[wildroast]

First of all, hi to all members, I am new here. I have an SD digital cable box/PVR hooked up via S-video (Monster), and F-pin coax (RG-6) to the ANT input of my TH-42PA20U. To my surprise, the picture seems to be a little better on the ANT input (RG-6) than the video input with S-video. Is that crazy or unusual? I've only had the tv for 2 weeks and have not done any sort of "tweaking" other than turning the contrast down. I was wondering if trying the RCA composite video instead of S-video would also result in an improvement. Has anybody experimented with this for best results? This is my first plasma tv, and was just wondering about this matter. Thanks in advance for any input.


Regards,


Roast

 

[3.1415926 pi]

wildroast, to my eyes, the rca composite video has a better resolution than the s-video on my 37" panny 6UY. The component connection gives even better rez.

Your eyes may see things a little better or worse, or as it's been noted, YMMV.

Cheers, VB

 

[wildroast]

Thanks for the info VB. Well, I went ahead and hooked up the composite RCA cable, and it really does provide a nicer picture than the S-video did. I wonder if I should try the same for cable & sat on my 65" RPTV...

 

[ThumperBoy]

It makes sense since the composite video input uses the superior comb filter in the display. Using s-video bypasses the comb filter and the external cable box is most likely inferior in that area. For interlaced video sources its usually better to use composite with a high-end display for this reason. Use component or DVI with sources like DVD players or HD STBs, but for composite/NTSC sources like NTSC cable boxes/tuners or VCRs - composite is better.

 

[BruZZi]

Yup. Composite also looks better on my (Now previous) TH-37PW5UZ even using a AudioQuest S-4 (Solid Silver - $450 List http://smilies.jeeptalk.org/cwm/cwm/eek5.gif ) for S-Video connection.

 

[FitzRoy]

Interesting. Now what about an s-video feed into an htpc containing a good s-video input card such as the flyvideo2000 or compro videomate. Use dscaler and output the video via dvi. Would that make s-video look good? I would think so.

 

[TomM]

Quote:

Originally posted by ThumperBoy It makes sense since the composite video input uses the superior comb filter in the display. Using s-video bypasses the comb filter and the external cable box is most likely inferior in that area. For interlaced video sources its usually better to use composite with a high-end display for this reason. Use component or DVI with sources like DVD players or HD STBs, but for composite/NTSC sources like NTSC cable boxes/tuners or VCRs - composite is better.

Interestingly enough, I brought this up while talking with one of Panasonics higher up application engineers (although he works exclusively on the commercial side of things) and he swears that the composite and Y/C (he refused to call it "s-video" and said its really Y/C ) on the Panasonics each go through the exact same image processor, which is why you may only use either Composite or Y/C at the same time. He explained that the image probably looked better because the composite input is softer to begin with, and the Y/C tends to be more "Blocky" and pixelated. He also pointed out some interesting advantages of composite over Y/C, in that any studio thats wired for video would be using composite cables, and not Y/C, since Y/C tends to degrade much more easily over distance than composite (due to the loss of sync between the separated signals in Y/C) and that if he wanted, he could easily send 600 lines of resolution over composite.


I know this goes contrary to the majority belief on this board, but this guy absolutely refuted it. Of course tifwiw, but I would tend to believe him. Talking to this guy for 45 minutes made 99% of the discussions I've read here seem like "light reading" if you know what I mean. He started going so far over my head I only started catching every 4th word , while here I find I can usually understand at least every other word when someone starts getting extremely technical


anyway, "flame on"


TomM

 

[wildroast]

I'll accept whatever explanation there is, as the switch to RCA composite on the plasma really does look better to my eyes. This being my first plasma tv, I did not know if this improvement was unusual. Do you guys think I could obtain similar SD improvements on my RPTV tv? That tv is a professionally calibratated 65" Mitsubishi Diamond. For SD sources, it has a cable box and D*tv sat receiver both through S-video.

 

[btwyx]

I have some theory and some practice. In practice I see a difference between composite and S-Video (on my Panasonic 37" 5UZ). For a good source, S-Video is better, sharper has more detail. (Edit: I use short S-Cables, so there is no signal skew, AVIA tells me that as well.) This is born out by AVIA's resolution test pattern, I see:


Composite 480 lines (line pairs) of resolution

S-Video 520

Component 540


There's no way I'm getting 600 lines, the patterrn can only show 540 (the theoretical limit on DVD acording to AVIA.)


The composite and S-Video suffer significant interference (miore I think) in the test pattern starting at 340 (S) or 320 (C) lines. Component does not have these artifacts.


The theory I have is depends on the signal you start with. A good signal is going to look good, there's no improving a bad signal.


Also you should take care to never down convert the signal in the signal chain, so it depends on what sort of signal you start off with. A component or S-Video signal *WILL* suffer by being sent composite. It is possible that this degradation will hide flaws in the signal. The signal is being degraded, but that may look better, again a good signal will look good, try to avoid bad signals.


