AVS Forum banner
Status
Not open for further replies.
1 - 16 of 16 Posts

·
Registered
Joined
·
6,092 Posts
Discussion Starter · #1 ·
HD and the LT150


I had the opportunity to give the LT150 a formal HD test signal workout courteous of Bill Cushman a professional equipment reviewer for Wide Screen Review who happens to be here in Houston.


His test equipment includes the Accupal and Sencore HD test signal generators, DTC100 as well as a Philips PM5639/00 color analyzer.


The main objective of the testing was to determine how the LT150 handles HD signals namely sampling and scaling. Instead of using a digital camera to capture the output image, I decided to utilize the LT150's capture card to get a look at the input signal.


As well as testing the scaling, we took the time to test other general performance aspects of the LT150 in the process. We also had Cushman's pristine 10HT as a reference. The test set-up included both a 97" X 52" Dalite high power and Stewart Greyhawk. Let the testing begin:


-----------------------------------------------------------------------------

TEST PATTERNS

Accupal component Overscan patterns:


1080i: " http://www.geocities.com/mrwigggles/1080i_overscan.jpg "


720p: " http://www.geocities.com/mrwigggles/720p_overscan.jpg "


First note that the image capture sizes are quite unusual. These images are how the projector Acquires the image not how it is displayed. This sheds some insight on how the LT150 "sees" the input signal. The input image size for 1080i is 1024 X 510 while for 720p the capture size is 1024 X 686. The unusual vertical resolution is because the LT150 allows for 5% overscan on each input. The 1024 number IS the horizontal sampling for both input resolutions and it comes into play in the next set of tests. Please note that while the captures sizes aren't 16 X 9 they were scaled and displayed on screen as 1024 X 576.


Accupal Multi-burst:


1080i: " http://www.geocities.com/mrwigggles/1080i_multi.jpg "


This is where things get interesting. The 5 bursts go in increasing frequency from left to right. The first set of lines are 5 pixels wide, then 4 pixels, down to 1 pixel wide for the final 5th burst.


The 1080I, which comes out of the Accupal at a full 1920 pixels, shows something very interest. The first four bands show pretty good representation of decreasing line width. However, the final set of lines are at nearly twice the horizontal resolution of the projector and should in fact be gray. They instead create a beat frequency and are clearly anti-aliased image of the input signal. If the projector had captured the input at 1920 and then scaled it to 1024, it would have averaged each white and adjacent black line together to form a accurate downsampled gray like I mentioned. This averaging for high res sources can also be done with analog anti-aliasing filtering – all of these capture with the analog "Video" filter enabled. The filter helps some but not enough to remove the aliasing completely. With it off, the fifth set of lines become completely black and white wide bars. This anti-aliasing on 1080i was one of the few shortcomings of the LT150 we did. Also consider that the Accupal's perfect high frequency output exposes this deficiency much more than most HD boxes would which don't quite have the high frequency output.


Another example: " http://www.geocities.com/mrwigggles/black_lines_on_grey.jpg "

There are four vertical lines in the test signal. The first line is one pixel and the second is two pixels. However, the LT150 has each of the lines as one pixel wide in the capture. The second should be one pixel wide (1024/1920 ~= ½). However, the LT150 got lucky when it sampled and caught the first one pixel wide black line exactly when it occured. It should be blending of the black and gray to produce a dark gray (not black) line instead.



720p: " http://www.geocities.com/mrwigggles/720p_multi.jpg "


This is the same type of test signal as before but the LT150's performance is much better do to the fact that 720p is only 1280 wide thus it is only slightly out of the LT150's 1024 capture potential. The LT150's video filter works really well on the 720p to help make the slightly aliased lines slightly gray like they should be.


The 720p did real well on the black lines test. Picture omitted because it was too perfect.


Interestingly, the slightly wavy lines in the capture weren't noticeable in the actual image at all. My guess it could be partial MPEG artifacts or exercising the card writer couples come digital noise into the analog section and/or effects its timing.


Sencore RGBHV torture image:


1080i: " http://www.geocities.com/mrwigggles/1080i_torture_test.jpg "


720p: " http://www.geocities.com/mrwigggles/720p_torture_test.jpg "


You'll notice with each pict the LT150 does a pretty good job of capturing the input. There is some slight noise around the edges but overall the image is quite nice.


These where all of the meaningful images that I was able to take. Unfortunately NONE of the Accupal or Sencore's test images had one pixel wide horizontal lines the minimum width was 2. This would have been very helpful for 1080i "bob" or bounce testing. As it was all horizontal lines were rock steady but the test wasn't all that difficult.


------------------------------------------------------------------------------

PERFORMANCE TESTS


General:


Light output (4X3) : 638 lumens

Light output (16 X 9) : 478 lumens

Peak On/Off contrast : 911:1


notes: white segment was "on". These numbers need to be reduced by roughly 1/3 for the "off" setting.


