AVS Forum banner
1 - 20 of 69 Posts

·
Registered
Joined
·
24 Posts
Discussion Starter · #1 ·
Hey guys, I just have a quick question. This is my first time calibrating anything manually with a Spyder2 (I know, it's not the most accurate but it's what I have ATM). I have a Panasonic TC-P42C1 plasma and have followed the "Color Calibration for Dummies" guide and have had a great time so far. I've managed to enter the service menu and dial in the WB-TEMP to D6500 almost perfectly from 10%-100% grey steps. The image is light years better than it was in color temp regards. I also dialed in a light output of 35 ftL which I find is a good balance for my ambient light and night viewing. Now the only thing left is straight up color calibration. I followed the steps for color calibration and recorded my Y value for 100% white (73) and then calculated 21% of it for my 100% R-G-B values (15). I should be dialing in my R-B-G to a Y value of (15) but I can only do it for one color, the other colors are WAY off. I understand now why people only buy TVs with CMS (I can only assume this is why). Here are my Y values at 100% light output:


HDTV REC 709

White = 73

Red = 15

Green = 75

Blue = 3


As you can see, Red is perfect, Blue is under-saturated, and green is off the charts. Is this normal or is this my Spyder2 showing it's inaccuracy? I'm sure it's not the Spyder2 because I can see with my own eyes that the Red and Greens have always been over-saturated from day one with this TV. After using the above numbers for color, I have to say that I didn't like how under-saturated blue was so I calibrated for Blue instead. This still leaves me with very over-saturated reds and greens which are VERY distracting in almost every movie I watch (other than CGI and cartoons). So what do I do? Return the TV which I JUST bought and save for a CMS one? If so, can you guys recommend an affordable CMS Plasma with good blacks? (good blacks are why I bought the Panny over the Samsung P42B450). Or am I doing this all wrong? Are my colors screwed because I'm using HDMI and not Component (I think Component is more tweak-able)? Thanks for any help you guys can offer. Cheers.


-Joe
 

·
Banned
Joined
·
20,735 Posts
Did you start with the blue-filter method? Is there a reason why you moved on to using the Y value?


Further, this is a recurring problem, but this method does not work unless your primaries are correct to Rec709. You have to actually plot the xy coordinates of your primaries, which is not very accurate on something like the spyder, then after that you have to calculate the proper Y value for those primaries. Just plugging a % referenced to white drawn from REC709 primaries won't be correct.


If you don't understand what I mean, I suggest you stick to the basic blue-filter method, being careful to check filter leakage first whether or not to trust the filter. If the filter leaks, it will bias your color saturation up a little bit too high, so you can compensate a little bit for that judiciously.
 

·
Registered
Joined
·
69 Posts
Ahm, Green is brighter than White?? Something has gone seriously wrong!


Also, the numbers you gave us don't tell you the saturation of a specific color, Y is the color intensity.

Quote:
You have to actually plot the xy coordinates of your primaries, which is not very accurate on something like the spyder, then after that you have to calculate the proper Y value for those primaries. Just plugging a % referenced to white drawn from REC709 primaries won't be correct.
I once, just for fun, modified a document created by Don Munsil in order to make this as easy as possible. So for those who don't have an application that fixes this for you, feel free to use it (it's just a massive copy paste work from the original document so it may be some erros in there, let me know if thats the case)


The lack of meter-accuarcy will ofcourse still be a big problem (not to mention the lack of controls in the display...).

 

Color_intensity_analysis.zip 14.8056640625k . file
 

Attachments

·
Registered
Joined
·
24 Posts
Discussion Starter · #4 ·

Quote:
Originally Posted by ChrisWiggles /forum/post/16917489


Did you start with the blue-filter method? Is there a reason why you moved on to using the Y value?


Further, this is a recurring problem, but this method does not work unless your primaries are correct to Rec709. You have to actually plot the xy coordinates of your primaries, which is not very accurate on something like the spyder, then after that you have to calculate the proper Y value for those primaries. Just plugging a % referenced to white drawn from REC709 primaries won't be correct.


If you don't understand what I mean, I suggest you stick to the basic blue-filter method, being careful to check filter leakage first whether or not to trust the filter. If the filter leaks, it will bias your color saturation up a little bit too high, so you can compensate a little bit for that judiciously.

I totally get what you mean about xy coordinates. Never thought about it before but now it makes sense. And yeah, I used the blue filter method but wasn't happy with the results. I used the blue filter that came with DVE HD Basic. After calibration the reds and greens had way too much "pop", they looked almost florescent. The "Y" method let me dial in red, which was my biggest beef and now it looks closer to like it should (but it could still be off of REC709 standards).
 

