Hoping to get a general discussion going about how specs relate to the actual colours you see on screen, and the information that various manufacturers release.

For example, I don't think I've ever seen Panasonic say that "our Viera processor features X-bit processing". What they did say in older literature is "the Viera is capable of producing 68 billion colours and 4,096 shades of gradation". 68 billion seems like an impossible amount, until you read that Hitachi plasmas have a 16-bit processor that can produce 281 trillion colours, but they do not mention shades of gradation. I believe I read that this year's Samsung plasmas have 18-bit processing. Sony says that it uses a 10-bit processor in its XBR LCDs. I don't really know what Pioneer does.

In its 2008 literature, Panasonic mentions an improved number for "shades of gradation" and also equips its sets with "Deep Colour". The accompanying description says that Deep Colour takes the colour system to 10-bit or higher. But 10-bit colour seems lacking compared to Samsung's 18-bit, or even Hitachi's 16-bit. But maybe this is a different comparison.

What do these numbers mean for the on-screen display? If Hitachi's 16-bit system produces 281 trillion colours, Samsung's 18-bit system must produce many trillion more. What is "shades of gradation" and would it even matter in this scenario? Wouldn't all those shades be included in the ultimate number of colours the TV can produce?

Any discussion would be appreciated.

Red Green Blue
4096 shades x 4096 shades x 4096 shades= 68,719,476,736 colors

Don't worry about these numbers, what counts is how the display looks when fed quality material.

Don't worry about these numbers, what counts is how the display looks when fed quality material.

Yes, I am aware of that. Otherwise, the TV with the highest bit processor would be the highest-rated where colour is concerned. I'm just interested in what all the numbers mean and why different manufacturers do not include the same information in their literature.

Yes, I am aware of that. Otherwise, the TV with the highest bit processor would be the highest-rated where colour is concerned. I'm just interested in what all the numbers mean and why different manufacturers do not include the same information in their literature.

Not really. It doesnt matter how good the processor is if the panel only uses 10 bit colors.

If the panel uses ADS driving then the number of possible gray levels per subpixel = 2 to the power of (# subfields).

If the panel uses Contiguous driving then the number of possible gray levels = # subfields +1

For instance most PDPs use 8 subfields. That would be 256 levels of gray for ADS method and only 9 levels of gray for Contiguous.Pioneer uses Contiguous subfield driving and the new 9G Kuros will have 14 subfields which means 15 levels of gray per subpixel. When you hear "billions of colors" they are referring to virtual gray levels produced by halftoning methods (error diffusion, dithering....etc)

Cheers

Not really. It doesnt matter how good the processor is if the panel only uses 10 bit colors.

But that's exactly what I mean. Sony's XBR LCDs are supposed to have excellent colour reproduction ability, yet Sony's literature indicates 10-bit colour.

All colors come to those who wait, in time. Such is the nature of temporal dithering.

If the panel uses ADS driving then the number of possible gray levels per subpixel = 2 to the power of (# subfields).

If the panel uses Contiguous driving then the number of possible gray levels = # subfields +1

For instance most PDPs use 8 subfields. That would be 256 levels of gray for ADS method and only 9 levels of gray for Contiguous.Pioneer uses Contiguous subfield driving and the new 9G Kuros will have 14 subfields which means 15 levels of gray per subpixel.
Cheers

Interesting .... what are the pro/cons of using ADS or contiguous method?
Why it seems the Pioneer 8G has much more colors than its competitors (specially in the grades department)?
The spacial dithering should be visible in the grades then, don't you think?

Interesting .... what are the pro/cons of using ADS or contiguous method?
Why it seems the Pioneer 8G has much more colors than its competitors (specially in the grades department)?
The spacial dithering should be visible in the grades then, don't you think?I can't tell you all the pro's and con's but here is what I've read so far.

ADS driving
pros - significantly more gray levels per subpixel
cons - false contouring - higher black level


Contiguous driving
pros - eliminates false contouring - very good black level is possible (depending on cell design)
cons - very few true gray levels per subpixel, requires many more virtual gray levels

What do you mean by virtual gray levels?
If you meant spatial dithering, it is quite terrible when there is color gradation ...
How could you explain the Kuros has the best colors? Especially when there are color gradation? That remains a mystery for me ...
Anyway thank you for your explanations :)

Yes, I am aware of that. Otherwise, the TV with the highest bit processor would be the highest-rated where colour is concerned. I'm just interested in what all the numbers mean and why different manufacturers do not include the same information in their literature.This doesn't make much sense to me. Digital display technology where PDP makers want us to do a forklift upgrade every year to take advantage of some additional features they release in dribs & drabs, but the manufacturers refuse to publish any "quantifiable" specs to help make an "intelligent" decision with.

