View Full Version : Calibrating LCD TV's with a Colorimeter


Steffche
10-30-07, 01:28 AM
Hi,

A while back I purchased a Sony Bravia LCD TV and attempted to calibrate it using both a Spyder2 and an EyeOne Display2 probe and HCFR.

The results were very disappointing because it seems that these sensors are being greatly affected by IR light or some sort of light emitting from the TV that is causing very inaccurate readings to be obtained.

For example the HCFR RGB results were indicating a fairly uniform D65 tracking grayscale, however when looking at the gray test patterns on the screen it was VERY obvious that there was way too much red in the picture.

So it appeared as though what the sensor/software thought was right, was infact way out. I used my professionaly calibrated Plasma as comparison in which case the overly red appearance of my Sony LCD was very noticable.

Has anyone done more research and come to any conclusions on this? Im sure there are heaps of people using either of these sensors and trying top calibrate LCD TV's...they must be experiencing the same issues??

Thanks in advance!

Michael TLV
10-30-07, 02:02 AM
Greetings

Correct. Calibrators are warned from time to time not to trust these probes on LCD flat panels.

The error seems to be about 500 to 800K to the red green. So usually shooting for 7200 range gets you closer to where you should be. On the diagonal toward blue.

Regards

scooper750
10-31-07, 01:34 PM
What's the x/y coordinate for 7200?

Michael TLV
10-31-07, 01:37 PM
Greetings

That depends on which 7200 you are trying to hit ... the one on the black body curve or simply in the direction of blue. The crosshair typically represents d6500 ... so move in the diagonal toward blue until it gets to 7200 ish ... and you are closer to where you really should be.

Remember that this is an approximation ... specific coordinates will not help.

Regards

TomHuffman
10-31-07, 03:11 PM
I've used the D2 with LCDs, and it seemed fine to me. Here's one result:

http://www.avsforum.com/avs-vb/showpost.php?p=9812011&postcount=82

I haven't seen the reddish bias you saw.

Steffche
10-31-07, 08:22 PM
Interesting results Tom. Which LCD TV was that just out of curiosity?

I have a spyder2 and the Display2 and have compared the difference between them while calibrating my Sony Bravia LCD TV and the D2 is much better than the S2, but I still found that the grayscale looked quite reddish with either (moreso with the S2 though).

I tried your suggestion again Michael as I think you have given me that advice once before (to shoot for around 7200 instead of 6500), but as a previous poster mentioned, what 7200 should we aim for?

For example, I found that have the following percentages on the RGB bargraph in HCFR while taking continous readings while trying to aim for 6500 "looks" about right...

Red - %92
Green - 101%
Blue - 111%

I emphasise the word "looks"!

The above figures resulted in somewhere around 7200K.

I also tried using a piece of clear glass between the D2 and the LCD surface (carefully so as not to scratch the panel!) to see if that would reduce the amount of IR interference, but it made very little difference. I know its not a proper IR filter, but I thought it might show up something!

Michael TLV
10-31-07, 11:32 PM
Greetings

That's the problem ... if we are not aiming for 313,329 ... then the science part tends to fall to the side. Moving in the diagonal direction toward blue gets you closer to where you need to be.

If one absolutely needs coordinates ... I suggest you that that diagonal route until it reads 7200 and mark down the x,y. Just remember that the 7200 is a ball park figure. Sometimes it is 7100 ... sometimes 7300 ... 7400 ... 7350 ... varies with the display type.

Regards

Steffche
10-31-07, 11:37 PM
Just remember that the 7200 is a ball park figure. Sometimes it is 7100 ... sometimes 7300 ... 7400 ... 7350 ... varies with the display type.

Regards

Can you comment on Sony Bravia's in particular?

Michael TLV
11-01-07, 12:06 AM
Greetings

I know on the current crop of bravia flat panels that I have worked on this calender year, that probes like the spyder stuff have left me in the 5500 to 5900K range too red.

So observing a 600 to 1000K variance ... I split the difference and say 700 in this case.

When the tools fail you ... consider building a small optical comparator to double check the measuring device.

Alan Brown's Ideal Lume back lights work great with a kodak graycard.

Regards

BobL
11-01-07, 06:55 AM
Some of the Sony's have red push secondary to the color decoder, which can be adjusted in the SM. I'm not sure if your problem is a probe error or a color decoding provlem. AVIA has a color decoding check on the disc.

Bob