3D Calibration info from Mike Wood (Samsung)

[Gregg Loewen]

Calibrating Samsung 3D TVs
At least for Samsung 3D TVs, calibration is somewhat easy to accomplish in the 3D mode.

We have a 2D to 3D upconversion mode that will activate the 3D setting (which in turn will activate the glasses) and will upconvert your regular 2D test patterns to 3D. No, the test patterns will not have much depth effect, but they will go through the 3D processing channel so that you can calibrate the full signal path, and it will put the TVs picture settings into the 3D setting memory. For what it’s worth, the 2D to 3D function works surprisingly well with high-res, high-contrast content. You might want to be sure you have a copy of the movie “Baraka” on hand to demo for your customers when you’re done.

As set from the factory, the TV compensates for the tint of the glasses. Any further calibration should be done with a meter pointed through one lens of the glasses. Most meters will compensate for the refresh rate. I don’t recommend using contact (tristimulus) meters. If you have no choice, you might be better off using a visual gray reference for 3D mode. Calibrate for 2D, then activate 3D, wear the glasses and compare the reference to the on-screen image.

General calibration tips:
While it’s not readily evident (the user menu looks the same) the TV does have a separate memory for many of the major 3D functions, like black and white levels and white balance, so you can calibrate for those separately. I don’t know if there’s a separate CMS for 3D. I’m working on updated calibration notes and hope to have them posted for you soon. For now, see my 2009 notes, as 95% of the functions are the same.

I’m told that all the TVs should have the Cal-Day and Cal-Night functions. You would activate them (with the TV in Standby) by pressing Mute-182-Power to turn on the TV in the service mode. Then go to Expert and turn N/D ADJ to “On”. Cycle the TV power to get back to the user mode and Cal-Day and Cal-Night will appear under Picture Modes. When you’re done calibrating the Day/Night modes, you can go back to the service menu and can “fix” or lock those settings so that the user can’t adjust them.

Always calibrate HDMI signals first.

In some cases, the Mute-182-Power command won’t bring up the Expert menu. If there’s no firmware update for that TV that makes the Expert setting available, then you might have to calibrate the TV using the user-menu’s Movie mode.

As I’ve said before, when calibrating the white balance, start with Gain first, then use Offset. The Warm 2 Color Tone should be closest to D65, so use that as a starting point as well. If your meter tells you that Warm 2 is more than a few hundred Kelvin from 6500, and if there isn’t enough range in the TV controls to make it correct, then your meter is likely incorrect. In these sets, use Offset sparingly, as only mild increases in red, for example, can make deep black turn reddish.

We now also have 10-point White Balance, if you really want to spend all day tweaking a TV. I recommend using the 2-pt WB first, to get the image close to D65, then use the 10-pt WB to dial it in. Unlike with the 2-pt WB, I recommend starting the 10-pt adjustment from the bottom (Interval 1 or 2) and work your way up. Each Interval seems to correspond with a 10x IRE setting, so Interval 2=20 IRE, 3=30 IRE, and so forth…

You can adjust the 3D mode next. It’s tough to make the gray scale super flat in 3D mode. Again, adjust the gain first, possibly with a 70 IRE signal, and look at where the low end goes on a full ramp (or something with good shadow detail). Use the Offset sparingly, if at all. You might have to adjust the top end a bit to get a better compromise with the low end.

Topic

[Anonymous]

Thanks Gregg, that’s helpful information as I ready for 3-D calibrations.

I’ll echo Mike Wood – use R-offset sparingly…one click too far changes the 10% and below red… do everything right on a Samsung LED and you can get 10% dE of 2 with the rest staying intact (ok, maybe just a bit of blue around 40-50%) 🙂 Better to have blue error than red!

[Anonymous]

Gregg, (or anyone who has done one already)

Any tips on keeping the meter and glasses together as you perform the calibration? Also, how close do you want your meter/glasses combo to the panel? I would think as close as possible, but just want to make sure. Thanks,

[Gregg Loewen]

Hi Chris
Yes you will need to the filters / glasses in the measurement chain.
this is still new / evolving and I am not sure of all the details.
I just got off a phone meeting with Quantum Data and we will be meeting again next week while in Vegas. Lots of 3D patterns need to be added to the 780.

[Anonymous]

Thanks Gregg. Which meter model is good for 3D calibration? How about i1 Pro and OTC1000?

[Anonymous]

@Gregg Loewen 1899 wrote:

Hi Chris
Yes you will need to the filters / glasses in the measurement chain.
this is still new / evolving and I am not sure of all the details.
I just got off a phone meeting with Quantum Data and we will be meeting again next week while in Vegas. Lots of 3D patterns need to be added to the 780.