Digital A/V equipment manufacturers should take a page out of Microsoft's/Intel's/Dell's playbook then maybe we'll be more "motivated" to spend another $10,000 :eek: or so... for the next greatest PDP, which will surely come along.

All colors come to those who wait, in time. Such is the nature of temporal dithering.

Slick. GROAN!!

What do you mean by virtual gray levels?
If you meant spatial dithering, it is quite terrible when there is color gradation ...
How could you explain the Kuros has the best colors? Especially when there are color gradation? That remains a mystery for me ...
Anyway thank you for your explanations :)The use of halftoning methods (error diffusion, dithering...etc) is widespread in the digital display industry. It is not just used on Plasma displays. As for how and why it works so well in the Pioneer Plasmas......... Good question. I've read a couple papers by Pioneer on the subject and to be truthfull I don't understand them (especially the math). What I took from them is that Pioneer uses a combination of error diffusion, spacial dithering, and dither rotation (4 frames). Using this combination minimizes the perception of dither noise and resolution loss.

#1 all manufactures inflate specs unfourtunately i myself very well informed in the high tech field aways takes specs with a grain of salt
#2 very and i mean very important for everybody here to note that the best source for hdtv right now for regular consumers is blu-ray dvd witch is mastered at only 8-bit color so if a display over 8-bit useless in my opinion
#3 most specs not revealed by company as how it was arrived at (everybody uses different measuring to arrive at there specs which are not standardized)
# 4 analogy amplifier specs when do you see spec shown as watts per channel driven from 20-20000hz plus or minus#db rms (root mean square) +distortion (thd)
#5 in general displays surpass source materials available today
#6 only reason to get higher quality display is to future proof one self these are just my opinions

I'm not 100% sure, but I'm 99.99% certain that Blu-ray is 8 bit color per each RGB element, which means 24 bit for each pixel and therefore (approx) 16.7 million discrete colors. If it were 8 bits for the whole pixel then that would be only 256 colors; equivalent to Windows safe mode. Clearly Blu-ray is way better than that...

I was kind of being vague with some assumption when listing the driving techniques (ADS versus contiguous). And I realize that the comparison is wrong.

ADS = Address display seperate ; it means all pixels are addressed and "then" displayed.
AWD = Address while display ; it means some pixels are addressed while others are displayed.

Conventional driving = ADS with binary subfield combinations
Contiguous(Continuous) = ADS with continuous emission (PWM)


Pioneer uses CLEAR driving and have for many years. CLEAR driving is Contiguous. It not only eliminates false contouring (from driving) but also allows for a dramatic increase in low level gray expression (I just learned this little tidbit :)).

In the figure below the conventional 8 subfield drive will have 256 possible gray levels due to binary combinations. The CLEAR drive will only have 13 possible gray levels.

Yoichi Sato, Kimio Amemiya, Masataka Uchidoi
http://i222.photobucket.com/albums/dd126/xrox/CLEAR.jpg

Pioneer uses CLEAR driving and have for many years. CLEAR driving is Contiguous. It not only eliminates false contouring (from driving) but also allows for a dramatic increase in low level gray expression (I just learned this little tidbit :)).


Easy now. I was following you but now you are venturing into fanboyism. I and others have seen false contouring and Pioneer plasmas including Kuro.
Perhaps you could argue that the technology you say they use minimizes the likelyhood of false contouring. That may be closer to the truth. :cool:

I am also under the impression that Panasonic uses conventional ADS driving. To improve low level gray scale Panasonic increases the number of subfields when the APL is low. I think this is the source of the floating black issue. I have no confirmation of any of this though.

Cheers

Easy now. I was following you but now you are venturing into fanboyism. I and others have seen false contouring and Pioneer plasmas including Kuro.
Perhaps you could argue that the technology you say they use minimizes the likelyhood of false contouring. That may be closer to the truth. :cool:The technology "eliminates" false contouring caused by the driving technique. Other sources are of course possible. Note that the improved low level grayscale also "minimizes" the false contouring in low apl scenes.

Of course this is what Pioneer claims......

^^^Fair enough.

There was just a very long (somewhat heated LOL) thread in the LCD section about this.

xrox did a very good job (better then I could) of explaining what dither is and why you use it. To put it even simpler dither "blurs" the lines you see when you get false contouring.

To some other points.

10+ bit processing - As was pointed out all consumer media (blu ray, HD-DVD, broadcast etc) are 8 bit so you ask way would you need more? Simple if your display played it with no processing you would be fine or if your display was set to the same spec as the master you would be also fine with 8 bit. To bad neither of those are possible.

