or Connect
AVS › AVS Forum › Display Devices › Digital Hi-End Projectors - $3,000+ USD MSRP › 3D Color Calibration Thread
New Posts  All Forums:Forum Nav:

3D Color Calibration Thread

post #1 of 35
Thread Starter 
Welcome to the 3D projector color calibration thread. This is a continuation of the discussion in the mini-shootout thread and the recent Sharp 30k thread.

We fuss over 2D color down the last dE, yet 3D color usually looks pretty weak with most out of the box settings with the various projectors. The color tint on the glasses doesn't help.

When 3D projectors were first released, there used to be comments like 'The projector manufacturers compensate for the glasses in the 3D modes'. Yeah right.. most aren't even close to correct color. cool.gif


This thread is for discussing the various methods of calibrating the projector to compensate for the color shift of the glasses.
This topic has little exposure in most projector reviews. David Mackenzie @ HDTV UK has been adding 3D calibration sections to his reviews.

http://www.hdtvtest.co.uk/news/sony-vpl-hw50es-review-201212042380.htm?page=Calibration

In this review, he is putting the glasses in front of their Klein K-10 colorimeter with the following results:

3d-pre-rgb.png

This is consistent with what I've seen on the HW50, Red is usually deficient. We have to pull back fair amounts of green and blue to compensate and it does cost lumen output.

brightness vs. near D65 color with the glasses on. pick your poison.

3d-post-rgb.png


boy that's a lot of glasses to calibrate. biggrin.gif

jvc-splitter1.jpg


DLP-glasses.jpg
Edited by zombie10k - 4/23/13 at 7:53pm
post #2 of 35
Thread Starter 
reserved
post #3 of 35
Hey Zombie, thanks for the info from the other thread about being able to put in a 3d movie and then just overlaying the patterns from the Mini. I was not aware we could do that. I still have not got around to pulling out the meter and calibrating my 7000 behind the glasses, but will get to it at some point soon here.
post #4 of 35
my rsvp for me wink.gif

* Fell asleep while the pj was warming up and was too tired to take measures when I woke up. redface.gif



Jason
- - - - -
Edited by DaGamePimp - 4/24/13 at 2:13pm
post #5 of 35
Thread Starter 
Quote:
Originally Posted by Toe View Post

Hey Zombie, thanks for the info from the other thread about being able to put in a 3d movie and then just overlaying the patterns from the Mini. I was not aware we could do that. I still have not got around to pulling out the meter and calibrating my 7000 behind the glasses, but will get to it at some point soon here.

One day I was using the custom 3D ISO calibration disk with a 100 IRE pattern and then went to measure the color with Chromapure. I forgot I had the mini-3D hooked up, and noticed it was the mini 3D overlay, yet the projector was still in 3D mode. I did some A/B measurements and it was identical. So the mini 3D can definitely be used for the 3D color cal.

hopefully Kris will provide details on his method so we can give it a try as well. I believe David Mackenzie is on this forum, I'd like him to join in and discuss his techniques as well.

our UK friends JonStatt and Manni01's input would be much appreciated too. just like the old days, talking hardcore color calibration, just in 3D this time. cool.gif
post #6 of 35
Some other ideas...

My Sony S780 does 2D to 3D conversion and an Nvidia/3Dtv Play equipped PC can enable 3D while showing 2D as well (of course this requires the LUT to be verified or pre corrected via calibrated ICC profile).

Jason
post #7 of 35
Quote:
Originally Posted by zombie10k View Post

One day I was using the custom 3D ISO calibration disk with a 100 IRE pattern and then went to measure the color with Chromapure. I forgot I had the mini-3D hooked up, and noticed it was the mini 3D overlay, yet the projector was still in 3D mode. I did some A/B measurements and it was identical. So the mini 3D can definitely be used for the 3D color cal.

hopefully Kris will provide details on his method so we can give it a try as well. I believe David Mackenzie is on this forum, I'd like him to join in and discuss his techniques as well.

our UK friends JonStatt and Manni01's input would be much appreciated too. just like the old days, talking hardcore color calibration, just in 3D this time. cool.gif

Hell yeah! This will be a great thread! biggrin.gif

One question I have for both you guys is what is the best method for getting the glasses in front of the meter? Are you strapping it to the tripod somehow so it is positioned in front of the meter?
post #8 of 35
Thread Starter 
Don't laugh.. it works... smile.gif I just use painters tape since it's hardly sticky but enough to hold the glasses in place. it doesn't leave any residue on the glasses.

