xvColor and AX2000

HDR-AX2000 / HXR-NX5 (2010).
User avatar
Stephan
Site Admin
Posts: 592
Joined: 20 Mar 2010 18:51
Location: Paris, France

Re: xvColor and AX2000

Post by Stephan »

Arkady, look at your crossroads shots: it's the whole vectorscope that expands, not just the outer rim. These vectorscopes mean one thing: increased saturation, picture-wide.

About the red flowers, I just checked on EDIUS some footage I had shot - people with bright red safety jackets, slightly overexposed under the daylight. Guess what, the vectorscope overshoots well beyond the 100% threshold, despite it's been shot with a regular HDR-HC3 (no xv.color here).

You missed Steve's point: our consumer cams have been shooting for years beyond the 100 IRE threshold, they shoot 0-108 IRE because they already take advantage of the full 8-bit space. 108% = 255/235 = [max 8-bit value] / [100 IRE value]. That's why I said "I don't think it can be transparent", and why Steve said "we have the standard gamut spread over the maximum 256 values" and "Increasing the range of colours will put things back to the 16-235 quality in terms of smooth gradations" (a good point he kindly explained 3 times already): that extra space beyond 235 is already taken and widely used for purposes other than xvYCC.

<moderator mode on>
This discussion's been quite interesting, but I think we've beaten the horse to death now. There's no way we can decipher the manufacturers intent and how we're supposed to use xvYCC. It's supposed to be compatible, and everything tells us instead it is incompatible with existing workflows, unsupported by available distribution media, and requires non-transparent settings on the playback and display devices. We're not going to resolve the seemingly inconsistencies of the manufacturers design, so I suggest we please let the discussion rest now.
<moderator mode off>
Arkady Bolotin
Posts: 60
Joined: 24 May 2010 16:46
Location: Beersheba, Israel

Re: xvColor and AX2000

Post by Arkady Bolotin »

To Steve
steve wrote:…I'm not sure that it is unless the Sony professional range is going over to consumer microphone sockets like 3.5mm. I doubt that pro users would take that very seriously…
Seriously, Steve, have you ever seen the NX5U? The B&H technical site reads about it the following: “The HXR-NX5U has dual XLR inputs carried over from the Z5U… The audio recording format on HXR-NX5U can be selected from either Dolby Digital stereo or Linear PCM stereo. LPCM audio recording provides uncompressed audio quality. Customers who need high quality audio, such as in music videos will be able to acquire CD quality audio data.”

Now tell me, if it isn’t professional audio, what is it?


To Stephan
Stephan wrote:Arkady, look at your crossroads shots: it's the whole vectorscope that expands, not just the outer rim. These vectorscopes mean one thing: increased saturation, picture-wide....
I think you are wrong here: if the picture saturation was increased, the vectorscope would look similar to this:
Saturation increased.JPG
Saturation increased.JPG (31.68 KiB) Viewed 18925 times
However, if you consider this discussion has exhausted itself, so be it, let close it.
Arkady Bolotin
Posts: 60
Joined: 24 May 2010 16:46
Location: Beersheba, Israel

Full (Extended) Range vs. Limited Range

Post by Arkady Bolotin »

I think that the most reason of everyone’s confusion in the previous topic was misunderstanding and misinterpreting of some basic terms and notions of analog and digital video.

So, allow me to clear some of them.

Let me start with the very basics. Analog video levels are typically measured in IRE. For NTSC composite signals (except for Japan), proper black level is at 7.5 IRE and proper white level is at 100 IRE. Correspondingly, for PAL, black level should be at 0 IRE and white level at 100 IRE.

At the same time, digital video formats that follow ITU-R Rec. 601 (DV, DVD, Blu-ray, etc.) store video in YCbCr form. Roughly speaking, the Y component determines the black and white levels of digital video, while Cb and Cr store the color information.

It’s very important to remember, that YCbCr is not an absolute color space, it is a way of encoding RGB information. Therefore, understanding how Y, Cb, and Cr signals are converted from R, G and B signals is imperative. The formulas encoding Y, Cb, and Cr from R, G and B are quite simple and yet elaborative, so instead of printing them here I just present some results.

Let us take red color, which in RGB color space is R=255, G=0, B=0. Then in YCbCr space we get Y=82, Cb=90, Cr=240.

