Theres an italian in the house! - Page 9

To me a multiplication of all levels is a dynamic gamma in the projector, but that is somewhat semantics as it depends on whether we are talking about before the iris or after (at least for an iris after the DMD where one before just changes the semantics and math more).

I'm not sure which CR you are talking about here. Sequential, intra-image, or both. If sequential CR improvement was due just to the smaller aperture they wouldn't need to have a dynamic system for higher CR. There is the issue of more ANSI lumens, but as far as CR, the dynamic system should be creating more sequential CR than they could get for native on/off CR just by closing the iris. But maybe your sentence meant something different than the way I interpreted it.

There are some fairly simple tests which seem like they could answer much of this about what they are doing. Put up a 20 IRE window on a black background and measure the window portion of the image with the DI enabled and with it disabled. If they are basically the same and the iris has closed down then the signal would have had to have been increased inside the projector to compensate for the more closed iris. If instead the absolute black level is 3x lower and so is the 20 IRE window, then no gamma adjustments are being made inside the projector.

I didn't measure the Planar, but I think they multiply levels to get 20 IRE basically back where it would have been if the iris hadn't been shut down. It will be interesting to see if SIM2 is doing basically that.

Any system that shuts the iris down could increase the sequential CR like that, but the key is largely whether you still get intra-image CR between things like 20 IRE and 0 IRE in darker scenes that is better than without the dynamic system. If not (there is no compensation to 20 IRE pixels at the imaging chips for the iris shutting down) then I don't think there would be much point to even doing a DI system other than marketing, at this point.

I agree that this is an issue, but maybe they aren't being aggressive enough to be much of an issue.

--Darin

Alan,

If you get time could you please run an experiment with the Lumis. Put up a 20 IRE window pattern and switch between DI enabled and DI disabled to see if it looks any different and then measure the 20 IRE window with the DI on and off. I'm not worried about absolute values as much as whether they differ, so as long as the measurements were being taken from the same place it would be good enough.

I'm guessing that your meter would have trouble measuring the black level around the window or that would be interesting too.

Thanks,
Darin

Does this come with the satin bag?

Here are some MC shots. I cant get them quite as sharply focused as I would like, but you can see that the convergence is excellent. CA is very good too. I tried to get as close as was possible to allow the full pixel structure to be seen.

This one is dead center of screen



Right hand side



Left hand side

CM,
when I talked to Sim guys at CES about getting a perfectly converged Lumis projector they told me that this is a concern of 'old technology'.
The Lumix has software built in to converge it by user like a CRT.
Is this true? Was your convergence at of the box or is did you use the menu system to converge?
Thanks in advance.
Ash

Thats OOTB.

I will be posting some more "Technical" shots later, I'll get some movie stuff posted soon.

CM,
Have you played with the convergence menus or can you confirm its existence for those who do not recieve perfectly converged Lumis?
Thanks again.
Ash

Yes it exists. It has horizontal and vertical for each color.

No one will ever get a perfectly converged machine, and the MC adjustment wont make it perfect either. Numbers like 1/4 pixel can be considered outstanding. A fair bit of MC is actually needed to visibly impact an image. It is far more important on a DLP machine as they are inherently capable of greater sharpness.

as far as i know the can shift it pixelwise but not in between.

cm the convergenze is 100% perfect.
think you like the unit

i will have my next week wednesday.

Alan,
That is awesome... thanks.
One more request. . . .
You posted width is 18 inches!
From the center of the lens, would you measure width to the right and width to the left?
My current projector Qualia 004 in the projector box has the lens in the center and I am wondering if I can fit the Lumis in my current box?.
Ash.

Its as close to perfect as my HT5000 was. If i could get in a bit closer and focus better, you would see some MC at the edge. I would guess at 1/4 pixel in one corner. Most is impossible to see.

I think its also important that people who will move to a 3 chip machine understand that some MC is unavoidable. Im sure we've all seen fantastic images produced my machines that have MC much worse that 1/4 pixel.

Look forward to reading your impressions.

The MC adjustment is in whole pixel increments.

That does look like great convergence. I'm looking forward to hearing how you think it compares to the HT5000 overall, especially for screen sizes where both provide enough light.

--Darin

Darin,

Read Greg's description in his review of the Planar projector. This is just one type of implementation and I don't know if the Lumis is like this or not.

You scale the highest level up to full white, adjust all the other levels so that the proportion between levels stays the same and close the aperture so that the highest level comes back to the original illumination. The result is that the gamma stays the same, except that black is now lower. Greg calls this no dynamic gamma. The second effect is that stray light is blocked by the smaller aperture, further increasing sequential CR.

Yea, what you said .

Art

Holy $hit.... the real deal indeed ,those look spectacular !

Art

Now for some resolution tests. These photos dont do this aspect full justice as the compression of the photos kills the sharpness a bit. The images show increasingly thinner lines expressed as a fraction of full rez....ie 1/5 rez is 5 pixel wide, 1/2 is 2 wide and full is 1.

