Capturing VHS through S-video - Anything wrong with this?

[echo1434]

Hi,

I haphazardly discovered that I have better luck with synch issues using S-video instead of composite while capturing VHS tapes.

Here is my gear:

JVC HR-S2901U S-VHS VCR
Monster 3 S-video cable
Monster 2 composite cable
ATI All In Wonder Radeon 9600


Is there anything technically "wrong" about using S-video for outputting standard VHS tapes? Or maybe any possible advantages?

I'm not sure if I see a difference or if it's just in my head. I don't see anything wrong, but I'm much more of an audiophile than a video expert, so maybe I'm missing something...

Thanks in advance.

[davideck]

Using S-Video should provide an advantage because it eliminates the needless mix and subsequent re-separation of chrominance and luminance required by the composite connection.

[lordsmurf]

Sometimes using s-video instead of composite, or composite instead of s-video, can lead to luma errors. It's almost random when it happens, changing wire types fixes it.

[edDV]

Forgive the brain dump but a little historical context needs to be added to this S-Video and TBC analysis. We need to get on the same page to advance the debate.

Pt1: Once upon a time and far far away ...

Original videotape formats (monochrome) recorded the video signal directly. The first commercially successful format was Quadruplex introduced by Ampex in 1955 and used 2" transverse recording. You can look up the details.

NTSC color was being broadcast before the Quadruplex VTR was invented. To achieve monochrome television V&H lock, Ampex developed "AMTEC" the first timebase corrector. This allowed the recorded signal to be played back without excessive horizontal jitter on the home TV.

To expand the format to color recording, Ampex, RCA, 3M and others had to refine the technology to record a full 4.2MHz bandwidth so that the 3.58 MHz NTSC color subcarrier and its sidebands would be recorded at full bandwidth and without excessive phase distortion. While AMTEC was able to produce acceptable stability for monochrome (microsecond stability), color reproduction required significantly greater phase stability (+/- 2 nanoseconds). This was achieved by Ampex with their COLORTEC TBC in the mid 60s and color video recording was achieved.

Direct (full bandwidth) video recording continued on with the analog 1" Type C format 70s-80s and later the 4xsubcarrier (4x fsc) sampled D2 digital composite broadcast format. These two were helical scan recorders that used brute force digital timebase correctors to achieve the high precision required for broadcasting direct color.

A parallel form of helical scan recording used the "color under" technique for recording the color portion of the TV signal. Initially helical scan technology was used for monochrome industrial applications. By the 60s, TV sets were able to accept a wider range of phase jitter. Helical recorders could be manufactured at lower cost with the tradeoff of much lower phase stability (jitter) than required for broadcast and less luminance bandwidth (~2.5MHz). These machines were used for educational and industrial applications.

In 1969, JVC invented the "color under" recording method that split the bandwidth for color and luminance recording. A heterodyne technique was used to down convert the 3.58MHz color signal into KHz bandwidth while luminance was recorded separately in the 0.5-3 MHz bandwidth. This was the first "component" recording technique. On playback, the chroma signal was upconverted back to 3.58MHz and added to luminance to restore the composite signal. The color under process destroyed the phase relationship between luminance and chrominance. Chroma was no longer tied to a precise relationship to horizontal scan frequency and was processed independently at the TV set.

This is essentially the technique used later for U-Matic, Betamax, VHS and 8mm formats. These formats avoided luma-chroma crosstalk difficulties by keeping luma and chroma in separate frequency space both on tape and on the output composite cable.

S-VHS and Hi8 recording formats were luminance bandwidth enhanced versions of VHS and 8mm respectively. While the benefit of increased luminance bandwidth (above 3MHz) was obvious, the tradeoff was going to be crosstalk between luma and chroma in the 3.1-4.2 MHz bandwidth space where the two signals would overlap in a composite NTSC signal. This would be especially bothersome in higher end home TV sets that were capable of resolving >3MHz (>240 lines of luminance resolution). Separation of luma and chroma (Y/C) in this frequency range requires a costly "comb filter". Even if a comb filter is used, the lack of a stable frequency lock between H and subcarrier (fsc) causes additional moving beat patterns when a luminance enhanced VCR is played to a high resolution TV set.

