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

· Registered
Joined
·
1,160 Posts
Discussion Starter · #1 ·
Recently I had William Phelps over, and he has developed a special calibration program that allows finer tuning for the D-ILA gamma curve.


I previously had my G15 calibrated by Richard Martin and the picture looked great, which can be confirmed by the many visitors who have come by to see my HT setup.


So I figured that William's software would simply show how well a job Richard did with his calibration, and indeed it discorvered that Richard had gotten the shading dead-on.

But William's Gamma program has the added ability to plot a full 1024 points for red, green, and blue for much finer gamma mapping.


Of course this would have been just another techie academic exercise, except for the fact that it provided a visible color quality improvement.


First off, I had previously set the brightness to -7 to get the black level right, but now 0 is best.

More importantly, the color is more saturated and richer, which can be seen both in brighter colors, and with fuller fleshtones.

But it's not like I just cranked the color up, because it looks more realistic, and not overblown.


It brings to mind the captured pictures of the 3 chip DLP vs. LCD thread that was here a while back, with my colors now looking similarly better. Both with the HTPC and HD material.


Anyway, I'm always surprised at improvements like this calibration and changing to the Radeon card in my HTPC because the picture already looked fantastic, and I wouldn't have guessed that there could be a visible improvement made. Well, it's not a day-and-night difference, but it is visibly better than before.


And yet, the quest for improvement still continues.

Next step, the new 2048 X 1536, 7K lumen JVC FPTV, and HD-DVD drive. http://www.avsforum.com/ubb/wink.gif

-Dean.
 

· Registered
Joined
·
760 Posts
Where Dean leads, I always make sure I follow along the way http://www.avsforum.com/ubb/biggrin.gif


This is great news! I am one of the lucky ones to've had his pj calibrated by R. Martin, so to hear this is better...wow!


So, is William taking appointments yet? I wonder what the turnaround time will be, considering it should be pretty fast as it's done via software.


Dean, can you tell us how much time it took and what were the steps involved?


Thanks,


Luca

 

· Registered
Joined
·
547 Posts
Dean, or william;


I hear alot about this software by william. is this something new that's competing with dilard? are you able to use this software yourself to get the gamma corrected? is william going to sell this software to dila users? looks interesting, I hope mark hunter's up comming gamma software upgrade to dilard will be as powerfull as this! http://www.avsforum.com/ubb/smile.gif


seng
 

· Registered
Joined
·
1,160 Posts
Discussion Starter · #4 ·
Luca,


William does his calibrations in-person. There are many stages in the calibration process which took a few hours. His software tools were designed for his use along with other interactive calibration measurement tools.


I don't know exactly whether he plans on selling the software for other calibrators to use, or as a stand-alone product in the future.


Since my shading was fine, he moved onto setting the 6500K temperature for red, green, and blue with the CA-1. Once these were set to nearly dead-on, he ran the gamma software and it ran many iterations as it smoothed out the variances off of optimal to create a gamma curve for all of the colors relative to the color temperature.


Then he did some more calibrations relative to different input resolutions (HTPC, HDTV).


I'm sure that some of the improvement was due to making the gamma calibration measuring off of my new Grayhawk screen, and mapping to optimize the picture to that new screen material.


-Dean.
 

· Registered
Joined
·
4,525 Posts
Quote:
is this something new that's competing with dilard? are you able to use this software yourself to get the gamma corrected? is william going to sell this software to dila users? looks interesting, I hope mark hunter's up comming gamma software upgrade to dilard will be as powerfull as this!
Hi Seng,


Good questions! I have heard good things about William's software, and it sounds like it does an excellent job. I have communicated with him about integrating this code into Dilard, but I don't think that there was sufficient interest to move forward. We decided that it would be best he would keep his software for doing calibrations, and Dilard would be designed for commercial use.


The software applications will probably be very similar. There is only one Gamma curve in the D-ILA projectors. It always has 1024 points, regardless of whether or not JVC's software uses all of the points. It can be adjusted automatically (or semi-automatically) with a light meter and a computer. Dilard will support multiple light meters, which is something that William's software does not need to do.


