or Connect
AVS › AVS Forum › Display Devices › Display Calibration › HD MPEG-2 Test Patterns
New Posts  All Forums:Forum Nav:

HD MPEG-2 Test Patterns - Page 10

post #271 of 290
Thread Starter 
Quote:
Originally Posted by mrMuppet View Post

Hi,
i'm looking for the HD IRE 0-100 patterns, but not as video. Because my german IP-TV mediaReceiver doesn't play any extern videos at the moment i need the patterns in stills (bmp?)

Can you help me?

You can download the source files here:

http://www.w6rz.net/allpatternsyuv.zip

and then convert them to .bmp files with this command line utility:

http://www.w6rz.net/yuvtorgb.zip

Like so:

yuvtorgb -b 1920 1080 irewindow100.yuv irewindow100.bmp

Ron
post #272 of 290
Sorry if this isn't the best place for this question, but between this thread and the "xvYCC calibration" thread that links to this thread, I figured this was the better place to ask ...

I downloaded the bars709_110 test pattern to play with as my Tosh 46LX177 has xvYCC support but as far as I know I have no devices that can output xvYCC until I get my Pany HDC-TM700 in two months (boo! to Canada getting them so late ) ... I fired it up on my PC and as expected the two top bars seemingly merged, displaying the same colour. I then dropped it into my Media Center library, went downstairs and fired up the extender on my 360 (@1080i connected via component) and played it ... to my absolute amazement I was able to see 3 shades of every colour (the 3 on the far right were much harder to tell apart, but the yellow, cyan and green were noticeably different).

Can someone explain to me what is happening here? I'm hoping the extended colours are somehow being passed through to the set as I plan on recording all my home movies on the HDC-TM700 with xvYCC turned on and don't want to have to play back from the camcorder exclusively to be able to enjoy the expanded colour space, but from all I've read and understand that isn't technically possible as it isn't an HDMI connection, and from what I know, the 360 has no xvYCC support. Any insight would be greatly appreciated
post #273 of 290
Quote:
Originally Posted by Polaris75 View Post

Sorry if this isn't the best place for this question, but between this thread and the "xvYCC calibration" thread that links to this thread, I figured this was the better place to ask ...

I downloaded the bars709_110 test pattern to play with as my Tosh 46LX177 has xvYCC support but as far as I know I have no devices that can output xvYCC until I get my Pany HDC-TM700 in two months (boo! to Canada getting them so late ) ... I fired it up on my PC and as expected the two top bars seemingly merged, displaying the same colour. I then dropped it into my Media Center library, went downstairs and fired up the extender on my 360 (@1080i connected via component) and played it ... to my absolute amazement I was able to see 3 shades of every colour (the 3 on the far right were much harder to tell apart, but the yellow, cyan and green were noticeably different).

Can someone explain to me what is happening here? I'm hoping the extended colours are somehow being passed through to the set as I plan on recording all my home movies on the HDC-TM700 with xvYCC turned on and don't want to have to play back from the camcorder exclusively to be able to enjoy the expanded colour space, but from all I've read and understand that isn't technically possible as it isn't an HDMI connection, and from what I know, the 360 has no xvYCC support. Any insight would be greatly appreciated

Not being familiar with the XBox 360, my guess would be that it's doing some sort of level expansion. It's taking the feed from a PC, correct? It may be outputting at PC levels (0-255) instead of video levels (16-235), giving more discrete brightness steps from black to full brightness and letting you see the color gradations. Theoretically, you'd see the same from your PC monitor. However what you actually get depends on how your video card is treating the signal-does it send it at video or computer levels-and how your monitor is set up. The monitor could be crushing levels and compressing the brightest colors through adjustment or performance limitations.
post #274 of 290
Quote:
Originally Posted by Rolls-Royce View Post

It's taking the feed from a PC, correct?