Enter the S-Video connection which puts luma and chroma on separate coax (4 pins). The S-Video connection separates luminance from chrominance so that luminance bandwidth can be increased without stepping on chroma, thus eliminating both crosstalk and the need for a sophisticated comb filter in the TV set.

Meanwhile TBC technology has split between the needs of broadcasters to restore a helical recorded signal to a complete lock of chroma to horizontal scan frequency (needed for post production mixing) and home users who just needed a more stable dub to a second VCR (or in latter years, to a capture card).

enough for now...

Hi,

I haphazardly discovered that I have better luck with synch issues using S-video instead of composite while capturing VHS tapes.
...

Is there anything technically "wrong" about using S-video for outputting standard VHS tapes? Or maybe any possible advantages?

...

Nothing wrong at all. Keeping luma and chroma separate is a good thing. While in theory, VHS luminance should be low pass filtered around 3MHz before recording, the reality is cheap VCRs use cheap filters that may generate noise above 3MHz and even good VCR's may put luminance harmonics into chroma frequency space. If you don't mix them, the noise crosstalk doesn't occur. For SVHS and Hi8 the actual video signals overlap and crosstalk unless kept separate.

The subject gets more complicated if composite is used upstream of recording and even more complicated if the tape is multi-generation.

[pfh]

Brain dump-

Does s-video make the encoder more effecient?

[lordsmurf]

Does s-video make the encoder more effecient?

No. I wouldn't say that.

[edDV]

Encoder or decoder before A/D capture?

If you feed composite NTSC to the capture card, it needs to decode NTSC into Y, U and V components. The hardest and most expensive part of decoding is separation of Y from C (UV). Capture cards are notorious for cheaping out on Y/C separators.

The cheapest way out for the capture card is to low pass filter at 3MHz and call everything under that Y. Then bandpass filter around 3.58MHz and call that C.

This results in tossing any luminance resolution over 3MHz except for the high frequency luminance that is crosstalking into chroma. It's the worst of both outcomes, but alas, that is what we get with a typcial composite input on a capture card. Their advertising motto should be "None of the high resolution but all of the chroma crosstalk noise".

The more expensive approach is to use an expensive comb filter for Y/C separation in the 3.1 MHz up range. New 3D comb filter chip technology has been developed for HD-Ready TV sets. This was needed because NTSC decoding errors are very visible and objectionable on HDTV. These comb filter chipsets are finding their way into capture cards (e.g. ATI TV Wonder Elite).

Now back to a SVHS recorder. The signal already is separated into Y and C on tape. Playing Y/C out of the recorder in Y/C through the S-Video cable bypasses the Y/C filters on the capture card and presents more pure YUV to the capture card's A/D converter-Nyqust filter. If the composite output is used, then Y and C from tape are mixed onto a single wire. Next, the cheap capture card Y/C filters are used to reseparate Y and C. For VHS this may or may not be a problem. For SVHS recordings this can seriously degrade the signal.

I'm still looking for a senario where use of the composite cable gets better results than the S-Video connection for VHS tape.

[pfh]

Yeah, I see what you mean edDV- encoder or decoder. For the decoder, your explanation makes good sense.
My thinking was that, given the fact the encoder would be technically getting a separated signal with less noise then it could do a better job of converting. It wouldn't have to deal with as much noise and the encode would actually be working on more of the true picture. Effecient isn't a good word but for lack of a better..............

[edDV]

Yeah, I see what you mean edDV- encoder or decoder. For the decoder, your explanation makes good sense.
My thinking was that, given the fact the encoder would be technically getting a separated signal with less noise then it could do a better job of converting. It wouldn't have to deal with as much noise and the encode would actually be working on more of the true picture. Effecient isn't a good word but for lack of a better..............

Yes, the MPeg2 encoder works from the YUV digital components after A/D conversion. Any artifacts from the NTSC or PAL decode will show up in the MPeg2 and DVD.

Basic rule: garbage in garbage out. Every step of the process is important.
In analog, repeated decode and encode is a bad thing.

[victoriabears]

I get better results from good tapes using composite rather than s-video connection, it might sound dumb, but could the s video be giving out too much information and over doing the image?