All TVs let you adjust brightness, color and such (processing the image). Very few TVs can be set to the "standard" the media was mastered at and even fewer TVs are set to a "standard". There is also more then one "standard" and who to say the master matches any "standard" many don't and we won't even get into broadcast (NTSC - never the same color)

What does all that mean? Well since you can't match the master (8 bit) you will need to proses the image and without getting into all the math, errors happen, a higher bit rate gives you the ability to reproduce the image with less banding and false contouring.

As to the panel itself 10bit is great BUT don't always believe what the manufactures claim. The panel could be 10+bit dither but they don't always advertise the dither part so the panel could be infact 8 bit.

As to the shads of gray.

8 bit is 256 shades of gray, each shade of gray is one shade of each primary color (red,blue.green) 256X256X256=16.8 million distinct colors or 24bit color depth.

12 bit would be 4096 shades for gray. 4096X4096X4096=68,719,476,736 colors :eek: This is just bragging rights because remember all consumer media is only 8 bit and even with all the processing errors more then 10bit really wont give you very much if anything.

Did you every notice that external VPs like the 4K Radiance XD "only" uses 10 bit processing? They do that because 10 bit is useful 12 bit is not.

nice post tower

10? 12? the new 2008 panny's have 18 bit processors...:eek: talk about overkill....

if only they stopped adding "bits" and bothered to fix their 3:2 1080i deinterlacing ;)

.....pio does it, the new sammy does...panny (85u) still no....

Thanks to all.

In its 2008 literature, Panasonic mentions an improved number for "shades of gradation" and also equips its sets with "Deep Colour". The accompanying description says that Deep Colour takes the colour system to 10-bit or higher. But 10-bit colour seems lacking compared to Samsung's 18-bit, or even Hitachi's 16-bit. But maybe this is a different comparison.

On the topic of Deep Color, one should note that these new TVs claim to be capable of supporting the new xvYCC color space (or x.v.Color as the industry has decide to call it). (Can they really? If someone knows please chime in). And that feature always seems to get a lot of flack from people here because all normal consumer sources (ATSC broadcast, DVD, BD) are in the rec709 color space and hence what's the point of a display capable of different color space when the source is not? So they say it's useless, a waste, etc....

In defense of xvYCC, while this is true for consumer sources, it should be stated for the record there are some new camcorders available that again claim to shoot in the xvYCC color space. If you have one of those (and assuming that both the cam and the display do actually work in the xvYCC color space correctly) then the feature does have value as long as you play back from the cam directly to TV. And of course wait for technology to catch up (i.e. be able to master BDs, PC files, AVCHD DVDs or whatever else to the xvYCC std and have the players (BD player (http://www.engadgethd.com/2007/03/28/sonys-bdp-s1e-improved-blu-ray-player-hits-europe-this-summer/), PC software, etc) be able to support it also, etc etc)

Lot of assumptions here guys so be gentle :) and please chime in if you have more info on the assumptions made....

Here is why one would care to use the new color space...

In a paper published by Society for Information Display in 2006, the authors mapped the 769 colors in the Munsell Color Cascade to the BT.709 space and to the xvYCC space. 55% of the Munsell colors could be mapped to the sRGB gamut, but 100% of those colors could map to the xvYCC gamut.[2] Deeper hues can be created - for example a deeper red by giving the opposing color (green) a negative coefficient.

source - http://en.wikipedia.org/wiki/XvYCC

I have been reading many papers on the subject later, and the more you read more questions are actually raised than solved. But xrox showed me a couple of papers that helped clearing out some doubts, still the mechanics of the plasma display is quite a mystery. You can even get the math, the the actual mechanical process done to represent all the colors i a bit obscure.

From the xvYCC color, the only diference i see is the greyscale, xvYCC would be the same as a FULL RANGE RGB 0-255, instead of the usual 16-235 NTSC range used in YCbCr. So it is an extension of the YCbCr color range. To a plasma display, it sounds to me, taking advantage of 256 possible shades per subpixel/subfield increases the temporal dithering capability in comparison to the usual 236. But, since all BDs that i know follow the NTSC standard of 16-235, i actually don't know how much advantage they can take of that without a proper source. Maybe some sort of haltoning technic...

Am i in the wrong track here?

I have been reading many papers on the subject later, and the more you read more questions are actually raised than solved. But xrox showed me a couple of papers that helped clearing out some doubts, still the mechanics of the plasma display is quite a mystery. You can even get the math, the the actual mechanical process done to represent all the colors i a bit obscure.

From the xvYCC color, the only diference i see is the greyscale, xvYCC would be the same as a FULL RANGE RGB 0-255, instead of the usual 16-235 NTSC range used in YCbCr. So it is an extension of the YCbCr color range. To a plasma display, it sounds to me, taking advantage of 256 possible shades per subpixel/subfield increases the temporal dithering capability in comparison to the usual 236. But, since all BDs that i know follow the NTSC standard of 16-235, i actually don't know how much advantage they can take of that without a proper source. Maybe some sort of haltoning technic...

