s-video artifacts?

[sanlyn]

zerowalker, if you use one of the aforementioned Toshiba DVD recorders as a TBC, you can feed it composite input and receive high quality S-Video as output. sanlyn has only good things to say about its Y/C comb filter, and my own DVDR3475 has a pretty good filter as well. I'm using the original camcorder for to play back my VHS-C tapes: It only has composite output, but the DVDR375 seems to eliminate all of the dot crawl and most of the rainbowing. The Toshiba's filter might be even better. In the end, I can get better quality from my composite camcorder than I can from my S-Video VCR!

I don't know that the Toshibas are "better" at filtering, tbc, etc., but I see no AGC effects through them. If their AGC is doing anything, it's keeping levels within a respectable range. There seems to be no effort at "normalizing" dark and bright scene changes. If the video has, for example, a bright scene followed by a dark one, followed by another bright one, I see no blooming or brightening/darkening, or pumping for those scene changes. In some respects that's sometimes a problem, as I get many cable broadcasts with really outta sight, grimy lower-than-zero blacks and burned-up brights. The Toshiba seems to effect a limiting factor at the extremes, but I've never seen it try to "balance" an incoming video to some norm in between. My Panasonic ES10 and ES20, I can't say: it's been forever since I used them. But I did record an old Marx Brothers VHS onto the ES20; that video had washout on some scenes, jet blacks in others, but the Panny didn't seem to try to clean it up--maybe because I had my PA-100 proc amp feeding it and had adjusted luma levels with it. Still, there was no attempt at adjusting across scene changes. Its tbc obviously did some work, though. It was a really grainy old tape, so the recording DNR cleaned it up a bit. But later I captured it to AVI with that DNR inactive (I don't think Panny's DNR works on pass-thru, but I turned it off anyway). The AVI post-processing obviously looked better.

[sanlyn]

You can probably factor this into two questions:

• Do the Y/C comb filters on PAL equipment work as well as the ones on NTSC equipment? Hopefully someone with PAL experience can weigh in here.
• Are there PAL versions of the DVD recorders we're talking about?

Hopefully the answer to both is yes, but I can't say...

If memory serves, WhatHiFi in the UK reviewed all of those old PAL Toshibas and gave them rave reviews. In fact, when I was researching DVD machines for purchase back in 2003/2005, I went by the European reviews. UK reviewers are more impressed by image quality than the boys at CNET over here, who favor features and buttons that don't do much for picture quality. That's a vast generalization, but since then some U.S. review sites have got into more critical detail. The NTSC-vs-PAL model numbers are similar, but I recall seeing a suffix letter or number after the model to distinguish them from the comparable NTSC versions. That goes for the "XS" versions as well.

As Orsetto has noted, later Toshibas after 2005 or 2006 are the same junk being sold today.

[lordsmurf]

- Panasonic ES10 DNR does work on passthrough when enabled.
- Toshiba DVD recorders as a TBC seems odd to me. There are better real TBCs.
- DVDR3575 is a nice ATSC HDD DVD recorder, but there nothing inside that I'd consider to be a TBC.

Threads like this are deja vu for me, because I know I've discussed the topics at length in the past.
Search digitalFAQ.com and search forum.videohelp.com

[Mini-Me]

- Panasonic ES10 DNR does work on passthrough when enabled.
- Toshiba DVD recorders as a TBC seems odd to me. There are better real TBCs.
- DVDR3575 is a nice ATSC HDD DVD recorder, but there nothing inside that I'd consider to be a TBC.

Threads like this are deja vu for me, because I know I've discussed the topics at length in the past.
Search digitalFAQ.com and search forum.videohelp.com

Threads like this are deja vu for me too, because I've read all of them, both here and at digitalFAQ.

lordsmurf, you're mistaken about the DVDR3575's TBC capabilities. My DVDR3475 has a TBC in almost every sense except for defeating copy protection, which it doesn't do (at least not reliably). Its TBC capabilities are as follows:

• It DOES correct waviness and horizontal jitter at least as well as my AG1980's field TBC, and it does so far better than my AVT-8710 (which I know you consider to be a true TBC). This is a TBC capability.
• It also DOES frame sync, and it does so far more reliably than my AVT-8710. This is a TBC capability.
• Unlike my AVT-8710, it doesn't tear the image at the top or side or anywhere else...at least so far.
• It MIGHT help with vertical jitter like the ES-10/ES-15, but I haven't tested that yet with any of my nasty tapes. All I know so far is that I haven't seen any vertical jitter while using it.
• I'm pretty sure it rewrites the sync information as well: If the DVDR3475 left the sync information unchanged, then chaining an AVT-8710 afterward would make the image geometry look much like it does using just the AVT-8710 alone, without the DVDR3475 (because the AVT-8710 would still have access to the original sync information). That's not the case; chaining the AVT-8710 afterward doesn't have a big effect on the image geometry, although it does reintroduce very slight, subpixel-level micro-jitter at sharp edges. The DVDR3475 may not clear all copy protection errors from the sync information, but that doesn't mean it's leaving the sync information alone. Instead, it may actually be rewriting it all and reintroducing the copy protection garbage. That's not ideal behavior, but it's not passive behavior either.