3D-color.jpg

3D-color1.jpg

How is that for a 3D projector show-down. The battle of the only 2 3D DLP's that work on the HP screen.

3D-color2.jpg
post #9 of 35
I use an industrial grade Velcro strap to hold the glasses.

Jason
post #10 of 35
Perfect! Thanks guys.

Zombie, you get some major style points for the painters tape. I love it! biggrin.gif
post #11 of 35
This post is dedicated to those of you who don't believe that the only way to calibrate for 3D is with the shitter glasses ON. Despite the fact that those are the conditions that everybody watches 3D in.

First I wanted to comment on 2 points made by some of the guys in the Sharp XV-Z30000 thread.

"I knew it couldn't be 80% the image looked too bright to be 1/5th the brightness!"
Look at what you do. You take absolute scale (lumens/lux) and compare it in % to your perception. Our perception works in logarithmic scale! 80% in absolute scale isn't equal to 80% in logarithmic scale! For example take loudness. Decibel is a relative (logarithmic) scale. The boost in 3 dB roughly translates to doubling the loudness compared to previous level. You cannot take an absolute scale which lumens/lux is and make an assumption that 1000 lumens is two times brighter than 500 lumens in your eyes. We may further discuss this topic here.

"With glasses on the meter is most likely averaging a lot of black frames (shutters closed) which would (in my opinion) hurt the results."
Doing otherwise will hurt the result. Basically you want to disregard time-domain. You need to take all those black frames into account. Do you know the term 'exposure'? Every shade of gray of different IRE you see with or without the glasses is the result of exposure, in other words an amount of luminance of a scene in a given time. 3D shutter glasses are liquid crystal based and operate in the 96-144 Hz region. Meters are susceptible to exposure and operate just like our eyes in this regard. If this weren't the case meters would be unable to take measurements from CRT monitors which use narrow electronic beam of very high light intensity to scan raster to create an image (Look up for videos on YouTube of a CRT monitor in slow motion). My CRT monitor is able to refresh at 160 Hz per second and I'm still able to get constant readings from my meter.

I think it is a well-known fact that shutter glasses operate by alternatively closing left and right eyes to present the picture with the left and right perspective at one time for each eye. While one eye sees the picture the other sees a blacked out shutter glass. This happens 96-144 times per second. In theory we can say that you will see black frame 50% of the time regardless of refresh rate. There are different techniques used by different technologies (DLP/IR/RF) and manufacturers to vary the time that each shutter glass remains in the on state (blacking out the picture). But I have yet to see ones that can preserve more than 60% of brightness for each eye. The brightness loss due to work of shutter glasses alone is no less than 40%. Do an experiment. Walk outside on a blindingly bright day. You're squinting, right? You're trying to reduce the amount of light reaching your brain by closing down eyelids. Now close one eye. Will your other eye still be squinting? No, it will be open to equalize the amount of light coming to your brain!

If this is not enough for you to believe that it is necessary to take readings behind the glasses when they are shuttering, since you watch a 3D image in such conditions, then I give up.

Shutter glasses 3D is an early technology and I believe some of its limitations (brightness) cannot be fully overcome.

@Zombie I think "Shutter Glasses 3D Calibration Thread" or "Active Glasses 3D Calibration Thread" is better suited to this topic.
Edited by Elix - 4/24/13 at 1:36am
post #12 of 35
Very interesting points Elix, thanks! I will do some comparison stuff this weekend or next week, don't have a lot of time this week for testing.
post #13 of 35
I must say that I have calibrated my RS40 in 3D with glasses in front of the meter. The results were nothing short of astounding. There is a scene in Ultimate Wave Tahiti where everyone is sitting around a campfire at night, and the color was so dead on that it looked just like real life, I literally could not tell the difference. Outside of that one scene, things may not have always looked perfect, but it was a huge improvement.
post #14 of 35
Thread Starter 
Quote:
Originally Posted by 3dmaven View Post

I must say that I have calibrated my RS40 in 3D with glasses in front of the meter. The results were nothing short of astounding. There is a scene in Ultimate Wave Tahiti where everyone is sitting around a campfire at night, and the color was so dead on that it looked just like real life, I literally could not tell the difference. Outside of that one scene, things may not have always looked perfect, but it was a huge improvement.

Thanks for your feedback. What meter / software are you using?
post #15 of 35
Thread Starter 
sharp5.jpg



Sharp 30k, through the glasses measurements with the glasses turned on / off.