I am most excited about a recent shipment that we received with an inexpensive sensor that is about 10 times more accurate than a CA-1 at low light levels. It is my goal to have a solution that allow you to do a calibration at home with this meter that is equal to or better than the results that William's software can achieve.


By the way, having multiple options like this is just excellent for the consumer, and hopefully soon there will be calibration options that meets the needs of everyone interested in the very best video quality...whether they are do-it-yourselfers or would rather have the service performed for them.
 

· Registered
Joined
·
835 Posts
Wow, that's kind of exciting...


I seem to remember Mark saying at one time that Dilard's gamma adjustment was superior to the normal one because you could manipulate all 1024 values in the gamma curve individually where you had 8 anchor points in the regular adjustment.


The exciting part is William Phelps automating the process, I think it might even be quicker than adjusting the 8 anchor points. Because you're directly manipulating each of the values in the gamma look-up table to match 6400K and a gamma of 2.2, you don't have to run the process over and over again to find the right values for the 8 anchor point so that it will give you the most accurate values for your grayscale.



Regards,


Kam Fung
 

· Registered
Joined
·
2,290 Posts
I got into this because someone else was trying to get the "Richard Martin" software working and not having much success. Since I am an experienced programmer (just completing 30 years with IBM on Monday), and also an experienced videophile, it was a fun challenge to get the software working, but once that was done, it was clear that more could be done. With the process as I received it, there was very bad "posterization" in low level signals, requiring hand adjustments of the gamma curve. It took considerable effort to discover and solve the cause of this - I wanted as automated solution as possible... Finally, with considerable help from Steve Johnson, the goal was reached.


I would have been happy if my calibration of Dean's G15 had been as good as Richard's, as Dean said his picture was already outstanding. To go beyond that and achieve a significant improvement is very satisfying.


I had seriously hoped to be able to share this with Dilard customers, but discussions along these lines were not fruitful. I'm still not sure what the final form of this package will take. At this time it is available only to myself and Cliff Plavin (Cliff is doing the "AVS" calibrations) - I have invested considerable effort (and cost!) in developing this program and need to recover some of that. However, it is certainly intended for "commercial use", it's just not clear yet what form that will take.


A word about accuracy and measurement instruments - my program currently supports 3 devices, the most accurate of which is the McMahan Lightspex, the most accurate spectroradiometer I have encountered to date for anything less than $20k. Unfortuately it is also rather slow. Once I had the Lightspex working with the program, I investigated using the Progressive Labs CA-1 as an alternate instrument. I found that I could get repeatable, accurate results; the CIE xyY numbers I get from the CA-1 are within 0.001 of what the McMahan reports with both probes placed side by side taking readings from the same projector, at all light levels. This is more than accurate enough, the error is less than the human eye can detect.


It's probably also of interest to the readers of these forums that proper calibration of any projector requires acccurate test signals. Dean's calibration was done with a PC that has an oscilloscope calibrated video card. Under development is support of an external signal generator, allowing "on site" calibrations with a laptop.


There is more to my calibration process than just measuring and adjusting the gamma curve for proper grey scale, much more, both in the software and adjustments that must be done by a skilled person. But in the end, the results speak for themselves...


Wm

 

· Registered
Joined
·
6,227 Posts

· Registered
Joined
·
842 Posts
A couple of corrections.


In the factory software there are 17 anchor points, two of which

are fixed. The last one may or may not be used depending on the

way you set up the peak white level. Richard martin changed the

X coordinates of the 3 bottom points to improve linearity for

the blacks, however this is not really enough.


When the factory software programs the projector it calculates

using straight line segments between the points and then uploads

to the projector a 1024 word table which is what is actually used

as the gamma.


Some of the problems mark and i had in integrating the ca-1 device

result from Kxy calibration files which really need to be generated

on a per device basis. A standardized Kxy file for a dila projector

just will not work on every ca-1 device at low light levels. Plus

the ca-1 drifts with time and temperature with a heavy emphasis on

the later. Making any tri-stimulus sensor work reliably for lcd

projectors is fairly tough.