Yes and no. The file is on the PC, and it's being displayed via essentially a remote desktop connection *to* the PC, but it's all rendered locally. I believe there is a 16-235/0-255 switch in the settings which I am fairly sure is currently set to 0-255. I figured I was viewing 0-255 on the PC locally though and unable to see the difference between the top two bars. I know for sure my monitor and video card do not have xvYCC support as there are very few monitors and video cards on the market currently that do.

Edit - Now that I'm home I checked the Xbox. It does have a black level setting, but it says it only effects 480p output and DVD playback.
post #275 of 290
It's absolutely mind boggling how confusing they've managed to make one simple little feature.

By "they" I mean just about everyone with anything to do with it.


The TV manual shows an xvYCC setting that is independant of the HDMI settings, in the "Video" menu, on/off/auto - but it then goes on to say "If a non-HDMI input is selected ... this feature will appear grayed out ..."

And then on the other hand there's this from the wikipedia xvYCC article that I have read no less than 20 times in the last 4 months (yet I still don't seem to understand how it works - and yes, I've read plenty of other things as well, not just that over and over and over ) - "xvYCC-encoded video retains the same color primaries and white point as BT.709, and uses either a BT.601 or BT.709 RGB-to-YCC conversion matrix and encoding. This allows it to travel through existing digital YCC data paths, and any colors within the normal gamut will be compatible."

But ... component isn't digital ... so I'm really no further ahead than when I started ... I missed the "digital" part of that statement when I first read it and thought I had figured something out, but now in copying it here, I realize I have not ... *sigh*
post #276 of 290
Quote:
Originally Posted by Polaris75 View Post

I downloaded the bars709_110 test pattern to play with

That test pattern does not truly test xvYCC. It tests how your video path handles RGB levels above 100% (colors that are brighter than usual but not more saturated than usual). xvYCC has more to do with how your video path handles RGB levels below 0% (colors that are more saturated than usual).
Quote:
Originally Posted by Polaris75 View Post

but from all I've read and understand that isn't technically possible as it isn't an HDMI connection

There are two types of xvYCC.

1. xvYCC video without HDMI's proprietary gamut metadata.
2. xvYCC video with HDMI's proprietary gamut metadata.

#1 can be carried over HDMI or analog YPbPr. #2 can only be carried over HDMI. Some devices only support #2.
post #277 of 290
Ah! Things are much clearer now, thank you

So does anyone know of a test pattern with negative RGB values? I know I can just test when I get my camcorder, but if I'm going to need to convert my xvYCC videos to sRGB for viewing (due to bad colour conversion/overflows/etc.) I'd like to get the workflow figured out *before* my baby is born and I have no time to play around with it and figure it out
post #278 of 290
Quote:
Originally Posted by Polaris75 View Post

So does anyone know of a test pattern with negative RGB values?

I don't know of any such test pattern. However, if it helps, someone posted a sample MTS file from a camcorder, although the colors are not very saturated. See this thread:

Finding and Decoding xvYCC HDMI metadata in AVCHD
post #279 of 290
Thread Starter 
Quote:
Originally Posted by Polaris75 View Post

Ah! Things are much clearer now, thank you

So does anyone know of a test pattern with negative RGB values? I know I can just test when I get my camcorder, but if I'm going to need to convert my xvYCC videos to sRGB for viewing (due to bad colour conversion/overflows/etc.) I'd like to get the workflow figured out *before* my baby is born and I have no time to play around with it and figure it out

I'm working on one right now. Stay tuned.

Ron
post #280 of 290
Thread Starter 
Okay, here it is.

http://www.w6rz.net/bars709_sat.zip

The layout is three rows of 75% brightness color bars with the top row equal to 120% saturation (except the blue bar which is only 115% because a 120% blue bar would require a Y value less than zero), middle row equal to 100% saturation and bottom row equal to 80% saturation.

Here's an image with the levels modified to show the layout.



And here's the image in regular 0 to 255 PC levels.



Note that the only bar that requires YCbCr levels beyond 16-235/240 is the blue bar.

Ron
post #281 of 290
Thread Starter 
The YCbCr and 0-255 RGB values in the pattern.

