CurtPalme.com Home Theater Forum

[garyfritz]

Register to remove this ad. It's free!

Lately the color-filtered HD144 and HD145 lenses have become really popular, and for good reason. They're supposed to have better optics, better corner focus, etc., than the stock HD8b's in Marquee 8500's. More importantly (IMHO anyway), Marquees and other non-color-filtered projectors really benefit from the color filtering in these lenses. In my opinion the stock 8500's colors are very dull, very lifeless. If that's what you're used to, you may not notice a problem. But if you switch from a color-filtered projector like an NEC XG to an 8500, you'll notice a big difference. I recently put a set of HD145's in my Marquee 8500 and the improvement is dramatic.

And it turns out there's a VERY inexpensive way to try out color-filtering to see if it's worth it to you.

Unfortunately my camera is a cheap digital and can't capture the differences in colors that my eye can see. But I have a colorimeter, and I can MEASURE the difference.

Some background: the CIE diagram (below) represents the colors that the human eye can see. You can represent any of these colors with an x,y coordinate, which is what the colorimeter measures. The triangle represents the three "primaries" (red, green, blue) defined by the SMPTE C standard, which (I think) is used for North American broadcast and DVDs. It's also very close to the standards used by HDTV and PAL/SECAM.

If you want to represent an image accurately, your projector must have the SAME primaries as the standard used in your signal source -- SMPTE C if you're in North America, very close to SMPTE C elsewhere. In other words, your CRT phosphors must emit exactly the same color as the SMPTE C color. If they don't, your colors will be wrong.

Why? Think of your projector as a blind paint-by-number artist. He's got 3 pots of paint (R, G, B), and he just paints whatever color the signal tells him. If the video signal says to paint pure green, he goes to his green pot and uses that. If the signal has a yellow color, he mixes red and green together.

Now imagine what would happen if somebody swiped his green paint and substituted a pot of pink paint. He's a blind painter, so he can't correct for the bad color. Obviously his greens would be wrong; grass would be pink. But ALL colors that included any green would also be wrong. Since all colors except pure red, pure blue, and shades of pure magenta include SOME green, this means that ALL other colors will be wrong.

What about the Marquee (and Barco and similar non-color-filtered projectors)? Obviously the green tube isn't pink, but the phosphors in the CRTs don't quite match the SMPTE C standard. You may have heard of "yellowish greens and orangish reds" with the Marquees. That's why.

There's nothing you can do (tuning-wise, e.g. setting your grayscale) to change the projector's primary colors. Those are just the colors that the phosphors produce. However you CAN modify the primaries optically -- changing the color that actually hits the screen -- by putting a colored filter in the optical path. That's what the HD144's/145's do, and you can also do it with inexpensive colored gel filters.

I measured my 8500 with its stock HD8b lenses, with stock lenses & colored gels, and with color-filtered HD145's. For the "colored gel" tests I used 3 Calcolor gel filters: 4430 (light green), 4460 (med green), 4690 (dark red).

I measured the color coordinates of the primaries, which are affected only by the phosphors and the filters. I displayed a 100 IRE window and turned off the other 2 CRTs so I was measuring the pure color of a single tube. I placed the colorimeter sensor in front of the screen, pointed at the projector so the color of the screen didn't affect the measurements. (Caveat: I don't know when this Minolta TV2150 colorimeter was last calibrated. I'm hoping it's reasonably accurate.)

This picture shows the results:



As you can see, blue was already pretty close to the correct color, just a bit more "vivid" than it needs to be. That (and blue's limited light output) is why color-filtered lenses use a clear lens for blue -- it's already good enough.

Green was WAY off, in a very "yellowish" direction. The first (4430) filter moved it a bit, and the second (4460) moved it a bit more. The HD145 lens moves it almost exactly onto the SMPTE C value, just where you want it.

I was a bit surprised by the uncorrected red measurement. With all the "orangish red" comments you hear, I expected it to be above the SMPTE C point and to the left, in a really "orangish" area. To my surprise it was actually pretty close to the correct color, certainly much closer than the green. The HD145's and the red gel moved it farther out into the "intense red" area, but kept the right shade of red.