[belloq]

This is my first post here. I've done some editing and burning of original DV footage onto DVD but am just now focusing on transfer to DVD of older VHS tapes and some laserdiscs that will probably never get to DVD. I didn't want to start a new thread unnecessarily, so here's my question:

I'm planning on transfering old standard VHS tapes to DVD through a Canopus ADVC-110 converter/PC or more likely to a Sony DVD recorder. For standard VHS tapes, will using a SVHS player to play the tapes through S-Video cable result in better video than using the standard VHS player/composite cable? Thanks.

[edDV]

In theory, not significantly for VHS tapes*, but it makes a major difference for S-VHS tapes.

In the real world a composite connection has more risk so if you have a S-VHS player it is better to use the S-Video output to the capture device even for VHS, but if you don't have a S-VHS deck, composite will be OK for VHS.

For Laserdisc, the opposite is true. Laserdisc is recorded natively in composite, so the composite output is superior in most cases. For laserdisc transfer, it is important to use a capture device with a high quality comb filter.

Ref:
http://members.aol.com/ajaynejr/vidcomb.htm

* since VHS recordings are low passed filtered below 3MHz, in theory there is no overlap of luminance and chrominance so simple filtering techniques can be used to separate luminance from chrominance for a composite connection. In the real world, since VHS Y and C are recorded separately on tape, it is better to deliver Y and C on separate wires to the capture device via the S-Video output. The capture device will then skip the separation filters and process Y and C directly.

[belloq]

Thanks for the information. I am new to capturing video and burning to DVD so any suggestions are very helpful. So I think I'll transfer the standard VHS with a S-VHS player. Even if it's a minimal improvement, the piece of mind that I'm doing my best is always welcome. I don't plan to transfer these again in the near future.

Regarding laserdisc, so you are saying I shouldn't even try transfer via S-video? You recommend composite only?

Also, how do I know if my capture device has a comb filter- does the Canopus ADVC-110 converter or the Sony GX315 DVD-recorder have a comb filter appropriate for transfer of laserdisc video?

Thanks again.

[edDV]

Laserdisc players output directly to composite from the disc. That is the native format of laserdisc. In later years of production a Y/C separator and S-video output was added to some laserdisc players but the reviewers of the time found the Y/C separators very poor relative to the comb filter on the typical high end TV set of the day.

Most "tuner" type capture cards have notch rather than comb filter. Most quality capture devices have a comb filter built in but the quality varies.

Ref: http://members.aol.com/ajaynejr/vidcomb.htm

Current HDTV sets have gone to very sophisticated digital 3D Y/C comb filters for NTSC/PAL inputs but few capture devices have these yet. The new ATI TV Wonder Elite is one of the few capture devices advertising a digital 3D Y/C comb filter and this would probably work best for capturing laserdisc.

Laserdisc is the only wide band composite source available to the home consumer other than direct over the air analog broadcast. Everything else (VCR, Camcorder, DVD, DVR, PVR, ...) is native component technology. In those cases, component connections generally yield better results. Laserdisc is the exception to the rule.

[davideck]

An interesting test would be to feed your laserdisc composite output to your S-VHS VCR (in EE loopthru mode) and then feed the VCR S-Video output to your capture device. You might find that the Y/C separator in your S-VHS VCR is the best one you've got.

[edDV]

An interesting test would be to feed your laserdisc composite output to your S-VHS VCR (in EE loopthru mode) and then feed the VCR S-Video output to your capture device. You might find that the Y/C separator in your S-VHS VCR is the best one you've got.

Also, some HD Ready TV sets have 3D Y/C comb filters and S-Video outputs of the program being displayed. These produce nice separated video but unfortunately introduce a video delay vs. audio.

[Suds-N-Spuds]

An interesting test would be to feed your laserdisc composite output to your S-VHS VCR (in EE loopthru mode) and then feed the VCR S-Video output to your capture device. You might find that the Y/C separator in your S-VHS VCR is the best one you've got.

Actually, this is how I capture DV from laserdisc. The comb filter in my Mitsubishi HS-HD2000U is significantly better than the one in my laserdisc player or the DV camcorder. To avoid a delay in the audio/video synch I also route the audio through the VCR.

[edDV]

An interesting test would be to feed your laserdisc composite output to your S-VHS VCR (in EE loopthru mode) and then feed the VCR S-Video output to your capture device. You might find that the Y/C separator in your S-VHS VCR is the best one you've got.