Again great geometry and MC is apparent and no CA visible at all. The one thing that's really missing from these shots is the impact of the very high ANSI CR. Its just trashing the camera and causing it to shut down.

This is an overall look. The compression of the photo is causing some interference pattern to be visible at full rez. The closer shots show that the highest rez lines are resolved perfectly.

Holy crap; those shots are amazing.

Thanks, Coldmachine!

I can live with that convergence . Now to get some comparative numbers on how bright the little monster really is .

Alan, when you took your AEMC 813 measurments, did you post the fc reading directly from the meter, or did you convert to lumens? Sure seems hot.

Cheers,
Scott

That will be interesting indeed.

The HT5000 still packs a significant optical superiority that may well reveal extra finesse with fine detail. I suspect the 5k will have less pixel noise too. Pixel noise isn't something you normally see when its there, its more like a veil being lifted when its not.

Also, the 5k doesn't loose any light when zooming.

How much light would you lose zooming with the Lumis? Could anyone measure? If I pick this up I'll be at max zoom with the T2 lens. Any idrea how much light i'll lose?

Thanks!

This can be looked at from either direction. The 5k doesn't gain any light when zooming to bigger like the Lumis does.

Seriously, it seems to me that a constant aperture lens is just a personal preference thing as far as whether it is good or bad. It should mean either getting more light and less CR at the long end of the throw or less light and more CR at the short end of the throw compared to not putting that in there (or both).

For people with things like a 2.05:1 screen without an anamorphic lens I think constant aperture can be a negative. Makes it harder to get close to the same ft-lamberts for 2.35:1 material as 1.85:1 material, where the non-constant aperture lenses tend to increase the light from the projector as the image is zoomed to bigger for 2.35:1 material and help counteract that the light is spread over a larger area.

I don't think it is a huge deal either way, but just that I don't see constant aperture lenses as a great feature as opposed to just a different implementation with positives and negatives depending on the particular setup. Unless I'm missing something else they provide.

--Darin

You can go find the spreadsheet I posted on the C3X1080 for Bruzonsky. (His thread, this forum, somethign about C3X). I measured the light output through the zoom range It's in a graph in a tab of that xls. I expect it would be similar, relatively. It was pretty linear in Steve's unit.

HTH, Scott

How do you define "max zoom"? I'm guessing you mean furthest away from the screen for the particular screen size (smallest image end of the zoom) since you are talking about losing light, but I've seen it used the other way too.

--Darin

Am in the same boat. The choice is either long end of the T2 or short on the T3. If I go to a T3 it's more construction on the ceiling when I thought I was already done with the ceiling platform which I purposely made huge to have some flexibility. Unfortunately, I've fallen into the gap (mind the gap) of ratio black hole between the T2 and T3.

My preference would be more contrast and less brightness as a tradeoff as I run a 127" 2.37 studiotek with 1.3 gain at around 20 feet from the screen. From the above it sounds like a T3 with constant aperture. Counterintuitive methinks. Can someone confirm this as accurate.

Also, the out of box convergence is spectacular. Can someone please let us know if convergence adj is same as C3X, one whole pixel at a time or have they gone to fractions of pixels like sony did..

TIA

I think of that as having a dynamic gamma since the signal to the chips has to be modified to counteract the iris closing, but I see what he is saying. The above approach basically has the ultimate brightness compression, i.e. clipping. But it can be controlled by not being too aggressive. They wouldn't want to jump between two very different iris positions if an image with say 60 IRE as the brightest all of a sudden had a couple 100 IRE pixels show and then disappear unless they were okay with visible image pumping, so clipping in some cases would be necessary.

In Planar's case the dynamic iris isn't in the lens, so I wonder how much it improves the static on/off CR. Also, SIM2's isn't in the lens and from what I've heard about it I'm not sure if the static on/off CR with the iris/mechanism closed down would be better than the static on/off CR with it open either. It might, but I'm just not sure given that it seems to be back near the lamp.

I've wondered if any DLP manufacturers would go with 2 dynamic irises to improve static on/off CR when they are closed down (like the basically dual manual irises in the Sharp 12k and beyond and the Marantz 11S2), but haven't seen it so far.

--Darin

I don't see how it would be technically possible to adjust DLP 3C convergence in less than 1-pixel increments. But both these samples look good to me.

Yeah, I really enjoy mine; but light output in Iris 1 has been the one sticking point (I use a 120" scope SMX). The Lumis seems like the major step up I've been looking for.


Thanks Alan... so you're using a ND.9 for 3 stops then? I'm partial to the B+W's multicoated myself. Might start with .3 (1 stop) and see if the brightness is acceptable. Will definitely keep the "angle trick" in mind.

Gorgeous stuff CM...

Will Smith says it best: I have GOT to get me one of deze!