Y Pb Pr vs. RGB for classic games

[zedrein]

At this point I think we all know the fundamental technical differences between a RGB and Y Pb Pr "component" signal. Most often I hear Y Pb Pr called "low-bandwidth RGB" but I've also heard it called "A not much better alternative to S-video" So, according to you devious minds, what sort of differences can a person with zero experience with either format expect to see between these two signal types? {EXCLUDING 480 PROGRESSIVE SCAN, JUST A 480i/240p SIGNAL}

Let's take the SNES for example: If I had a television equipped with both RGB and Component video inputs, what would be some fundamental differences between the RGB and Y Pb Pr signal's picture quality coming from that system? {I know that the SNES doesn't natively produce a Y Pb Pr signal, but for the sake of the question lets assume it does}

Or an even better example: Let's take the Nintendo Wii's Virtual Console service which allows you to play emulated "retro" games. The PAL version of the console can do both RGB and Y Pb Pr...so on said television equipped with both types of inputs, what would those signals look like compared to eachother? {once again, speaking strictly about NON 480 LINE PROGRESSIVE SIGNALS}

I think this is a fair question because it will help a lot of people decide if they want to simply purchase a RGB to Y Pb Pr transcoder for their television with only that input, or go to the trouble of buying a hard to find quality display and creating a cable to get RGB from their consoles.

Thanks for your time!

[NFG]

Interlaced or non-interlaced makes no difference at all for the purposes of this discussion.

Generally speaking, you won't notice any difference between component and RGB on a quality TV.  In fact, you're better off spending money on a good TV with component support than a shitty RGB monitor.  I got way more quality from my old JVC TV with S-video than all the RGB arcade monitors I could find.

[Fudoh]

Interlaced or non-interlaced makes no difference at all for the purposes of this discussion.

it doesn't matter in terms of quality, but it can matter in terms of compatibility.

While low-res progressive (240/288p) is inside the RGB video specs, the official Component specs do not include it. That's the reason why **MANY** TVs and other video devices cannot display a 240p YUV signal. The screen stays blank. This includes many of the Samsung LCDs, most video capture cards and a lot of the newer video processors.

CRTs should be fine, but in case you're considering a flatscreen TV, remember that this CAN cause problems.

[viletim]

While low-res progressive (240/288p) is inside the RGB video specs, the official Component specs do not include it. That's the reason why **MANY** TVs and other video devices cannot display a 240p YUV signal. The screen stays blank. This includes many of the Samsung LCDs, most video capture cards and a lot of the newer video processors.

No only that, low res non-interlaced video has not been included in any official TV video standard at all! It has always been a hack.

[zedrein]

Just to clarify I was speaking about both 480i and 240p on a CRT display. I only made sure to say "non-progressive" cause alot of people will say "Well duh, the difference between RGB and Y Pb Pr is that Y Pb Pr can do 480 lines progressive" I just wanted to get that out of the way before someone brought it up trying to be all snarky...


With that out of the way, I am very interested in what Lawrence had to say about there being very little to no differences between these two formats on a quality display, and him having better results with S-video compared to the RGB arcade displays he was able to find. I suppose that makes sense considering that it doesn't matter what input capabilities your display has if it is of poor quality.

With all this information I have gathered, I may set out on a quest to locate the last surviving quality CRT displays equipped with Y Pb Pr inputs to test out my consoles with a transcoder--is there is such a thing as a lossless RGB to Y Pb Pr transcoder anyways?

[Jon8RFC]

I was having some noise issues with my setup (RGB to component via APTUS1, displayed on a Sony SXRD) but after fixing that, the picture looked much better.  I'll have to do some more comparisons, but the s-video on my SNES has a sharper image than the component...once I get a new power supply for my N64, I hope to see an improvement of component over s-video.  Both are a huge improvement over composite though.  When I find my Jaguar, I'll finally take some pictures to compare all three systems with composite, s-video and component.  I think the problem for me is the APTUS1...it's almost like it's a software-based conversion rather than purely hardware...I know it uses hardware, but software RAID on a computer uses hardware even though it's software-driven.  The biggest improvement for me is that the SXRD does phenomenal deinterlacing, so the s-video looks great and VERY clean compared to a 1080p DLP RPTV...I never bothered connecting my older consoles to this set until I started working with the idea of RGB, but the SNES s-video (I use monster rather than the oem cables) looks incredible compared to my CRT.  The problem with a CRT is scan lines though, even if it's a progressive set.

[Jon8RFC]

Bump...oh, the Sony CRTs all have solid vertical phosphors and I've never seen another brand which has them, and they reduce the visible scan lines and provide a more "complete" picture in my experience, especially when a progressive picture is displayed.  The unfortunate disadvantage of a Sony CRT (despite their superior quality) are the two lines...I won't describe any further because if you don't already see them and have very sensitive eyes, you probably never will see them unless they're pointed out to you or you actively look for them.  High-end home theater salesmen often don't see or know of them.  The late-model Samsung CRTs are pretty good, and if you can get your hands on one, a Mitsubishi CRT is great as well.

