AVS Forum banner
Status
Not open for further replies.
1 - 18 of 18 Posts

·
Premium Member
Joined
·
126 Posts
Discussion Starter · #1 ·
While talking with Cliff Plavin two months ago, he mentioned that he would be getting software to alter the gamma look-up tables for NEC projectors including my recently purchased LT150. He asked if I was willing to be a guinea pig, and I agreed to let him have my projector since I was in the middle of changing jobs and could be without one for a while.


I got my pj back recently, and now that I've spent much of this weekend testing it, I can say that I am extremely pleased with the results.

Before calibration...


I found the picture to be quite good and was very pleased with my purchase (thanks AVS members for the recommendation!). I was particularly impressed by the shadow detail and the smoothness of the picture. I was also lucky enough to have seen an LT150 before I purchased one for myself (always a good idea if you have the opportunity!).


However, I felt there were a few short comings. First, the "rainbow effect" was a problem, since I'm sensitive to it (I even see it on the Sharp Z9000U), but considering the price vs. performance of this little pj, it wasn't a major concern for me and I can easily live with it.


Second, the colors seemed off and the whites were somewhat blown out above 90 IRE. No matter how I adjusted the white balance, I could not get the colors to look as good as my RPTV (a Mits 65905). When I set the 0 IRE and 100 IRE values by adjusting the white balance settings, either the bright colors had too much blue or the midrange was much too red. I adjusted it as best as I could to balance out the two extremes. To fix these problems, the gamma look-up table needed to be adjusted. Fortunately, this was exactly what the factory software that Cliff has can do.

Calibration...


What Cliff did was to first adjust the picture controls (brightness and contrast), the color temp control and the white balance settings on the unit to the correct levels. He found that when the 0 IRE and 100 IRE levels were close to D65, the midrange was off. This agreed with what I was seeing (above).


Before making any changes to the gamma look-up table, he took some measurements using the default settings. My LT150 had about 100 hours on it so the color temp of the bulb was most likely stable, but he found that it had too much green and blue at the top and bottom of the gray scale. The result is that the contrast ratio was artificially boosted to 550:1 (with the white segment off). Cliff found out from NEC that the red part of the light output on a new bulb is higher than on a stable bulb. This might explain the lack of red in measurements that he took.


Once he had the correct factory software for the LT85/150 (and I believe the zoom versions as well), he was able to create new gamma look-up tables and install them on my pj. He then measured the color temp at each IRE level with his spectroradiometer to verify that the settings were accurate. In place of "Natural 1", he installed a 2.2 gamma look-up table and in place of "Natural 2", he installed a 2.4 gamma look-up table. He did this for both RGB and for video. Cliff can fill in the details of this procedure for those that are interested.


After finishing the calibration, the pj's color temp tracked very close to D65 at all IRE levels. Cliff measured a post-calibration contrast ratio of 335:1. He said the lower values are due to the green and blue being lowered to make the color temp track D65.


Cliff sent me a list of settings that he made, as well as some data (see below), so that I could verify what he changed.

After calibration...


Once I received my LT150 back, I eagerly set it up and made sure that the values were set as Cliff had described. It was still light outside, but I couldn't wait for the sun to go down. Since I've got pretty good light blocking shades, I went ahead and watched a few DVDs that I normally use for testing.


The thing that I noticed immediately was that the flesh tones are so very much more accurate! It was wonderful to see that kind of accuracy -- especially considering that I had become used to seeing the colors on my RPTV.


Next, I waited very impatiently ;) for it to get dark and then ran through my test material. I tested DVDs and HDTV using the 2.2 and 2.4 gamma tables over then next several nights. I compared the results to my RPTV, which is setup in the same room. For the DVD, I used both my HTPC (with an ATI Radeon card and WinDVD) and my Toshiba 5109 (which allowed me to view the same image on both the LT150 and the RPTV). The HDTV source was from a Time Warner HD cable box.