If that doesn't count as TBC performance, I don't know what does. Am I 100% sure that it technically outputs continuous sync at all times? No, I'm not, and I'd need an oscilloscope or something to test it. What I do know is that it corrects waviness completely (i.e. actual time base errors) and is generally more effective at frame syncing than my AVT-8710...and that's what really matters in practice, is it not?

I understand that a "true" external TBC is generally much more hostile toward copy protection and much more inclined to wipe out the garbage, but if that's your litmus test, then Sima machines would count as TBC's, but the field TBC on the AG1980 and the line TBC's on JVC decks would not. For the purpose of correcting geometric distortions and frame syncing, the DVDR3475 is a strong contender, and multiple people have said the same of Toshiba RD/DR models. The biggest downside of the DVDR3475/3575/3576 is that it requires a proc amp to reduce the input contrast, or its AGC starts acting up.

In the case of Toshiba DVD recorders, what better real TBC's would you suggest? I know from personal experience that the AVT-8710 does not fully correct geometric distortions in the image, and I know from video clips on this forum that the Datavideo TBC-1000 suffers the same mediocre performance here...and once you use them, they bake the remaining waviness into the image, so it can never again be corrected by a better performer. In the case of the TBC-1000, here is a video comparison with and without it: It helps, but it only delivers C+ performance, and this is the same level of half-hearted correction I get from my AVT-8710 (when my DVDR3475 eliminates every bit of waviness I throw at it). In terms of line TBC performance, both of the commonly recommended full-frame TBC's are lacking...but the DVD recorders we're talking about are not.

In fact, my AVT-8710 doesn't even properly sync frames: If I take multiple captures of a tape, every single one will have a different number of frames, and I have to carefully step through the entire video to figure out where it's messing up (not to mention all the tearing and reordered frames in dicey sections). I know the frames are being lost/duplicated by the AVT-8710 for two reasons:

• Virtualdub shows 0 dropped/duplicated frames.
• If I take multiple captures and pass them through the DVDR3475 instead, they usually stay in sync. The only place where it ever drops frames is during unrecorded/static sections of tape. That's totally inconsequential if you're only taking one capture, and it's child's play to find those sections and realign your clips if you're wanting to take an average or median of more than one capture. In contrast, my AVT-8710 may technically "output continuous sync," but that means absolutely nothing to me when it gets easily overwhelmed by errors in the signal and drops/duplicates frames wherever it feels like it. It's totally unpredictable in the areas that actually matter, using the same capture PC.

I will concede that my AVT-8710 might be defective/broken in terms of its frame syncing underperformance, and it's possible that a properly working one would give perfect frame sync all day, every day. If that's the case, I'd agree that a properly working external TBC would be marginally better than the DVDR3475 in this area. Nevertheless, that still wouldn't change the fact that I've never seen a full-frame TBC correct geometric distortion/waviness/jitter even half as well as my DVDR3475 can. There's just no contest. Highly recommended VCR's can do it almost as well. Certain MiniDV camcorders can do it at least as well. The AVT-8710 and TBC-1000 can't even come close.

Maybe a full-frame external TBC exists that flat-out beats these DVD recorders in every category, but I haven't seen it, and it's not the AVT-8710 or TBC-1000. If it exists at all, it's probably going to be a much more boutique and expensive higher-end model. I've read an enormous number of your posts on this forum and digitalfaq, and I'm not ignorant of your view. I'm sure there are DVD recorders that are only "TBC's just because it says so on the box," as you suggest. Perhaps this is even the case for most of them. However, this characterization simply does not apply to the models we're discussing. They're not called TBC's because it says so on the box. They're called TBC's because of demonstrably effective - and even demonstrably better - performance. They may not eliminate copy protection, but in every other measure of TBC performance, my DVD recorder is getting the job done...which is more than I can say about my full frame TBC.