** I'm running in 3D frame packed mode with a 100% IRE screen for these results. **


1st measurement, glasses are turned on, factory setting in 3D Stage mode.

newsharp-glasses-on-precal.jpg

2nd measurement, glasses are turrned on, negative gain on blue and green to compensate for the loss of red.

newsharp-glasses-on.jpg

3rd measure, glasses are turned OFF, same calibration as above. There is a relatively minor dE between ON / OFF.

newsharp-glasses-off.jpg
post #16 of 35
Interesting. Almost be fun to do a full calibration thru the glasses (either on or off) and then just run the full measurements again in the opposite condition and see what the overall difference would be. If the delta is low enough, I would probably stick to calibrating with the glasses off to preserve as much light as possible for the measurements. With that much light loss (glasses on) you have to wonder if the meter sensitivity in low light would cause more issues than its worth.
post #17 of 35
I did a new 3D Cal last night and experimented with glasses/meters/position since I am using DLP Link and the glasses do not like to stay sync'd when closer to the screen.

I used BenQ D3 glasses with a w1070 and ended up using my i1 Display Pro instead of the i1 Pro spectro (the Display pro was profiled).

I took reads with the glasses on and off and the difference was honestly fairly minimal, there was more of a dE shift going to different glasses.

I used the 2D to 3D conversion on my Sony S780 blu ray player (set to low) with the AVS HD 709 blu ray and had excellent post Cal results...

6528K

Highest CMS dE was 6.9 on Blue which could not be corrected by built in user menu controls (I did not try the ISF controls since it is said they do not save on firmware 1.04 and of course I have 1.04).

All other colors were under 3 dE.

Greyscale is excellent after going back and forth several times initially and then a few more times once the CMS had been done.



Jason
post #18 of 35
Quote:
Originally Posted by zombie10k View Post

Thanks for your feedback. What meter / software are you using?

I use Calman 5 with an i1Display (the newish meter) profiled with a Colormunki spectrometer. I put the 3D glasses in front of my spectro and used that to profile my i1Display (without glasses on it), this resulted in a high offset, but the results were great.
post #19 of 35
Hi,

I read in the Spectracal forum the next recomendation for 3D Calibration:
Quote:

The recommended way is to either profile your colorimeter without the glasses to your spectro through the glasses, and then use the colorimeter without the glasses for the actual calibration, or if you don't have a spectro/meter combo to profile your meter without the glasses (field) to your meter with the glasses (reference), so that again you can do the calibration without the glasses to avoid interference from the shutters, especially when reading low light patterns like dark blue or the low end of the greyscale.

Method for the profile:

- With Spectro/colorimeter combo (ex. i1 Pro/i1 Dipsplay 3)

1. Read 100 IRE WRGB with the i1 pro spectro (= reference) WITH glasses
2. Read 100 IRE WRGB with the i1 display 3 or C6 (= target) WITHOUT glasses
3. Calibrate WTHOUT glasses using the i1 display 3 or C6

- With only colorimeter or spectro

1. Read 100 IRE WRGB with the spectro (= reference) WITH glasses
2. Read 100 IRE WRGB with the spectro (= target) WITHOUT glasses
3. Calibrate WTHOUT glasses using spectro

Best Regards
post #20 of 35
So Spectracal has the same worry I have, simply not enough light to get good readings when going thru the glasses. This is why I mentioned that not turning on the glasses may be better in the long run for calibration purposes and the lesser of two evils. I would also suggest that if you are going to calibrate thru the glasses that you zoom the image to the smallest size possible to give yourself the most light to work with. You could also just point the meter at the lens and forgo whatever shift you may get from the screen (I still prefer to measure off the screen)
post #21 of 35
I used the Spectracal method initially but then discovered I had enough light to take the reads through the glasses and went with it, 3D image now looks almost as good as 2D.

Jason
post #22 of 35
Quote:
Originally Posted by Kris Deering View Post

So Spectracal has the same worry I have, simply not enough light to get good readings when going thru the glasses. This is why I mentioned that not turning on the glasses may be better in the long run for calibration purposes and the lesser of two evils. I would also suggest that if you are going to calibrate thru the glasses that you zoom the image to the smallest size possible to give yourself the most light to work with. You could also just point the meter at the lens and forgo whatever shift you may get from the screen (I still prefer to measure off the screen)

How about just calibrate for the projector/glasses, and not the screen. With a white diffuser on the sensor and the glasses attached to the front of the sensor/diffuser and then with the sensor positioned just a few feet (e.g., 3 ft.) in front of the projector with the sensor pointed toward the projector, you could certainly get plently of light into the sensor. Of course this would not calibrate for any color shift caused by the screen, but for 3D this may a reasonably good trade-off. Also some active 3D glasses may not sync up in this position.