DVD: 5th Element (superbit), Dark City, Gladiator, Shakespeare in Love, The Patriot, AVIA


HDTV: PT demo loop (PBS-DT), UT/SC football game (CBS-DT), The Patriot (HBO-HD)


To test the colors, I looked at the standard color bars pattern on the AVIA DVD. The cyan, green and blue bars were approximately the same on both the LT150 and RPTV; however, the RPTV's colors appeared to be somewhat more saturated than the LT150's when producing the the same sized image. I could see a larger difference in the magenta bars (slightly too blue) and yellow bars (slightly too green), and the red is very obviously orange-red on the LT150. I compared the LT150's and the RPTV's colors while watching The 5th Element and Shakespeare in Love, and they appeared to be very accurate, but again the LT150's colors did not appear to be quite as saturated as on the RPTV. The flesh tones on both also appeared to be nearly the same.


I'm still amazed that this pj can nearly display all of the shadow details that my RPTV can. I was worried that the lower contrast ratio after calibration would adversely affect the picture, but I was pleasantly surprised to see that the shadow detail appears to be just as good as it was before calibration. As expected, the LT150 cannot produce the jet blacks that the RPTV can, but it does an admirable job. The LT150 was able to handle the dark scenes in both Dark City and Gladiator very nicely. In addition, I can finally tell that the red capes in the opening scene of Gladiator are actually red! Before calibration, they appeared to be more gray than red.


I watched both DVD and HDTV versions of the The Patriot a few days ago, and both DVD and HDTV versions of Toy Story this evening. I had the DVD running on the RPTV and HDTV on the LT150, and then I switched them. The colors on the DVD and HDTV are very close to being the same. The detail is MUCH greater on the HDTV version of The Patriot than on the DVD version. Unfortunately, there was not much additional detail in the HDTV version of Toy Story -- I'm not sure why.


As for 2.2 vs 2.4 gamma tables, I like the 2.2 gamma table best since it allows me to discern slightly more shadow detail.


Unfortunately, I was not able to make measurements with a colorimeter since mine is out of calibration. I compared the gray scale to my calibrated RPTV, and the color temp of both (as seen in gray steps and ramps from AVIA) appear to be very similar. When I have a colorimeter again (one is being calibrated and the other is still on order), I will make measurements and post them here, if people are interested.


I've also watched SW:TPM and Driven all the way through to see if my initial impressions were correct, and I believe they are. I'm extremely pleased with the results of the calibration. Not everyone will be as picky as I am about color accuracy, but I think everyone will see a definite improvement, if they have their LT150's calibrated.

Misc data...


The color temp values that Cliff measured after calibration using the 2.2 gamma setting are:
IRE.....x.......y.......K

100.....0.31....0.33....6537

90......0.31....0.33....6557

80......0.31....0.33....6491

70......0.31....0.33....6418

60......0.31....0.32....6805

50......0.31....0.32....6721

40......0.31....0.33....6626

30......0.31....0.33....6617

20......0.32....0.33....6365

Cliff measured the red, green and blue chromaticity values of the LT150:
Color...x.......y

Red.....0.6187..0.3684

Green...0.3131..0.5818

Blue....0.1502..0.0663

Comparing the values above to the standard SMPTE-C chromaticity values, you can see that the blue is pretty close, but the green is a little too yellow and the red is quite a bit too orange.


If you have any questions or comments, please let me know. I hope some of this info is useful to the forum. The info I've found on the forum has been very helpful to me, and I'm glad that I can contribute something back!


Cheers,

Kevin
 

·
Premium Member
Joined
·
25,495 Posts
" Cliff measured a post-calibration contrast ratio of 335:1. He said the lower values are due to the green and blue being lowered to make the color temp track D65."


Why not do what Steve Smallcombe does with the 10HT - use a red filter to cut blue and green, as well as lower black level/increase contrast? This could be done in conjunction with the new gamma tables to get the color spot on.
 

