Jagged Image in DV 30p FCP capture?

[Ben Holmes]

Have you viewed the footage from FCP on an external monitor? The Canvas is not the most reliable means of judging quality. It looks like you have footage playing using RT in FCP, and it’s dropped the resolution to give smooth playback, or you have the RT quality set to ‘low’. Alternatively, do you have any render bars in the timeline? Is the sequence set up for 16:9 as well?

Edit Out Ltd
—————————-
FCP Editor/Trainer/System Consultant
EVS/VT Supervisor for live broadcast
RED camera transfer/post
Independent Director/Producer

https://www.blackmagic-design.com/casestudies/detail.asp?case=therydercup

[Chris Babbitt]

It does sound like a field order issue, however, I recall that the XL2 does not have 16×9 chips, which means that it does a pseudo 16×9 by blowing-up the 4:3 image. This would result in a softer image, and possibly some jaggies.

[Chris Babbitt]

Sorry, I was thinking of the XL1. The XL2 DOES have 16×9 chips.

[Alan Okey]

This is not a configuration problem, it’s simply a case of subject matter that highlights the failings of the DV codec.

It’s glaringly obvious from your screenshot that you shot on DV. The visible stairstepping is not a result of incorrect capture settings, it’s simply that you have footage that is the worst possible subject matter for the DV codec – in this case, highly saturated reds on a black background. What you are seeing is the 4:1:1 chroma sampling of DV compression, which in this case is manifested as highly visible vertical blocking in the reds. DV’s chroma resolution (U and V) is only one quarter that of its luma resolution (Y). That’s what 4:1:1 means – in YUV video, for every 4 samples in the Y (luma) channel, you only get one sample in the U and V (chroma) channels. This is precisely why DV is such a poor choice for greenscreen work.

More details here:

https://www.neopics.com/bluescreen/

Ignore the second part of the article, it’s out of date.

You can mitigate the jagged chroma subsampling artifacts to some extent by editing in a ProRes 422 sequence and applying a 4:1:1 chroma smoothing filter to your footage, but if you export back to DV tape you’ll see the same thing all over again.

[Emily Millay haddad]

Hey there Alan,

Thanks for this — I knew that shooting in such low light, and effectively with a chroma key light only, was likely to present a problem. I just wasn’t expecting this! I have some a beginner’s understanding of how this works, but I appreciate you taking the time to explain it.

Here’s a question — if I desaturated the footage and basically removed all chroma information, focusing entirely on the more densely packed luma information, would that solve the jagged edges and go to a smoother image? I’m going to try it, but I’m curious about what you think theoretically.

I’m also curious about why the footage looks fine in my camera’s viewfinder (in playback) and on the clamshell’s screen. The footage was captured onto a MiniDV tape and encoded at that point onto the tape into a DV codec (right?), so why does it change and become so crappy looking just because I move the data from the physical media onto my computer? I can see through my camera and the clamshell that the data has the capacity to look fine — so why can’t I work with /that/ image in my editor? If you or other folks have time to shed some light on the subject, I’d be grateful.

Thanks again for the help.

[Emily Millay haddad]

Hey Ben,

Thanks for responding! My sequence is definitely set for 16:9, as were my capture settings. I’ll check the RT settings as well, but I think that Alan’s answer below about the YUV 4:1:1 DV capacities is probably more at play, since the entire show was low red light only.

The only exterior playback monitor that I have is my GV-HD700 clamshell. Interestingly enough, when I use the clamshell as an external AV device, and send this wacky jagged picture to it, it looks fine. As in smooth and perfect. As in… what?? So… hmmm… Can I edit this blocky thing and swallow my distaste, then output it in FCP in some way that the jaggedness isn’t also exported? If this is somehow just some weird artifact that’s occurring in FCP that will cease to occur once I properly output, I’m happy to do that.

Any thoughts?

[Emily Millay haddad]

Hey there Dave,

Thanks for taking the time to respond — I did a few searches in Creative Cow forums for folks handling problems similar to mine before I posted. I believe I started with clips that were set to “None,” or possibly “Not set.” Then when I saw other posts, one of the pieces of advice was to set the field dominance to “lower” for everything, so I tried that too. I’ll try again and see what comes up, though I’m going to give some of the other pieces of advice a shot, too.

Thanks again!

Emily

[Emily Millay haddad]

😛 You almost gave me a heart attack. I bought this camera about five years ago, and the 16:9 chip was one of the big selling points for me then (along with the Canon glass and interchangeable lenses). But it’s been so long since I’ve did that research that I thought — for a split second — that maybe I had been wrong!

Oy vey. Anyways — thanks for the reality check. 🙂

[Alan Okey]

[Dave LaRonde] “Isn’t there a 4-1-1 chroma smoothing effect in FCP?”

Yes, that’s what I was referring to, although I didn’t make it explicit that it was an FCP filter.

[Emily Millay haddad]

Found the filter, applied it and it seems to have solved my problem for now.

Goodness, that is weird watching it change in FCP and do almost nothing in the “reference monitor” of my GV-HD700.

Thanks for all the help, guys.