View Full Version : Another Svideo to composite video prefilter


wa2ise
04-03-2016, 04:47 PM
Here's a better pre-filter for taking S-video from a DVD player, CECB or digital cable box and making composite video that will look good on vintage and more modern color TVs without comb filters, known as notch filter chroma-luma separator sets.

You can see the difference looking at these test patterns:
http://www.wa2ise.com/radios/regularcomposite2notchfilterTV.jpg
http://www.wa2ise.com/radios/prefilteredcomposite2notchfilterTV.jpg
The schematic:
http://www.wa2ise.com/radios/Notchfiltercomposite.gif
and simulations of the response of the luma and the chroma to the output:
http://www.wa2ise.com/radios/Notchfiltercomposite-chroma.gif
http://www.wa2ise.com/radios/Notchfiltercomposite-luma.gif
The object here is to keep the high frequency luma from contaminating the chroma in notch filter TVs. To avoid the crawly colors on the ref's shirt in football games. Notch filter TVs don't display the high frequency luma anyway, so you won't miss it.

benman94
04-05-2016, 01:07 PM
I built the previous pre-filter you had posted into my composite video adapter for my CT-100. Now I'll have to update ithe with this version. Thanks wa2ise! :thmbsp:

tubesrule
04-05-2016, 03:31 PM
What is the group delay distortion and return loss for this filter?

NewVista
04-07-2016, 01:51 AM
You could also add a variable Crispener centered at 3.2 MHz?

wa2ise
04-07-2016, 09:45 AM
What is the group delay distortion and return loss for this filter?
It gets to be +-100nsec, but most of that is in the skirt of the stopband. Which means that the frequencies with the severe group delay is also attenuated, so you won't see it.

tubesrule
04-07-2016, 09:20 PM
It gets to be +-100nsec, but most of that is in the skirt of the stopband. Which means that the frequencies with the severe group delay is also attenuated, so you won't see it.

Nice!

wa2ise
04-07-2016, 11:10 PM
I took another look at the group delay, it's worse than I had remembered. :no: I updated the diagrams above. About 250nsec for luma and 300nsec. However, the spectrum that would now be chroma only gets some group delay predistortion and thus it still looks okay on a TV set.
My other filter's frequency and group delay response:
http://www.wa2ise.com/radios/notchfresponse.gif
And the circuit:http://www.wa2ise.com/radios/svideo2notchfilterTV.gif
It's a tradeoff between frequency and group delay.
You'll have to decide which looks better. :scratch2:

wa2ise
04-24-2016, 11:47 PM
More on this thread, about group delay. While the luma and chroma are still separate (as S-video) one can use long lengths of 75 ohm coax to introduce delay in a selected signal to make the luma and chroma match up on the TV set's CRT. Signals running inside coax have delays of about 1nsec per foot of coax. I used about 150 feet of RG6 to delay the luma to get the chroma aligned with the luma, as sen on the CRT. The chroma signal looked to be a bit too far to the right of the B&W, about 1/4 inch as seen on a 24 inch CRT. The 150nsec delay on the luma made it match a lot better. Only real downside is having a big spool of RG6 siting behind the set, feeding the luma into the video modulator.