[secnarf refohs]

I am looking at contrast ratio specs on the six or seven brands available with the new HD2 DLP panels (Yamaha 1000, Marantz S2, Sharp10000, ...). Can anyone tell me why these machines have spec-sheet contrast ratios that range from 900:1 through 2600:1? That seems to be a wide range for the same panel technology. I feel bulb wattage is an absolute while CR is a ratio (right??? CR should be same), so why the big difference?


Why I ask: I went to a local home theater store last night and looked at a Sharp 9000 and a Sony 11Ht, as well as four RPTVs (Elete, Sony). I want a front projector rather than a crt-based RPTV, but those crt-based RPTVs seemed so bright and did NOT appear washed-out on night scenes.


Both projectors I saw (the Sony ...11Ht LCD more so than the Sharp 9000 DLP) appeared washed out on night sceans (Spiderman was the common DVD), while the RPTVs seemed to have black blacks and bright whites.


So ... do you think the newer HD2 DLPs with very high CR and reasonable lumens will be a big improvement over the HD1 / Sharp 9000 generation? Also, the Sony ...12HT spec says CR of 1000:1 - so will this be a noticable improvement over 11ht?


Thanks all.


Fran

 

[MrWigggles]

Ms. Shofer,


I think all HD-2's are above 900:1 and most will have at least twice the contrast of the Sharp 9000 HD-1 you are familiar with.


It is possible that the Sharp you saw had the brightness set too high which caused the "blacks" to be gray. (this happen at local store that I went to so it could have happened to you as well.)


Or you could be very sensitive to the issue becuase most people think the blacks on the 9000 are quite good.


In any case, contrast ratio is obviously a very important thing to you and I would stick with the HD2 projectors that promiss the highest CR's. That would put the Yamaha, Marantz and Sharp at the top with the Infocus technically at the bottom.


However, the Infocus should still have twice the contrast of the Sharp you saw and that should help the blacks a whole lot.


bottom line: you are probably going to like the new HD2's and if contrast is what your after than DLP is the best digital technology for you.


-Mr. Wigggles

 

[secnarf refohs]

Mr. Wigggles - thank you for the quick reply. contrast and black level appears to be more important to me than shere brightness. Your feedback helps me and confirms my suspiciions. Hope this helps others considering HD2's.


I still wonder why the same panel technology could produce such a wide range of CR specs accross Mustang HD2 manufacturers. How would one select a new HD2 product, I mean, what do I trade off with contrast ratio? Let's say I buy an InFocus HD2 rated at 1:1400 instead of a Marantz HD2 rated at 1:2600. Do you think I am trading off dollars? Circuitry? Bulb type?


I read an article about the Sharp 10000 (HD2) --> ( http://www.smr-home-theatre.org/cedi.../page_02.shtml ) --> that says, in part, "Sharp offers a switchable brightness/contrast function; in High Contrast mode, it delivers 2500:1 contrast (vs. 1500:1 in High Brightness mode), but the increase in contrast comes at the cost of a significant drop in brightness (from 1500 lumens to 500)."


So sharp has a SWITCH that allows you to choose between contrast and brightness??? This can't JUST be a bulb brightness switch -- as I can't reconcile a RATIO (here meaning contrast) versus and ABSOLUTE (meaning brightness).


Do you - or anyone reading this - know how circuits can trade off CR and brightness? Do you think circuitry is why the six or seven manufacturers have such a broad range of CR offerings for identical panel technology? Or is it MSRP? I think Sharp's HD2 product offerring is beneficial in that they give the USER the CHOICE of better CR or better lumens.


What's your thoughts? Again, thanks!


Ms. FranCES

 

[noah katz]

Ms. Frances,


The CR/brightness tradeoff is done optically, not electronically. The new highly regarded NEC HT1000 also has this feature, implemented by a variable sized lens aperture, which gives 2000:1 CR at full brightness of 1000 L or 3000:1 CR at 800 L.


Blocking the outer light rays of the beam bundle with the greatest angle from the centerline increases CR because those are the rays most likely to hit the inner walls of the optical path and scatter around, eventually exiting he lens and hitting parts of the screen they're not supposed to.

 

[xb31]

Quote:

Originally posted by secnarf refohs HD2 DLP panels ... why these machines have spec-sheet contrast ratios that range from 900:1 through 2600:1?