Am i in the wrong track here?

Ycbcr uses the full range, it just doesn't display it all.

Ycbcr uses the full range, it just doesn't display it all.

chris, everything that i have read, even the simple text available in wikipedia states the opposite. YCbCr can only use 0-255 with the xvYCC gamut. So it goes beyond the BT.709 sRGB specs that takes use of the 16-235 shades only, therefore taking advantage of the full RGB 8-bit range.

http://en.wikipedia.org/wiki/XvYCC

And you can see some math here:

http://en.wikipedia.org/wiki/YCbCr

chris, everything that i have read, even the simple text available in wikipedia states the opposite. YCbCr can only use 0-255 with the xvYCC gamut. So it goes beyond the BT.709 sRGB specs that takes use of the 16-235 shades only, therefore taking advantage of the full RGB 8-bit range.

http://en.wikipedia.org/wiki/XvYCC

And you can see some math here:

http://en.wikipedia.org/wiki/YCbCr

With Ycbcr/video levels there is data in the 0-15 and 236-255 range, but the display is calibrated so that it isn't visible. I'm pretty sure that xvYCC is more than just PC levels, which is how you're describing it.

"With digital TV signaling, there is no undershoot or overshoot, and the values from 0-15 and 236-255 can be used to represent colors." <---this isn't really correct, because there most certainly is headroom and toe room in digital video. Now it may not be NECESSARY, but it's definitely there.

With Ycbcr/video levels there is data in the 0-15 and 236-255 range, but the display is calibrated so that it isn't visible. I'm pretty sure that xvYCC is more than just PC levels, which is how you're describing it.

Actually i am saying that this is the only difference i can get and read anywhere. Deep Color is indeed something else, 30 bit minimum color space, but the only thing i ever read about the specs of xvYCC is the 0-255 in comparison to the usual YCC gamut.

And the math part of "normal" YCbCr range shows that everything bellow 16 is clipped as well anything beyond 236. Actually there is an wider range in the CbCr (not the Y) that goes to 240:

"When representing the signals in digital form, the results are scaled and rounded, and offsets are typically added. For example, the scaling and offset applied to the Y' component per specification (e.g. MPEG-2[1]) results in the value of 16 for black and the value of 235 for white when using an 8-bit representation. The standard has 8-bit digitized versions of Cb and Cr scaled to a different range of 16 to 240. Consequently, rescaling by the fraction (235-16)/(240-16) = 219/224 is sometimes required when doing color matrixing or processing in YCbCr space, resulting in quantization distortions when the subsequent processing is not performed using higher bit depths.
The scaling that results in the use of a smaller range of digital values than what might appear to be desirable for representation of the nominal range of the input data allows for some "overshoot" and "undershoot" during processing without necessitating undesirable clipping. This "head-room" and "toe-room" has also been proposed to be used for extension of the nominal color gamut."

And the math bellow this text in wikipedia proves what is written there. The "head-room" and "toe-room" proposal the text mentions is the xvYCC extended gamut.

The xvYCC may be something else, or more than just 0-255, but i couldn't grab any tech info about that.

Actually i am saying that this is the only difference i can get and read anywhere. Deep Color is indeed something else, 30 bit minimum color space, but the only thing i ever read about the specs of xvYCC is the 0-255 in comparison to the usual YCC gamut.

And the math part of "normal" YCbCr range shows that everything bellow 16 is clipped as well anything beyond 236. Actually there is an wider range in the CbCr (not the Y) that goes to 240:

"When representing the signals in digital form, the results are scaled and rounded, and offsets are typically added. For example, the scaling and offset applied to the Y' component per specification (e.g. MPEG-2[1]) results in the value of 16 for black and the value of 235 for white when using an 8-bit representation. The standard has 8-bit digitized versions of Cb and Cr scaled to a different range of 16 to 240. Consequently, rescaling by the fraction (235-16)/(240-16) = 219/224 is sometimes required when doing color matrixing or processing in YCbCr space, resulting in quantization distortions when the subsequent processing is not performed using higher bit depths.
The scaling that results in the use of a smaller range of digital values than what might appear to be desirable for representation of the nominal range of the input data allows for some "overshoot" and "undershoot" during processing without necessitating undesirable clipping. This "head-room" and "toe-room" has also been proposed to be used for extension of the nominal color gamut."

And the math bellow this text in wikipedia proves what is written there. The "head-room" and "toe-room" proposal the text mentions is the xvYCC extended gamut.

The xvYCC may be something else, or more than just 0-255, but i couldn't grab any tech info about that.

