s-video artifacts?

[themaster1]

The checkboard pattern is only visible on the menu (normally), sorry to burst to your bubble yabrra/zerowalker the 1st day of 2k13.
If you have a checkboard pattern when playing a tape then your s-video out socket and/or cable may be faulty , it happen a lot especially when the vcr is old, the socket become loose somewhat.

[zerowalker]

Oh, well if it´s only in the menu, it´s just how it sends the picture, if i remember correctly, JVC tends to send out half resolution pictures and stuff, which can mess up.

I thought it was the checkboard i get on my PS2, Toshiba DVD and Xbox, cause those are as i wrote, Composite through S-Video.

[ybarra]

Actually, I've done more research into this and I think the artifacts are comb artifacts.

No, it's not a broken port on the s-video. I have two SVHS decks that do the same thing. Apparently the JVC 3d-comb is only engaged when recording from the s-video ports, passing material through, or using the tuner. Simple playback of tapes does not engage the comb filter. The reason the composite source looks a little better is because the checkerboard is softened during the merging of the signal. I ended up capturing using s-video and virtual dub and then using a comb filter and the checkerboard is gone.

For the record, you can look at captures from others posted in other threads using JVC decks (like the thread on the digital r3 on the 9911 deck) and you can see the comb effect on the edge of his captures.

[NJRoadfan]

Actually, I've done more research into this and I think the artifacts are comb artifacts.

No, it's not a broken port on the s-video. I have two SVHS decks that do the same thing. Apparently the JVC 3d-comb is only engaged when recording from the s-video ports, passing material through, or using the tuner. Simple playback of tapes does not engage the comb filter.

The comb filter is only active on the VCR when using the onboard tuner or composite inputs. Output doesn't require one since VHS stores luma and chroma data separate on the tape. The checkerboard artifacts only show up on "240p" (288p for PAL/SECAM) sources like the VCR's menu, older computers, or video game consoles using Y/C video outputs.

[ybarra]

I wondered if it was menu-only as well, but it is not.

Some of my SVHS footage was done using Panasonic professional decks when I used to work in video production (VITC time code and all). That being said, we tended to record 1 minute of SMPTE color bars at the beginning of the tapes. The upside is that it makes it incredibly easy using clrtools to import my footage because I could nail the colors, brightness, and saturation using the virtual vectorscope. The downside is that I could immediately see the same checkerboard artifacts in the actual video signal that I saw in the menus all up and down the solid color bars.

Zooming to 300% on any area of captured footage that has some solid-ish color areas and you can see the artifacts as well. I don't have any other VCR's to compare with, but both my 101 and 7600U have the exact same artifacts using the s-video out ports on playback regardless of the TBC, DNR, or digital r3 settings. I never noticed it back in the days of CRT's probably because of the actual analog noise + interlace masking it. It's very easy to see now in the 1080 LED world.

I'm going straight from the deck into a Hauppauge 2250 card (artifacts are present when recording into my standalone LiteOn DVD recorder or my DV camcorder so it's not the Hauppauge).

So far I'm using dotcrawl.vdf with both extreme settings and the checkerboard is removed.

[NJRoadfan]

Looks like we are referring to two separate issues. The checkerboard on the VCR blue screen problem I mentioned above is unrelated to VCR playback as I have seen screenshots of SNES captures using S-Video with the same problem.

Does the checkerboard on the color bars go away when you turn off the TBC/DNR system? If so, its likely an artifact of the noise reduction. I can very faintly see it on my captures but one has to really look for it when the video is paused. Its not really visible during playback. I wouldn't be surprised if de-interlacing makes it worse. JVC designed the decks around playback on CRTs, not flat panels.

I guess I should find some VHS tape and record some bars. I have 3 Digipure decks and a AG-1980 I can test.

[ybarra]

I tried a capture with the TBC off and the checkerboard is still there except now it's more zig-zag (probably because the signal gets a bit unstable). I'm curious to see what you get on your decks. Both my digipure decks are 2MB buffer models.

I agree on the not really being bad part. On moving scenes you will not detect it. It does make h.264 compressions larger unless you do some kind of noise-reduction to smooth them out (even something as simple as deinterlacing).

One other question... How do you get the timing right on the interlacing. For instance, sometimes when I capture during a cut scene 1 field will be cut and the other field still on the previous scene causing a mess when I deinterlace or use noise reduction for those specific frames. If I recapture, sometimes it will land perfectly where each field pair match for the scene cut. I know I can split the fields, throw one away, and recombine and it fixes it, but wasn't sure what exactly causes it unless it's just part of the whole analog to digital process.

[NJRoadfan]

One other question... How do you get the timing right on the interlacing. For instance, sometimes when I capture during a cut scene 1 field will be cut and the other field still on the previous scene causing a mess when I deinterlace or use noise reduction for those specific frames. If I recapture, sometimes it will land perfectly where each field pair match for the scene cut. I know I can split the fields, throw one away, and recombine and it fixes it, but wasn't sure what exactly causes it unless it's just part of the whole analog to digital process.