Gamma:


100 IRE = 13.01 ft-L

75 IRE = 8.51 ft-L

50 IRE = 3.25 ft-L

25 IRE = .73 ft-L


Calculated Gamma = 2.1


note: Natural 2 was used. Natural 1 will be slightly higher in value.


Colorimetry:


100 IRE:

x = .277

y = .315

temp = 9189 K


75 IRE:

x = .278

y = .314

temp = 9059 K


50 IRE:

x = .282

y = .306

temp = 9057


25 IRE:

x = .282

y = .284

temp = 10171


notes: I have my red and green contrast up 2 clicks each.


Uniformity: Flawless.


------------------------------------------------------------------------------

CONCLUSIONS:


Cushman was impressed to say the least. This was the highest contrast ratio he had ever measured. It was also one of the first DLP's to come pretty close to its light output specs. I have about a 100 hours on my unit so it is probably would be very close to its 800 spec brand spanking new. He loved the gray scale tracking, consistent white balance, uniformity, shadow detail, very good test pattern behavior, geometry, honest light output, etc. However, He did notice rainbows, thought the colors were decent, didn't like the aliasing on 1080i, could have a quitter fan, and wished it was WXGA like the Sony.


To put into perspective, the LT150 outperformed the Marantz 8100 by an incredible margin. His words about the Marantz weren't pleasant. He feels that the unit is about the quality of the Yamaha based on his initial (not formal) impression of the Yamaha.


What did I think, The Da-lite high power is awesome. The Grayhawk helped the 10HT look better in the dark scenes; but the high power was the better screen for the LT150. I did notice slightly more rainbows with the hi-power but the resulting colors were well worth it. It is counter-intuitive but retro-reflective screens don't hot spot from ANY reasonable viewing position. The intensity changes somewhat as you move around but the screen doesn't hot-spot. To get the full benefits, you'll need to table mount, but ceiling mount will still be pretty darn good.


I used to have a 10HT so seeing one again was a trip down memory lane (a whole year and a half ago!). Cushman's unit perform just like the one I had used to. On bright HD images it looked very good, on dark images it looked pretty bad. The LT150 looked pretty good all of the time. I do need to point out how well the 10HT handled the test images. It did about as good as you could imagine, the 1080i multi-burst being a dazzling image. The slightly scary thing is after almost 2000 hours of use, slight 4 X 3 burn-in was noticeable, sorry.


As far as my suggestion are concerned, Try to get a 720p component source if possible. While 1080i technically has more resolution, the LT150 actually captures 720p with more. Let your source do a digital 1080i to 720p conversion if possible. If it isn't possible, make sure the video filter is engaged or if you are adventurous you can make a better external analog "T" filter using two inductors and a capacitor (good luck). The other option is to break down and get a Vigatech HD scaler to give you pixel perfection XGA output for 1080i or 720p data, its relatively pricey but could be very useful for this application. My HD box at home has the luxury of 720p or 1080i output. The 720p looked better to me; these tests confirmed what I suspected.


I know it is crazy but I actually like using the clear section to get a little more contrast. When Thumper does finally get his modifications finalized, I will get the best of both worlds.


Sort of off-topic, most 1080i stuff looks very poor as soon as motion occurs, especially sports on any display. It is like the image downgrades to 480i when motion happens. More things need to be shot in 720p. For 99% of the HD displays out there, it is the max resolution possible and it gives much better motion. 1080 at 30p would work better than 1080i. NEC/Plus need to come out with a 720p projector. Take a LT150 slap in a 720p chip and a faster color wheel and we'll be all set for ~$4.5K street.


Any questions?


-Mr. Wigggles
 

·
Registered
Joined
·
819 Posts
Quote:
Originally posted by MrWigggles
HD and the LT150

NEC/Plus need to come out with a 720p projector. Take a LT150 slap in a 720p chip and a faster color wheel and we'll be all set for ~$4.5K street.
Yes, it's called Sharp XV-Z9000, it costs $8000+ and it also has problems with fast motion :)


Seriously though, what we really need is a projector you described, but with DVI/PanelLink/HDCP and a built-in de-intercer/scaler that REALLY works on all input material (Genesis+Faroudja, anyone?)


The wait continues. I'm betting it still won't happen next year at the price level you set.


Regards,

Halcy


PS My 5 minutes old LT-150 projector that I bought as an interim solution blew up due to badly installed electricity outlets. Now I'm looking for a replacement...
 

·
Registered
Joined
·
4,072 Posts
Quote:
Originally posted by MrWigggles
Accupal multiburst:


1080i: " http://www.geocities.com/mrwigggles/1080i_multi.jpg "


This is where things get interesting. The 5 bursts go in increasing frequency from left to right. The first set of lines are 5 pixels wide, then 4 pixels, down to 1 pixel wide for the final 5th burst.


