CurtPalme.com Home Theater Forum

[draganm]

your right I misread, 1920 vertical lines is your Horizontal resolution. Still, I have a hard time thinking it's impossible. AFAIK no one has ever taken a serious look at the deflection circuitry of a MArquee and tried to really tweak it. I remember the VDC Hi rez 8500 with special tubes fully resolving 2500 x 2000 and TSE posted the patterns to prove it.[/b]

[mp20748]

SC and Gary, I'm more than agreeing with on this subject.

A CRT projector can be made to properly do 1080P with the 1:1 pixel test. which is the actual test for CRT bandwidth. But with that, it can't do perfect rise and fall timings. In other words, it can't produce perfect pixels because of the rise and fall timings that make up a perfect square wave.

If a CRT projector could produce perfect square waves, it would be able to PERFECTLY resolve 1080P or whatever rate.



However, with the 1:1 pixel pattern test, it's puts us close enough.


The main test of that six section pattern is the top left section, which would be the top vertical lines (lines going vertical). When those lines are being shown clear and sharp matching the horizontal lines to the left (or right) of it, then the 1920 scan is going well. And that is the part where the scanning moves from left to right and the electronics has to be fast enough to switch on/off 1920 times during each scan. That is where the vertical lines show how well that is being done. If not being done properly, the top vertical lines would show darker than the top horizontal lines, or the vertical lines would also be blurry or not clear at all.

The clearer the vertical lines, the truer it is to being resolved. But again, the test does not show how square the square wave would be, and that is an area that could not be corrected with CRT. It's just not going to produce perfect pixels like digitals...


it just somehow still looks better, or more film like..

[stefuel]

You guy's friggin kill me. There are people in this country (natural born Americans) who are starving and homeless and you have the nerve to piss and moan about only having a 8" crt projector. I can see it all now. Someone sitting on a street corner with a tin cup and a sign that reads "Deprived videophile needs money for projector upgrade"

[ecrabb]

Chip... Lighten up, man! If anything, most of us in this thread are pretty damn happy with what we have! Besides, if you go down the "people are starving and homeless" road, well, then the very fact that this forum even exists is unreasonably lavish. Never mind a home theater - compared to most of the world, just the very fact that most of us own two cars and a garage to put them in makes us ultra-wealthy.

SC

[WanMan]

[garyfritz]

What SC and said. I believe that since the vertical-line sections were significantly darker than the horiz-line sections, that implies that the projector didn't have enough bandwidth to approach the full-on/full-off levels of the vertical lines. In the horiz sections the horiz lines are fully bright white, and the gaps are pretty much fully black. There's no bandwidth challenge in those areas. But in the vert sections the lines/gaps are alternating levels of gray, causing the vert areas to be darker. I believe that's because the bandwidth is insufficient to pass that high-frequency signal. It's chopping off the square wave and turning it into a sine-like wave that never reaches fully white or black.

[kal]

[kal]

[Kiev Savoie]

the interesting nugget i got from this discussion so far is that thing about maximizing raster size. when i was playing with the the design of the 2 mirror system and how to fold the projection cone into the most compact configuration possible, my models led me to a set up where the back end of the unit was higher than the front, putting the lenses pointing dead center of the screen at the zero axis. this made the raster almost perfectly square allowing me to bring the sides right to the very edge of the tube with no keystoning. the only drawback to this was that the bottom part of the screen would not completely focus during setup, I still haven't figured out what to do about that, but i WAS using every bit of raster available from those tubes.

[Brian Hampton]

Hi,

I have no doubt a 9" pj is the ultimate but I never got that far.

I had a Sony 1271 and Marquee 8500 and then G70.

Having looked at various rez on the G70,... I went with 1080p.

Was it fully resolved? Maybe not. Was is very sharp, Absolutely.

I personally think 1080p is best for G70 and Marquee 8500.

-Brian

[cmjohnson]

I recall my own FrankenYoke experiments and development. Using an 8000 as a testbed first, then an 8500, and then a 9500LC, I was
able to determine that the HD-8B lenses themselves are incapable of fully resolving 1080p even though I could fully resolve up to 2048x2048 in 4x3 mode, with maximized rasters, using the FrankenYoke Mark 1s, on the CRT face. The HD-8B lenses are limited to
about 1440x1080 (4x3) or 1440x810 (16x9).

Upgrading to the 9500LC for the Mark II trials, with HD10GT17 lenses, I was able to achieve full system resolution of 2048x2048 from
end to end. I believe the tubes are good for 2500x2000, with optimized Mark II FrankenYokes, but I think 2500 pushes past even the
GT17's optical resolution limit. I was able to simulate that resolution by shrinking the raster by the calculated amount.

To achieve that test resolution (2048x2048) I was using a Sencore monitor analyzer/pattern generator.

I say that for an 8000 or 8500, the optimum resolution is 1440x810 in 16:9 mode as it's at the limit of the lenses.


CJ

[garyfritz]

But if you're using HD145's instead of HD-8B's, you've got a lot more optical oomph to work with. I'm not entirely certain but I don't think 1080p is stressing my HD145's much. The horizontal lines sure resolve well.

[cmjohnson]

HD145s WILL resolve 1080p. That's well established. But the far more common HD-8Bs can't do it.


CJ

[ecrabb]

[WanMan]

You know, most newbies such as myself focus on the limitations of clearly resolving resolution in the raster, and never considered that of the lenses. I wonder what the stock lenses on an XG1351LC are not capable of.

[ecrabb]

Wan, your XG should be very comparable to my G70. They should be identical optically since both have the same lenses and C-elements - HD18's.

At 1080p on my machine, horizontal on/off lines are nice and sharp. Vertical are very visible, but just not sharply resolved. I'd say real horizontal resolution on my machine is probably limited to around 1500 or 1600 horizontal. That's definitely a video-chain limitation because vertical elements are MUCH sharper than horizontal elements. Still, it's not like I can see 1080p scan lines...

Your XG might even be a tad better on horizontal res/bandwidth - I think the XGLC is rated around 150mhz compared to the G70's 120, though I'm not sure if the XG was over-rated by NEC. Maybe somebody can confirm.

Dude, you're in for a real treat - Just set the damn XG up well and watch some movies!

SC

[Mark_A_W]

NEC made up the bandwidth specs. Different models with different bandwidth specs all use the same hardware.


Real world the NECs and a G70 are about the same.

In fact the only differences are:

- Possible better corner focus on G70 due to shithot flapping system.

- More robust deflection on G70 (less raster ringing).

- Better geometry on NEC due to shithot convergence/alignment controls.

- Better memory management on NEC.

- No AKB/ABL line on NEC.


The only difference that may actually be visible is the AKB line.

[skippyar]

[perisoft]

[garyfritz]

[Kiev Savoie]

I dunno man, Optical components have to be made in a reasonably precision manner. but, that's just my assumption.