CurtPalme.com Home Theater Forum

[Hfuy]

First things first. I started earning a living almost exactly at the point where CRT started going out of favour, so my experience of them is as a projector that:

- Was expensive, and therefore underspecified, and because of this
- Was constantly driven at absolute maximum output,
- Was put in the wrong position, minimising best use of CRT area, and
- Was not actively maintained in all its 25,000 hours, during which it had
- More or less constantly displayed the local sports station's lower-third graphics, and would therefore be displaying a shadow of them for the rest of its life.

Now, to be blunt, I assume that we're all here because a well-maintained CRT projector is capable of better results for the money than an LCD/SXRD/LCOS/DLP one, at the cost of being the size and weight of a transatlantic cargo ship and requiring regular maintenance. Would this assumption be correct?

The reason I'm interested is twofold. First, CRT projection was all the rage when I was forming an adolescent interest in such things, and I'm greatly dismayed that modern projectors don't have that vivid multi-coloured porthole array on the front (go on, stand close to the screen! Make a rainbow-coloured rabbit shadow!). Nostalgia is a wonderful thing.

Secondly, though, I'm now involved in the film and TV industry, and this is an area where we have a bit of a problem with display devices. Until recently, nobody was doing critical work on anything other than a CRT monitor, because (despite constant manufacturer attempts to improve things) no LCD technology has nearly high enough contrast ratio. Now Sony and co. have stopped making CRTs, there's an issue with exactly how you grade an image.

CRT projectors offer very high contrast ratio. This is very, very interesting.

Are these reconditioned CRTs capable of producing an image that doesn't look... archaic? My memory of CRT projectors is that they don't hold a candle to the modern stuff, but then, the ones I remember were ancient, mis-aligned, and worn out. Can I put a sharp 1080p image, scanned at 72Hz, onto an eight or ten foot wide screen, and achieve something that does't look like it was made in 1982? Can I do this and achieve proper Rec 709 colour aims, contrast, and luminance? Can I show it to a client without them going "hey, he's using a crappy old projector"?

If I can do this with a CRT projector for $10k (the most expensive option advertised on this site) then that's quite interesting. I'd put up with a weekly re-converge, for that.

-H

[Nashou66]

Yes, id get Curts VDC marquee MP9 he has for sale this is a very current model and has the highest "actual" bandwidth of all PJ's. Sure some of the others stae its higher but the Marquee easily does 1080p, I have a lower end model the Marquee 8000 i have some screen shots of on a 12 foot screen running 1080p@48 Hz



And here is my latest Marquee, a 2004 build date 8500 ultra that was to be used in the Apache longBow simulator till they decided to upgrade tot he marquee 9500LC Ultra which is exactly like the VDC madigral i mention above.

this is a dirty set up as I just got this and have not done any extensive calibration or fine tuning of the Focus magnets.
1080p@72Hz about a 10 foot wide screen shot.




So yes they can do 1080p and there area few mods that you can get done to them for even better colors and resolving
power.

Also VDC still makes the Marquee and is continually advancing its abilities.

Edit, also i maybe d a re convergence every three months or less, depends on how anal i am

Athanasios

[Hfuy]

Interesting, although you could probably make it look like anything in a photo!

If anyone has a representative device hooked up to a computer, I'd be interested in seeing it do a zone plate, and perhaps some single pixel checker patterns, photographed close enough to see the scanlines. Usually when setting up DLP, I like to back the focus off just enough to lose the pixel edges, and I wonder how well that works with scanlines on CRT projectors.

On another tack, it strikes me that the best thing to do would be to get one that was rather overspecified, then you'd have enough blue, and you'd be able to run it at something other than maximum output. Sensible?

P

[Nashou66]

Ok here is the one on one off of a marquee 8500HR, this was taken by a VDC engineer who frequents these forums. and remember this is an 8 inch tube, a 9 inch tube will resolve even better.
But this is at 60hz not 72Hz



And i have done nothing with those pics. But I agree its hard to see what a PJ looks like form a screen shot, and it depends on if the display your viewing the pic is calibrated or not. But in person these PJ are phenomenal. One of the members here is having a meet in Indiana, he is doing a triple stack of Sony G90's... now that is a site to behold.