·
Registered
Joined
·
4,525 Posts
Very interesting stuff, Kevin,


My findings were quite similar to yours. I took some pictures to show what the LT150 looks like.


Here are the primaries. I forgot to take a screen shot of the "raw data", but you can see that the chromacities are very similar to those in your description above. I measured the green as a tiny bit greener than your projector, but it's pretty small.

http://www.dilard.com/dilard/images/...50CIEChart.jpg


Here is the color temperature histogram. This shows the temperature throughout the entire gray scale, with black on the left and white on the right:

http://www.dilard.com/dilard/images/.../LT150temp.jpg



But here is the really interesting part...these are 100% completely stock projector readings! I've haven't even done minor, end-user tweaking (I've been too wrapped up with the "other" projector lately).


Out of the box, the projector runs a tad to the blue...but it's fairly flat across the grayscale. This is common as it helps manufacturers to look their best to potential customers.


If I get the chance, I will measure it after just some end-user tweaking and repost the pictures.


Do you notice the decreased contrast after the calibration much?
 

·
Registered
Joined
·
4,525 Posts
Here are Kevin's chromacities again so you can compare to the picture above:

Color...x.......y

Red.....0.6187..0.3684

Green...0.3131..0.5818

Blue....0.1502..0.0663


You will have to "plot" the points on the graph above to compare the out-of-box unit with the calibrated one.
 

·
Premium Member
Joined
·
126 Posts
Discussion Starter · #5 ·
Noah,


Good idea! Does Steve Smallcombe read this list? I'd be happy to work with him to figure out the best way to try this idea. I'll send him some e-mail.


Thanks,

Kevin
 

·
Premium Member
Joined
·
126 Posts
Discussion Starter · #6 ·
Mark,


Good to hear from you! I hope the product releases are going well!

Quote:
Originally posted by milori
Very interesting stuff, Kevin,


My findings were quite similar to yours. I took some pictures to show what the LT150 looks like.


Here are the primaries. I forgot to take a screen shot of the "raw data", but you can see that the chromacities are very similar to those in your description above. I measured the green as a tiny bit greener than your projector, but it's pretty small.
Excellent! It's always good to get independent confirmation of the data. If you get a chance to remeasure the LT150, I would be interested in the raw data.


Quote:
But here is the really interesting part...these are 100% completely stock projector readings! I've haven't even done minor, end-user tweaking (I've been too wrapped up with the "other" projector lately).


Out of the box, the projector runs a tad to the blue...but it's fairly flat across the grayscale. This is common as it helps manufacturers to look their best to potential customers.
Interesting... what setting on the LT150 did you use? All of my measurements were using "Natural 1" prior to calibration. There was too much green and blue, and when I tried to compensate for it with the white balance, the red became too strong in the middle.


Also, how many hours do you have on your lamp? NEC said that the red burns off after a while, which might affect your measurements. My lamp currently has about 180 hours on it, which should be stable.


BTW, have you looked at the gamma curve? Before calibration, it didn't quite fit the 2.2 or 2.4 standard. Mine had a significant bump near 90 IRE, which crushed some of the whites.


If you are interested (since you're only a short distance away), I'd be happy to bring my LT150 over to compare to your pre-calibrated one (or have you over at my place). I would find the comparison very interesting. However, I know you're quite busy right now with product releases, but afterwards, let me know if you are interested.


Quote:
Do you notice the decreased contrast after the calibration much?
I thought that I would have noticed it, but I really haven't so far. I think this is where the comparison of a pre-calibrated LT150 and my post-calibrated one will be most interesting!


Kevin
 

·
Premium Member
Joined
·
126 Posts
Discussion Starter · #7 ·
Quote:
Originally posted by kem
Good idea! Does Steve Smallcombe read this list? I'd be happy to work with him to figure out the best way to try this idea. I'll send him some e-mail.
Noah,


Do you have Steve's e-mail address? I didn't see it on his website. I sent him a PM instead.


