Why is S-Video capture looking better than firewire?

[Stephen Mann]

Why is the Canopus in the workflow when you are using the DSR-11? Just go DSR-11 –> firewire –> PC. Not that the Canopus in the circuit should change anything since Firewire is really a file copy, not a capture. Also, which Canopus? If you are using the ADVC-300 then there’s a TBC in the loop that can sharpen analog video. I am also not seeing Apples to Apples comparison here since the HV30 doesn’t record to tape. If you aren’t starting with the same original footage, then how can you make your conclusion?

Steve Mann
MannMade Digital Video
http://www.mmdv.com

[John Romein]

Thanks Steve.

The Canopus is in the chain because I’m trying to set up a general video to PC path and the Canopus handles the composite and s-video and firewire connections.

I’ve tested the path using the Canopus ADVC 100 and ADVC 300. I have a ADVC 110 but I have not used it in the testing. What I did observe is that the firewire thru the Canopus (either 100 or 300) to the PC or the firewire from the Canon HV30 direct to the PC does NOT show any difference in image. I suspect that the Canopus does nothing for firewire…just a pass through? I could connect the DSR-11 directly to the PC but I suspect that it would be identical to the HV30 to PC…especially since DSR-11 -> Canopus -> PC currently equals the footage of HV30 -> PC.

I’m puzzled as I always thought that firewire was better (over composite, s-video).

I realized also that going from DSR-11 via s-video or composite would be converting to analogue and then the Canopus back to digital. So, in reality I would never have connected the DSR-11 to the Canopus…but I just wanted to test the different paths and now I’m puzzled with the results. The pathway originated because the Canopus is also being used for VHS, S-VHS, Betamax, and Hi8 vcrs…and that is the path used for these VCR’s.

Still puzzled!
John

[Bill Burnette]

The “milky” description sounds like reduced contrast, which could also be vaguely perceived as less sharp. DV records with safe color range. Is the canopus converter expanding the color range? Just guessing.
Posting a screen grab might help someone diagnose it.

[John Romein]

Here is a screen grab of the same image from different capture file setups.

DSR-11 -> Composite -> Canopus ADVC300 -> firewire -> Vegas Pro 12 DV capture:
DSR-11 -> S-Video -> Canopus ADVC300 -> firewire -> Vegas Pro 12 DV capture:
DSR-11 -> firewire -> Canopus ADVC300 -> firewire -> Vegas Pro 12 DV capture:
DSR-11 -> firewire -> Vegas Pro 12 DV capture:
Canon HV30 camcorder -> firewire -> Vegas Pro 12 DV capture:

Here’s a slightly different test using an 8mm video and the same screen grab using different setups…this still shows the firewire looking not a good as s-video. And the Canopus 100 or 300 look the same for firewire:

Sony GV-D200 -> composite -> Canopus ADVC 100 -> firewire -> Sony Vegas Pro 12 DV capture:

Sony GV-D200 -> composite -> Canopus ADVC 300 -> firewire -> Sony Vegas Pro 12 DV capture:

Sony GV-D200 -> firewire -> Canopus ADVC 100 -> firewire -> Sony Vegas Pro 12 DV capture:

Sony GV-D200 -> firewire -> Canopus ADVC 300 -> firewire -> Sony Vegas Pro 12 DV capture:

Sony GV-D200 -> s-video -> Canopus ADVC 100 -> firewire -> Sony Vegas Pro 12 DV capture:

Sony GV-D200 -> s-video -> Canopus ADVC 300 -> firewire -> Sony Vegas Pro 12 DV capture:

My observations are that:
– the Canopus boxes when inputting via firewire seem to do nothing to the single…pass thru?
– the S-video seems more crisp and better colour than firewire
– firewire, whether direct, thru the Canopus, or from different player always appears “milky” and not as clear
– to compare these I took the screen shots into ACDSee Pro and did a side by side compare of 2/3/4 images and then zoomed into them to compare details

Your thoughts or observations….I always thought firewire should be better?

[Stephen Mann]

Your firewire images *are* better. Much better. Your analog images have been processed in either the camera or ADVC or both by adding edge enhancement, crushing the blacks and increasing saturation and contrast. (I.E., they are crap. What do the scopes show?) Look at the pants the girl is wearing in the first images. Note the detail in the shadows – In the analog images the pants are black with no detail. The Firewire images show some shades of gray and wrinkles in the fabric.