Code:
120% Yellow
Y = 166, Cb = 27, Cr = 137
R = 190.792, G = 191.399, B = -38.695

120% Cyan
Y = 138, Cb = 151, Cr = 27
R = -39.012, G = 190.973, B = 190.640

120% Green
Y = 125, Cb =  50, Cr = 36
R = -38.014, G = 192.578, B = -37.849

120% Magenta
Y = 39, Cb = 206, Cr = 220
R = 191.713, G = -38.880, B = 191.548

120% Red
Y = 25, Cb = 105, Cr = 229
R = 191.546, G = -38.439, B = -38.105

115% Blue
Y = 5, Cb = 225, Cr = 119
R = -28.942, G = -28.697, B = 192.094

100% Yellow
Y = 168, Cb = 44, Cr = 136
R = 191.328, G = 190.635, B = -0.455

100% Cyan
Y = 145, Cb = 147, Cr = 44
R = -0.384, G = 190.918, B = 190.341

100% Green
Y = 133, Cb =  63, Cr = 52
R = -0.015, G = 190.595, B = -1.073

100% Magenta
Y = 63, Cb = 193, Cr = 204
R = 190.974, G = 0.363, B = 192.032

100% Red
Y = 51, Cb = 109, Cr = 212
R = 191.343, G = 0.040, B = 0.617

100% Blue
Y = 28, Cb = 212, Cr = 120
R = -0.369, G = 0.322, B = 191.414

80% Yellow
Y = 171, Cb = 61, Cr = 134
R = 191.235, G = 191.569, B = 38.948

80% Cyan
Y = 152, Cr = 143, Cb = 61
R = 38.242, G = 190.862, B = 190.042

80% Green
Y = 143, Cb = 76, Cr = 67
R = 38.519, G = 191.473, B = 38.031

80% Magenta
Y = 86, Cb = 180, Cr = 189
R = 190.864, G = 37.910, B = 191.351

80% Red
Y = 77, Cb = 113, Cr = 195
R = 191.141, G = 38.521, B = 39.341

80% Blue
Y = 58, Cb = 195, Cr = 122
R = 38.147, G = 37.813, B = 190.435
Ron
post #282 of 290
Quote:
Originally Posted by dr1394 View Post

The layout is three rows of 75% brightness color bars with the top row equal to 120% saturation (except the blue bar which is only 115% because a 120% blue bar would require a Y value less than zero), middle row equal to 100% saturation and bottom row equal to 80% saturation.

Thanks.

I would like to point out that these saturation percentages are calculated in R'G'B' space (gamma-encoded), not xyY space (gamma-decoded). So when you measure the xy chromaticities on an actual display, both 80% and 120% saturations will be closer to 100% in terms of relative distance from the white point. For example, on a CRT with 2.35 gamma, 80% will be 97.7%.
post #283 of 290
Thread Starter 
Quote:
Originally Posted by Erik Garci View Post

Thanks.

I would like to point out that these saturation percentages are calculated in R'G'B' space (gamma-encoded), not xyY space (gamma-decoded). So when you measure the xy chromaticities on an actual display, both 80% and 120% saturations will be closer to 100% in terms of relative distance from the white point. For example, on a CRT with 2.35 gamma, 80% will be 97.7%.

Here's the xy values for the 120% (and 115% blue) saturated 75% brightness bars. Looks like the red bar is a little too aggressive, while the yellow, magenta and especially cyan bars could be more aggressive.
Code:
0.423300 0.511833
0.222160 0.328733
0.299254 0.615901
0.321224 0.147670
0.677477 0.330081
0.142989 0.048387


100% saturation.
Code:
0.419321 0.505255
0.224658 0.328741
0.300000 0.600000
0.320920 0.154154
0.640074 0.329971
0.150017 0.060007


80% saturation.
Code:
0.412859 0.494571
0.228809 0.328753
0.301150 0.575496
0.320426 0.164693
0.589220 0.329820
0.164604 0.084124