[sanlyn]

I use the Toshibas strictly as line-tbc and y/c comb filter for composite. Whatever its tbc is doing (or not), it ran circles around my old JVC 7600 and 9900. Beyond that, I use a full-frame external tbc when needed. Granted, using a DVD recorder as pass-thru isn't the ultimate in tbc's. But it's here, it's paid for, it's also used as a regular DVD recorder. I've been to the pro video department at B&H and looked over the prices for studio tbc's. Just couldn't afford that second mortgage to buy one. But I've no doubt they're superior to any pass-thru device. Now, if I were running a pro shop for video restoral, I'd be in the same waiting line with LS to buy the latest and greatest. No doubt about that.

[sanlyn]

sanlyn can you give me an example of direct composite to your card, and composite-s-video through a DVD Recorder please?

And well, my card is very bad with the Y/C Separator in my eyes, as i get lines of dot crawl that isn´t even possible to reduce.

And well, as i see it S-video from the start must be the best, but of course cheap stuff is different, i want quality, but not a fortune.
But if i can see some examples i might be able to see if it´s worth it

If you think turning on a y/c filter is "in your eyes" and creates dot crawl, you have it in reverse. Meanwhile, I'm uploading some samples for those who have the time (and who doesn't have time, hmmm? ). These caps and clips are samples using composite, s-video, and Panasonic and Toshiba as pass-thru tbc's. The tape is copy protected (The African Queen), is 20 years old, is all beat up and has been played to death. Not old captures, I made these today. No luma oir color work, no denoise, just straight capture to YUY2 huffyuv AVI.
Player: Panasonic PV-4662, circa 1996. Before Panasonic got into excessive over-sharpening. I tried to get some frames with wiggly lines, etc., but the 4662 tracked too well. My PV-8661 would likely have wiggles and oversharp, but I didn't have time to dig it out of storage. In a post below I'll show a sample of repairing tears on damaged tape.
Composite: Acoustic Research Performance Series ('blue jacket").
s-video: My favorite cheapo generic ("RCA" branded) cable. No time to try others.
Capture: VirtualDub, no filters, ATI All In Wonder 7500 Radeon 64MB, 640x480.
Pass-thru: Panasonic DMR-ES20 (refurb, purchased 2003) or Toshiba D-KR2 (refurb, made in 2004)

First, a note about using the Panasonic ES20 for pass-thru. The darn thing puts out some vague rolling lines. They change rolling speed now and need. I mentioned this earlier. The lines show up in dark areas and large bright areas such as sky. I see this on the ES10, ES15, and ES20. Here is a cap:


Direct link to about 6 seconds of this noise (NTSC/MPG):
http://dc451.4shared.com/download/jGGZkxXH/B1_Panny_noise.mpg

This image is from the capture using composite cable, straight into the capture card. No tbc. Indeed, it is definitely copy-protected. Straight playback with no tbc makes it look dingy and noisy. Typical effect.



A fairly clean frame using composite-only:

NTSC/MPG video of this scene, composite-only, no tbc (copy protection effects):
http://dc452.4shared.com/download/sSlm6D3n/A_straight_AR.mpg

The Panasonic ES20 wasn't affected by copy protection. NOTE: It don't always happen this easy. Depends on the tape. Sometimes the full-frame TBC is required. ES20 with composite-only, tbc pass-thru, DNR on:

NTSC/MPG video of the same scene:
http://dc387.4shared.com/download/LUfsQ845/C_Pan_AR_DNR.mpg


ES20 as pass-thru, composite in/svideo out, DNR on:

NTSC/MPG of the same wscene, composite+svideo:
http://dc430.4shared.com/download/eeoX5UGD/D_Pan_AR_SV_DNR.mpg

The Toshiba wasn't affected by copy protection. Toshiba D-KR2 as pass-thru, composite-only, no DNR, y/c turned on:

NTSC/MPG, Toshiba as pass-thru, composite-only:
http://dc302.4shared.com/download/CJBLyYDv/E_T_AR.mpg

Toshiba D-KR2 as pass-thru, composite in/s-video out, no DNR, y/c turned on:

NTSC/MPG of the same scene, Toshiba as pass-thru, composite+svideo:
http://dc384.4shared.com/download/KBzH-X8o/F_T_AR_SV.mpg

Sorry for the 22Mb size, people, but a mere few seconds doesn't say much about the differences. In many respects, they look alike. In other respects, they don't. Depends on the tape.