Thanks,

Kevin
 

·
Registered
Joined
·
707 Posts
As Milori's color temperature histogram points out the projector tracked linearly at roughly 1000 degrees Kelvin too Blue. The problem is that when you try to reduce the overly Blue color temperature to a proper D65 by utilizing the color temperature slider and utilizing the White Balance controls, that you will not get the same results due to the lamp having an inability to generate enough Red at the top and bottom end of the Grayscale. By making the appropriate changes in the OSD adjustments you will always come up with a compromise for the midrange where the fleshtones suffer. I recorded a similar pre-caibration histogram and was not able to get a perfectly uniform calibrated response from the projector without adjusting the gamma look-up table.


The projector does give very good performance out of the box in the User 1 RGB mode but can show susbstantial improvement via calibration. The contrast ratio does show a significant reduction due to the reduction of the Green which is substantially overdriven to produce the light output which all of the manufacturers consider first in their specmanship with each other. This is very typical performance from most digiital projectors made today. Being unable to reduce the Black level by a similar amount as the peak White causes the contrast ratio to be reduced. The picture quality post calibration as Kevin noted tracks substantially better without losing any detail in the Black.


I believe that by utilizing filters to compensate for the color gamut that the projector produces you may experience a greater light output loss than the calibrated projector presently provides. This would further reduce the contrast ratio! The filters could improve the colorspace and still not correct for the improper gamma look-up table which is mistracking due to the lamps spectral output being reduced in Red content. The filters will provide a linear change while the mistracking is due to a non-linearity taking place!


For those interested in having this gamma look-up table modification performed on their LT150 you may contact me anytime at my office at 201-342-1235 and I will provide you with additional details.


Cliff
 

·
Registered
Joined
·
4,525 Posts
Thanks for the information, Cliff.


As the graph shows, the temperature was pretty flat across the grayscale out of the box, but I will try adjusting the blue slightly down and running the tests again to see if this changes.


Mark


PS - Maybe I just got an exceptional unit, but my out-of-box LT150 contrast (full on/off) was significantly higher than either the before or after contrast measured on Kevin's unit. Untouched, the projector measured 838:1.

http://www.dilard.com/dilard/images/...50contrast.gif
 

·
Premium Member
Joined
·
25,495 Posts
Kevin,


Sorry, I don't know Steve's email.


Cliff,


"... that you will not get the same results due to the lamp having an inability to generate enough Red at the top and bottom end of the Grayscale."


I don't understand. The lamp itself is always on putting out whatever its spectral characteristics are. How could the red level relative to the rest of the spectrum be different at low or high IRE?


"I believe that by utilizing filters to compensate for the color gamut that the projector produces you may experience a greater light output loss than the calibrated projector presently provides. This would further reduce the contrast ratio!"


I thought filters were available in various densities, so this would be true only if the mildest filter cut output more than calibration. And since optical filtering cuts black level also, contrast should not be eaffected. Compared to calibration, it should be increased, since the latter doesn't lower black level at all.


Thanks
 

·
Premium Member
Joined
·
126 Posts
Discussion Starter · #11 ·
Quote:
Originally posted by milori
PS - Maybe I just got an exceptional unit, but my out-of-box LT150 contrast (full on/off) was significantly higher than either the before or after contrast measured on Kevin's unit. Untouched, the projector measured 838:1.
Out of the box, the white segment is turned on, which will further boost the contrast ratio. If you didn't switch it off, then ~800:1 should be possible.


The 550:1 before calibration was measured with the white segment turned off in software. This is in line with another post made by Thumper where he measured ~600:1 on an LT150 with the white segment turned off. I don't remember what he measured on that unit with the white segment turned on.


Kevin
 

·
Premium Member
Joined
·
126 Posts
Discussion Starter · #12 ·
Quote:
Originally posted by ghibliss
I believe that by utilizing filters to compensate for the color gamut that the projector produces you may experience a greater light output loss than the calibrated projector presently provides. This would further reduce the contrast ratio! The filters could improve the colorspace and still not correct for the improper gamma look-up table which is mistracking due to the lamps spectral output being reduced in Red content. The filters will provide a linear change while the mistracking is due to a non-linearity taking place!
I think what Noah and I are interested in trying is to see if it is possible to use a CC filter on a post-gamma-calibrated LT150 to compensate for the lack of red in the lamp output. This way we shouldn't have to push the red part of the white balance quite so high.