Ron
post #284 of 290
Quote:
Originally Posted by dr1394 View Post

Looks like the red bar is a little too aggressive,

120% red is even more aggressive when you consider the camera. The Rec709 OETF (extended for xvYCC) maps it from (x=0.7731,y=0.3299), which is very far beyond the spectral locus, so the camera should never create it in the first place.
post #285 of 290
Thread Starter 
Quote:
Originally Posted by Erik Garci View Post

120% red is even more aggressive when you consider the camera. The Rec709 OETF (extended for xvYCC) maps it from (x=0.7731,y=0.3299), which is very far beyond the spectral locus, so the camera should never create it in the first place.

This is why I've been reluctant to create this pattern. It's difficult to come up with something that's useful given that you have to fit in to a real-world gamut (either coming from a camera or being able to be displayed by a panel).

Any suggestions for something better are welcome.

BTW, do you have the equations (or tool) for converting RGB to XYZ that includes the negative gamma in xvYCC?

Ron
post #286 of 290
Quote:
Originally Posted by dr1394 View Post

BTW, do you have the equations (or tool) for converting RGB to XYZ that includes the negative gamma in xvYCC?

You can use my color calculator. It has the equations in JavaScript.
post #287 of 290
Thread Starter 
Quote:
Originally Posted by Erik Garci View Post

You can use my color calculator. It has the equations in JavaScript.

Very nice, but I believe there's a bug in the RGB to XYZ conversion below. Shouldn't the

if (form.re.value <= -0.081) {

be

if ((form.re.value/255) <= -0.081) {

since form.re.value is in 0 to 255 range?

Code:
if (form.gamma.options.selectedIndex == 0) {
   if (form.re.value <= -0.081) {
      form.rd.value = -Math.pow(((-form.re.value/255+0.099)/1.099),(1/0.45));
      }
   else if (form.re.value >= 0.081) {
      form.rd.value = Math.pow(((form.re.value/255+0.099)/1.099),(1/0.45));
      }
   else {
      form.rd.value = (form.re.value/255)/4.5;
      }

   if (form.ge.value <= -0.081) {
      form.gd.value = -Math.pow(((-form.ge.value/255+0.099)/1.099),(1/0.45));
      }
   else if (form.ge.value >= 0.081) {
      form.gd.value = Math.pow(((form.ge.value/255+0.099)/1.099),(1/0.45));
      }
   else {
      form.gd.value = (form.ge.value/255)/4.5;
      }

   if (form.be.value <= -0.081) {
      form.bd.value = -Math.pow(((-form.be.value/255+0.099)/1.099),(1/0.45));
      }
   else if (form.be.value >= 0.081) {
      form.bd.value = Math.pow(((form.be.value/255+0.099)/1.099),(1/0.45));
      }
   else {
      form.bd.value = (form.be.value/255)/4.5;
      }
   }
Ron
post #288 of 290
Quote:
Originally Posted by dr1394 View Post

Very nice, but I believe there's a bug in the RGB to XYZ conversion below.

Yes, that's a bug. I corrected it. Thanks.
post #289 of 290
Thread Starter 
Here's the corrected CIE chart for the 120% saturation (115% blue) bars. As Erik pointed out, 120% red is way out of natural gamut. Even 115% blue is too much also. I'll have an updated pattern soon (the previous one had a little bug in it anyway, the small white bar on the right is at 235).



Ron
post #290 of 290
Quote:
Originally Posted by dr1394 View Post

Okay, here it is.

http://www.w6rz.net/bars709_sat.zip

Awesome! Thanks so much! Oddly, I think I can see 3 bars in this one from the 360 via component too! I got my HDC-TM700 today, so now that I've got something I know for sure can output xvYCC, and control over the recordings so I can be sure there is xvYCC data in the clips, I can test to my heart's content!
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Display Calibration
AVS › AVS Forum › Display Devices › Display Calibration › HD MPEG-2 Test Patterns