[NFG]

The vertical phosphors are often a byproduct of the aperture grille TV technology that Sony's Trinitron line pioneered.  Any TV that uses an aperture grille will have vertical phosphors.

[viletim]

With all this information I have gathered, I may set out on a quest to locate the last surviving quality CRT displays equipped with Y Pb Pr inputs to test out my consoles with a transcoder--is there is such a thing as a lossless RGB to Y Pb Pr transcoder anyways?

Not lossless - all signal converison is the analog domain is lossy to some extent. If you change that to 'negligible loss' (ie, no noticable image degradation) then the answer is yes, absolutely. But if you buy the cheapest converter you can find then don't expect such quality!

A typical CRT TV will convert the Y-Pb-Pr bank into linear RGB to drive the picture tube so you get double the loss when you use an RGB to Y-Pb-Pr converter. It shouldn't matter much in theory but in practice there are lots of variables.

[zedrein]

Wow, I am surprised that Y Pb Pr has such a great reputation for computer generated video, most often I've heard people claim that it's "basically identical to S-video" or "VASTLY inferior to RGB in theory and practice" I am intrigued and pleased that Y Pb Pr can yield such great results.

[NFG]

Component video is an unnecessary format, designed to create something that could be less useful than the existing standard (RGB) so that Macrovision DRM could be added.  There's no technical reason component video should exist as an on-the-wire signal.  For storage or transmission, sure - it makes sense to conserve bandwidth.  For images transmitted between devices it's totally pointless.

And that's where a lot of the anger comes from.  Component video isn't that bad but it's a pointless, stupid step backwards.  We're saddled with this less-capable format simply because Sony (and others) thought we'd all copy their DVDs if they let us use RGB.

Hey Sony, we're copying them anyway.  Fuck you very much.

[zedrein]

Component video is an unnecessary format, designed to create something that could be less useful than the existing standard (RGB) so that Macrovision DRM could be added.  There's no technical reason component video should exist as an on-the-wire signal.  For storage or transmission, sure - it makes sense to conserve bandwidth.  For images transmitted between devices it's totally pointless.

And that's where a lot of the anger comes from.  Component video isn't that bad but it's a pointless, stupid step backwards.  We're saddled with this less-capable format simply because Sony (and others) thought we'd all copy their DVDs if they let us use RGB.

Hey Sony, we're copying them anyway.  Fuck you very much.

Interesting to say the least. I didn't realize that component video's start came from fear of piracy, I guess I shouldn't be surprised though. Lawrence, while I have you hear I'd like to ask you a question completely unrelated to the thread:

Yesterday I conducted a picture quality test between my Super NES and Wii. I had both systems connected to my 27" Samsung CRT via composite lead. I ran several titles including Legend of Zelda: Link to the Past, Donkey Kong Country and Super Castlevania IV on both systems simultaneously and would change between the sources to observe significant differences. I was pretty surprised to find that the Wii's composite output was far superior to the actual SNES composite output in terms of picture quality. Why would this be? Does the Wii just have a much better video processor and therefore better quality video output signals in general? Would the age of my AV connector or even power supply for my SNES system be cause for significant differences between both versions?

Thanks for your time.

[NFG]

I've never played a retro game on the Wii, so I cannot speak with any authority about the quality of its output, sorry.

[viletim]

Yesterday I conducted a picture quality test between my Super NES and Wii. I had both systems connected to my 27" Samsung CRT via composite lead. I ran several titles including Legend of Zelda: Link to the Past, Donkey Kong Country and Super Castlevania IV on both systems simultaneously and would change between the sources to observe significant differences. I was pretty surprised to find that the Wii's composite output was far superior to the actual SNES composite output in terms of picture quality. Why would this be? Does the Wii just have a much better video processor and therefore better quality video output signals in general? Would the age of my AV connector or even power supply for my SNES system be cause for significant differences between both versions?

Composite video isn't really so bad, it's the low cost implementation of it in older game consoles which does the damage. I'm pretty sure all current generation consoles do the video encoding entirely in software and output through a fast DAC. This allows for some pretty tricky filtering to be done in the digital domain at no extra cost. Methods like this simply weren't an option in 1990.

[zedrein]

^Splendid! thanks for the reply, Timothy. I was concerned for a moment that my SNES may of somehow lost some fidelity after all it's years of use, but that is apparently not the case.