Thinking about this, the new Quantam Data 708 PDF says it has the 3D Graphics available to it now:

http://www.quantumdata.com/pdf/780_DS.pdf

The question being, as the unit only has HDMI 1.3 and not 1.4, how can it signal a TV to go into 3D mode? :confused:

[Anonymous]

@Gregg Loewen 1890 wrote:

Calibrating Samsung 3D TVs
Any further calibration should be done with a meter pointed through one lens of the glasses. Most meters will compensate for the refresh rate.

Seems like a lot of room for error aiming the meter through glasses unless one is able to look through a viewfinder on the meter. Aim the meter first for 2-D, then put the glasses in front… But I’m just thinking about hanging the glasses over the front of the meter. That won’t work on either of my meters. I’d have to physically hold the glasses for each measurement and make sure my meter’s lens is in focus each time. Any wobbling of my hand might screw a reading up?? My hand has never been super-steady. Should be interesting to try…

…any recomendations as to how to mount the glasses? Maybe I need to custom make a bracket for one distance and no motion all of the time…?

[Anonymous]

@Mike Osadciw 1992 wrote:

Seems like a lot of room for error aiming the meter through glasses unless one is able to look through a viewfinder on the meter. Aim the meter first for 2-D, then put the glasses in front… But I’m just thinking about hanging the glasses over the front of the meter. That won’t work on either of my meters. I’d have to physically hold the glasses for each measurement and make sure my meter’s lens is in focus each time. Any wobbling of my hand might screw a reading up?? My hand has never been super-steady. Should be interesting to try…

…any recomendations as to how to mount the glasses? Maybe I need to custom make a bracket for one distance and no motion all of the time…?

Here’s a question that is bugging me in that regard – according to some reports – the 3D field collapses unless you are directly in front of the TV – and the glasses MUST be @ 90 degrees to the set – in other words you cannot be laying down or twisting your head and using 3D glasses to get a 3D effect. This *MIGHT* only be true with certain screens/glasses combinations – the subject is so new, its hard to make overall generalized statements.

But if it is true that certain combinations need very precise angles between the screen and the glasses – it seems that the glasses need to in exactly the same configuration during calibration as they would be for proper 3D viewing – otherwise, you will run into issues.

So how could you even begin to mount a pair of glasses correctly on a meter?
:confused:

I’ll be the first to admit I am speculating on this – as this subject is so new -but if you cannot use 3D glasses except in very exact positions/angles – why would using them during calibration be any different?

Mike Wood @ Samsung has been a Godsend for getting info to calibrators – I hope that he will give more info very soon on these issues.

[Gregg Loewen]

Below is some new information from Mike Wood.

· The meter does need to measure the TV through the glasses (only for 3D). Contact meters (aka “hockey puck”, or tri-stimulus meters) will not work. Then again, they’re not that accurate anyway (hate me for being the messenger if you want, but you should really hate the company that sold you on it). You need a spot meter. The OTC100 might be fine. PR650s or higher, and KM CS-200/2000s are good. Otherwise, just accept the factory-calibration for 3D as “close enough”.

· Cloth gaffing tape seems to hold my glasses on to my PR-650 just fine. How they are held to the meter doesn’t really matter, other than how it might leave a mark on the meter. The glasses should be horizontal and parallel to the screen.

· The meter should be far enough back that the meter can take a large-enough sample of the screen, and far enough back that the glasses are activated by the emitter in the TV. The PR-650 has a dark spot in the viewfinder. The meter only measures the area behind the dark spot, not the entire field of view of the viewfinder.

· One could measure the delta of the measurement through the glasses versus the measurement from the TV directly, and calibrate the TV directly, but it would be a good idea to double check the results when you’re done. In other words, if the TV measures 5500K without glasses (in 3D mode), and 6000K with glasses (in 3D mode), then you could calibrate the TV to 6000K, and be reasonably certain that the result will be close to 6500K when seen through the glasses. The delta may NOT be 500K; I pulled that number out of thin air for demonstration purposes, and it may vary from set to set, so check it, and go from there. Confirm that both the high and low levels have the same offset, as well.

[Gregg Loewen]

Also…
Quantum Data has just released a 3D update for their 780 generator. the generator has also just had a price decrease!!
http://www.thxvideotech.com and then surf to the store.

[Anonymous]

thanks for the update!

On the Quantam Data 780, I guess a question that will need to come up eventually is what 3D format does this output(s)?

Checkerboard?
Side by Side?
Top/Bottom?
Frame Packing?

