Who else sees rainbows on DLPs?

[Drewbert]

I do... and im so anal I think it might bother me. But Im willing to get over it for that super clean DLP projection....
I sell HT stuff all day... so I see a lot of TVs. No one else sees it but one of my other employees... so do you see it?
Just curious

[David Meek]

I haven't noticed it, but my exposure to DLPs has been spotty.

[Kevin P]

I've seen it on older DLPs during fast motion scenes, or if I sweep my eyes/head across the image, but newer DLPs are a lot better in that regard.

If you find it bothers you, even with the newer sets, you probably won't be happy with DLP.

[Shane Martin]

I've seen then even on the newest 1080P sets. They do bother me.

Oddly I do not see them hardly at all on the 720 HD2+ DC3 FPTV's like the Marantz 12s4.

[aud19]

Saw them faintly once on a fast motion scene while turning my head but otherwise no :lol:

[Brandon B]

I see them in all single chip DLP products so far (and I have seen many). Doesn't give me eyestrain or anything, but I find it too distracting to live with.

The only place I have seen it significantly reduced was during a demonstration by a company (Genoa) using 5 or 6 primary colors where they had modified two DLP projectors to have two different sets of primaries to have a bigger color gamut. Since there were two images overlaid with two different sets of color fringes, and I don't know that they were even in sync, timing-wise, it almost went away for me. I mostly could not see the specific colors in the fringing, just was aware of a sort of flickering associated with it. On standarg RGB color wheels, I vey easily see the individual colors, even on 5X wheels.

TI's new 1080 DLP products are alleged to maybe have something like this, so that would be a really promising development.

BB

[Shane Martin]

TI's new 1080 DLP products are alleged to maybe have something like this, so that would be a really promising development.

The Wobulated 1080P chips still have this issue.

[Brandon B]

Yes they do, plus another issue specifically to do with wobulation (which is why you will not see it in FP). But there is something called BrilliantColor alleged to be appearing in the forthcoming 1080 products that may solve the rainbow problem.

And it is a problem for TI, whether or not it affects more than a small percentage of people, as all the other display technologies have latched onto it and are harping on it in their advertising. They have to counter the negative PR at some level.

The stuff in this link is a bit ambiguous, but it does sound like the Genoa process, and that definitely improved things. I am more easily able to see rainbows than anyone I know, and I would probably be able to live with a single chip DLP with something like this process. Especially with the added benefit of the color enahncement. I haven't seen definitive research or proof that it maintains color fidelity, but the increased gamut is really beautiful.

BB

[George Bellefontaine]

When I first got my NEC HT1000 dlp Front Projector I saw rainbows on certain types of scenes. They were so infrequent and so fast that they never really bothered me. Funny thing is, I rarely see them now unless I force myself to watch for them.

[Brandon B]

They don't bother most people. I would put the ceiling on the number of people who see them at 15% or less, the number for whom it is a problem at less than half that, just guessing from everything I have read/seen/heard.

But if they can get it down to 1-2% or less, better yet.

BB

[George Bellefontaine]

But if they can get it down to 1-2% or less, better yet.

BB

I agree, Brandon. The best and easiest way would be to get 3 chip dlp down to today's 1 chip projectors.

[Brandon B]

TI seems determined/unable to go that route yet. Next generation LCD and Sony's SxRD push are going to put lot of pressure on that front though.

DLP is going to maintain the ANSI contrast advantage for a bit more, but onff is going to become much more comparable in the next couple of years. Epson's new C2Fine technology and Sony's Bi:Na both promise not only huge boosts in contrast (into the 10,000:1 arena) but remove the organic blue polarizer plate deterioration issue. The remaining weak point for HTPS LCD will be its low fill factor (~50% for these new panels), but at 1080p, SDE becomes a pretty moot issue. And given LCDs maturity and ability to hit really low price points, TI is going to have to work some to counter the perception they are over priced in their 1 chip DLP offerings.

Sony is going to hit them with superior fill factor and 3 chip smoothness at the mid and high end (Ruby and Qualia), and affordability with full resolution at the low end (1080p LCD HS51 replacement). And others will be joining them in the LCD bread slice. We'll see if JVC joins them at the other end. Right now they have offered a new 1080p machine with 2 scaler choices at $15K and $20K. Unfortunately for JVC, it is probably dead out the door. Compared to Ruby, 20% dimmer, 1/2 to 1/8 the contrast (no dynamic iris) and 50% - 100% more in price.

So better ANSI (which is nothing to dismiss for sure, it definitely translates into dimensionality of the image) and the promised color capability of BrilliantColor are going to be TI's talking points, not 3 chip at single chip prices, until and if they find that ain't working.