From these measurements you'd think the green is the biggest contributor to improved colors with color filtering. But it may be that our eyes are extremely sensitive to color variances in red, so the slight red error may cause more visually apparent difference than you might guess. I'm not sure. This picture (a Marquee red CRT with a red filter over one half of the tube face) sure makes it look like the filter makes a huge visual difference, even if the measurement shows a small change.

So, enough geekery, what does this look like on the screen? I found skintones to be much more lifelike. Greens and reds are much more vivid and beautiful. The opening scene of LOTR:FOTR (with Frodo reading in the woods) seemed really dull without filtering, but it's *G*R*E*E*N* and beautiful (not artificially green, just vividly alive-looking) with filtering. The Peony Pavillion scene in "House of Flying Daggers" is stunning.

Here are the numbers:

The ideal SMPTE C RGB primaries:

R: .630 .340
G: .310 .595
B: .155 .070

My 8500's measured primaries without filters:

R: .647 .346
G: .352 .566
B: .133 .051

My R/G primaries with filters:

R: .668 .327 (with 4690)
G: .326 .588 (with 4430)
G: .312 .615 (with 4460)

My R/G primaries with color-filtered HD145 lenses:

R: .668 .330
G: .304 .588

Have you wanted to add filtered lenses to your projector, but you hesitated because of the cost? A set of HD144's or HD145's will run you somewhere around $200-250 with shipping, and you'll probably need a set of adapter plates (another $250 or so) to mount them on your projector. How do you know if it's worth it?

It's easy to find out, and very cheap. Order some Roscoe Calcolor filters. I found a place on eBay (seller jayplayer) that sold me 3 colors of filters for about $11 delivered, and that was enough to make at least 4 sets of filters. You want 4690 (Dark Red) and 4460 (Medium Green), though you may want to experiment with 4430 (Light Green) or 4660 (Medium Red) as well. Cut a piece of filter to fit on the front of your CRTs. Remove the R & G lenses, clean off the front glass of the CRTs with a good glass cleaner and a lint-free cloth, and mist a little cleaner onto the filter. Stick it onto the CRT glass and work out any bubbles. (You probably want to cover it with a tissue while you do this, so you don't get fingerprints on it.) The surface tension will keep the filter attached for a LONG time, without adding any new surfaces to increase reflections or haloes. Re-mount your lenses.

Then (**** VERY IMPORTANT ****) rebalance your grayscale to D6500 !! The colors will look TERRIBLE until you adjust your CRT for the filters. You'll have to drive your red and green a bit harder to make up for the light lost in the filters. (You'd have to do the same with the HD145's.)

But once you do, I think you'll see a dramatic difference! Pick a movie you know well that has colorful scenes. Watch it before you add the filters, and again afterwards. I'll bet you'll like the change!

If you want to go with a budget solution, you may decide to just leave the gels in there. The colors are fine, and it's dirt cheap. If you're happy with it, you've just improved your colors for almost nothing! Only problem: the gels are designed for tinting lighting, not for optical clarity. You may think the image is a bit softer, or you might notice a bit of haze in the image. If so, then you might decide the sharper HD144/145 lenses & adapters are worth the money.

(Thanks to KennyG, cmjohnson, and others for doing some of the initial research on this!)

Enjoy,
Gary

(image attached so it gets archived with the forum)

[Gino]

Thanks for the informative write up Gary, nice to see actual measurements etc...

[jask]

Thanks Gary, great info.
I was suprised to see the red values on the 145 so far out,but I imagine there is some very good (objective) optical basis for that value.

Nice graph too, nothing works like visual aids!

[Bighitter]

Very good write up!

Just a couple points of clarification, the human eye is most sensitive to green errors +/-, then RED is next and BLUE being the hardest color of the primaries to detect errors.

While the way you isolated the projector via using the Minolta pointed at the PJ is great for showing the changes the lenses made, for best results in the end calibration should be obtained with sensors pointed at the screen as the screen is obviously part of the video system and will obviously effect final PQ.

Again great job on the write up!

[mrking]

Thanks for agreeing with me.
This is exactly what I think of this color filtering and it is the same with 9" only they are a bit different.

Green is definately needed on 8" AC but red? NOT if you want to be as accurate to D65 as possible!

I will see if I can dig up the gamut of my 9500LC without red C to compare with your findings.

In my opinions red c is wrong because it over saturates the red making the picture simply put. Too darn red!

And this does like you said affect every color except pure blue and pure green.