.
Edited by Ron Jones - 4/26/13 at 5:18pm
post #23 of 35
Thread Starter 
Quote:
Originally Posted by DaGamePimp View Post

I used the Spectracal method initially but then discovered I had enough light to take the reads through the glasses and went with it, 3D image now looks almost as good as 2D.

Jason

With HP screen, the Display 3 is getting enough light with the glasses on and off. The bold above is the reason I started the thread. With several techniques available, folks with calibration equipment can significantly improve the color in 3D as seen through the shutter glasses.

thanks for everyone's input so far.
post #24 of 35
I calibrate through the active glasses directly from the lens. My setup is a JVC X35, Display 3 pro, lumagen mini and Chromapure with 125p autocalibrate. Doing so allows me to get enough light for lower greyscale and black readings, which makes it possible to run the BT.1886 gamma. BT.1886 is relying on an accurate black level reading, and since it's a JVC, this is really low. Besides having the High Contrast High Power screen I see little benefit of measuring from the screen anyways. Slight changes in angle significantly affects the measured results making screen readings rather moot. Still looks good by eye though no matter where I'm sitting in the sofa.

The end result is a night-and-day difference and even if it doesn´t measure 100% perfect it certainly looks stunning!





post #25 of 35
Quote:
Originally Posted by Drexler View Post

. Slight changes in angle significantly affects the measured results making screen readings rather moot. Still looks good by eye though no matter where I'm sitting in the sofa.

I agree, with the HP screen (at least with my meter), it might be better to measure from the lens, the angle screws with the RED grey scale readings too much because the meters have trouble with the way the light reads from the HP. There is another way to do it where you can actually create an offset for the HP and still calibrate from the screen but taking the angle error into account, but it's complicated and requires too many measurements (not that complex I guess, I'm just too busy doing other things). That said, you can probably get decent good doing it from the screen (but it's somewhat of a gamble). Calibration is a life or death endeavor (only kidding).
Edited by coderguy - 4/27/13 at 5:49am
post #26 of 35
Thread Starter 
Quote:
Originally Posted by Drexler View Post

I calibrate through the active glasses directly from the lens. My setup is a JVC X35, Display 3 pro, lumagen mini and Chromapure with 125p autocalibrate. Doing so allows me to get enough light for lower greyscale and black readings, which makes it possible to run the BT.1886 gamma. BT.1886 is relying on an accurate black level reading, and since it's a JVC, this is really low. Besides having the High Contrast High Power screen I see little benefit of measuring from the screen anyways. Slight changes in angle significantly affects the measured results making screen readings rather moot. Still looks good by eye though no matter where I'm sitting in the sofa.

The end result is a night-and-day difference and even if it doesn´t measure 100% perfect it certainly looks stunning!

Hi, how close to the lens are you placing the meter? Are the glasses on or off?
post #27 of 35
Quote:
Originally Posted by zombie10k View Post

Hi, how close to the lens are you placing the meter? Are the glasses on or off?

About 3 ft or so. The glasses are on.
post #28 of 35
The HP does color shift so if you are not reading off the screen you will not be as accurate as you could be.

I owned the 2.8 HP for years and calibrated with/without it many times and could always see the shift if I did not account for the screen.

Jason
post #29 of 35
I believe this is the opposite of what THuff recommends, he says calibrate from the lens with the HP.
I don't know, I'll leave it up to the people in here.

The best way to do it though, is probably to calibrate from the screen first by hanging a slip of neutral screen material over the HP, then measuring from the lens, measuring the HP, and measuring the neutral material, and then create the offset based on those three, but it's a lot of work.
post #30 of 35
Hi guys
i would like to do 3D calibration for my PJ but have problem , my Oppo 93 and Optoma hd 83 have no conversion 2d-3d is there some posibility to get HCRF pattern ore any other pattern disc in 3d , it can be sbs for example ? any help be apriciate
New Posts  All Forums:Forum Nav:
  Return Home
AVS › AVS Forum › Display Devices › Digital Hi-End Projectors - $3,000+ USD MSRP › 3D Color Calibration Thread