But that simply isn't true -- look at any "below black" test pattern on a calibration DVD (assuming that your DVD player and TV don't clip below black) and you'll see that there is data in the 0-15 range. It is not necessarily clipped, though many devices and displays do clip it, and it is a part of the ycbcr standard.

I probably won't be able to do the test the way you would like me to. All my players are xvYCC, including PS3. My TVs, both LCD and Plasma are also xvYCC. So the only test i can do, and already performed, was using the PS3 with the YCbCr mode selected and Super White ON (Super White means xvYCC gamut is passed according to Sony's own explanation). I used both the DVE HD Basics and the AVS HD 709 i downloaded from the forum. In both the 0-255 range could only be displayed with the Super White mode ON on my PS3, if i turned it OFF, it didn't matter how much brightness was applied to the display, everything above 235 and bellow 16 was clipped.

I also tried the opposite, and the whole thing got more interesting. With the PS3 set to "Super White MODE ON" i turned the xvYCC to OFF on my TV (It has two options, AUTO and OFF), and it still displayed the 0-255 full range. The only thing i can get right now from all this is that xvYCC requires the 0-255 range but it obviously doesn't mean that if you can display the full greyscale range you will be able to expand the YCC color gamut. The xvYCC means 0-255 shades are needed to use its color space, so with the TV xvYCC mode set to ON, the Color Space is changed from the 709 (that is limited to the 16-235 range) to the xvYCC. It obviously, doesn't mean color depth, that would be related to the Deep Color thing.

So, like the wikipedia states, it is a color space that takes use of the full greyscale range. It looks like 'till now, at least...

I think these guys wrote it all in a quite simple way:

http://www.audioholics.com/tweaks/calibrate-your-system/hdmi-black-levels-xvycc-rgb

I probably won't be able to do the test the way you would like me to. All my players are xvYCC, including PS3. My TVs, both LCD and Plasma are also xvYCC. So the only test i can do, and already performed, was using the PS3 with the YCbCr mode selected and Super White ON (Super White means xvYCC gamut is passed according to Sony's own explanation). I used both the DVE HD Basics and the AVS HD 709 i downloaded from the forum. In both the 0-255 range could only be displayed with the Super White mode ON on my PS3, if i turned it OFF, it didn't matter how much brightness was applied to the display, everything above 235 and bellow 16 was clipped.

DVE HD Basics and AVS HD 709 are encoded in ycbcr, like all HD content. The fact that you can see a "full range" really proves that digital ycbcr sources can and do contain data above and below 16-235. If they could not, where were those extra levels coming from? The PS3 didn't create them from nothing, and neither did the TV.

So, like the wikipedia states, it is a color space that takes use of the full greyscale range. It looks like 'till now, at least...

Both ycbcr and xvycc use the full range. The difference is that a properly calibrated display doesn't show the 0-15 or 236-255 picture informaiton.

I think these guys wrote it all in a quite simple way:

http://www.audioholics.com/tweaks/calibrate-your-system/hdmi-black-levels-xvycc-rgb

This article is confusing because it says two different things:

"ALL video stored on modern discs, be it DVD or Blu-ray, are stored as YCbCr with a range of 16-235."

No!

"though it [ycbcr] technically can do 0-255"

Yes!

Ycbcr uses the full 256 level range possible in 8-bit video. So does PC-style RGB. The difference is that in PC RGB, 0 is black and 255 is white. In Ycbcr (and video RGB), 16 is black and 236 is white, BUT THERE IS OFTEN DATA CONTAINED ABOVE AND BELOW BLACK AND WHITE.

Is this below and above black area wasted? Arguably.

In today's world of digital video and fixed pixel display devices would it be better to have a system that doesn't devote 15% of its dynamic range to non-visible data? Seems likely.

This article is confusing because it says two different things:



No!



Yes!

Ycbcr uses the full 256 level range possible in 8-bit video. So does PC-style RGB. The difference is that in PC RGB, 0 is black and 255 is white. In Ycbcr (and video RGB), 16 is black and 236 is white, BUT THERE IS OFTEN DATA CONTAINED ABOVE AND BELOW BLACK AND WHITE.

Is this below and above black area wasted? Arguably.

In today's world of digital video and fixed pixel display devices would it be better to have a system that doesn't devote 15% of its dynamic range to non-visible data? Seems likely.

chris, i'll first please ask you to not crop the quotes, this is what is actually written in the link i posted:

"YCbCr has a native data range of 16-235 (though it technically can do 0-255, but more on that later)."

Meaning he will get into xvYCC later on the article, which is an extended YCbCr space.

And again:

"Basically, in YCbCr mode, full range signal (0-255, which it can do) is not allowed and limited range values are specified (16-235 for 8-bit color sources). Initially, analogue sources & displays had something called under- or overshoot which in essence took into account the 1-15 and 236 to 255 values so YCbCr was limited to 16-235."