There are at least 3 principle optical engine designs within the

industry into which a single HD2 DLP panel can be used. Combine

that with lens quality, number of lenses, and whether the interior

walls of the optical engine have been coated or otherwise designed

to minimize light scatter, and you can easily explain the range of

speced CRs.


Also, once you start getting into high 3 digit and any 4 digit

CRs, the black lumen level becomes a big magnifier. CR is

nothing more than the ratio of the brightest-bright and the

darkest-dark falling on a screen. The HP xb31 is rated 1500

lumens and 1800:1 CR. That means the brightest-bright is

1500 lumens, and the darkest-dark is 0.83 lumens. 0.83

lumens isn't much, and if you tweak it just a little up or down,

it has a huge influence on CR. Halve 0.83 at the blackest-black,

and we'd be talking a CR of 3600:1. For that matter, if you kept

0.83 as your darkest-dark and wanted 3600:1 CR, your brightest

bright would have to be 3000 lumens.


Within the HP projector team, we've started to debate whether

all you HT fans really want a higher (the higest) CR, or just a

reasonably darkest-dark. Put another way, is a sub 1 lumen

darkest-dark dark enough? Would going from 0.83 to 0.415

lumens on the dark-side really result in a more stunning video

experience, as oppoesed to going from 1500 to 3000 lumens

on the bright-side?


--xb31

 

[mark haflich]

XB31. Great post. Extremely informative. Thanks.


How much of a reduction in the sub 1 lumen level would result if TI is able to eliminate the dimple flaw in it's chip design?

 

[MrWigggles]

Quote:

Originally posted by xb31 There are at least 3 principle optical engine designs within the industry into which a single HD2 DLP panel can be used. Combine that with lens quality, number of lenses, and whether the interior walls of the optical engine have been coated or otherwise designed to minimize light scatter, and you can easily explain the range of speced CRs. Also, once you start getting into high 3 digit and any 4 digit CRs, the black lumen level becomes a big magnifier. CR is nothing more than the ratio of the brightest-bright and the darkest-dark falling on a screen. The HP xb31 is rated 1500 lumens and 1800:1 CR. That means the brightest-bright is 1500 lumens, and the darkest-dark is 0.83 lumens. 0.83 lumens isn't much, and if you tweak it just a little up or down, it has a huge influence on CR. Halve 0.83 at the blackest-black, and we'd be talking a CR of 3600:1. For that matter, if you kept 0.83 as your darkest-dark and wanted 3600:1 CR, your brightest bright would have to be 3000 lumens. Within the HP projector team, we've started to debate whether all you HT fans really want a higher (the higest) CR, or just a reasonably darkest-dark. Put another way, is a sub 1 lumen darkest-dark dark enough? Would going from 0.83 to 0.415 lumens on the dark-side really result in a more stunning video experience, as oppoesed to going from 1500 to 3000 lumens on the bright-side? --xb31

We want contrast any way we can get it.


If I want good blacks, I put my lens cap on.


-Mr. Wigggles

 

[lovingdvd]

Quote:

Originally posted by xb31 There are at least 3 principle optical engine designs within the industry into which a single HD2 DLP panel can be used. Combine that with lens quality, number of lenses, and whether the interior walls of the optical engine have been coated or otherwise designed to minimize light scatter, and you can easily explain the range of speced CRs. Also, once you start getting into high 3 digit and any 4 digit CRs, the black lumen level becomes a big magnifier. CR is nothing more than the ratio of the brightest-bright and the darkest-dark falling on a screen. The HP xb31 is rated 1500 lumens and 1800:1 CR. That means the brightest-bright is 1500 lumens, and the darkest-dark is 0.83 lumens. 0.83 lumens isn't much, and if you tweak it just a little up or down, it has a huge influence on CR. Halve 0.83 at the blackest-black, and we'd be talking a CR of 3600:1. For that matter, if you kept 0.83 as your darkest-dark and wanted 3600:1 CR, your brightest bright would have to be 3000 lumens. Within the HP projector team, we've started to debate whether all you HT fans really want a higher (the higest) CR, or just a reasonably darkest-dark. Put another way, is a sub 1 lumen darkest-dark dark enough? Would going from 0.83 to 0.415 lumens on the dark-side really result in a more stunning video experience, as oppoesed to going from 1500 to 3000 lumens on the bright-side? --xb31

Interesting you are thinking ths way. If I'm following correctly, InFocus was thinking this way to and decided the brightness was more important. Sound right?