If you are using an external TBC, you basically can't. The tearing you see is due to the difference in timing between the source video's sync and the TBC's sync. Its a trade off of the full frame units. If the timing difference from the source and the TBC's clock drifts far enough, the TBC will add or drop a frame to keep up. The only way to try and avoid this is to use a VCR that can be directly genlocked, a feature limited to professional units. Even then you may have the problem if there are enough breaks in the signal.

[ybarra]

Ahhhhh... That's right. I remember now. It's been a LONG time since I've worked with industrial VTR decks. I remember we used to genlock them.

I'll wait for your tests on one or more of your digipure decks and your Panny. I'll mess around with mine a bit as well and try to record some colorbars and play it back to see what happens.

[ybarra]

I ran some tests using my camcorder and am getting the pattern on it as well. I am suspecting my cables and crosstalk.

http://preservation.bavc.org/artifactatlas/index.php/Poor_Y/C_Decoding_(S-Video_crosstalk)

[ybarra]

NJRoadFan:

Looks like the problem came down to cabling. I dug through my cable bucket and got out 5 different cables. All of the thinner gauge cables exhibited the checkerboard pattern over s-video. The medium gauge cable only showed checkerboard on the magenta / red / blue side of the color bars. I only had one think and short cable in my bucket and sure enough it was clear all the way across. It appears the the lower gauge / lower grade cables were causing the checkerboard because of crosstalk.

[jagabo]

Looks like the problem came down to cabling. I dug through my cable bucket and got out 5 different cables. All of the thinner gauge cables exhibited the checkerboard pattern over s-video. The medium gauge cable only showed checkerboard on the magenta / red / blue side of the color bars. I only had one think and short cable in my bucket and sure enough it was clear all the way across. It appears the the lower gauge / lower grade cables were causing the checkerboard because of crosstalk.

Could you post some samples of the worst and the best? Other tests have shown minimal differences between various s-video cables:

https://forum.videohelp.com/threads/221249-Test-Caps-various-composite-and-s-video-cabl...ight=composite

I've never seen any difference myself either.

[sanlyn]

I see differences in several brands and types of s-video cable (BetterCables having the most noise, Monster having the worst color, etc.) but no s-video cable I ever used has been accompanied by the noise described here -- unless it was coming from elsewhere.

[ybarra]

Here's the results. DSC00099.jpg is the original file in the memory of the camcorder connected via S-video to Hauppauge. med1 and med2 have the worst crosstalk. thin1 has mild crosstalk. med1gold and thick have no crosstalk.

[jagabo]

Thanks for the samples. There's definitely a difference.

[vhelp]

ybarra, i don't know..i wan't to believe you but this looks too good to be true. ever since i started capturing from my directv dvr, i have the similar checker box artifacts demonstrated in some of your images. however, when i disconnect from the dvr and go directly to my vcr, the symptom is gone. so that leads me to believe it is the dvr unit causing the artifacts. the results vary from different capture cards, pci and usb2 when connected to s-video leads.

so, i would ask, what model s-video leads are you using and where can i pick up a pair to test on my dvr and other capture cards ? i will need approx 20 feet or i can split them using a 4-button splitter box that i use currently. i will order them first thing tomorrow if they can be purchased through amazon, thank you in advance.

[sanlyn]

Thin cheap cables of every type always have problems of some kind, especially the skinny stranded-core stuff. I still use my old standbys, but they've disappeared from the market. The best I've found to replace them are a new design I've just started using, with very clean results. Not that cheap, but I just ordered 2 more.
http://www.bluejeanscable.com/store/svideo/index.htm

[ybarra]

vhelp:

Keep in mind the checkerboard I am seeing is coming from my Digipure JVC SVHS decks going into either a Hauppauge card or DV passthrough from my camcorder. In such cases, the signal should have been clean from the JVC circuitry across the Y/C cable and recombined at either the Sony DV camcorder or the Hauppauge. The checkerboard ended up being from crosstalk in the cable. As for your DVR, I cannot guarantee that the signal is clean coming out the s-video ports. Also, I'd be skeptical of a 20' run or any kind of splitter. I'm using a 2 foot run to do my captures.

[sanlyn]

Also, I'd be skeptical of a 20' run or any kind of splitter. I'm using a 2 foot run to do my captures.

I use a 25-foot s-video from my digital cable box to an extra TV across the room. No problems. Depends on the cable.