[RGB32E]

Well, the Kramer FC-14 does an excellent conversion from RGB to component and does not soften the picture nor does it add noise... a very clean unit.  However, it's priced higher than the cheesy CSY-2100 I bought off of ebay (and is >4x the cost).

I've performed a similar comparison between the Wii VC and SNES using S-Video.  Yes, the picture is a little different, but I still prefer the SNES' output over the Wii.

As far as filters in software, I'm not sure that there is any public fact as to any of the filters programmed in software for the Wii (certainly is possible, but is speculation as to the specific application).  Furthermore, the 360 has it's HW video scalar that handles those related things.  For instance, comparing SFIV via 480i on the 360 and PS3.  The PS3 does not employ any means for flicker filtering and appears to render at a lower resolution.  Whereas the 360 appears to render at a higher resolution (specific res can vary from game to game) and downsamples the image with flicker filtering.  So, I'm sceptical of the "video encoding entirely in software" quote... 

[viletim]

Perhaps software isn't the correct word, I'm not too familier with how digital signal processors work. By filters I was refering to digital luma/chroma filters which reduce interference without sacrificing bandwidth.

[zedrein]

I've got to wonder if that's the case because it's {the Wii's} video signal was noticeably better. But then again, there could be many variables at play.

[RGB32E]

Perhaps software isn't the correct word, I'm not too familier with how digital signal processors work.

Either way, I'm sure the Wii has better video encoding than the OLD SNES systems...  The HW involved would be undoubtibly better.

By filters I was refering to digital luma/chroma filters which reduce interference without sacrificing bandwidth.

In the display settings on the PS3, there is a cross-color filter setting, but I think it only affects the composite video output (does not appear to change anything on S-Video or above, so I leave it off  ).

[Jon8RFC]

I guess I'll have to give the Kramer FC-14 a try.  The FC-4 looks like it would be similar to the APTUS1 I have, but the FC-14 looks like the RGB>component king even compared the more expensive Kramer products!

Is anyone running their signals through an HQV chip?  I'd like to know how the quality is on a 1080 set (720 sets will do a great 1:1 upconversion, but I don't have a 720 tv) with an HQV chip in the mix compared to using your tv's native upconversion...I'm not interested in upconversion using a Faroudja chip.  I've been having a difficult time deciding if I'll invest more in my new receiver to get one with HQV for my consoles, because I love my BD-P2550's HQV upconversion of standard DVDs.

[RGB32E]

I guess I'll have to give the Kramer FC-14 a try.  The FC-4 looks like it would be similar to the APTUS1 I have, but the FC-14 looks like the RGB>component king even compared the more expensive Kramer products!

Yup, the FC-14 is supposedly fully analog (no AD->DA), so it is less expensive than Kramers higher end units.  Hence, it does not have compatiblity problems, or digital artifacts on the component output.  But like I've stated before, it doesn't accept composite video as sync.  I think this missing feature allows the unit to have wide operational compatibility, and does not add cost.  Besides, Kramer sells SCART RGBcv to component converters outside of the US for that purpose (and only support standard def).

If I ever do obtain a digital camera... I will post pictures on the Wiki!

[Fudoh]

From the little testing I've done so far, the transcoder included in the XSelect-D4 isn't any worse than the FC-14. On both units I can't spot any obvious differences between the original RGBs signal and the YUV output signal. From a videogaming standpoint the XSelect-D4 might be the better choice, as it has a Scart input (japanese layout though) which accepts composite sync as well and it has a dedicated VGA input for progressive transcoding. Both inputs are transcoded to the same output. At the same time it does transcoding in the other direction (from progressive YUV to VGA).

Lawrence has a review hidden somewhere in his website jungle.

[RGB32E]

http://nfggames.com/games/xselectd4/



This review doesn't appear to cover RGB to Component conversion at all.  Besides a couple of points between the XSelect-D4 and FC-14...
-I could (and did) purchase the FC-14 new for a similar price as the XSelect-D4
-I still haven't found a XSelect-D4 for purchase (new or used... and yahoo jp isn't an option  )
-I don't benefit from using using the 21-pin JP input (unless I decided to use one of my SHVC-010 cables)  
-All inputs use BNC connectors (aside from the HDDB15 RGB input tied in parallel with the BNCs)
-The FC-14 has easy to access push buttons for 4 sync configs (RGBHV, RGsB, RGBS, and RsGsBs)

Though the XSelect-D4 appears to have its benefits over the FC-14 (as Fudoh mentioned).  To use the XSelect-D4 as a RGB to component converter, it looks like you'd need a D-Terminal to 3 RCA breakout cable (FC-14 outputs on BNCs, for a pro/commecial environment). 

However, if all you want is a high quality RGB->Component converter, then you couldn't go wrong with the FC-14.

[RGB32E]

This is funny...  (from 2005)

http://nfggames.com/forum2/index.php?topic=1527.0