 

[xb31]

Quote:

Originally posted by mark haflich How much of a reduction in the sub 1 lumen level would result if TI is able to eliminate the dimple flaw in it's chip design?

I suspect if we investigated the source of remaining light in

the blackest-black, it would be 1/10-ths of lumens here and

there, and the dimple would just be one contributor. Knocking

off those final 1/10-ths of lumens could be expensive.


--xb31

 

[xb31]

Quote:

Originally posted by lovingdvd Interesting you are thinking ths way. If I'm following correctly, InFocus was thinking this way to and decided the brightness was more important. Sound right?

I'm not sure what InFocus is thinking, but I think its fairly obvious

that even in the darkest room, the darkest-dark looks pretty darn

black these days, and yet the whole viewing experience can be

destroyed by sun light of indoor lights given the lack of brightness.


--xb31

 

[Jim Ferguson]

I vote for lower absolute black versus more lumens if there is a trade-off to be made. You can pooh-pooh CRT-blacks, or refer to them as "lens cap blacks" if you want, but this is one area where many of us still feel that digital projectors have a long, long way to go.

 

[barrysb]

Quote:

Originally posted by Jim Ferguson I vote for lower absolute black versus more lumens if there is a trade-off to be made. You can pooh-pooh CRT-blacks, or refer to them as "lens cap blacks" if you want, but this is one area where many of us still feel that digital projectors have a long, long way to go.

I for one, would like to see more discussion about how the various projectors handle small changes in video signal level at both ends of the gray scale range. It seems to me to be more important to know if projectors are crushing whites or blacks rather than what is the CR range. In other words, what is the gamma response of a particular projector and how does it compare to its competition.

 

[JHouse]

I've been asking a question for a long time that no one has really answered, but I think we are tickling it here.


So you've got 1800 lumens available light, and a 2 lumen floor at your blackest black. That's 900 to 1 in contrast. So you find a way to keep the 1800 and reduce the 2 lumens floor to 1 lumen. Presto! You now have 1800 to 1 in contrast ratio. Numerically, that sounds twice as good.


But it just can't be that big of a deal visually. In one case the projector has 1800 lumens to spread across the sprectrum from lightest to darkest portions of the image. In the other case it has 1799. Who would notice?


If you are working with a 1000 lumen machine, you have 800 fewer lumens for use in spreading the total contrast range of the material. Now THAT has to hurt.


I don't care if the 1000 lumen machine has a black level of .5 so that it has a 2000 to 1 contrast ratio. It still only has 999.5 lumens to work with.


If this RANGE from brightest to darkest is greater, then each differentiation between brightness levels will be seperated by more lumens. That has got to make a substantial difference. If the spread between this and that is 1 lumen on one maching and 2 lumens on the other, the contrast between the this and the that will be much greater (twice as great).


Such a difference must show up. And this type of difference would be visible.

 

[Jim Ferguson]

The problem doesn't show up when you have an image which spans the range from from the brightest to the darkest ends of the range. In that circumstance even very gray blacks will appear jet black. I don't think anyone has ever complained about the blacks on a checkerboard test pattern.


The problem comes in scenes which are overwhelmingly black on a percentage basis, with parts of the image that are supposed to be just above black. This is where the "black" and the "near black" all fade into the same undifferentiated level of gray.

 

[barrysb]

Quote:

Originally posted by Jim Ferguson The problem doesn't show up when you have an image which spans the range from from the brightest to the darkest ends of the range. In that circumstance even very gray blacks will appear jet black. I don't think anyone has ever complained about the blacks on a checkerboard test pattern. The problem comes in scenes which are overwhelmingly black on a percentage basis, with parts of the image that are supposed to be just above black. This is where the "black" and the "near black" all fade into the same undifferentiated level of gray.

The last two posts are finally getting to the crux of the problem. But now, how do you measure the gamma performance of a projector. What is acceptable, what isn't?