On the Blue issue, the marquee has a gamma circuit on the blue to even out to gamma. But most here do a slight defocusing of the blue to increase output with out over driving the tube, it flattens out the typical blue hump of CRT's.

Athanasios

[AnalogRocks]

[AnalogRocks]

[Nashou66]

here is my attempt at the same pattern, I do not have this PJ set up optimally yet and i know i can do much better once i adjust the spot shape via the CPC magnets.
Also i did not have a tripod for the pic so it was just a point and shoot pic.
this was done at 1080p@72Hz I need to also adjust the front and back porch along with the syncs to et a better vertical line, i am having some distotion i need to adress.



Athanasios

[Nashou66]

[AnalogRocks]

Oh, I sit corrected. I though he was in the Chicago 'burbs

[AnalogRocks]

Oh, I sit corrected. I though he was in the Chicago 'burbs

[Hfuy]

Well that's not half bad, is it. Noticeably better resolution on the horizontal lines, but what're you gonna do, it's a CRT.

A zone plate would be the acid test. I suspect it'd probably make you folks very depressed, because it's the one test that the digital stuff tends to pass with flying colours and everything else fails miserably - you'd never quite get the convergence that microscopically perfect.

I've converged quite a lot of CRT monitors, including winding the scan down to a 16:9 image where possible. I presume converging projectors is similar.

And as regards our host working on them - I guess that a CRT projector is probably similar electronically to a monitor, just more so, and three times. That must be a labour of love.

P

[bbfarmht]

[perisoft]

Hfuy, the biggest drawback of CRTs vs. digitals aside from brightness is ANSI contrast. They're far better at on/off - full blacks - but atrocious at ANSI. Without knowing what your goals are, though, it's hard to say whether they're a better choice than the new digitals, which I presume aren't sh*t at color the way crappy LCD monitors are.

If all you need is to wow visitors, a CRT can do that easily... but so can a digital. If you're after color-critical stuff, I'd stay away from a CRT just because the low ANSI will wash the whole image out if there's any big bright areas. This is horribly obvious with air coupled machines, but isn't that great with LC either, as far as I'm aware. They look fantastic for watching movies, but that's clearly not entirely what you're after...

[Hfuy]

That sounds like a serious downer, although I'd have to read up on exactly what you're talking about. In my experience the problem with any CRT technology is the EHT regulation, which causes geometry, brightness, and contrast issues with respect to peak image brightness. Really good CRTs minimise this to tiny levels, although the EHT on projectors must be brutally high to get the output.

It sounds like you're talking about blooming, though. I assume that the purpose of the liquid in LC projectors is to avoid refractive index jumps between the face of the CRT and the lens, which would minimise blooming due to internal reflection, depending how closely the RI of the liquid is matched to that of the lens. Any device with a lens can bloom; is there some reason CRT projectors are particularly bad? Is it just an artifact of a CRT being driven hard or are the lenses junk?

The idea - cautiously expressed - was to apply one of these things to a colour grading suite. There is no perfect technology for this, since you're usually trying to match an image that will eventually be burned out to film on a laser (or even CRT!) film recorder, and there's a million and one considerations to that beyond the display device. While everyone understands that an absolute 1:1 relationship between what you see while grading and what you see on film is not always possible, this is the goal, and there is a certain degree of contractual responsibility involved in presenting a representative image. I would absolutely expect to drive the display via something like a Blackmagic HDlink with a full 3D colour look up table in it, for really bang-on calibration, but that won't account for optical effects or anything that's a characteristic of the CRT itself.

No transmissive LCD technology has even nearly good enough blacks - most of the colour analysis systems used in the industry will refuse point blank to calibrate them. LCOS/SXRD has similar problems, fades with age, and often has wierd primaries. DLP is possibly nearest ideal, but horrendously expensive to do well; single chip is excluded and even the best high end 3-chip DLPs don't have such good blacks as a CRT monitor.