Now, let’s take white color (R=255, G=255, B=255), we will get Y=235, Cb=128, Cr=128.

For black color (which is R= G=B=0 naturally), the result is Y=16, Cb=128, Cr=128.

Thus, we get that the range of the Y component is [16;235], and it wholly corresponds to RGB range [0;255].

Again, we get that the black level (0 IRE, or 7.5 IRE for USA) is Y=16, and the white level (100 IRE) is equal to Y=235, that’s it. There are no other Y values like less than 16 or greater than 235, period.

Since the equations defining YCbCr are formed in a way that rotates the entire nominal RGB color cube and scales it to fit within a larger YCbCr color cube, there are some points within the YCbCr color cube that cannot be represented in the corresponding RGB domain (at least not within the nominal RGB range).

For example, let’s take a particular point in YCbCr color space with coordinates Y=10, Cb=90, Cr=240. We will find no place for such point in the standard RGB space because the conversion formulas give us the negative results R=172, G=-83, B=-84.

But this is the very idea of xvYCC color space: using the Y values less than 16 or greater than 235 (i.e. negative RGB signals) to extend the color gamut.

Once more time, we are not talking about the extension of the RGB range from [16;235] to [0;255] but about the extension of the Y range of the YCbCr from [16;235] to [0;255]. Here is the root of all the confusions and misconceptions.

Black and white levels are either full range or limited range. While the RGB values may be either full [0;255] or limited [16;235] (as it is required by the CEA-861-D video formats), the Y value of the YCbCr color space is always in the limited range [16;235] (at least it was before xvYCC proposal).

Nobody took advantage of the full YCbCr color space before simply because nobody knew how to correctly interpret and display those YCbCr signals. All video stored on modern discs, be it DVD or Blu-ray are stored as YCbCr with the range [16;235].

And only with xvYCC, this range is extended to full 0-255.
Last edited by Stephan on 27 Jul 2010 21:13, edited 1 time in total.
Reason: merged from duplicate thread
steve
Posts: 74
Joined: 28 May 2010 10:04
Location: UK

Re: xvColor and AX2000

Post by steve »

Arkady Bolotin wrote:To Steve
steve wrote:…I'm not sure that it is unless the Sony professional range is going over to consumer microphone sockets like 3.5mm. I doubt that pro users would take that very seriously…
Seriously, Steve, have you ever seen the NX5U? The B&H technical site reads about it the following: “The HXR-NX5U has dual XLR inputs carried over from the Z5U… The audio recording format on HXR-NX5U can be selected from either Dolby Digital stereo or Linear PCM stereo. LPCM audio recording provides uncompressed audio quality. Customers who need high quality audio, such as in music videos will be able to acquire CD quality audio data.”

Now tell me, if it isn’t professional audio, what is it?
Firstly Arkady an apology. I was on another forum dealing with the NEX-VG10 and I obviously didn't clear my thoughts when I read your post that said:

"As to why xvCC isn’t included in the professional cameras’ specifications, you’re mistaken. And the point in the case is the new Sony NXCAM HXR-NX5U professional camcorder"

My comments were aimed at the NEX-VG10 which most definately bears the design features of the consumer division including a 3.5mm microphone socket and it is being marketed as an enthusiast's camera.

Now back the the NX5. Yes I have used this camera a few weeks ago and it is quite an improvement over my 5 year old FX1, (as it should be). Well this is also the latest in a long line of 'prosumer' models with pro type features bolted on. Ever since the TRV900/PD100 and VX2000/PD150 days, the Broadcast division have taken designs from the consumer parts of Sony and added some features that might be useful to a professional. The timeline includes cameras such as the FX1/Z1, HC1/A1, FX7/V1, FX1000/Z5 and now the AX2000/NX5. All of these cameras are basically consumer devices, adding a box with a pair of XLRs and a 48V pahntom source does not make the camera a professional machine. Even the inclusion of SDi outputs, TC facilities and GPS are just putting low cost chips down on printed circuit boards, laid out ready for them. The rest of the 'professional' features are just firmware enhancements. That's not to put them down, these cameras are very popular with enthusiasts and the lower end of the wedding market, but they don't get much exposure (sorry about the pun) in the corporate and broadcast market. Probably the cheapest true professional camcorders currently marketed by Sony are the EX1R, the S270, and the EX3, and their prices reflect this.