Actually, this is how I capture DV from laserdisc. The comb filter in my Mitsubishi HS-HD2000U is significantly better than the one in my laserdisc player or the DV camcorder. To avoid a delay in the audio/video synch I also route the audio through the VCR.

Delay through analog comb filters is manageable at a line to three lines but these digital 3D jobs delay in fields and frames.

[belloq]

Thanks again. I'm going to purchase a new low-priced consumer JVC S-VHS player for the VHS transfers.

Regarding the comb filter for the laserdiscs, what would you recommend as a reasonably-priced comb filter? Or are there no worthwhile filters for someone on a tight budget? Does a low-priced consumer JVC S-VHS player (current model 2902) have such a comb filter? (play laserdisc through the S-VHS player into the DVD recorder) Thanks.

[edDV]

You need to read the specs for the various JVC models. The upper range will have a simple comb filter. Since laserdisc is composite video, quality will depend on the comb filter.

I think you should consider computer capture and authoring to DVD.

Capture cards sometimes have simple comb filters to separate YUV. The only one with a fancy 3D comb filter that I have seen is the ATI TV Wonder Elite

http://www.ati.com/products/tvwonderelite/features.html
http://www.ati.com/products/theater550/index.html

This card has the most advanced hardware for Y/C separation but early reports question the software. I question what they do about video delay vs audio (probably nothing). So consider this card leading edge but you may need to solve technical problems.

[davideck]

I'm going to purchase a new low-priced consumer JVC S-VHS player for the VHS transfers.

Regarding the comb filter for the laserdiscs, what would you recommend as a reasonabley-priced comb filter? .

The JVC SVHS VCRs with TBC/DNR DigiPure have comb filter Y/C separators.
I suspect that you would be better off with a 7600, 9600, 9800, etc., than with a 2902.

[edDV]

Reviews of Theater 550 Pro cards

http://www.firingsquad.com/hardware/ati_theater_550_pro/
http://www.beyond3d.com/reviews/ati/550pro/

[belloq]

Thanks for the reviews. Are there composite or s-video inputs on those cards with the ATI 550? If not, how would I get the A/V into the PC- the coax input? If so, will that degrade the video or audio signal any?

Regarding the video delay caused by the processing, is there a device or software in which you can synchronize or adjustably delay the audio so it syncs with the delayed video?

Also, regarding devices with comb filters, though it doesn't seem to list it in its current description, among other things, the Canpous ADVC 300 used to list digital "3D" Y/C processing- any idea if this is a good standalone device to get the video filtered and converted to digital without dealing with a PC card? Or would you still recommend a PC card?

http://www.canopus.com/products/ADVC300/index.php

[edDV]

Thanks for the reviews. Are there composite or s-video inputs on those cards with the ATI 550? If not, how would I get the A/V into the PC- the coax input? If so, will that degrade the video or audio signal any?

The 550 Pro cards are supposed to come with a break out box with S-Video, composite and audio input connectors. Since audio is processed through the card, delays should be internally handled.

Regarding the video delay caused by the processing, is there a device or software in which you can synchronize or adjustably delay the audio so it syncs with the delayed video?

If the audio paths are separate sync can be difficult.

Also, regarding devices with comb filters, though it doesn't seem to list it in its current description, among other things, the Canpous ADVC 300 used to list digital "3D" Y/C processing- any idea if this is a good standalone device to get the video filtered and converted to digital without dealing with a PC card? Or would you still recommend a PC card?

http://www.canopus.com/products/ADVC300/index.php[/quote]

The ADVC-300 also has some kind of 3D Y/C separator and also internally processes audio. Canopus is very secretive about how their box works and I have no first hand experience with it.

As far as separate box vs. PCI card, that depends on connection method. IEEE-1394 should work OK. USB2 is nonstandard and proprietary.

BTW I ordered a Sapphire 550 Pro and will soon (mid-Jan'06) have first hand experience with that. I'll be evaluating it with BeyondTV4 and GB-PVR and am eagerly awaiting ATI's next MMC that is supposed to work with the 550 Pro. I would never have purchased the card based on ATI software only. Fifty times burned.