Svideo x Composite

[ADLANCAS]

I'm a little bit (or so much) confused!
Maybe someone could explain me.

Basic question:
Since Svideo output produces better signal to capture card...In theory, should a Svideo captured file be bigger or smaller then a Composite captured file, both using of course the same source and for example Huffy compress in YUY2?

Thanks,

Alexandre

[tedkunich]

It depends on if you are saving with compression or not - If no compression is used, they will be the same size, if compression is used typically the S-video capture will be larger since there is more information in the signal. As as example, using the same LD, capture app, card, and settings (Huffyuv) I get a 3.1:1 compression ratio with the composite signal and roughly 1.8:1 with an S-video connection. You can visually see that the S-video connection has more detail than the composite by looking at the raw captures.

[ADLANCAS]

Very clear...
The question comes from a new (maybe broken) VCR. You gave a good example.
Thanks a lot.

Alexandre

[ADLANCAS]

Something is really wrong in my VCR.
I tested with a new tape VHS from the market recorded at professional way, results:
Clip range, Capture card, Huffy, aplication and settings are the same:
Svideo - Compression ratio = 2.7
Composite - Compression ratio = 2.9

Just a little difference in values. Difficult to find a diference in AVI.

Some one has an idea ?

Thanks,

Alexandre

PS: before someone post that it´s not allowed make a copy , and so on...I´m only making a 1 min file for testing my new vcr:D
I hope that is no problem to check my eletronic device!

[FlimsyFeet]

Just a thought, maybe the composite capture has more noise in it?
(noise doesn't compress well)

[ADLANCAS]

I really don't think so, because I tested composite capture with another VCR and values are nearly the same.
Ideas ?!

[Arachnotron]

The difference for VHS tape is not as big as for LD or DVD.

VHS tape only has details up to 3 MHz. The color carrier frequencies for NTSC (3.85) and certainly for PAL (4.43 MHz) are above that, so not much detail is lost in separating luma and chroma from a composite VHS VCR signal.

LD, DVD or S-VHS tapes on the other hand have luma frequencies up to 5 MHz or even higher. So the chroma carrier signal is right in the middle of the luma fine detail, and separating them out will allways damage details to some extend. This does not happen with a s-video signal.

A small complicating factor in this is that separating chroma and luma from a composite signal by a capture card will generate some extra single pixel noise on sharp color changes, and this noise probably will compress badly.

[ADLANCAS]

Do you mean that, in theory, it can be accepted this little difference betweeen compress ratios ?

Thanks,

Alexandre

[Arachnotron]

Do you mean that, in theory, it can be accepted this little difference betweeen compress ratios ?

Yes, for VHS the difference will not be as large as for other sources. This does not mean however you should not use a s-VHS deck and s-video connectors if you have them available, since there still is a quality difference. It is just not as big as with high bandwidth sources.

By the way, compressability is not a very reliable way to asses noise levels or image quality, so don't get hung up on these numbers.

@tedkunich

LD is perhaps not the most lucky example to illustrate this. Unlike (S)-VHS and DVD, LD actually stores a composite signal. Which means that for LD, there is a big difference between composite and s-video signal capping.

In the first case, the composite signal is passed on unchanged to the capture card, where it is digitized and combfiltered into separate chroma and luma values.
In the second case, the combfilter in the LD (if it has one) does the luma chroma separation, and the capture card only has to digitize the separated luma and chroma signals.
So with LD, you are not only looking at the difference between s-video and composite; you are also comparing the difference between the combfilters. Considering the age of some LD players, it may even be that a modern cap device has the better combfilter and you are better off using the composite.

[FlimsyFeet]

I suppose when capturing from Laserdisc the choice of S-video or composite would also depend on the quality of your interconnects - with cheaper cables composite would have a higher signal loss than s-video.