[sanlyn]

Sample1: working with bad tape that displays tearing. The VHS tape is Swing Time (1934). Tearing along the top. All played on the PV-4662. I fiddled with tracking, but it was no use. The Toshiba cleared almost all of it. Composite in / svideo out on all:

G: AVT-8710

NTSC/MPG video of this scene: http://dc260.4shared.com/download/7q2vE55g/G_Swing_AVT.mpg

H: Panasonic DMR-ES20 as pass-thru:

NTSC/MPG video of this same scene: http://dc365.4shared.com/download/8paBbmRV/H_Swing_Pan.mpg


I: Toshiba D-KR2 used as pass-thru

NTSC/MPG video of this same scene: http://dc193.4shared.com/download/xFR77wdE/I_Swing_Tosh.mpg

Sample2: A better fix (sometime you just can't explain why). A football broadcast on ESPN, a VHS recorded at 6-hour rate probably on a Panasonic over SD digital cable. Abused tape, football fanatics like to play with fast-forward/replay/repeat and ruin tapes. These captures also used composite-in/svideo-out (same cables as above). No color or levels work, I had to get this VHS transfer to my nephew for his graduation, and the noise and damage were enough to handle on their own. This was 4 or 5 years ago. The borrowed tape was played on a Panasonic PV-8661. Pass-thru: Toshiba RD-XS34. No full-frame tbc required:

The "bad" sample clip was reduced in size and given borders, so the full damage wouldn't be hidden by a CRT's overscan. 352x480. The "fix" sample was not resized (704x480 NTSC). The only filter used on the "fix" version was NeatVideo.

Two images from the "bad" initial capture, no tbc:


2 images from the fixed version, with Toshiba tbc pass-thru


damaged video: http://dc201.4shared.com/download/XkrgWcwB/Sample2_bad.mpg
fixed version: http://dc432.4shared.com/download/Lh8bM50i/Sample2_fix.mpg

[zerowalker]

Wait so the passthrough can´t help with dot crawl?
You can see my dot crawl in the other Thread in all the clips that are from composite, and it´s very bad if you ask me.

[sanlyn]

In itself, a line-tbc has little effect on dot crawl. Using s-video output will help, and a y/c filter helps more.

[zerowalker]

Well, i olnly really need a good reduce in dot crawl;S
So i thought if it had high quality yc seperator, it would help alot, but if it´s pretty much the same as my card, it´s not what i look for currently

[sanlyn]

Your capture card has a y/c comb filter? The Toshiba in the samples above has a y/c comb filter. You can see plenty of edge problems in the ESPN video sample posted earlier. That was badly damaged video that should have had much more cleanup that I gave it (no time for it, though).

[zerowalker]

I don´t think it had that;S
but both your capture looks nice, atleast on the dot crawl parts, i got lines all over the place;S

[sanlyn]

Every VHS capture will have some noise, and more than one kind. Your "S-VHS" avi posted earlier had edge ringing and color bleed. There's no pass-thru or tape player that will catch all of those problems, unless you can afford studio equipment. I sure can't !! .

[zerowalker]

Well the only thing i complain about is the dot crawl really;O
the other can be "fixed" one way or another, pretty good atleast;D

[sanlyn]

Dot crawl can be fixed, too. Take a look at the black-and-white video made with the AVt-8710. Plenty to clean up in that one.

[zerowalker]

well not afterwards, not with mine atleast;S
that is why i though the passthrough would pretty much clean it up.

[sanlyn]

tbc's aren't noise smoothers, de-ringers, or de-blockers. They are used to correct sync, line and frame timing issues. Those are problems that can't be fixed "later". If you capture a video with bad tears like those shown, you can't fix them with filters. The samples I posted don't display any cleanup of major color problems (except those caused by Macrovision). If you watch them compared to the original, you'll see a smoother flow of video, less flickering on some of them, and other visual differences. Very few of those problems can be fixed after capture. One of the biggest glitches with the videos I posted is that they're telecined. A tbc won't fix that, either, but it will help clean some of the combing and jaggies that come with bad frame timing on interlaced video.

[sanlyn]

well not afterwards, not with mine atleast;S

Yes it can.

[zerowalker]

Is this the one you are reffering to?
as i am not talking about deblocking or stuff like that, that was for an MPEG2 capture.

https://forum.videohelp.com/attachments/10892-1329060287/Normal.avi

[sanlyn]

Normal.avi has a lot of problems, mostly from oversharpening. Play Normal.avi and watch the straight lines of doors and corners along the walls. Notice anything about them?

[zerowalker]

yeah, the halos around everything;S
Well isn´t the straight lines dot crawl from composite?

[sanlyn]

The ahlos are oversharpening artifacts. That's obvious. Look at the lines of the doors and corners. They wiggle.

[sanlyn]

It's Valentine's Day and I'm taking the wife to the movies. See you maniacs a little later tonite.

[zerowalker]

Well i guess that will be solved with TBC, but that isn´t the issue, it´s the lines on the floor and all over the place;S

Have Fun

[lordsmurf]

lordsmurf, you're mistaken about the DVDR3575's TBC capabilities.

