Slightly OT: Component v RGB v S-video

alfredsvideo wrote on 3/27/2004, 1:36 PM
Component connections are appearing on many of our newer toys these days and I have often wondered what is the difference between Component and RGB. after all, they have te same color terminals. Also, if there is a choice available, which gives the better result out of these two and also s-video connections?


RBartlett wrote on 3/27/2004, 2:33 PM
Component is strictly not enough to describe anything much on its own.

Component YUV or Component RGB is my reference terminology.

The component terminals in the US consumer equipment are typ. RCA=phono connections and these are in the form of component YUV, with "composite sync" on luminance. DVI-I but with only the digital mode available, and HDMI ports are likely to sort out these tangle of triple screened wiring and the bundles as they breakout. HiDef and WXGA will help this drive.

Terminals in EU etc include a special SCART/Peritel/21pin connector that natively has the optional support of RGB, with again optional sync on green, or more commonly extracted from the composite video terminal. However SCART can support YUV, and in the EU, the use of additional ports on TVs and A/V decks can provide RCA=phono YUV terminals too. DVI-I and HDMI are likely to kick SCART out over time.

I say YUV, but they are YCrCb. Some folks wanted YIQ specifically to be on the names and in the encoder.

There are advantages to each format of YUV and RGB. YUV is more like the current range of media delivery and broadcast formats. RGB is more like the sensor transducer and display electronics ("the from and the to"). However YUV is also quite like the rod/cone stimulus of the eye, so the battle over which is truly component video continues!

SVIDEO is the luminance+sync and the two vectors of chrominance mixed.
Interestingly a camcorder can have the closest to uncompressed output on the composite video terminal. This strangely can be the best terminal on your camcorder. However your nearest digitiser for it may have a sub-optimal decoder that will still give a better rendition from Y/C (S-Video). If these consumer analogue terminals are your only option to DV. It comes down to where the analogue signal is generated from, sometimes 4:1:1 colour subsampling has occured and even some of the DV compression artifacts are introduced at the SVIDEO terminal but not the composite video port.

Which is better between Component YCrCb and RGB? Well it depends if your better encoder electronics are in your set top box or in the decoder of your TV. If they are the best in both, I think you could throw a coin.
JJKizak wrote on 3/27/2004, 2:34 PM
I have a Sony KV-34HS510 HD tv set with all of the jacks and to tell you the
truth I can't see any difference between the RGB and the Component, and I mean absolutely no difference and I have tried them both many times. The distances involved are 36 ft for both. The "S" video is restricted to 480i so I cannot use it for HD. Not saying that a side by side test in a lab will be the same but in my opinion for us uninformed consumers there is no difference.

TVCmike wrote on 3/27/2004, 9:35 PM
In practical terms, you shouldn't see a difference.

A video signal is fully reconstructed from YCrCb/YUV space to RGB and back anyway. There's a particular formula depending on the video system, but it's all there. Your monitor, for example, will give the LCD controller or CRT controller signals in RGB format, even though they may have been in YCrCb at the inputs. Most of the visual information in an image is in the Y or luminance, which is why we often use 4:2:2 color sampling or lower the vast majority of the time. The color of the connectors/terminals has nothing to do with anything. There is no direct correspondence between the YUV and RGB color spaces in terms of the connectors.

S-video separates the luminance (Y) from the chrominance (CbCr) information, but the chrominance remains bundled together. There's some benefit to using component versus S-video, but you won't see a huge difference depending on the source - again, because most of the information is in the luminance. Definitely use component if you have it versus S-video. There is, however, a big jump from composite, where the luma and chroma are all together on the incoming signal.