Component vs. RGB color quality

Xan · Post by **Xan** » Sat Jul 05, 2014 12:03 pm

I've always wondered about this, is there any discernable difference in color quality between the two?

I guess Component is always limited range while RGB can use full range, so the latter should be superior in theory, but the question is which systems take advantage of that full RGB color range...

kamiboy · Post by **kamiboy** » Sat Jul 05, 2014 12:06 pm

Prepare to split hairs.

Lawfer · Post by **Lawfer** » Sat Jul 05, 2014 12:12 pm

Xan wrote:I've always wondered about this, is there any discernable difference in color quality between the two?

That depends entirely on the hardware you are using.

Xan wrote:I guess Component is always limited range while RGB can use full range, so the latter should be superior in theory, but the question is which systems take advantage of that full RGB color range...

Are we talking about SD Gaming or HD gaming? CRTs or LCD/LED/OLED/RGB LED and PLASMAS?

In HD gaming Full-range RGB is about having the best shadow effects and deepest blacks with 0-255. I can get 0-255 with my Playstation 3 and 4, but not with Xbox 360 (I get crushed blacks instead), the Wii U is apparently limited range of 16-235.

Xan · Post by **Xan** » Sat Jul 05, 2014 12:25 pm

Classic analog sources. As a display you can assume a PVM/BVM with YPbPr and RGB inputs.

Lawfer · Post by **Lawfer** » Sat Jul 05, 2014 12:31 pm

Xan wrote:Classic analog sources. As a display you can assume a PVM/BVM with YPbPr and RGB inputs.

If its a High End RGB Monitor, RGB is going to be much better than Component.

Fudoh said that this (RGB being better than component) is especially to the Playstation 2 you can even do RGB at 480p if the monitor sync on green, because the native output is RGB and the Playstation 2 transcodes the native RGB to component, apparently the transcoding of the Playstation 2 is pretty crappy.

Fudoh · Post by **Fudoh** » Sat Jul 05, 2014 1:56 pm

The PS2's component output is pretty amazing. It just contains more noise than the RGB signal. The BVMs have a near perfect LPF though, so I doubt that this brings any problems.

As far as I can tell (and I discussed this up and down elsewhere) chroma subsampling is only a problem on digital signals, so the only downside to YPbPr compared to RGB is that you only have two color channels (versus three on RGB). This said, if the source is doing it right, I think the visible differences are negliable.

There aren't too many systems where you got a choice for your 15khz output. This example for the Wii's output shows that there's hardly any difference and the luma offset is easily compensated on the monitor itself. http://retrorgb.com/wiirgbvscomponent.html

Xan · Post by **Xan** » Sat Jul 05, 2014 2:29 pm

FWIW a while ago on the PS2 I did a little comparison for this with Crash Bandicoot 2 on my BVM-1454D and the colors did come off as slightly duller to me with Component. It was noticeable with the bright orange fonts in the game especially. And for some reason the dithering became subjectively more visible with Component... of course this could be a matter of different calibration on my monitor. I also did test my PSP-2000 hooked up via Component and these differences seemed even more pronounced there.

neorichieb1971 · Post by **neorichieb1971** » Sat Jul 05, 2014 2:47 pm

I'd use component just so I didn't have to hear the hum (audio) of crappy shielding most SCART connectors use.

Mind you I usually use audio equipment with my gaming setup, which pronounces the noise.

EmperorZelos · Post by **EmperorZelos** » Sun Jul 06, 2014 5:18 am

Considering RGB uses sync seperated in composite signal, while component has it on the green one, I'd tkae a guess on more bleeding aswell, just a guess

Artemio · Post by **Artemio** » Sun Jul 06, 2014 8:31 am

Xan wrote:Component is always limited range while RGB can use full range, so the latter should be superior in theory, but the question is which systems take advantage of that full RGB color range...

This is not the case, the wii for example uses full range over component.

And regarding which systems use full range in rgb, you can assume all can. But this is developer dependant, for instance even the Genesis has a color in its palette that goes below 7.5 IRE.

Thrill · Post by **Thrill** » Sun Jul 06, 2014 2:51 pm

Artemio wrote:the wii for example uses full range over component.

It does? Jesus christ. I guess that explains why I was so confused when I was toggling between 0-255 and 16-235 on my capture card. I could swear that 0-255 looked right with the Wii but I told myself no since it was YPbPr and HAD to be 16-235.

Lawfer · Post by **Lawfer** » Mon Jul 07, 2014 9:17 am

Thrill wrote:
Artemio wrote:the wii for example uses full range over component.
It does? Jesus christ. I guess that explains why I was so confused when I was toggling between 0-255 and 16-235 on my capture card. I could swear that 0-255 looked right with the Wii but I told myself no since it was YPbPr and HAD to be 16-235.

Thats interesting, what about the Gamecube over component? The Wii U can only do 16-235 (though I dunno if its only over HDMI or if its over HDMI AND Component).

Artemio · Post by **Artemio** » Mon Jul 07, 2014 7:06 pm

The thing is that there is nothing hardware wise that limits you to use 16-235 or 0-255.

The hardware can output full range if you draw that way, depends on your assets. Unless Nintendo enforced it through quality control or had some tools in the official sdk specifically to limit color range, it is up to the developer what to use. When coding the 240p suite I used 0-255 assets and when checking the output with the oscilloscope everything was from 0 mv to 700mV, if I remember correctly. I will confirm these values and annotate them here in a few hours so we can have this clear.

Artemio · Post by **Artemio** » Tue Jul 08, 2014 4:21 am

I did some tests, here are the results:

The 240p suite in the Wii via coponent, the grayramp pattern. Goes from 0mv to almost 700mV.

Pattern:

Scope:

The texture starts at RGB 9 and ends at RGB 255.

And checking out the definition for YPbPr, it says luma levels go from 0 to 700mV, with no setup.

Please let me know if I am making invalid assumptions on this

Fudoh · Post by **Fudoh** » Tue Jul 08, 2014 9:05 am

I would say that the RGB range (9-255) refer to the Wii's internal graphic rendering, but the video encoder translates that to proper YCbCr levels. Since any decoder on the display's end can't tell what's the original rendering, the resulting number of shades (grey or colors) on the display should be the same.

Lawfer · Post by **Lawfer** » Tue Jul 08, 2014 9:55 am

So because of this werent Wii games developed/designed around 16-235? If that is the case and knowing this, then even if you somehow managed to get full range RGB, would you gain anything out of it?

Joelepain · Post by **Joelepain** » Tue Jul 08, 2014 7:24 pm

Artemio wrote:I did some tests, here are the results:

Please let me know if I am making invalid assumptions on this

Maybe a good test to do to validate this is to mesure the Component output of a DVD player. In this case we will be sure that the digital source is "limited range", and we will see if the DAC still outputs in the full 0 to 700mV range.
If it's the case, we will be sure that this full/limited range problem is just another crappy hdmi non sense, and doesn't apply to analog consoles.

shmups.system11.org

Component vs. RGB color quality

Component vs. RGB color quality

Re: Component vs. RGB color quality

Re: Component vs. RGB color quality

Re: Component vs. RGB color quality

Re: Component vs. RGB color quality

Re: Component vs. RGB color quality

Re: Component vs. RGB color quality

Re: Component vs. RGB color quality

Re: Component vs. RGB color quality

Re: Component vs. RGB color quality

Re: Component vs. RGB color quality

Re: Component vs. RGB color quality

Re: Component vs. RGB color quality

Re: Component vs. RGB color quality

Re: Component vs. RGB color quality

Re: Component vs. RGB color quality

Re: Component vs. RGB color quality