Steve
Arkady Bolotin
Posts: 60
Joined: 24 May 2010 16:46
Location: Beersheba, Israel

Re: xvColor and AX2000

Post by Arkady Bolotin »

I hear you, Steve; you do not have to apologize. Usually I care only for what I did personally but the NX5U/AX2000 isn’t my creation (I wish I would make something similar in my life). I just purchased the camera, why should I defend its design and merits.

So, let’s continue our lecture on xvYCC colors and calculate the saturation.

According to the following formulas for hue (h), saturation (s), and volume (v)
HSV_HSL Color Model.jpg
HSV_HSL Color Model.jpg (24.31 KiB) Viewed 18907 times
the saturation of any given color from the nominal YCbCr color space cannot be greater than 1 (or less than zero).

However, in our previous example of the color described with negative RGB components R=172, G=-83, B=-84, the saturation would be s = 1 + 84/172 = 1.49, that is greater than one!

This is what the xvYCC official explanation means when saying that xvYCC is the color gamut with 1.8 times more reds, greens, and blues of RGB colors. It means that theoretically applying xvYCC color gamut we would have colors with saturation up to 1.8.

Thus, in theory, all we need to prove that we really have xv-colors in the picture is a vectorscope which plots every individual color in the video signal as a dot on a color wheel so that a dot’s distance from the center of the wheel represents its saturation.

In practice, it’ not that simple because usually vectorscopes that can be found in various video software do not display the absolute saturation (which has the range [0;1]) but only the relative one (showing the saturation of a given color in comparison with broadcast-legal color saturations).

Nevertheless, we may try and if, in the vectorscope display, we find a dot at the distance from the center, say, more than 120% it would mean the presence of a xv-color in the image.

With this purpose in mind, I’ve shot (in xvYCC) another short movie “The road to Big (Beersheba mall)”, you can watch it here http://www.vimeo.com/13699523.

On the approx. 26th second of this movie, you will see the scene with traffic lights; one of those lights is of a LED-empowered kind giving the pure cyan color. Look at the screenshot of the vectorscope monitoring this scene:
Traffic light vectorscope.JPG
Traffic light vectorscope.JPG (85.87 KiB) Viewed 18889 times
As you can see, the saturation of that cyan-colored traffic light is about 1.44, far beyond of any conventional colors, which speaks in favor of the xv-color.
FoxAdriano
Posts: 50
Joined: 08 Dec 2010 01:27
Location: Rome, Italy

Re: xvColor and AX2000

Post by FoxAdriano »

Very interesting these posts ... but I'm sorry, my english is poor and I don't understand everything. Can you tell me how have I set my Sony AX2000 precisely?
I need precise settings for getting a little bit more saturated colours.Thanks a lot.
acgold7
Global Moderator
Posts: 400
Joined: 14 May 2010 23:03
Location: Woodinville, WA, USA
Contact:

Re: xvColor and AX2000

Post by acgold7 »

I know you'd like a different answer than what you've already received elsewhere, but there really isn't one. You shouldn't be doing this in the camera, you should be doing this in post. But if you must do it in the camera, all you need is in your picture profile settings, which you can learn about by reading the manual. You can start by changing COLOR LEVEL and go from there.
Adam
FoxAdriano
Posts: 50
Joined: 08 Dec 2010 01:27
Location: Rome, Italy

Re: xvColor and AX2000

Post by FoxAdriano »

acgold7 wrote: You shouldn't be doing this in the camera, you should be doing this in post.
... but if I will do it I will lose quality of image (AVCHD color space is 4:2:0). If I will do it in the camera I will not lose that quality.
acgold7
Global Moderator
Posts: 400
Joined: 14 May 2010 23:03
Location: Woodinville, WA, USA
Contact:

Re: xvColor and AX2000

Post by acgold7 »

I don't think that is true, or else there wouldn't be so many people spending so much time doing color grading and correction in post. Nonetheless, if you insist on doing it in camera, you've been told what tools you have at your disposal and how to access them. Please let us know the results of your experiments.
Adam
Post Reply