·
Registered
Joined
·
24 Posts
Discussion Starter · #5 ·

Quote:
Originally Posted by my_pacman /forum/post/16918456


Ahm, Green is brighter than White?? Something has gone seriously wrong!


Also, the numbers you gave us don't tell you the saturation of a specific color, Y is the color intensity.



I once, just for fun, modified a document created by Don Munsil in order to make this as easy as possible. So for those who don't have an application that fixes this for you, feel free to use it (it's just a massive copy paste work from the original document so it may be some erros in there, let me know if thats the case)


The lack of meter-accuarcy will ofcourse still be a big problem (not to mention the lack of controls in the display...).

Thanks a bunch! I will check this out.
 

·
Banned
Joined
·
3,587 Posts
I thought you said White measured 35 fL for 100%?


Did you measure with a full field or window pattern? All plasma measurements should be made with window patterns - never full-field. If your measurements are right, it sounds like you used full field to set the 35 fL level, then went to window patterns when doing the CMS measurement/calculations because you are NOW claiming 100% white measures 73 fL??? What's up with that? 73 fL for a window pattern is WAY TOO BRIGHT for a dark room.


That said... when you have color errors and there are no CMS adjustments, anything you do to the TV to try to make colors better is just going to royally mess up something else - as you have discovered. So put your controls back where they were before you tried to improve blue and stop worrying about it. Blue errors are the least obvious errors - green errors are the most obvious. So, if anything, you should be trying to minimize green errors, not blue. But again - anything you do to try to improve the colors in any way is just going to make other things much worse. Grayscale errors are easier to see than color errors (unless the color errors are REALLY) bad.


There's only 1 plasma with really good CMS controls... Samsung B650 or higher (i.e. B750, etc.). You may not think the black level is as good as current Panasonic models (like G10 or V10) though. And the Samsungs can be off (grayscale-wise) quite a bit at 20% and 10%.


By the way - I wouldn't trust ANY reading you get from a Spyder at 10% - your eyes will tell you more at that luminance level than the Spyder will. Just because a meter gives you a reading, doesn't mean you can trust the reading.
 

·
Registered
Joined
·
24 Posts
Discussion Starter · #8 ·
Great info, thanks so much for taking the time Doug. Things are getting clearer and clearer regarding my understanding of video calibration. I didn't know anything about 2 weeks ago ... and to some I still don't but meh
.


I think there was a misunderstanding and some error. As far as the patterns, you were bang on correct and I fixed things today. I originally used full screen images for the colors and windowed for the white. As for the misunderstanding, I meant a Y value of 73, not ftL. These are my luminance readings now with white at 39ftL.


White - Y134

Red - Y25

Green - Y88

Blue - Y6

Magenta - Y28

Cyan - Y95

Yellow - Y126


I calibrated the colors using the excel spreadsheet in "Cinema" mode. It worked like a charm in Cinema mode but not so much on the others (Standard and Custom). In Cinema I punched in all my recorded numbers and then calibrated to the suggested "Y" values for each color. After calibrating for red, ALL of the other colors were almost bang on almost to the decimal ... it was ODD in a good way. The other modes weren't so accurate but I think it's because Cinema is the only one with a smooth curve? Don't know if I'm right about this though. Anyways, Cinema looks amazing ... a bit washed out but very realistic. For ****s and giggles, I calibrated Custom mode the same way and while it wasn't as smooth, the image has some realism with punch. The image almost looks 3D. So I think I'm sticking to Custom mode even though it's not as accurate.


The one thing I have noticed now is that the image has a pink hue no matter what I do. After shutting off the TV and looking at it, I noticed my reflection is pink ..... almost like there is a pink coating in the glass itself. Really annoying. Is this what Panasonic does to all their models or is it just because I bought the cheapest one? I think I am gonna return it, possibly for an S1 but if it has the same pink AR, I don't want it. Can anyone confirm this?


Thanks so much!
 

·
Banned
Joined
·
3,587 Posts
In xyY space, Y is in foot-Lamberts or cd/m2. It sounds like you are saying Y has some units other than foot-Lamberts. Are you mixing fL and cd/m2 (if so, don't do that!)?


You can't have 100% white=39 fL then have Y=135 for 100% white.


And furthermore... you should really use 75% stimulus for measuring white and colors on plasmas (and other TVs) anyway, not 100% as 100% measurements put you all the way out at the maximum end of the luminance scale while 75% stim measurements put you closer to the middle of the luminance range. The same proportions should be present at 75% stimulus - the key is to use 75% white for the reference.


So if you measure 39 fL for white (with a 100% window pattern), you should measure quite a bit LESS than that when you measure a 75% white window.
 

