VTR-Project:Umatic DUB to Y/C conversion

Hi !

Since I have a couple of Umatic-VTRs around, I wanted to integrate them in my Y/C-video setup.

The S-Video Y/C-format uses Y & C signals, where the C-part is encoded @ 4.43 MHz (PAL)

To those who don't know - the 2 Umatic - studio VTRs came with DUB-connectors. That was a 7-pin port, that provides

1. separate Y-line (the luminance-signal), which is fully S-Video Y compatible (4.43 MHz)

2. a "color-under" coded C-line, which,depending on the tape-format works at different frequencies (<-and this is the trouble-maker - it's running at 688 - 924 kHz)

The Umatics provided these ports for better-quality machine to machine dubs, in order to circumvent the composite en/de-coding circuitary, which generates quite a loss in vid. quality.
The signals present at those dub-ports are basically as they get recorded on tape..

So I made a bit of investigation :

* According to my handy TV-technician-Cookbook, *all* home VTR are using (recording with) the color-under technique.
A VHS (PAL) for instance 626.9kHz, a Betamax 685.5 kHz & 689.4 kHz

* All VHS & Betamax VTRs are treating it's signal in the following manner:
1. they split it up into the Y & C-components - using BP/notch-filters
2. they modulate the C-component down to 626.9kHz

So all I need to do is use the same technique that a VHS-VTR uses to de/-modulate the FBAS 4.43 into it's internal subcarrier-format & vice versa - then I'm done right ?

* Reading further on, I found out, that the C-conversion is done rather quite simple - by applying some average FM-voodoo.

1. the C-signal gets filtered trough a BP-filter, to ensure we have a clean 4.43 MHz *modulated* signal present.
2. then the signal gets converted DOWN by simply mixing it with
a) a 4.434MHz color-carrier
b) an Oscillator, which is running exactly with a difference of 626.9kHz above the C-frequency (5.061 MHz for instance)
- which leads to a new subcarrier @ 626.9kHz !

* This voodoo is reversible - by simply applying a 5.061 MHz oscillator to the 626.9kHz color-under-coded (modulated) signal !

(- and maybe cleaning it up a bit with a filter - though in theory it's ready to use !)


- since I'll start to build this converter firstly for Y/C (S-Video)-ports, I don't need to worry about filtering, since the C-signal is already present in a "clean format".

- So the only thing I'd need to build is a precise 5.0xx MHz oscillator...

Am I correct about my conversion theory ?

Anyone made a Y/C-to composite AND composite to Y/C-converter - or something alike ?

Anyone out there using umatics, wanting to do (or have) the same thing ?

Anyone made an *precise* 5.00xxx MHz (and I mean precise at the "x"-ses :)) oscillator ?

Should I go the PLL-route - or fiddle around with a quartz ?

How about building a low(er)-frequency oscillator & using clock-multipliers ?

Thanks for any hints in advance & happy holydays to everyone !

In the meantime, I'll play around a little bit with my oscilloscope and some transistor HF-oscillators..
U-matic dub cable

Hello JanK!

I've just come across your ancient posting about a dub cable from U-matic to S video, [located vcia Google] as I'm wanting just such a thing to transfer many tapes to DV.

Did you ever make this device, and if so, did it work?

Are you able to provide me with necessary info or sell the cable/box to accomplish this? (I'm not as tech savvy as you!).

I look forward to hearing from you!

A solution

The archive I work for asked me for a solution to allow U-matic component dubbing. The unit I made has been successfully deployed to digitise some important collections at a better quality than previously possible using composite connection. It was apparent that no off-the-shelf solution exists so I made some more. The are available at the Keystrobe web site called "Dub-optimiser": KeyStrobe quality essentials
It's not rocket science, but it does have carefully tailored filtering for Y and C components for minimal artefacts.

Moderators, please move this post if it better suited to a different forum. I posted it here because it directly related to the thread. Thanks.
Yes you can

As designer of the product you mention, it's entirely feasible to DIY a solution, indeed the product developed from such a situation.
The concept is to extract Y and C before they have been remixed into a composite signal. The reason for this is that they have been processed separately in the machine and are no longer coherent. When added together there is a loss of resolution since the C includes an amount of unfiltered Y. An added gain is that it's possible to extract the Y component before the rather primitive noise coring in those machines.

So what do you need:

The Y signal can be taken from the Dub connector where it can directly drive a 75 ohm line. You could use that directly as Y but it would be better to filter some of the out of band components if you intend to digitise to a lossy format. If any of the playback is of Hi-band tapes (in suitable machines) you will need some gain as the DUB Y signal drops from 1V P-P to 0.5V P-P with these tapes.

The C signal is bit more problematic. You will probably have to dive into the machine and pick up the 4.43 or 3.58MHz subcarrier just before it is added to the Y. You will doubtless need to build a 75 ohm cable driver. The amplitude of this signal is not so critical as the burst is used as a level reference.

Once you have the Y and C signals, make sure they stay apart. The worst thing you could do is send them down a cheap S-Video cable where often the signal cables share a common shield.

The above is how we started out. Unfortunately there are so many different U-Matic models, each with its own locations for chroma takeoff, and you have to go inside.
To make a "plug-n-play" product we extract C from the composite output and pass it though an active filter. It took a lot of effort to get the results as good as the direct C takeoff, so I don't recommend this for DIY.
So there is a little trade off in chroma to achieve a universal plug and play product.
With a good enough filter (which we've achieved) there is no visible tradeoff.

Have you tried building a chroma booster from the "chroma under" pin in the dub connector ...... ?
Yes but it requires a lot more circuitry. The (typ 688KHz) colour under signal from tape is unstable in phase and frequency due to the nature of VCR mechanics. To stabilise the frequency you have to multiply it by ∆fH. Then convert it to 4.433/3.58MHz and stabilise the phase. Inside the machine this is achieved by a Sony custom chip set that's unfortunately no longer available.
There was a product many years ago by another company that used this method but by the size and cost of it I'm sure they used the Sony chip set.
In all PAL machines, the conversion from the low-frequency subcarrier to a regular PAL chroma is made internally.
It is then added to the Y to form the composite signal.
All you have to do is to pick this signal up just before the mixing, maybe add a buffer to be able to drive 75 ohm, and bring it to an external connector