xvColor and AX2000

HDR-AX2000 / HXR-NX5 (2010).
Arkady Bolotin
Posts: 60
Joined: 24 May 2010 16:46
Location: Beersheba, Israel

xvColor and AX2000

Post by Arkady Bolotin »

I have been thinking, what the heck, let us go beyond 4:2:0 sampling format right into xvYCC extended color space and see what we will get.

I have made some tests and here are their results.

I shot the Munsell Color Cascade (a calibration chart)
Fragment.jpg
Fragment.jpg (41.7 KiB) Viewed 19610 times
and recorded a scene with a blooming bush with and withot x.v.Color function enabled; the screenshots are following.

As you can see, the results speak for themselves. The bottom line is that if you ever experienced the xvYCC-gamut color, you would never want to go back to the conventional sRGB-limited color.
Attachments
Color test.jpg
(92.99 KiB) Not downloaded yet
halfpipe
Posts: 36
Joined: 24 May 2010 18:51
Location: Canada

Re: xvColor and AX2000

Post by halfpipe »

impressively improved image
what is this xvcolor (sounds like generation x stuff) anyway and how come i never heard of it before. Is it specific to the ax2000 and once enabled does it require shooting a calibration chart each time?
Arkady Bolotin
Posts: 60
Joined: 24 May 2010 16:46
Location: Beersheba, Israel

Re: xvColor and AX2000

Post by Arkady Bolotin »

Hi halfpipe,

There’s nothing special about xvColor, actually it’s the new extended-gamut color space for video applications (technical name is xvYCC).

Probably you are already aware of such examples of object surface colors beyond sRGB-gamut as a tropic ocean, a scarlet-red car, petals of the rose, green-illuminated signboards, lilac flowers etc. These colors cannot be reproduced by conventional sRGB-based signals.

xvYCC is an expanded version of older YCC color gamut, with 1.8 times more reds, greens, and blues of sRGB.

Sony lunched a series of camcorders that use the xvYCC standard: HDR-HC5, HDR-HC7, HDR-UX20, HDR-SR10, 11, 12, and also NX5U/AX2000 (the last ones record in AVCHD format).

Panasonic also has xvYCC-compatible HD camcorders HDC-SD9 and HDC-HS9 (Panasonic’s version of xvYCC is called Digital Cinema Color).


Nota bene: I’ve introduced some improvements into my tests: to make the difference more evident, I shot the Munsell Color Cascade beaming from the screen of a full HD xvYCC PC monitor.
Attachments
Screen color test.jpg
(52.6 KiB) Not downloaded yet
User avatar
Stephan
Site Admin
Posts: 592
Joined: 20 Mar 2010 18:51
Location: Paris, France

Re: xvColor and AX2000

Post by Stephan »

From what I mildly understand, xvYCC takes those colors from the YUV space that are illegal once converted to RGB (because some RGB channels would have negative values or greater than 255) and makes them legal. Instead of having clipped RGB channels on display, you get more colors.
steve
Posts: 74
Joined: 28 May 2010 10:04
Location: UK

Re: xvColor and AX2000

Post by steve »

The downside is that the transitions between colours (as in graduations) may be more likely to show stepping, particularly on 8 bit per colour systems.

Steve
jbeale
Posts: 2
Joined: 22 Jul 2010 16:57
Location: San Francisco, CA

Re: xvColor and quantization steps

Post by jbeale »

steve wrote:The downside is that the transitions between colours (as in graduations) may be more likely to show stepping, particularly on 8 bit per colour systems.
Steve
Is that true? According to Sony's description, the xvColor (xvYCC) standard is identical to the older HD color scheme (Rec.709 standard) for all legal values of color under the old system. What is new is that previously "illegal" values (Y,Cb,Cr coordinates that map to one or more negative values of R,G,B) are now used to expand the color gamut. If that is true, there should be exactly no difference in the ability to display all the colors you previously had under Rec.709.

See also: Figure 2, "Technical Description of xvYCC Color Space" at
http://www.sony.net/SonyInfo/technology ... cc_01.html

...and also http://en.wikipedia.org/wiki/XvYCC
http://en.wikipedia.org/wiki/Rec._709
Arkady Bolotin
Posts: 60
Joined: 24 May 2010 16:46
Location: Beersheba, Israel

Technical information on xvColor

Post by Arkady Bolotin »

As we all are aware, any color can be represented as a sequence (tuple) of three (or four) components. Those components (called primary colors) depend on the particular color model used: they are red (R), green (G), and blue (B) in the RGB color model, or cyan (C), magenta (M), yellow (Y), and black (K) in the CMYK color model.

