I would expect that it would be easy to know how much the images had been dimmed when the LEDs were reduced, where a system with something like one manual iris and one dynamic iris with a DLP could have problems with the projector not knowing how much the light had been reduced by the dynamic iris (although this likely depends on the particular design).
I wonder if they have a similar thing to what this Optoma does, where some of the intra-image CRs are higher than the measured dynamic on/off CR would normally indicate, even for a projector with as much native on/off CR as that.

It would be interesting to know. This goes to show the value of supplementing on/off and ANSI with something that shows intra-image contrast when the contrast enhancement is engaged.

Quote:


I may have missed it, but did you measure full screen white with the DB enabled? I saw that you got around 15% more for a small amount of white on black with DB on than DB off, but was wondering if full screen white was that much brighter or even more than that.

I didn't post the numbers, but I did take the data. The full white reading was essentially identical given the variability of the probe and I always take multiple readings and average the result. Here it is:
Code:
         DB=OFF   DB=C1  DB=C2
white  555        551      556
black  .28        .09      .07
on/off  2003     6482     7831
New Posts  All Forums:Forum Nav:
  Return Home