[mrking]

I found my old post at the bottom of page 2 in this thread.

http://www.curtpalme.com/forum/viewtopic.php?t=2809&postdays=0&postorder=asc&&start=20

[mrking]

Your gamut REALLY contradicts what Ile posted and what VideoGrabber claimed in that thread because their red was WAY out of whack in comparison to yours and they blamed it on the MEC tubes having poorer phosphor than the Sony tubes?

According to your gamut that is inaccurate since they seem pretty much inline with this gamut and I seriously doubt you have Sony tubes in your 8500?
Then again MAYBE and this is a big maybe the phosphor is different in 180DMB22 than 180DVB22?








Like I've stated earlier.
I would be VERY intreagued to see the primaries of a projector where you have replace the red lens with a clear 145 lens.

[Person99]

[garyfritz]

My personal observation is that the reds are NOT too intense for my tastes -- in fact I feel a bit more oomph in the red wouldn't hurt. The HD145's move the red primary out to a bit more "intense" area but keep basically the identical shade of red, and I think that's a Good Thing. So for my case, the red HD145's do NOT cause a problem at all.

[mrking]

I mean the contradiction is that the MEC tubes measures basically the same as the Sony tubes?

[garyfritz]

I don't know about that. I've only measured the MECs in my Marquee.

[draganm]

[AVphile]

Can we conclude that only the green filter was really neccessary (more true with Sony tubes)?

It appears that filtering the red may not be worth the reduction in clarity?

[garyfritz]

I never measured it with just the green filter, but you might be right. The unfiltered red is pretty close, and the direction it's off should make less difference than the green error.

On the other hand I gotta think these companies checked it out before shipping red filters on their projectors, and they probably wouldn't have done it if there wasn't enough improvement to justify it.

[AVphile]

It would also be cool to see what a medium red filter would do.

I have an outgoing Barco with Sony 07MSP tubes and tinted glycol. I was happy with the colors, but sometimes see an unrealistically intense red (something I've always been sensitive to with CRTs). I should measure the gamut and post it up.

I am getting a "newer" Barco with Sony 07MFP2 tubes, NO tint, and afraid I will need to add some filters. How different do you think the results will be compared to your MEC application?

[AVphile]

I think I will remove the green filter for LOTR FOTR EE and many other blu-rays

[AVphile]

[garyfritz]

I'm not sure how to interpret that Luminance chart, but your gamut looks a bit hot. The green especially is waay out there. The HD145's made my 8500's green almost spot on the target, and yours is way past the green target. I'd be concerned that it might result in an almost "cartoony" over-blown green

But the ultimate test is what it looks like to you. Set the grayscale and see how it looks.

[AVphile]

Thanks Gary. The Luminance chart is also saying my green is hot (black bars and dots are my actual readings). I can confirm that something is up with my green as the color does not look near as natural as the ole 801. Greyscale is OK in the highlights but all over the map everywhere else. However I'm not sure if it's due to the green filter or some other problem.

(My set blew its HV power supply upon installation so I never got to see the 808 before. After fixing it I had to do a full ground-up restoration of everything mechanical, magnetic and electrical. In the process of flattening the greyscale I maxxed out all the CUTs and MIDs, which has led to a crazy gamma - not much of a curve!

Full report attached.

I've tried 3 different methods for setting G2, so this is very time consuming, but I plan to try again with initial conditions posted by zGman 31-March-2008.)

[garyfritz]

You've got two different issues there. For one, your green primary is "more green" than it should be. I don't know how visible that will be on a properly-adjusted projector but it's a property of your phosphor and your lens/filter.

Then your grayscale needs a lot of work. The green is way hot in the mid/lower IREs. That's not caused by your filter, that's caused by the G2/CUT/MID/whatever settings. I suspect that's going to result in greenish skintones along with other problems. I'm astonished your color temps run so close to 6500, according to the lower-right chart.

I have no experience with Barcos, and it's been over 2 years since I last fooled with my grayscale so I'm going on foggy memories here. But I think you need to work on your grayscale some more to lower the green, raise the red/blue, and try to pull them down to the right gamma curve. You may prefer the end result a bit above the gamma curve (lower gamma) -- that way the shadow detail is much better, not so dark -- but I don't think you want it as high as it is now, especially not at the mid/higher IREs.