Look, we already talked about the signal processing headroom and toe room, i even told you that the xvYCC is the proposal that became reality to use the range taken by the head and toe room and therefore take advantage of the full 0-255 range into a useful color space.

You seem not to get, that the BT.709 color space does not admit 0-15 and 237-255, it is not part of it. YCbCr DOES NOT ALLOW 0-15 and 237-255 into its gamut, it is clearly specified into its standard, and the math more than proves it. Headroom and toe room is not supposed to be seen! Movies ARE NOT MASTERED in 0-255 when dealing with the YCbCr standard, the WHOLE 0-255 originally contained in a movie coming from a tape (analog) is COMPRESSED into 16-235 range, leaving the 0-15 and 237-255 for transient signals, the headroom and toe room, how they are called.

xvYCC have NO HEADROOM OR TOE ROOM! It takes real use of the 0-255 range, and this is the point. Since to take real use of that, YOU NEED A NEW COLOR SPACE. That is what xvYCC is, a YCbCr expanded color space that takes use of the 0-255 entire range, with no headroom or toe room.

So, YCbCr standard, does not allow digital content to be mastered in 0-255. You are not supposed to see anything beyond 16-235 in something that has been mastered using YCbCr standard, the player should clip everything, the same way as your display. If you ever select YCbCr mode in a player IT CAN'T PASS ANYTHING BEYOND 16-235, because it is obeying the standard specs that only allows what is contained and intended to be contained in the movie (that follows its specs) to be seen; not the transient data

Please do not make a confusion to what can be passed through an HDMI cable. To pass the full 0-255 rage through YCbCr mode, you have to specify xvYCC standard.

So, to simplify, xvYCC allows a 0-255 encoded digital movie to be transmitted without any compression in the greyscale and with a broader color space that the movie MAY IN THE FUTURE POSSIBLY take advantage.


More here:

http://avisynth.org/mediawiki/ColorBars_theory
http://www.poynton.com/notes/colour_and_gamma/ColorFAQ.html#RTFToC29

No, you are simply not correct. You can and do get data using standard ycbcr in the head and toe areas. I don't know what else to say. Maybe someone else who knows about this can come and argue about it too.

I have TVs that don't support xvycc in any way, and yet I can see above white and below black data if I screw around with the brightness and contrast controls. Some TVs clip that information, some don't. Some DVD players clip it, some don't. That doesn't mean that the data doesn't exist, just that they don't use it.

I'll again refer to the DVE HD and AVS HD 709 discs -- if they're encoded in ycbcr, and ycbcr can't have information outside 16-235, then how is it possible that the discs HAVE information outside those ranges? It's possible because ycbcr uses (or rather can use) the full 8-bit range.

No, you are simply not correct. You can and do get data using standard ycbcr in the head and toe areas. I don't know what else to say. Maybe someone else who knows about this can come and argue about it too.

I have TVs that don't support xvycc in any way, and yet I can see above white and below black data if I screw around with the brightness and contrast controls. Some TVs clip that information, some don't. Some DVD players clip it, some don't. That doesn't mean that the data doesn't exist, just that they don't use it.

I'll again refer to the DVE HD and AVS HD 709 discs -- if they're encoded in ycbcr, and ycbcr can't have information outside 16-235, then how is it possible that the discs HAVE information outside those ranges? It's possible because ycbcr uses (or rather can use) the full 8-bit range.

chris, that is the whole problem. It seems i can't explain to you, that this data beyond 16-235 does not belong to the movie encoded in YCbCr standard itself. This is transient data, you can use it to transmit data, but it won't be part of the movie encoded obeying the YCbCr standard. The problem seems that most people don't actually know how the mastering process works.

The DVE and any other test DVD (even the THX tests that come with DVDs) place black bars in this "headroom" space, so you can calibrate your display and so you can AVOID seeing this range during playback. What you don't get is that the YCbCr encoded movie has NO relation to the data beyond the 16-235 range. You can place any transient data beyond this range in a DVD, it just shouldn't be visible in the movie encoded. The movie DOESN'T TAKE ADVANTAGE OF THE RANGE BEYOND 16-235, the color space in which the movie were encoded DOES NOT USE THE RANGE BEYOND 16-235.

You are confusing the headroom and toe room with the the range you should see in the encoded movie itself. Your display and player may display a range beyond 16-235, but it shouldn't, because the YCbCr encoded movie ONLY USES 16-235.