William has a real meter to calibrate his ca-1 against. He can recalibrate

his ca-1 whenever he wants. I am sure his software is excellent, anyone

with 30 years perserverance at IBM deserves some kind of prize. results

from people who have had william calibrate their projectors are excellent,

time will tell whether the 8 bar adjustments stay put.


Mark and I have received our devices from an alternate vendor. I will

be spending the entire weekend in the calibration room and will be

comparing the new device with a ca-1, lightspex, photoresearch, tek

and an old jarell-ash instrument of absolute accuracy.


As far as calibrating with a video card of known accuracy, this is

certainly important too unless you are only going to be driving the

projector from that video card, in which case all non-linearitys

in the chain are taken into account. A lot of the higher-end video

cards out there use brooktree d/a converters, and these can be

spec'd as low as 7 bits accuracy, and as high as 10.


Ought to be a fun weekend.
 

· Registered
Joined
·
760 Posts
Thanks Dean!


This is such wonderful news! Reading about the process made me eager to want to have mine calibrated. Hopefully, William will get up and going soon so we can start booking him for a visit to our homes.


William, could you tell us at this time what your services will cost?


It is also exciting that Mark is working on his own calibration setup in Dilard. This can only bring good to us HT fans, to have so many talented folks that are so enthusiastic to keep striving to make digital projection the absolute choice for us, by continuously raising the standards to which these systems must adhere to.


Digital technology has come a real long way in such a short lifespan. While already superior in certain aspects to more proven technologies, it has also become obvious that it lacks in certain others with respect to the same ones. It is nice to know that there are folks out there that care enough about quality that they tap into the hidden, powerful resources within the technology to make it a better, more polished offering.


Luca
 

· Registered
Joined
·
1,836 Posts
I can also vouch for the results of this cal. In my case Williams software was used by Cliff (Doing AVS cals). The results were really spectacular. I continually find myself re-watching DVDs - Last night was Gladiator; to see the improvements. Color richness and accuracy in all scenes without fail, Deep blacks without loss of visibility (contrasat), complete lack of posterization. I would recommend any DILA owner or purchaser get this calibration (William or AVS-Cliff). Of course, Dilard is also a must have - whether you use it for cals or not.


Thank You - William, Steve, Cliff, Mark as well as predicessors Mark Foster, Richard Martin. We've come a long way... JVC should grant these guys some sort of lifetime use of the latest DILA - or at least a lifetime supply of Xenon bulbs!(that's a small fortune by itself)


ken
 

· Registered
Joined
·
2,290 Posts
This is not a "correction" - what Richard Martin did has nothing to do with what my calibration does now. We don't use the 17 points at all now.


Wm


Originally posted by kevin gilmore:

A couple of corrections.


In the factory software there are 17 anchor points, two of which

are fixed. The last one may or may not be used depending on the

way you set up the peak white level. Richard martin changed the

X coordinates of the 3 bottom points to improve linearity for

the blacks, however this is not really enough.


When the factory software programs the projector it calculates

using straight line segments between the points and then uploads

to the projector a 1024 word table which is what is actually used

as the gamma.

 

· Registered
Joined
·
842 Posts
I was not implying that williams software uses the 17 points because

it certainly does not. I was correcting the person that said it

was 8 points. The dilard calibration wizard does not use the 17 points

either. This does however point to a problem which means that in the

unlikely event you have to send your projector in for repair, and

for whatever reason the repair facility uses the factory tool to change

bias, black level, peak white, or gamma, your calibration just went

out the window. Same thing with changing any of the boards which

force a recalibration. Better backup your $700 calibration...
 

· Registered
Joined
·
2,290 Posts
Actually, your calibration is already backed up. Both Cliff and I take full backups before and after calibration, and we archive these. Should it ever become necessary, I can provide a copy of the backup and a special purpose utility program to restore it.


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