I agree that, based on the measurements that you and I both made, the CC filter wouldn't fix the problem that the pre-calibrated gamma look-up table caused. I.e., you still need to fix the gamma look-up table.


Anyway, if nothing else, it should be an interesting experiment.


Kevin
 

·
Registered
Joined
·
707 Posts
Noah,


Your statement is correct that the lamps output is constant (after approxiamtely 50 hours of use) whatever its spectral characteristics are. However keep in mind that when performing a White Balance with the software you are capable of only adjusting the Black and White (2) points. By aligning these two points to a given color temperature (D65) you are depending on the gamma of the display to keep the Grayscale accurate through out the range. The projectors gamma is not linear based on the spectral characteristics of the lamp when it has decayed to its stable state hence the need for the calibration.


Kevin is correct that the contrast ratio measurements were made with the White segment turned off so that we could eliminate the crushed White problem associated with that feature. This feature is simply a means to increase the peak brightness measurement (generally designed for data/graphics presentations in mind rather than video) but at the expense of color accuracy.


I seriously doubt that the contrast ratio can be increased by virtue of using a filter. The filter will reduce the light output at peak White by the same percentage as at Black thereby maintaining the ratio and not improving it.


Cliff
 

·
Registered
Joined
·
4,525 Posts
Kem,


Sorry. I mistated above. I haven't taken Avia, Video Essentials or my laptop to the projector yet. I have done the "normal procedures" of switching to Natural 1, running at native resolution from an HTPC and turning white segment off.


I haven't touch any of the color controls. I also only have about 30 hours on the unit, so I haven't seen any changes in the spectral output. I will continue to measure it and track changes over time to see what happens.
 

·
Registered
Joined
·
873 Posts
Do I recall correctly that the Thumper mod would increase the contrast to 1000+?


Who will be the first to have a Thumperized and calibrated LT150? To have measurements before and after each step would be really valuable.


I volunteer, but Kevin seems to be halfway there already!


Am I crazy to think it would be better to see measurements of the Thumper mods first, then the calibration?


Kelly
 

·
Registered
Joined
·
2,091 Posts
I received an email when inquiring about calibration and Cliff recommended having the Thumper mods done first since Thumper's modifications affect the black level.
 

·
Premium Member
Joined
·
25,495 Posts
Cliff,


"I seriously doubt that the contrast ratio can be increased by virtue of using a filter. The filter will reduce the light output at peak White by the same percentage as at Black thereby maintaining the ratio and not improving it."


It would improve relative to the reduced contrast from calibration, which does *not* lower the black level. I guess a better way to say it is that the filter method would lessen the severity of loss in contrast that the calibration method incurs.
 

·
Registered
Joined
·
707 Posts
Milori,


"I haven't touch any of the color controls. I also only have about 30 hours on the unit, so I haven't seen any changes in the spectral output. I will continue to measure it and track changes over time to see what happens."


The spectral output of the lamp will change after it has approximately 50-60 hours of use. The measurement data provided on Kevin Martin's projector was taken with 70+ hours on the lamp. The contrast ratio was derived with the projectors White segment turned off and the projectors contrast and Brightness adjusted to a Pluge pattern and the White Balance adjusted to get close to the D65 target values.


A new lamp (with or without the White Segment turned on) will provide higher light output as well as greater Red content yielding a better contrast ratio than what was listed in our measurements. The data we provided reflects what a typical user would have measured had their display been calibrated without utilizing a revised gamma look-up table with a lamp that has reached its maturity point and become stable spectrally.
 
1 - 18 of 18 Posts
Status
Not open for further replies.
Top