I don't know why this is so confusing. Let me try to put in a different way:

1) I get a movie to master and originally it comes to me in 0-255 full range. That is how the original transfer to tape is recorded.
2) In the mastering process the original 0-255 range in which the original file came encoded is compressed to 16-235 to be distributed in a BD or DVD, for instance, obeying the YCbCr/BT.709 standard. The range beyond 16-235 (which contains ALL the movie useful info and color space) is available for any transient data. So, this way, room is made available for this extra data; that i repeat, is not part of what you should see in the YCbCr encoded movie itself!

The whole movie and color space is contained between the 16-235 greyscale range, as determined by the Bt.709 and YCbCr standard.

For an encoded movie in YCbCr to take use of the 0-255 full range, it has to be encoded in the xvYCC color space, because BT.709 can't accept anything beyond 16-235.

So that is, again, what the xvYCC stands for, a YCC expanded gamut that takes advantage of the full 0-255 without head or toe room, everything is destined to the movie and the color space to which it belongs. In a way of saying...

I think the whole debate is kinda locked around the idea that you have for headroom and toe room, i think that the problem is that you MAY have a misconception about it.

The DVE and any other test DVD (even the THX tests that come with DVDs) place black bars in this "headroom" space, so you can calibrate your display and so you can AVOID seeing this range during playback. What you don't get is that the YCbCr encoded movie has NO relation to the data beyond the 16-235 range. You can place any transient data beyond this range in a DVD, it just shouldn't be visible in the movie encoded. The movie DOESN'T TAKE ADVANTAGE OF THE RANGE BEYOND 16-235, the color space in which the movie were encoded DOES NOT USE THE RANGE BEYOND 16-235.

There is often meaningful picture information above 235 (clouds, highlights in white shirts, etc). I can accept that this indicates a poor job of mastering the movie, but many people actually calibrate their displays with 235 slightly below peak brightness to get this extra bit of information at the high end.

I don't think I've ever heard of a movie containing real information below black, though I have read that displays may use this information to process the image though it isn't visible. I don't know if I believe that.

There was a very long and contentious thread about this on AVS maybe three or four years ago but I can't find it.


You are confusing the headroom and toe room with the the range you should see in the encoded movie itself. Your display and player may display a range beyond 16-235, but it shouldn't, because the YCbCr encoded movie ONLY USES 16-235.

I realize that you shouldn't see this information, but that doesn't mean that a ycbcr signal doesn't have 256 levels. It does, but if a movie (or game or video or anything else encoded in ycbcr) is mastered correctly this information either shouldn't exist or at least shouldn't be important.

Most of the time there's probably nothing (or at least very little) outside the "legal" range but sometimes there is.

I am not saying that you're supposed to see anything outside that range, or that it's useful, or that it wouldn't be better to have a standard that always used the full 256 levels, like xvycc or PC-style RGB.

I think I found the old thread though I actually recommend not reading it -- http://www.avsforum.com/avs-vb/showthread.php?t=494606. It's endless.

here's another germane but headache-inducing thread from the past: http://archive2.avsforum.com/avs-vb/showthread.php?t=416292&page=1&pp=30

There is often meaningful picture information above 235 (clouds, highlights in white shirts, etc). I can accept that this indicates a poor job of mastering the movie, but many people actually calibrate their displays with 235 slightly below peak brightness to get this extra bit of information at the high end.

This is definitely a poor job of mastering, but highlights are common pratice in the toe room.

I don't think I've ever heard of a movie containing real information below black, though I have read that displays may use this information to process the image though it isn't visible. I don't know if I believe that.

It is also a common practice to include filters in the transient data, i don't know if you mean the use of data like that...

I realize that you shouldn't see this information, but that doesn't mean that a ycbcr signal doesn't have 256 levels.

It does, beyond that it doesn't belong to the YCbCr encoded data, it is something else. It becomes confusing because people just don't obey standards as they should. If you set a player to YCbCr mode it should never pass anything beyond 16-235, otherwise it is not obeying the standard.

here's another germane but headache-inducing thread from the past: http://archive2.avsforum.com/avs-vb/...2&page=1&pp=30

I didn't read the whole thing, but lots of confusion there about some stuff, like IRE levels.

I think I found the old thread though I actually recommend not reading it -- http://www.avsforum.com/avs-vb/showthread.php?t=494606. It's endless.

ChrisWiggles article is pretty good, i didn't read the thread but his article is in touch with what we have been debating.

Anyway, something like xvYCC would just end all these issues with 16-235.