 

[Greg Young]

I have a question. Since we are talking about contrast and black levels why can't I adjust my television and get the kind of black level on a crt if it operates on the same principal of a crt. The blacks are not much better than on my ht projector. Thanks G.Y

 

[JHouse]

Quote:

Originally posted by Greg Young why can't I adjust my television and get the kind of black level on a crt if it operates on the same principal of a crt. The blacks are not much better than on my ht projector.

That is exactly what I see. Supposedly black on a direct view crt actually shuts off the signal, so it should be pitch black. However, a pure black screen with no signal lights up my bedroom (32" ProScan). I don't get that.

And blacks in the program material are plenty good, but that shadow detail is better on my projector.

 

[JHouse]

Quote:

Originally posted by Jim Ferguson The problem doesn't show up when you have an image which spans the range from from the brightest to the darkest ends of the range. In that circumstance even very gray blacks will appear jet black. I don't think anyone has ever complained about the blacks on a checkerboard test pattern.

Good point, but when you say:

Quote:

The problem comes in scenes which are overwhelmingly black on a percentage basis, with parts of the image that are supposed to be just above black. This is where the "black" and the "near black" all fade into the same undifferentiated level of gray.

If you a greater difference in lumen level between adjacent patches of near black, then you should have more shadow detail. I think shadow detail is what you are talking about. The actual lumen level of the particular range of grays (whether it is brighter or dimmer) shouldn't play that big of a role in the amount of shadow detail. In fact, I would think that the darker it was, the more difficult it would be to see the differences between different lumen levels (since our eyes apparently become more sensitive to such distinctions as the scene brightens). What would seem to make the shadow details more easily discernible would be greater differences in lumen levels between those details.


Is there any chance that where there is no real information, that a blacker black gives us the confidence to ignore it, whereas if it seems just a big patch of charcoal gray, we think that there must be something there that we aren't seeing?


Is it possible that on the little screen, undifferentiated black patches are small enough to be ignored, but on the big screen, they are just too big not to draw our attention? I find myself frustrated in movie theaters all the time by undifferentiated dark scenes with no shadow detail. That is one of the things that critics also tend to complain about (that the movie is shot too dark to see anything). I find it very irritating, and on the REALLY big screen, it is even more frustrating.

 

[Sean Max]

Joe, I don't think lumen output has anything to do with shadow detail. If this were the case, then CRT's, with their anemic ANSI lumen output, would have terrible shadow detail, since they have the fewest lumens to work with. But, they also have the best black levels, giving them an enormous contrast range in any given picture, regardless of the level of illumination of the picture. Because the contrast ratio of a CRT is not compressed in low IRE scenes, they are rendered faithfully, provided the CRT is set up correctly. The fact that newer HD2 DLP's are getting better at shadow detail, even as lumen outputs remain the same or decrease, is a function of improved contrast ratio which manifests itself in dark scenes.


Of course, if you had a perfectly gamma ramped LCD and proper bias lighting, you might be able to achieve a similar effect. I don't know, I don't own an LCD, and I've never seen a fully tweaked one in a proper setting. I'd think that this would only be possible on one of the newer, high contrast LCD's. Some of the older ones simply don't have enough contrast to work with.

 

[JHouse]

Quote:

Originally posted by Sean Max I don't think lumen output has anything to do with shadow detail.

Imagine the different levels of gray are depicted on the folded edges of an accordion. Seems to me if you stretch the accordion wide open, the differences between each level will be farther apart, creating bigger differences, and thus the differences will be easier to perceive.


Or equidistant dots on a rubber band. Just stretch it out, and the dots get farther apart and easier to distinguish.


If the pj is calibrated, higher lumens have to improve shadow detail.


Take a dark scene, one where it is really hard to make out the details, If you jack the brightness way up you will see a ton of detail that was invisible before. Details that were picked up on the film that you probably aren't actually supposed to see. Yes, the picture will be all "washed out" but ironically, there will be more detail in the shadows. I don't think actual shadow detail is helped by darker blacks. I think it is a psycho-perceptual phenomena, that when things get to a certain level of darkness, our mind/eye no longer tries to make out any detail there. However, if the absolute level of the black is not really really black (whatever we think that is) we still hunt for details there and have the feeling that we can't see them even though we think we should be able to, given the level of illumination.