·
Registered
Joined
·
12,648 Posts

Quote:
Originally Posted by ChrisWiggles /forum/post/16917489


this method does not work unless your primaries are correct to Rec709. You have to actually plot the xy coordinates of your primaries, which is not very accurate on something like the spyder, then after that you have to calculate the proper Y value for those primaries. Just plugging a % referenced to white drawn from REC709 primaries won't be correct.

I find it easier to use the Rec. 709 method and then aim at minimizing the dE's in question.


For example, if I want to focus on blue, I set it to the correct % given the Rec. 709 standard and then see if going lower or higher results in a lower dE. Then, I'll keep on going in the correct direction until that dE is becomes as low as possible.
 

·
Banned
Joined
·
20,735 Posts

Quote:
Originally Posted by PlasmaPZ80U /forum/post/16927026


I find it easier to use the Rec. 709 method and then aim at minimizing the dE's in question.


For example, if I want to focus on blue, I set it to the correct % given the Rec. 709 standard and then see if going lower or higher results in a lower dE. Then, I'll keep on going in the correct direction until that dE is becomes as low as possible.

I don't understand. This method makes no sense. The whole point is you can't use the rec709 values unless your primaries actually are rec709. If they're not, then you're hitting the wrong target, and doing so will give you the wrong result. You have to use the values for the primaries you're actually measuring, you can't just plug in rec709 primaries just because. Or SMPTE C primaries just because. Doing that is wrong. Using dE doesn't matter, if it's referenced to 709 because that's not what your target should be. A low dE to the wrong target gives you a bad calibration. You have to calculate dE off your actual primaries, not 709, unless of course your primaries actually ARE 709 primaries (which they're not.)
 

·
Registered
Joined
·
12,648 Posts
"Rec. 709

x y Y

R 0.6400 0.3300 0.2126

G 0.3000 0.6000 0.7152

B 0.1500 0.0600 0.0722

Y 0.4193 0.5053 0.9278

C 0.2246 0.3287 0.7874

M 0.3209 0.1542 0.2848

W 0.3127 0.3290 1.0000



Of these definitions, only the xy coordinates for the primary colors and white point are absolute. The secondary colors and luminance values are DERIVED from the primaries and the white point. If your primary colors measure according to these standards, then this list correctly states the proper specifications for brightness and secondary hue/saturation. However, with a different set of primaries, you would want to shoot for a slightly different set of brightness and secondary hue/saturation targets. The math required to figure out these relationships is too complicated to go into here, but good calibration software should take all of this into account."


According to Tom Huffman's guide, the calibration software should take this all into account. Therefore, wouldn't minimizing the dE's give you the desired result?
 

·
Banned
Joined
·
20,735 Posts
What software are you referring to? Unlikely...


I posted a link to Greg Roger's calibration calculator which does exactly that, so just use that directly to check. Someone else also posted the spears/munsil excel spreadsheet that does the same thing as well.
 

·
Registered
Joined
·
3,440 Posts
ColorHCFR delta E calculates to the selected reference, while I believe CalMAN now has a way to do custom color points you can use if you want. My display measures very close to Rec 709 so I get the same color setting regardless if I use a custom reference or Rec 709, but on my last TV it would have been best to use a custom reference because it wasn't close any default colorspace. The copy and paste spreadsheet from #9 up the page with ColorHCFR is along the same lines as using the Accupel program, it's just far quicker to use the spreadsheet on multiple runs to see which is the better setting (lowest difference in Y).
 

·
Registered
Joined
·
12,648 Posts

Quote:
Originally Posted by ChrisWiggles /forum/post/16929897


What software are you referring to? Unlikely...


I posted a link to Greg Roger's calibration calculator which does exactly that, so just use that directly to check. Someone else also posted the spears/munsil excel spreadsheet that does the same thing as well.

I was referring to ColorHCFR. Anyhow, I did try the spreadsheet in post #9 and got the following results:


Difference in Y from target 4.7% 0.1% -7.6% 2.9% -5.4% -15.1%


What am I supposed to do with this data? Do I try minimizing it like I was doing previously with the dE's in ColorHCFR?
 

·
Banned
Joined
·
20,735 Posts
You measure and enter in the xy coordinates for each primary, and the Y for white, and then target the calculated Y values for color saturation. Again, this is only if the blue-filter method isn't working on your display.
 

·
Registered
Joined
·
3,440 Posts
The spreadsheet calculates the alternate colorspace like the prior quote talks about. The "Difference in Y from target" line just calculates how close the measured Y comes to the calculated Y. Personally I would probably just split the difference for red being high and blue being low on the results you listed (turn up color and measure again).
 
1 - 20 of 69 Posts
Top