I'm going to quote a portion of the article (while trying to avoid any talk about IRE levels which are confusing and not really relevant):

However, even with ‘perfect’ mastering, data regularly extends outside reference black/reference white. Peak white data allows for highlight details to be maintained, which is common in clouds and other bright objects. BTB data helps prevent image anomalies from hard clipping of the analog waveform at black when converted to analog. BTB can also sometimes become visible as the actual black level on a CRT display floats up and down with image content because black level retention on CRTs is not perfect. The mastering engineer is viewing on a CRT display and actively changing the encoded levels so that they are imaged correctly on the display. This reverse-float compensation in black is allowed with BTB footroom. This reverse float compensation should be the only times data encoded below black is visible in the final image. If you are using a PLUGE pattern with BTB bars to calibrate, you should calibrate so that the BTB data is not visible. BTB data also helps define dithering duty patterns on DLP projectors. Lastly, BTB and peak white data is quite useful for any image processing/scaling etc applied to an image. For all these potential reasons, video engineers advocate preserving full BTB and peak white data whenever possible. Lastly, this data is helpful for any image processing that is applied on an image, such as sharpening, scaling, etc.

From this it really looks to me like there is and should be BTB and WTW data in ycbcr.

When designing the digital ycbcr standard, if the 0-15 and 236-255 areas weren't meant to be used, why were they included at all? It seems like a real waste to simply throw away all that dynamic range if it wasn't even meant to be used for some purpose.

I'm going to quote a portion of the article (while trying to avoid any talk about IRE levels which are confusing and not really relevant):



From this it really looks to me like there is and should be BTB and WTW data in ycbcr.

When designing the digital ycbcr standard, if the 0-15 and 236-255 areas weren't meant to be used, why were they included at all? It seems like a real waste to simply throw away all that dynamic range if it wasn't even meant to be used for some purpose.

It is meant for something extra, but it is not meant for the actual movie data. The article touches this point as well, the movie is 16-235, but since there is room, people just take advantage of it, some in a more "legal" way others just abuse it. But what matters is that the color space (BT.709) does not contain data beyond 16-235, so for the movie to take real advantage of the 0-255 and finally get rid of all this BTB and Super White mess, changing to the new xvYCC color space (all the way from the beginning of the mastering process) is the only way.

But the interesting part, is that my first question was answered by this article and the links it contains, in a way. I was curious about the subpixel/subfield behavior in relation to this data beyond 16-235. It seems PDPs take advantage of this info to improve dithering, even when the visible range is limited to 16-235. Which raises another doubt, wouldn't it be better to always let the players (PS3 for instance) to pass the full range (Super White ON) and just calibrate your display to clip the BTB, while allowing the data to exist in the invisible range?

It is meant for something extra, but it is not meant for the actual movie data. The article touches this point as well, the movie is 16-235, but since there is room, people just take advantage of it, some in a more "legal" way others just abuse it. But what matters is that the color space (BT.709) does not contain data beyond 16-235, so for the movie to take real advantage of the 0-255 and finally get rid of all this BTB and Super White mess, changing to the new xvYCC color space (all the way from the beginning of the mastering process) is the only way.

I think the real way forward would be to move to a 10-bit system. I think this is "deep color"

But the interesting part, is that my first question was answered by this article and the links it contains, in a way. I was curious about the subpixel/subfield behavior in relation to this data beyond 16-235. It seems PDPs take advantage of this info to improve dithering, even when the visible range is limited to 16-235. Which raises another doubt, wouldn't it be better to always let the players (PS3 for instance) to pass the full range (Super White ON) and just calibrate your display to clip the BTB, while allowing the data to exist in the invisible range?

Yes, it's always best to have the player output the full range, 0-255, and calibrate the display so that BTB isn't visible. You may or may not want a little bit of the WTW area visible.

Yes, it's always best to have the player output the full range, 0-255, and calibrate the display so that BTB isn't visible. You may or may not want a little bit of the WTW area visible.

Yes, i am convinced of that now.

How much better does Deep Color on a Deep Color plasma display being fed with Deep Color material from a Deep Color camcorder look versus non Deep Color?

Can you easily tell a difference?

I'd like to see deep color. Has anyone here seen it?

How much better does Deep Color on a Deep Color plasma display being fed with Deep Color material from a Deep Color camcorder look versus non Deep Color?

Can you easily tell a difference?

I'd like to see deep color. Has anyone here seen it?

Pretty much every single HDSDI camera delivers true 4:2:2 10bit color depth video, but the color space is not xvYCC. There are many cheap cameras that have xvYCC (meaning 0-255 8-bit) but no deep color (10 bit). Like this:

http://www.sonystyle.com/webapp/wcs/stores/servlet/ProductDisplay?catalogId=10551&storeId=10151&langId=-1&productId=8198552921665291500

But if you wanna see deep color get a camera like PMW-EX3 and save the file through the HD-SDI connection. You can try to get some sample files from here:

http://www.dvinfo.net/conf/

Lots of 10bit and 12 bit stuff here:

http://www.redrelay.net/

You will need to download some codecs from here:

http://www.red.com/


You can only see the difference if you have the proper equipment, like a video card that can send the 10 bit video to your deep color TV.