The 1080I, which comes out of the Accupal at a full 1920 pixels, shows something very interest. The first four bands show pretty good representation of decreasing line width. However, the final set of lines are at nearly twice the horizontal resolution of the projector and should in fact be gray. They instead create a beat frequency and are clearly anti-aliased image of the input signal. If the projector had captured the input at 1920 and then scaled it to 1024, it would have averaged each white and adjacent black line together to form a accurate downsampled gray like I mentioned. This averaging for high res sources can also be done with analog anti-aliasing filtering ?all of these capture with the analog "Video" filter enabled. The filter helps some but not enough to remove the aliasing completely. With it off, the fifth set of lines become completely black and white wide bars. This anti-aliasing on 1080i was one of the few shortcomings of the LT150 we did. Also consider that the Accupal's perfect high frequency output exposes this deficiency much more than most HD boxes would which don't quite have the high frequency output.
Thanks for a GREAT review! :)


About this 1080i sampling issue, isn't current LT150 imperfect "anti-aliased" image gives a better image in high frequency by showing more resolution (more black/white transition)? In stead of the "better" way you said by just make it a "gray" field? I were always keep the video filter OFF in 1080i HD viewing. I think this imperfect "anti-aliased" image should look better than a perfect "anto-aliased" plain gray field.


Sorry, I don't really know the full technical detail here. Just my guess.


Btw, you should also check LT150 HD with the aspect ratio OFF (Normal mode, not Cinema mode). It should give more vertical resolution though you need a anamorphic lens setup. Or according to your finding, the LT150 only samples 1024x510 for a 1080i input signal, so there's no point to use the full 1024x768 panel for better resolution?


regards,


Li On
 

·
Registered
Joined
·
6,092 Posts
Discussion Starter · #4 ·
LiOn,


I think a little aliasing is O.K., but by the time the multi-burst reaches 1920 lines across the output needs to be grey.


I knew someone was going to ask the aspect ratio question.


These image are not digital pics they are input captures as they appear on the PC card.


Changing the aspect ratio to normal didn't change how the image was captured just how it was scaled. For 1080i it either scales the 510 up to 576 or up to 768 (it might do some odd field/ even field offset in the process but I'm not sure). However, the 686 lines of 720 lose resolution when it compresses them down to 576, but won't lose any when expanded to 768 thus facilitating anamorphic lens.


-Mr. Wigggles
 

·
Registered
Joined
·
2,027 Posts
I have compared running the 1080I output from my HD receiver versus scaling it to 1024x768 with the Vigatec. I actually prefer the "straight" image. There was slightly more aliasing with 1080I, but the Vigatec created a slightly disconcerting "blur" effect whenever there is any motion. Different parts of the image blur at different points. I found it distracting.
 

·
Registered
Joined
·
195 Posts
Mr Wiggles,


Wonderful info! Thanks for your work. Couple of questions..


You recommend the HighPower over the Greyhawk. Can you elaborate on your observations here? I think you are the first to look closely at these two screens side by side witht the LT150. Were the black level improvements from the Greyhawk not as beneficial as the the improved colors? Would you still prefer the HighPower over the Greyhawk in a ceiling mount install? I have a Greyhawk, is it worth "upgrading"? :)


You mention using a 720p component signal - why not RGB?


And finally, do you really have a Cold Fusion patent?


Thanks,

jeff
 

·
Registered
Joined
·
1,563 Posts
Probably a stupid question given your experience and the experience level that I'd assume your WSR friend had, but you completely recalibrated all brightness/color/etc. settings for each screen, correct?


Scott
 

·
Registered
Joined
·
6,092 Posts
Discussion Starter · #8 ·
Jeff,


One thing that is often missed when considering what is the best screen choice is the increased perceived resolution as the brightness gets to a high enough level.


At a 54" X 96" image size on a .8 or so gain greyhawk screen, the light intensity reflected back from a 478 lumen image is only 11 ft-Lambarts range (ft-L = lumens/sq ft). TV's are about 20 ft-L but the best images come when the output is even greater than that at about 30 Ft-L for projectors with good contrast ratios (Cushman thinks upwards of 50 ft-L are ideal). Regardless, the high power more than doubled the 11ft - L to somewhere in the 25ft-L range. The image became much sharper and the colors became much more alive.


On the 10HT which was at about the same output as my LT150 (2000 hours running in cinema black mode), the high power look the best normally with most material; but it raised the black level too much for my taste with dark material.


The key elements are for the resulting image are high contrast and good image brightness. Absolute blacklevel only becomes an issue with dark material on projectors with poor contrast that being shown at high gain screens.


Please bear in mind that their are many good high gain screen out there from Stewart, but if you want something that is around 2 or more, nothing beats the high power.