[perisoft]

Geometry can be made pretty damned good, and if you're referring to the propensity of CRT monitors to change geometry depending on total brightness (ie, full white screen is bigger geometrically than just a white border) they do quite well. My Barco 808s, which is hardly high-end, seems nearly perfect in this regard - a test pattern showing a white border with a flashing, inset solid white block shows zero movement of the border.

But with a CRT, the brightness of any individual white pixel is dependent on the rest of the image - a CRT projector is far less bright with a full white screen than with a 10% white screen. Put up a little white block in the left half of the screen, and its brightness will change if you light up the whole right half of the screen.

I'm not the right one to ask about the limiting factor in CRT ANSI contrast, since I'm not terribly experienced in that area. I just know people say it's a big weak point of CRT. I do know that it's not due to bloom on the CRT itself - it's an optical issue.

One other note: Beware of glowing praise for CRTs on this forum. While there are a lot of rational guys who will give it to you straight, there are some who would recommend CRT to someone who came in and asked for the easiest-to-setup, brightest, most portable projector available.

[km987654]

[Hfuy]

Heh - thanks for that. Yes, I appreciate that CRT video projectors require painstaking setup - I've done it a couple of times, and I've done convergence on normal monitors a couple more, although never at the top of a stelpadder! I also fully appreciate that they have sufficient mass to cause local gravitic anomalies, open wormholes to alternate dimensions, and snap grown men like twigs if they're daft enough to try and attach them to a ceiling.

There is always the issue, with CRTs, that they don't do horizontal contrast nearly as well as they do vertical contrast, especially when run at high bandwidth - it just can't switch the beam current on and off fast enough. This is clearly visible in the (otherwise very reasonable) first example shot above, which probably has a third less contrast between the vertical lines than the horizontal lines at the one-pixel rate. I suspect these things were never really designed to scan 1080p72. Any CRT device will do the same thing.

That's a slightly different thing to just optical blooming, though, which just implies that the lenses are... how can I be polite... not very good. They're comparatively large, and they're projecting an image that's quite close to their own size (most projectors use lenses much bigger than the imaging element, so you use only the optimal centre area of the glass).
They're probably not compound achromats, but then the output from the tubes is narrowband, so they shouldn't have to be. Making an extremely good lens that big would probably cost more than is workable. Also, the face of the tube seems to be greyish-white, so any scattering that does go on will bounce around a fair bit.

But I'm just guessing, really. Can anyone recommend some literature, or where I might find some numbers on static contrast?

Thanks,

-H

[ecrabb]

The reason ANSI contrast on CRT's is inferior to digital is because with any brightly lit phosphor, some percentage of the emitted light bounces off the back of the lens and illuminates the rest of the tube face... and raises overall black level in the process. With an air-coupled projector, the effect shows up as large "halos" or balls of light around small, bright objects. It can be pretty distracting and even quite obnoxious depending on the source material. With a liquid-coupled projector, instead of well-defined halos, the reflected light is reflected off the convex C-element and spread over the entire tube face. So, halos are essentially eliminated, and overall ANSI contrast is reduced. It's better than an AC-projector, but still not like a good digital projector.

Another problem with CRT's is non-linear output amplifier power. In short, peak brightness is much higher with low average picture level (APL) material than it is with high APL material. So, with a 20% white window, you might be doing 12 fL. Go to full-field white, and you might only be doing 6 fL. The effect manifests itself in that small, point sources of light (headlights, city lights, stars) are nice and bright, while a fully-white snow scene from Fargo is relatively dim. Of course, the effect gets worse and worse as screen size goes up, and the non-linearity is exacerbated.

What's your target screen size and brightness?

As you've pointed out, all display technologies have their strengths and weaknesses. A 9" LC/EM could certainly be calibrated to a target profile with excellent linear greyscale, SMPTE primaries, very good geometry, and good sharpness/resolution, but you'd need an outboard processor with a CMS (color management system) to do it.

I think something like a JVC RS2 or RS20 and a RadianceXD (full CMS control over primaries, etc.) would also be a reasonable solution. You get much more brightness (and linear), perfect geometry, correct primaries, linear greyscale, 1080p resolution, and superior ANSI contrast. An RS2 and RadianceXD could be had for well under $10,000 - potentially close to half that.