Nope. I've had one since 2007. Ran all sorts of tests on it multiple times.
Great ATSC recorder, terrible for working with VHS tapes. It doesn't correct jitter, chroma, frame sync -- nothing.

I will concede that my AVT-8710 might be defective/broken

You may have one of the known defective models.

[jagabo]

Well isn´t the straight lines dot crawl from composite?

You need to be more specific when you ask questions. There are all kinds of "lines" in your sample. Which ones are you talking about?

This motion is horizontal jitter caused by time base errors with video tape (2x point resize, reduced to 4 fps):



The door should have been perfectly straight and still. That is what a line time base corrector is supposed to fix. Full frame TBCs may or may not include a line TBC. The TBC in the Panasonic ES15 DVD recorder includes a line TBC:

https://forum.videohelp.com/threads/319420-Who-uses-a-DVD-recorder-as-a-line-TBC-and-wh...=1#post1983288

I'm not sure what these slightly diagonal lines are (4x point resize):



But I think they are related to dot crawl. Maybe from a poor dot crawl filter. Using some other device to perform the dot crawl removal (and then running s-video from that device to your capture card) might help -- if that device has a better dot crawl filter than your capture card. The best solution, of course, is to use an S-VHS deck with s-video output so the video never goes through a composite stage.

Dot crawl cannot be 100 percent removed in software or hardware without damaging other parts of the picture. That's why it's best avoided in the first place.

The bright line to the right of the black line is a halo (2x point resize):



That bright white halo (and the slightly bright halo to the left of the black line) wasn't in the original picture. It was caused (mostly) by a sharpening filter in the VHS deck. VHS is inherently low bandwidth (horizontal resolution) so many decks include a sharpening filter in an attempt to make the video look sharper. The result is oversharpening halos.

[Mini-Me]

lordsmurf, you're mistaken about the DVDR3575's TBC capabilities.

Nope. I've had one since 2007. Ran all sorts of tests on it multiple times.
Great ATSC recorder, terrible for working with VHS tapes. It doesn't correct jitter, chroma, frame sync -- nothing.

That's EXTREMELY weird, because my DVDR3475 absolutely does, and Davideck's experiences with the DVDR3575 are more like mine than yours. You're too experienced to make a beginner's mistake like putting another TBC in the chain beforehand, so I don't know how to explain your results. Maybe your particular unit has defective circuits?

Directly capturing a home movie from my camcorder causes all kinds of waviness, whereas passing it through my DVDR3475 results in a rock solid picture without a hint of it. The difference is all over the image, but it's a totally "night and day" contrast if you look at the timestamp in scenes where the camcorder was in motion while recording.

I will concede that my AVT-8710 might be defective/broken

You may have one of the known defective models.

Quite probably, yes...I agree with you that my AVT-8710 should be much better at frame syncing. Nevertheless, that still doesn't change the fact that external TBC's like the AVT-8710 and TBC-1000 are only mediocre at best at correcting image waviness from time base errors.

[sanlyn]

Well i guess that will be solved with TBC, but that isn´t the issue, it´s the lines on the floor and all over the place;S

Have Fun

There are filters that will help clear the noise along the bottom border (that noise isn't dot crawl). The halo's and edge artifacts are pretty serious, but they can be ameliorated somewhat, at least they won't be so obvious. What can't be helped is the line wiggling and shimmering, to the point where it looks as if "wrinkles" are continually swimming through the video. Without a line-level tbc in the player or in a pass-thru device during capture, those distortions are permanent visual elements of the capture, and they'll look worse when played on tv.

[sanlyn]

The door should have been perfectly straight and still. That is what a line time base corrector is supposed to fix. Full frame TBCs may or may not include a line TBC. The TBC in the Panasonic ES15 DVD recorder includes a line TBC:

As does the ES10 and ES20 -- along with RF noise from their poorly filtered power supplies. I don't know if anyone noticed (probably not) but the Toshiba's line tbc did better work fixing tears than the Panny did. I noticed that about the high-end Toshiba VCR's from the late 90's, whose built-in TBC's made JVC, et al, look sloppy -- and without eating tapes. I should never have sold mine, but . . .

[jagabo]

The door should have been perfectly straight and still. That is what a line time base corrector is supposed to fix. Full frame TBCs may or may not include a line TBC. The TBC in the Panasonic ES15 DVD recorder includes a line TBC:

As does the ES10 and ES20 -- along with RF noise from their poorly filtered power supplies.

Mine doesn't seem to have that problem. Not like your earlier sample. You can see lots of sample caps in the thread I linked to.