-Mr. Wigggles


Ps. Geocities isn't too happy with my current traffic load. Be patient with the "brown-outs" that might arise when downloading the images.
 

·
Registered
Joined
·
6,092 Posts
Discussion Starter · #9 ·
Scott,


I used settings that fit my tastes. There are sacrifices when trying to achieve 6500K, etc.


However, my settings for brightness and contrast to effect the black level and peak white level respectively didn't change throughout the test, there was simply no reason for temp to.


And since the High-Power and Grayhawk are perfectly nuetral colorwise there was no reason to adjust color either.


-Mr. Wigggles
 

·
Registered
Joined
·
1,563 Posts
Mr. Wiggles, at the least I would think that the LT150's brightness level would be set lower when projecting on the Hi-Power vs. the Grayhawk, no?


Scott
 

·
Registered
Joined
·
6,092 Posts
Discussion Starter · #11 ·
Scott,


I set the black level and white level for peak performance at all times. There is no sense raising the black level for no good reason and there is no sense lowering the white level either really.


-Mr. Wigggles
 

·
Registered
Joined
·
733 Posts
Thank you Mr. Wigggles! I feel a lot better now. I too leave the white segment on for the good transfers. Blazing contrast with wonderful punch from the Hipower. This combination rocks my world!!! Even better with an ISCO II.


Saw KBK's awesome setup last night - my marriage wouldn't survive it! Oh well, maybe in my next life. He gave me a screen sample (grey) to test with the LT150 - should be interesting.


Cheers,


Grant


Favourite setup - LT150 + Hipower + AVS HTPC + ISCO II.
 

·
Registered
Joined
·
6,092 Posts
Discussion Starter · #13 ·
Jeff,


I missed your earlier question about component 720p.


Using component 720p will let you have color saturation control. If your going to be a bad boy like me and leave your white segment on, you'll need to raise your color saturation a little to compensate. It's not a perfect solution but it is pretty darn good.


-Mr. Wigggles
 

·
Registered
Joined
·
707 Posts
Just a note that the Philips analyzer is not capable of measuring non-CRT based displays accurately. The matrix for the colorspace is defined for only SMPTE-C or EBU phosphor sets neither of which are used in any digital displays. The chromaticity coordinates as well as the color temperature information is all therefore highly inaccurate. Please see Kevin Martins review for accurate spectroradiometric data of this display. The Philips is off by 2-3,000 degrees (high) for the color temperature readings.


When properly calibrated to D65 which requires turning off the White segment and adjusting the color temperature slider down to increase the Red output as the lamp is spectrally deficient after 50 hours of use in this primary you will have a usable contrast ratio of closer to 350:1 with excellent Grayscale tracking. With the White segment on it is impsossible to reproduce a stairstep pattern without experiencing White Crush at the top end of the scale. The midrange will suffer from excessive Red after a White Balance is performed due to non linearities in the gamma look-up table created for the lamp when it was brand new.


This can be corrected by having the look-up table reflashed with a corrected look-up table designed for the projector with a lamp that has stabilized after 50 hours of use. The difference in picture quality is significant!


For information on having the LT150 look-up table modified please contact Progressive Labs at 201-342-1235.
 

·
Registered
Joined
·
6,092 Posts
Discussion Starter · #15 ·
Cliff,


I don't understand your comments. If the Philips reads 2000 to 3000K high, that would mean that I'm at 6500 K! Why would I need a firmware revision?


However, sadly I don't think that I'm at 6500 K, I would guess that I'm closer to 9000K. Heck my image almost looks identical to my Princeton Graphics monitor when it is set for 9300K.


As far as the capacity for the reviewer and or the operator's ability to measure color temperature accurately. I will use Cushman's review of the G20 as a guide. The JVC factory sent it to him at 6500K and guess what he measured it with his Philips at 6500K.


The color analyzer subject has been a hot one for a while now. If people would like to study more on color analyzers', people are welcome to visit projectorexpert.com to read Peter Putman's review of 6 different units. Each one had their own interesting quirks.


This is all I will say on the matter. This thread wasn't intended to be a critique of Cushman's methodology or his equipment. Cliff, you've had your say and have made your plug. If this topic continues along these lines, I will delete the thread and post the information again in a locked one.


Good day,


-Mr. Wigggles
 

·
Registered
Joined
·
707 Posts
Mrwiggles,


The reason why I gave a range from 2-3000 degrees Kelvin variance is due to the data you listed ranging from 9057 to 10,171 K. To be more precise I should have said that your data was off between 1657 and 2,571 degrees Kelvin. I am not disputing the methods employed to take the measurements. I was only commenting on the limitations of the instrumentation used to collect it with.


The LT150 has a great picture which can be improved upon is all that I wished to convey. I am not trying to promote a specific instrument for measurement here.


Cliff
 
1 - 16 of 16 Posts
Status
Not open for further replies.
Top