[vhelp]

yeah, unfortunately, i can't move the vcr from where it is. it is my main unit for connecting to and watching everything from

my satelite dish inside my apt is in my living room corner window, the dvr is next to it, but i have it connected to it, so in order for me to watch tv on my pc, i have to run a long cable but it was easier for me to interconnect a few via switch box. actually, i have two 4-way switchboxes in order to make it to my pc. its clumsy, i know..but works for me. i'll post a few images of the problem in case the error/artifact is determined something else. and i still watch everything on my 25 year old 20" tv set.

also, i forgot to mention, if i set my resolution on my pinnacle studio av/dv card to 352x480, the artifact is gone. i have so many capture cards from pci to usb...pinnacle studio, haup 2250, pinnacle usb, stony trv22 cam, and so on. its my hobby.

[sanlyn]

Wait a minute, are you referring to those manual push-button a/v switch boxes?

No wonder.

[Brad]

Still looking for a reasonably-priced, good S-Video cable locally. I could swear the same brand of dollar store cable worked fine in another environment (RV park vs apartment downtown) so I guess it depends on the local EMI.

AVISource("Sony STK1160 - S-Video.avi")

AddBorders(2,0,2,0,color_white)
StackVertical(StackHorizontal(UtoY(),VtoY()))
Histogram()

ConvertToRGB().AddBorders(0,20,1,0,color_white).Subtitle("DVP-NS50P -> STK1160 S-Video (dollar store)",size=16,x=2,y=0)

Trim(175,0)

AVISource("Sony STK1160 - better S-Video.avi")

AddBorders(4,0,0,0,color_white)
StackVertical(StackHorizontal(UtoY(),VtoY()))
Histogram()

ConvertToRGB().AddBorders(1,20,0,0,color_white).Subtitle("DVP-NS50P -> STK1160 S-Video (grey spiral)",size=16,x=2,y=0)

Trim(122,0)

[Brad]

Apparently one has to be very careful shopping for video cables in a computer recycling/thrift store.

The "grey spiral" cable is in fact an Apple ADB cord, which Wikipedia claims may be used but could deliver lower performance because of the impedance. And I got another one that certainly looks like an S-Video cable, but it's only giving me a black and white image.

[jagabo]

That's some really bad echoing on your cheap s-video cable. That would not be caused by local EMI.

[pandy]

Guys if you can't buy cable - there is always way to make it by your self - buy required length for RG-179 coaxial cable, two S-Video plugs and a bit of time, solder, soldering iron, flux and this will be best cable ever.

To have one cable made from two parallel RG179 you can use heat shrinking tube/s.

[SixFiftyThree]

Save yourself the trouble and go with Blue Jeans Cable, as Sanlyn advised earlier. They're unbeatable for the price, if you can afford it.

[sanlyn]

The Bluejeans Y/C s-video seems pretty clean to me. They sure look better than the cheapies I tried, and they did a dance around noisy and much more expensive stuff. My favorite s-video was a budget job with a unique connector that was the only thing that distinguished them from others. But they seem to have disappeared from the market. When I saw this happening a while back, I bought up every one I could, from Circuit City, Amazon, and monoprice. I also tried the Acoustic Research Pro-II, which had decent color and contrast and very little noise, but the image is a bit soft to my taste. The cheaper blue-jacket AR's are a no-no. I've been thru so many s-video wires from $0.99 to over $300 (I returned the later immediately) that I filled two big plastic milk cases with them. They were all rejects that I tried again and again on different material and different video and capture setups, and have been discarded. The worst s-video I've seen were from Monster and BetterCables. But you're welcome to try.

Test patterns do give some important info about a wire. In the end, playback of real-world video is the final test. The same is true of wire, cameras (ever see an analysis of the noise in Canon's popular digital cameras?), amplifiers, and everything else. The old Dynaco line of amps would test out with high levels of IM distortion. But they sell for up to $10,000 today, are still used in many cinema houses, and are prized by many audiophiles. Test patches certainly do tell you something, but not everything.

[vhelp]

i looked closely, and the one marked 'store' had more noticable hatches than the cheaper ones, though both have problems. not saying that you have poor cables. but i don't know your method/process with respect to your test pattern..how it is played back and recorded and captured. my problem with test patterns these days is that people are using "used" patterns, i.e., jpegs made into mpegs etc., and have mosqueto and other artifacts or noise in the test patterns. patterns that have flaws in them imho gives skewed results about a captures capabilities. these patterns need to be properly tuned for accurate results. i'm working on one pattern now, (cleaning it up, rebuilding some of the details to remove the noise etc) hopefully will have it finished today so i can test it with some equipment.

[jagabo]

i looked closely, and the one marked 'store' had more noticable hatches than the cheaper ones

Those are echos of the main signal. Contrast enhanced:



What I don't understand is how they are at 2x the frequency of the signal.

[SixFiftyThree]

At first sight they deliver an image with a "look" that seems a bit tame, until you realizie that they aren't really "doing anything" to the signal. They just pass it.

Precisely what makes a good cable for the purpose of video capture.