Steve Mann
MannMade Digital Video
http://www.mmdv.com

[John Romein]

Thanks Stephan for your observations.

I had a look at the detail in the dress and I see what you mean…however, if I zoom into the dress I do see more detail in all images.

I went back and looked at the scopes….maybe I don’t fully understand what I’m seeing. What I see is that the firewire on the waveform is between 10-110% and the histogram starts around 19-20 except for the blue. For the non-firewire the waveform is 0-100% and the histogram starts at 0. I always thought that the waveform should be between 0-100…not sure what the histogram should be. What should the ideal settings be?

Here’s the scopes for the different setups:

DSR-11 -> Composite -> Canopus ADVC300 -> firewire -> Vegas Pro 12 DV capture:

DSR-11 -> S-Video -> Canopus ADVC300 -> firewire -> Vegas Pro 12 DV capture:

DSR-11 -> firewire -> Canopus ADVC300 -> firewire -> Vegas Pro 12 DV capture:

DSR-11 -> firewire -> Vegas Pro 12 DV capture:

Canon HV30 camcorder -> firewire -> Vegas Pro 12 DV capture:

Thanks, John

[Stephen Mann]

The difference you are looking at may be the 7.5 IRE Setup from the camera which is why the Firewire levels are higher than the analog. Since the whites are not blown out, I suspect that either the camera or the scope settings include a black level setup.

But the 7.5 IRE setup confuses even video engineers who worked with scopes for 50-years, and digital further muddies the confusion. (All I know for sure is that in digital, you want no IRE setup.)

It’s not unusual to have levels above 100 IRE. If you are taping a performance against a black curtain, any stage lights will most certainly be above 100 IRE. You want the details that you want to see between 0-100 IRE.

Steve Mann
MannMade Digital Video
http://www.mmdv.com

[John Romein]

Very interesting stuff about the IRE settings. It definitely makes a difference setting the Canopus dip switch to IRE 0 or 7.5 (even if the control is set to the PC software…which is available and I use for the ADVC300). Here’s the scopes for some scenarios using different IRE settings:

DSR-11 -> Composite -> Canopus ADVC300 – IRE 0 -> firewire -> Vegas Pro 12 DV capture:

DSR-11 -> S-Video -> Canopus ADVC300 – IRE 0 -> firewire -> Vegas Pro 12 DV capture:

Image:

DSR-11 -> firewire -> Canopus ADVC300 – IRE 0 -> firewire -> Vegas Pro 12 DV capture:

Image:

DSR-11 -> firewire -> Canopus ADVC300 – IRE 7.5 -> firewire -> Vegas Pro 12 DV capture:

Image:

Interesting how the scopes and images for composite and s-video and firewire are almost the same with IRE 0. Why would they all look so similar? Your thoughts.

Cheers,
John

[Bill Burnette]

A plain way to look at it is that the S-video and composite analog outputs are meant to be plugged into a TV, which is (for historical reasons) not supposed to be fed signals below 7.5 IRE. You are looking at these images on a computer, which can display the full range of RGB, 0-255. The Canopus converter is giving you the choice by using the 7.5 vs 0 switch of whether you intend your analog capture to be carried through a DV workflow to a DV tape to be played back on a TV versus for final display on a computer. As you have discovered, viewing an image coded in DV for TV without any correction on your computer, it is lighter (milky), and the variation of shades of black is greater (not crushed).

Your histograms show that the total range of (digitized) color from black to white is less with the DV (and 7.5) capture than the 0 IRE capture. So it looks like less contrast on a computer. It’s an interesting thought whether you could find a more satisfying correction (e.g. curves or levels in Vegas) from DV/7.5 than the simple IRE transformation done by the Canopus.

[John Romein]

Thanks Bill.

Obviously not simple. It seems that if I’m capturing for the computer I need to use IRE 0 and if I’m capturing for DVD I should use IRE 7.5. Now the tape itself may have been recorded using IRE 0 or IRE 7.5…more than likely 7.5 if NTSC.

Cheers,
John