If you start off with a composite source, then it depends on which end has a better comb filter. If the source is not composite, there is no way a perfect comb filter can help you after downconverting the signal. The common composite signals are analog TV and VHS. (There's no saving VHS.)


The original question was about SD cable, if this is analog, then its a composite source, composite input may be better. You have the comparison between comb fitlers to worry about. (The original mentions digital cable, but some channels may actually be sent analog. Things may be different depending on whether you're watching channel 2 [analog probably] or channel 200 [digital].)


The original was also a DVR. The DVR probably stores the signal as component. It has already goine through the comb filter, S-Video should give a better signal on play back. However, the introduced digital artifacts may look better on a softer (composite) input. Again, you should try to work with good signals to begin with, there is no way a $450 cable can recover information which has already been lot.


If the original was digital cable, then you're talking about a component source. (The head end has already done any comb filtering necessary). The situation is a lot like the DVR case.


My conclusions are never down convert a signal if you can help it, and work with the best source signal you can. There's no helping a bad signal.

 

[TomM]

Quote:

Originally posted by btwyx There's no way I'm getting 600 lines, the patterrn can only show 540 (the theoretical limit on DVD acording to AVIA.)

Just to clarify, he wasn't saying he could get 600 lines from a 540 line DVD, but that a composite cable could handle it. He is also playing with some really nice equipment of course, and that being the case, I doubt a consumer device would compare. I think his point was, using the right equipment, he can send a 600 line signal through composite. Of course, given the chance and the $$$, he could probably do alot of things. He was just saying composite isn't all that bad. I guess.


As Rogo always says, go with what looks best to you.


TomM

 

[GerryL]

btwyx,


I am a relative newcomer to this forum because I am only just in the process of buying my first plasma.


Excellent explanation - one small correction though. For a VHS tape (or SVHS of course) the video signal is split into Y and C - using a comb filter - before recording to tape. They are recorded as separate signals (actually as Y, R-Y and B-Y) with the color part at extremely low quality. Hence when replayed they come off the tape as separate signals and have to be recombined to form the composite video output. Hence using the "S-video" output on a VHS deck should theoretically(!!!) give better results than the composite. The same comb filter is of course used to give the S output when you go direct from the NTSC tuner on the deck without recording. That comb filter STILL may be inferior to the one in your own display of course.


I presently have a Sony 35" XBR trinitron and the comb filter is superior to any in my other decks except DVD (i.e. Hi8, SVHS, DV tape and DirecTV) so I always use composite to link up because it gives cleaner results than S. My 50" panny is on its way to me so I will soon be able to check whether the same thing applies to that!


GerryL

 

[btwyx]

Quote:

Originally posted by GerryL one small correction though. For a VHS tape (or SVHS of course) the video signal is split into Y and C - using a comb filter - before recording to tape. They are recorded as separate signals (actually as Y, R-Y and B-Y)

As far as I know, VHS is written to tape as a composite signal, if you can point to any evidence to the contrary I'd be interested. My understanding is a composite signal is the Y information with the colour signals (R-Y, B-Y or wahtever) impressed on the top as 2 colour carrier signals in quadrature, its all one signal. This signal is modulated onto an RF carrier for bradcast. All VHS did was record this composite signal to a single track on the tape, at quite limited bandwidth. It would be substantial extra, unecessary, work to separate the signals, record them, play them back and combine them. VHS is so cruddy because they were pushing the boundaries of available technology at the time. The last thing they needed was extra complications..


The big noise about S-VHS was that the Y and C signals were recorded separately, and probably at higher banndwidth. This gave a substantial improvement in quality for JVC to crow about. In S-VHS the 2 colour components are still recorded as one composite colour signal. If the colour compoents are recorded separately you're talking about component signals, which S-VHS is not as far as I know, its using 2 tracks.

 

[MikeSer]

I found a technical post on the Internet that claims that C signal is recorded in the same way on S-VHS machines (claims FORTY lines of horizontal resolution). The Y signal has higher bandwith with resolution of 400 lines vs. 220 lines.

I've owned JVC S-VHS machine for a number of years. I could tolerate commercial VHS recordings, but home recordings always looked dreadful to me.


Mike

 

[GerryL]

btwyx


It's been a long time since I delved into this. I dumped so much of my literature a couple of years after I retired from designing chips for TV and monitors and I've forgotten a lot of it!


The total bandwidth of the VHS system is only about 2.4Mhz. This means that there is no way to record the composite signal "as is" because that would require at least 4MHz to contain the 3.58MHz C (chroma) subcarrier in addition to the Y signal. The signal is therefore split into Y and C. The Y (luma) component is frequency modulated onto a carrier at about 5MHz (can't remember the exact frequency!) and the 3.58MHz subcarrier for the C signal is frequency shifted downwards to about 600kHz and recorded directly (not frequency modulated) with about 250kHz bandwidth in the space below the fm'd Y signal.