I would *assume* one would need the ability for all formats to be totally prepared and check all possibilities? :confused:

FWIW, from Mitsubishi’s site:

A12: There are several different formats for 3D signals possible; however the HDMI 1.4a specifications defined some types as mandatory for transfer over the HDMI cable. Most signals will be one of these types of mandatory signals. These fall into 3 general categories:

1.Frame Packing — a very high resolution 3D signal that requires a lot of bandwidth — this is not suitable for most broadcasting over Cable TV systems or Satellite systems. It is very suitable for Blu-ray disc and gaming systems. For the USA, there is 1080p at 24 frames per second — sent by Blu-ray players and 720p at 60 frames per second expected to be used by future gaming consoles.

2.Side-By-side — this is where a normal video frame actually contains two frames (one for each eye) and if not processed as a 3D signal would look like two pictures side by side on the screen of the TV. In the USA most of these signals will be 1080p 24 or 30 frames per second, 1080i 60 fields per second, or 720p 60 frames per second. These signals are suitable for broadcast applications.

3.Top-Bottom — this is similar to Side-by-Side where the normal video frame actually contains two frames except the normal frame is divided in haft top to bottom. Again the in the USA most of these will be 1080p 24, or 30 frames per second, 1080i 60 fields per second or 720p 60 frames per second. These types of signal are also suitable for broadcast.

The Checkerboard format is really designed as method of displaying 3D on the TV and not expected to be used to send 3D signals to your home. The Mitsubishi 3D TVs display 3D using the checkerboard format so the Mitsubishi 3D Adapter is designed to convert the Frame packing, side-by-side and top-bottom signals mentioned above to the checkerboard format.

http://www.mitsubishi-tv.com/3D.html

[Anonymous]

It’s the long weekend for us in Canada…so I MAY get out to a store and see what it’s like to measure with glasses in front (not to calibrate ’cause I don’t have a 3-D pattern). My Minolta CS-1000A has the “dark spot” as well, so it’s easy to focus on a spot…but I’d like to see how I can focus the Minolta’s lens through the glasses, while still being able to see the 3-D image on the TV perfectly focussed through the lens. Just thinking about it in my head, I’m wondering if the Minolta will clearly focus on the lens of the glasses, but not the image through the glasses…?

Also…this is my food for thought… creating 3-D “content” vs. reasons we calibrate…

We calibrate because we adjust TVs to the standard that the production community follows when they create content and make new colour choices, etc etc etc… on TVs that are supposed to be calibrated. Now…my view is that the 3-D content we have right now is still created on 2-D monitors following our standards…

…that means what content creators choose as being red, green, blue, and all colours mixed to suit whatever mood they are aiming for, are probably chosen in 2-D.

I doubt the production community is wearing 3-D glasses all the way through creation to final film/video-to-video transfers.

So…are they making all of their original colour/level choices through 3-D glasses, on 3-D monitors, that have been calibrated through a 3-D pair of glasses…? Probably not.

Then…if everything – all decisions in post is done in 2-D…there is no compensation on the “finished product” for colour shifts (when glasses are put on) in the source material. (Did I read somewhere that the brightness of the content might be the only difference…so when viewed in 2-D, the 3-D version of the film might be a bit brighter on the source?)

I doubt the movie on the BD will be “color corrected” so that when viewed through 3-D glasses on a TV that has been properly calibrated through a pair of glasses, it will look like our 2-D HDTV system.

My guess is that content is created. Glasses are put on. Whatever happens, happens!

So, that begs the question…are we incorrectly calibrating a TV through the glasses and altering the intended image??

[Anonymous]

I have just been challenged by a potential client concerning a statement I made to him that there is as yet no standard for home reproduction of 3D as there is for 2D. He pointed to the THX website which shows that there are THX certified 3D TVs with THX 3D modes in them. Am I wrong? And if so what is the standard that we are calibrating towards now? Please help as I need to email the guy and explain myself! Thanks.

[Anonymous]

The same standards from 2D apply to 3D. User controls to be set right and then do grayscale and cms where applicable. Just through the glasses. Sure there are two types of 3D systems in terms of TV types but for our purposes, that is not too relevant. The approach is the same aside from glasses mounted on the passive side must be level as the user wears them or else the meter does not see the correct image.

nonstandard stuff is like how the sony glasses don’t work with the panasonic sets or samsung not working on sharp and so forth…

calibration is still the same but with all the limitations i mentioned in class about 3D

regards

[Anonymous]

Thanks Michael. So when THX are including a 3D mode in certified TVs does this mean that it will measure close to a THX 2D mode? Just trying to get my head around what THX are doing here.

[Anonymous]

FYI, from the THX website: THX 3D Cinema: Sets luminance, color, blacks, gamma and video processing for 3D broadcasts and Blu-ray Discs.

[Author]

Replies