Check out Craig Rounds' RS20/Radiance thread for some more info:
http://curtpalme.com/forum/viewtopic.php?t=14999

SC

[kal]

From what I understand, even the high end digitals (like the RS20) don't have colours that are stable enough to do colour grading for movie work. We've seen at least one other person come in here asking the same question in the last few years. See this thread:

https://www.curtpalme.com/forum_archived/viewtopic.php@t=4485.html

Pick up Curt's newly listed Barco Cine 9 ($10.5K) that is basically brand new and you'd be able to do what you want. You won't have to reconverge every week either. A minor (10 minute) tweak once a year at most. That's all I ever do and I'm fussy.

Kal

[kal]

From what I understand, even the high end digitals (like the RS20) don't have colours that are stable enough to do colour grading for movie work. We've seen at least one other person come in here asking the same question in the last few years. See this thread:

https://www.curtpalme.com/forum_archived/viewtopic.php@t=4485.html

Pick up Curt's newly listed Barco Cine 9 ($10.5K) that is basically brand new and you'd be able to do what you want. You won't have to reconverge every week either. A minor (10 minute) tweak once a year at most. That's all I ever do and I'm fussy.

Kal

[stefuel]

I stopped reading this thread half way through because it's starting to sound more like trolling than legit questions. I'f I'm wrong, I'm sorry.

[ecrabb]

Kal, I assume you're referring to changes in the bulb's emission that would affect color temp and brightness. While that does change over time, I've heard the changes are minimal over shorter periods of time. I would think you could do a quick display calibration say, monthly to account for small changes in bulb brightness and color temp. For color critical work, you want to calibrate a display at least once a month anyway - usually more. I don't know if the OP's color grading work is an 8-hour-per-day kind of thing, but you could replace the bulb even quarterly if you wanted more consistent performance. You could sell the used bulbs with 500 hours on them on the used market to recover some of the cost of keeping a newer bulb in the projector and have to spend maybe $500-600/year to keep newer bulbs in the projector.

But, your Cine 9 suggestion is also very workable I think - as long as the OP doesn't need to use it on too large a screen. Anything over an 8'-wide screen is probably not going to work well, since in that other thread, that poster was shooting for 14-16 fL. The Cine 9 would probably just barely do 14 fL on an 8'-wide screen - and even then only in a 20% window. Brightness would probably drop to nearly half that with a high-APL scene.

The HDLink Pro looks like it has great color management, so perhaps no expensive outboard CMS will be necessary - so that's good.

SC

[Hfuy]

I've met Robert Houllahan, from that other post, and regularly read his posts on another forum - small world!

No, I'm not "trolling". I'm asking the questions I know are difficult for this technology, perhaps with a bit of tongue-in-cheek humour at the expense of the keenness for this older gear that is being displayed, but definitely tempered with an appreciation for the work that's going into getting the best out of it. I'm afraid the rainbow-coloured rabbit shadow trick will always be a favourite of mine! Seriously, the fact that you now have two people looking seriously at this gear for grading work is quite a compliment.

The reason Robert was asking for 12-16fL is simply because that's the SMPTE spec for screen brightness, a fact of which I suspect many of the more technically-minded calibration engineers will be well aware. That's usually measured open-gate on a film projector, that is, with no film running, so active peak white tends to be a little less. Human beings are quite good at seeing relative brightness, but quite bad at absolute brightness, so there is some flexibility in this. However, the idea that the image might literally halve in brightness depending on APL is a dealbreaker - is the high voltage power supply on these things really that poorly-regulated? Perhaps someone needs to make more capable EHT supplies for them, so they're actually capable of delivering enough beam-watts to illuminate the entire active tube area at full output. As it is, even these secondhand CRTs don't feel like they're providing much better brightness-for-the-money than recent DLPs, although of course they have other advantages.

Is this dimming less bad if you simply don't drive it as hard? Could I stack two of these things one atop the other and get around it that way?

Kal, for what it's worth, I read your calibration article, and having worked for a company which produces a calibration system for the industry, it seems you're doing pretty much the same thing for the same goals. It's amusing, as an engineer, to walk into a TV showroom and see soccer games going on with a dozen different shades of green!

-H