The only difference with S-VHS is that the Y signal is modulated onto a higher frequency carrier - around 8MHz I believe - so that the available bandwidth is greater (about 4MHz). The C signal is the same.


The bandwidth available with VHS is so measly that there really is no reason to have a "S" connector. The reason for keeping Y and C separate is that the high-frequency components of the Y signal can be demodulated as part of the C signal (hence the rainbow patterns on a picture with closely-spaced vertical lines - the herringbone suit effect) and the C signal can be seen as crawling crosshatch lines. With such a limited Y-signal bandwidth there is no real interaction with the C signal. From a VHS recorder the C signal is also at rather low amplitude so that interaction with the Y signal is minimal.


S-VHS changed all that. Now the Y bandwidth was much greater (over 4MHZ) and could really interfere with the C demodulator. Thus the "S" connector could give a real benefit by keeping Y and C separate. The biggest benefit of this was on high-quality TV's (i.e. lots of video bandwidth) before good comb filters were commonplace.


The only recording system that I know of that actually records the composite signal is LaserDisc. All others - analog and digital do that separation first. The higher quality analog (e.g. Betamax) and all the digital systems also separate R-Y and B-Y (Cr and Cb) for recording.


Long message. Hope this helps.


GerryL

 

[nuke]

Quote:

Originally posted by btwyx As far as I know, VHS is written to tape as a composite signal, if you can point to any evidence to the contrary I'd be interested. My understanding is a composite signal is the Y information with the colour signals (R-Y, B-Y or wahtever) impressed on the top as 2 colour carrier signals in quadrature, its all one signal. This signal is modulated onto an RF carrier for bradcast. All VHS did was record this composite signal to a single track on the tape, at quite limited bandwidth. It would be substantial extra, unecessary, work to separate the signals, record them, play them back and combine them. VHS is so cruddy because they were pushing the boundaries of available technology at the time. The last thing they needed was extra complications..

No, in NTSC VHS, the luminance and chroma information are recorded as seperate signals.


The luminance is recorded as a frequency modulated carrier, from 3.4 to 4.4mhz (sync tip to white clip). The chrominance signal is downconverted to 629 kilohertz and recorded as an amplituded modulated signal, using the FM carrier as the bias signal. The carrier is recorded by helical scan using a pair of heads offset at about 6 degrees.


S-VHS uses the same system, but the FM carrier is pushed up to 5.4 - 7.0mhz and a higher quality of tape. The chroma signal remains the same as VHS.


Because of the down conversion, the chroma signal loses a significant amount of its original resolution, winding up around 40 lines at playback.


The S-Video (Y/C) output came about because the deck has to mix the y/c together to form the composite output, send that to the TV, which turns around and separates it. This can impose a significant loss of quality, particularly since SVHS luma bandwidth can exceed the 3.58mhz NTSC subcarrier, requiring a well-designed Y/C separator in the TV, to avoid the "Johnny Carson's Jacket" (rainbow) effects.


In the mid-80's, digital comb filters were not very common place yet. So it made more sense to supply video output in y/c format for display on the higher-end.


Even today, really well designed y/c comb filters are not all that common.


Combining the y/c to make composite also requires some filtering, so that will also cut into the potential of composite vs y/c.


So what gerryl has said is correct.


What looks good on any particular set depends on the quality of what's in the equipment and of course the eye of the beholder.


Up close, my DirectTivo looks blotchier (mpeg artifacts) on s-video and smoother on composite. But indeed, from a few feet back, the picture is WAY better on s-video. Probably points to a mediocre y/c separator in my TV.

 

[GerryL]

Thanks nuke - you were able to fill in all the details of actual carrier frequencies etc that I had forgotten!!!


GerryL

 

[nuke]

Hehe, Gerry I had to cheat a little to remember all that stuff too!


Back when I was in college I worked in a TV/VCR shop and I had my fingers in this stuff all the time. Back when VHS machines had huge mechanical piano-keys and we still fixed TV's that had tubes in them. Actually made decent money fixing those old boat anchors and U-Matics.


Still remember the theory from way back then, just had to hunt around on Google till I hit the right site.


Check out this one for a blast from the past!


Just about every VTR format I **EVER** heard of and about 5 gazillion obscure formats I never knew existed.

http://www.lionlmb.org/quad/format.html


Worth exploring:

http://www.lionlmb.org/quad/format.html

 

[Brucer]

Even though this diagram is a couple generations old, I imagine the path is pretty much the same.

 

[GerryL]

Hey, nuke - you really dug up a treasure trove of tape formats!!! Lots of useful info.

I'll keep that web page in my bookmarks.


GerryL