Practicality of S-Video out
 

Hey. I read recently that the analogue image that comes out of the XL2's S-video port looks quite abit better than the image recorded to either tape, or via fire-wire. Now, how feasible is it to record onto a laptop live, through an S video port. What type of external graphics cards can be hooked up to do this, and wouldnt then running the signal through the 1394 port look the same as doing it through the firewire in the first place?

am I making sense? Harder to type more detail, but hand is still broken.

Matt

From experience the difference between the S-video and DV signals on consumer cameras is negligible. Especially when viewed on a consumer monitor (tv set).

I'm not sure why you want to record via S to a laptop when firewire is the best and most readily available option ? You'll need to install an analogue capture card with an S-video (Y/C) port, as well as having the necessary analogue audio capture card with phono/RCA throughputs.

I've seen visual proof that capturing direct to an analog capture card via s-video out of the XL-2 looks better than firewire from a chroma smoothness standpoint. I was discussing this some time back and another member sent me video clips. There's an un-confirmed rumor that the s-video out is pre-compression like the analog component outputs of the Sony Z-1.

-gb-

If this is true though, one would have to capture straight to disk in order to get a pre-compression signal, since compression would occur in the signal chain before the camera writes to tape. Right?

I've read a few posts around here that would suggest that even if the video has been recorded to tape, analog captures via S-video still contain more chroma information than digital captures. Why would this be? I haven't been able to figure out how that could be possible.

I'm sure it's as much of a mystery to anyone else as it is to me, but any theories? Seems kinda strange.

After it's on tape, the compression has definitely taken place. It's possible, but unconfirmed that the live s-video signal never goes through the compression. I have seen side by side comparison of a shot taken via analog capture, and the same one from FW. The analog picture had much less color banding and smoothness on edge detail of the talent's shirt collar. IOW, it wasn't such a minor improvement that I need to use scopes to see the difference. It was visible to the naked eye.

-gb-

So Greg, when you say "analog capture," you're only talking about direct-to-disk with no tape involved, right? That I could understand as a possibility. But I could swear I've heard people say that analog capture from tape still looks better than digital capture, even if direct-to-disk is even better. Maybe I've misinterpreted what these people were saying.

At any rate, I believe you when you say that direct-to-disk via s-video provides more chroma information. You strike me as a guy who knows what he's talking about. :) I'll definitely have to consider picking up an analog capture card before I shoot another project involving chroma key.

Hmmm... you have me interested.

That is what I was talking about :)

So, I would need a comp with an analogue capture card in it... what about going from an external capture card to a laptop via 1394? Would compression still take place? Or has the S video bypassed some of the compression and the 1394 is just transmitting the signal?

OK, I tried it, and here's a couple of screen grabs for comparison. The YC grab was captured uncompressed (hence the UC in the filename) via my DV Storm card and the DV grab (DV in the filename) was captured from tape via firewire. I exported the images from Vegas 6 with all attributes set to square pixels to avoid resampling artifacts. Let me know if you have any problem downloading the files.

It's worth opening the images up in Photoshop so that you can zoom in to see the details. 300% or 400% shows up SD resolution in all its glory.

To my eyes, and on my LCD monitor, the 2 images are not that much different. The YC grab's a bit less saturated, but I didn't adjust the proc amps before capturing so that's probably not too significant. It's possible that better capture hardware would yield better results from the YC input but I'm not sure how much difference it would make in practice.

What I found interesting is that the vertical ghost lines next to the mic stands are more prominent on the YC grab. If anything, I thought the DV compression would have made that worse but obviously not. Anyway, I'd like to hear your comments if you have any.

Richard


http://www.jaegercat.com/~richard/XL2UCSquarePixels.png
http://www.jaegercat.com/~richard/XL2DVSquarePixels.png

Thanks for that Richard.

Although I am dying to see some difference, i really cant detect any! Anyone else got a conflicting result? Maybe we can see some video to compare them.

Richard,

I notice some difference in the gray area on the sides of the speakers.

Which one most closely represents the actual coloration?

Thanks,

I think the yellow thing on the wall looks richer in the svideo shot. the dv one looks a little washed out. There also seems to be slightly more noise in the dv pic on the speakers.

Hi Jesse. The speakers are actually black, but there is probably some light reflection that makes them look otherwise.

Richard

Hi Kelly. The UC image is slightly darker overall, which will affect the richness of the colours. I had the camera set to +6dB gain (maybe next time I should switch on some lighting!) so that would explain the noise in the DV image. It could be that if you brighten the UC image in Photoshop to match levels with the DV image, the noise will also be more evident. On reflection, I really should have taken more time to brighten up the room and allow the XL2 to operate in its "happy range". I don't think it performs too well in lowish light conditions.

Richard

Here's another couple of grabs with brighter lighting, so I was able to set the gain to 0dB and the lens to F2.8. I changed to 4:3 mode for this to avoid the camera resampling the 960 pixels per line down to 720 that occurs in 16:9 mode. I also switched off the custom presets so that there is no (or minimised) in-camera processing. Sorry about the shadows - my camera position was constrained due to the length of the YC cable attached to my PC, and also the black cat around the legs of the tripod.

From what I can see, I have doubts whether the YC output is uncompressed after all. In any case, based on the small differences I can detect, I think I actually prefer the DV image.

Richard


http://www.jaegercat.com/~richard/XL2YUYV4_3.png
http://www.jaegercat.com/~richard/XL2DV4_3.png

I've hear anecdotal reports... (Perhaps even via Adam Wilt?) That you can pull a better green screen off the S video connection. Might want to do a little compositing and see how the two hold up.

I can't see a significant difference between the two. And if there is I'd say it's pretty negligable.

I didn't see significant difference either, just visible difference in certain aspects of the pictures I saw.

I would expect there to be differences in quality between various analog capture methods also.

I can't find those emails at the moment and I never posted them because the sender had lifted them from a commercial ad and didn't have permission to pass them around.

-gb-

actually, in this second one, I see less noise on the left speaker on the YC than the DV and also on the cabinet beneath it. I also see less CA on the left speaker's white area.

Hmmm... If i had an analogue capture card I would try some out...