The main difference between these color models is how a given color is produced: either by adding primary colors (RGB model) or by subtracting them (CMYK model). The former model is used for radiating media (TV and computer displays) and the latter is used for light reflecting media (printing).

For example, in the RGB color space, orange is represented as the tuple 223R+141G+57B, while yellow-green can be described as 173R+197G+78B.

The Y’CbCr color model (widely used in video acquisition) is very close to the RGB model; Y' stands for the luma component (i.e. the electronic – voltage of display – brightness) and Cb and Cr are the chrominance (color) components. Thus, there’s simple transition between RGB and Y’CC models: if you know R, G, B tuple of the given color then its Y’CC coordinates will be Y’=16+(65R+129G+25B); Cb=128+(-38R-74G+112B); Cr=128+(112R-94G-18B).

Now, imagine that we want to go beyond the standard RGB color space (sRGB was created by HP and Microsoft in 1996 for use on monitors and the Internet).

In fact, the main drawback of sRGB model is that it allows only about 55% of the horseshoe shape of visible colors.

We have two options: we may increase the purity of the primary colors (to enlarge the triangle covering the range of possible colors), or we may use negative color signals values (to represent colors outside the triangle without changing chromaticity of the primary colors).

Obviously, the second option is more achievable technically than the first one.

According to the second approach (which actually is the xvYCC color space standard, or just xvColor), cyan, for example (that is outside the triangle) can be represented by mixing positive signals for G and B and a negative value for R: Cyan+Red=Green+Blue, hence Cyan =Green+Blue-Red. Other colors outside of the triangle can be represented by mixing negative G and B signals.

But how to do it?

Usually for signal quantization (transferring from analog to digital format), 8 bits per color is used. That gives us the data range from 0 to 256. However, in the sRGB color space only the data range of 16-235 is allowed. Signals less than 16 are called blacker-than-black, and the signals higher than 235 are called whiter-than-white ones.

With xvYCC, the range is again extended to 0-256, as is expected, since digital TVs have no under- or over-shoot, as did analog television signals. Therefore, using values between 1 and 15 and between 241 and 254 as picture signals the color gamut can be expanded.

Consequently, to be able to watch video recoded in the xvYCC color space, you should simply instruct your computer’s video card to use the full range 0-256 signal.


Right now, I am rendering my video shot entirely in xvYCC. I hope to share it with you later.

Meantime, below you can find the screenshots of the horseshoe shape of visible colors (along with the sRGB triangle) I made with and without xvColor function.
Attachments
sRGB1.jpg
sRGB1.jpg (50.25 KiB) Viewed 19506 times
xvColor1.jpg
xvColor1.jpg (49.73 KiB) Viewed 19506 times
Arkady Bolotin
Posts: 60
Joined: 24 May 2010 16:46
Location: Beersheba, Israel

Re: xvColor and AX2000

Post by Arkady Bolotin »

Here is the video I shot in xvYCC color space: http://www.vimeo.com/13543487.

Please take a look, watch it, try your monitors and video cards. Tell me the results.
User avatar
Stephan
Site Admin
Posts: 592
Joined: 20 Mar 2010 18:51
Location: Paris, France

Re: xvColor and AX2000

Post by Stephan »

It's hard to say. Although I understand you didn't, it looks to me like you simply increased color saturation. That's what I thought in your first side-by-side test (Color test.jpg) and I see the same in the video.
steve
Posts: 74
Joined: 28 May 2010 10:04
Location: UK

Re: xvColor and AX2000

Post by steve »

The 16-235 range restriction is a legacy of composite video when used for broadcasting to prevent over modulation. Since digital video and editing has been in general use, there has been the option to spread the (albeit limited) gamut over the full 8 bit (0-255) range. The removal of this unnecessary reduction of colour resolution has helped to mitigate the limitations of 8 bit video digitisation which is exacerbated when adjustments such as gamma and knee adjustments are made post digitisation.
Increasing the range of colours will put things back to the 16-235 quality in terms of smooth gradations. It's one thing comparing colour wheels side by side, but debateable whether the additional colour depth will be noticeable in prectice. For instance, Blu-ray does not support extended gamut encoding and I don't hear complaints about the colours in commercial discs. The loss of greyscale smoothness can however be detected after some compression artifacts have taken their toll where the footage is originated from 8 bit digitisation.

Steve
Post Reply