BB

[Drewbert]

Brandon that post rocked, thank you kind sir!

[DifferentLee]

I see the rainbow effect.

SXRD is clear and so is DILA but the black levels on DILA suck bigtime.

[Brandon B]

They are getting better. On all microdisplays. JVC will probably be the lagger on this, they have not show any sign of intention to use dynamic iris techniques, even in their expensive FP systems.

What's the contrast on Sony's newer SxRD's? Haven't read up on them so last one I saw specs on was the qualia TV.

BB

[aud19]

I believe JVC went the high light output aproach to theirs which naturally means worse black levels. I bet it stands out in stores though :roll:

[DifferentLee]

What's the contrast on Sony's newer SxRD's?

HT Mag measured it at 13,000:1! No joke. See November issue with red cover.

[Trevor Schell]

I don't think I see rainbows.
However now I am not sure.
I do see something but don't think it is the rainbow
effect but maybe something else.

Yesterday I had a good look at a Toshiba 62" Talen.
This set was feed a Hi-Def signal via a Motorolla
receiver so it was likely Hi-Def digital cable.
Anyways..
I don't see rainbows as such , but I do see lots of sparkels
all over the screen. They seem the produce a 3D effect, meaing they
look deep into the screen.
Really weird. I seen this consistantly and especially
during the white parts of the screen. Say a white background during
a commercial as an example.

So, is that the rainbow effect or something else.

[Brandon B]

No, that's not RBE. Sounds like maybe DVI pixel dropouts, or maybe the artifact some report seeing with the new 1080 wobulated sets. Is this one of those? RBE is literally where you see flashes of red blue and green, usually at the border of a very light/white area and a dark/black area, and usually in combination with your eyes shifting fovus to another part of the screen, or moving your head or something.

I would not expend any effort trying to see them. Certainly no good reason to learn to see something that doesn't bother you now. Of course, doing so is an integral part of this hobby.

13,000:1? So I am guessing the SxRD TVs have an iris system in them as well. Although, a measurement like that is a bit suspect on an RPTV, as the black level measurement would be too low for most measuring devices to do accurately.

BB

[Trevor Schell]

13,000:1? So I am guessing the SxRD TVs have an iris system in them as well. Although, a measurement like that is a bit suspect on an RPTV, as the black level measurement would be too low for most measuring devices to do accurately.

BB

Black Level was .006

[Shane Martin]

13,000:1? So I am guessing the SxRD TVs have an iris system in them as well

Yep auto iris. Even if it's suspect, the SXRD I saw this last Sunday blew the doors off of the 1080P Wobulated set sitting next to it.

[Brandon B]

Oh I agree about blown doors. Sony no doubt has something on the order of 3X the contrast with an auto-iris system and their superior chip level contrast to start from. But here's the deal with measuring onff CR:

Most meters, even the good ones, do not have the sensitivity to measure accurately at the low end. With a front projection system, you can accomodate this weakness by measuring max white level of a full size screen, then using a smaller screen to measure black level in the range where the meter is still accurate, then factor in the difference in screen size to do your calculation.

On an RP, you can't do that (without breaking open the box). If they are quoting a measured black level of .006, was that the reading on the meter? Was it 0.006000? Was it 0.006? Say it was the latter. That means they measured 78.000 or close to that (nits or FtL or whatever units they measured) for peak white and 0.006 for black. Now what if that reading was actually exactly 0.0056 rounded to the display's accuracy? 78/.0056 = 14,000:1. What if it was exactly 0.0064? 78/0.0064 = 12,000:1. And that assumes the meter was exactly correct.

Now what if the meter is not exactly correct at these low light levels, and that 0.006 is ±.001? Then you are down to ±2000 of your 13,000.

That's what I mean by suspect. Most professional meters cannot accurately read at the 0.01 FtL light level, which is .03426 nits. They are quoting a figure 1/5 that as accurate if they were measuring nits, 1/2 if Ft-L. If someone has read the review, and can shed light on A) what meter they use and B) the method they use to measure the black level, I'd be interested.

BB

[KeithM]

I can see it horridly well. I get a headache whenever I watch a DLP TV.

[Brandon B]

OK, found out HTM uses a Minolta LS-100 luminance meter to do their black level readings (and white level).

From the specs of the Minolta:

Accuracy: 0.001 to 0.999cd/m2 (or fL): ±2% ±2 digits of displayed value
1.000cd/m2 (or fL) or greater: ±2% ±1 digit of displayed value

So for a reading of 0.006, which falls into the first range of less than 1 Ft-L, ± 2 digits of displayed value.

I personally would disregard their black level reading based on that. And therefore their contrast calculation. I'd instead go with "better than anything else I have seen" or something to that effect.

BB