PsF clip "converted" to Progressive in Inspector needs rendering in Progressive Project?

Ian R. Brown wrote:

I felt like saying, "I told you so", but eventually resisted the temptation.

Ah, so the bigger man did come out on top after all!

Dave

So, I did another test (I'll get around to some proper video one day!) and after the posts about 10.0.4 now being able to handle 50/60P I was hopeful, but, alas.....

It still doesn't see 30PF or 24PF as progressive. Both come in as 1080i / 29.xx fps.

It still renders when you change the project / clip as per the provideo link.

One very small difference is that when I changed clip properties (field dominance from Upper to Progressive) it now shows up straight away as 1080P, whereas before I had to quit the programme and restart it.

What a nuisance.

I'm with you Ian, I think this "PF" nonsense is a marketing trick.... I wouldn't mind if my camera had cost $100 but it didn't, and I really don't like being hoodwinked by Canon! I feel a call to Canon Customer Service coming on.....😁

Thanks Ian. I still haven't called Canon, but I will.

For now I'm not going to bother with the PF settings, I'll stick with 60i and 24P. 24P looks quite good, to be honest. But I'm wishing I'd bought the Panny TM900 for $500 less than my Canon now... especially as FC 10.0.4 does 60P! Too late to take it back.... technology (or shiny gadgets) is a cruel mistress!

I read this thread with interest as my Canon HF100 clips, although shot in "PF25" as it says, sure enough are considered interlaced 1080i by FCPX. I tried ClipWrap after reading this thread half way through, only to discover that it does not help. As the rest of the thread explains nicely.

Now, could somebody explain perhaps what PSF (or PF25) actually is? If not two interlaced fields of the same frame shot once every 1/25", instead of two interlaced fields shot at 1/50" interval?

Because if so, then the material is interlaced and no clip-wrapping could make it progressive: you'd have to decode the frame using the full h.264 artillery, and re-encode in the format of your choice. Very likely not h.264 then. That's not just setting a bit in a header.

And in fact, I don't actually mind that FCPX maintains my project as interlaced, if it then is.

It should just realize that, at the time it actually needs to de-interlace for cuts, colour correction, transitions, or export to progressive, you name it... that at that moment it can weave (and not blend) the two fields like a still picture at no loss or extra cost, without any adverse effects such as combing or colour skews, no need to guess or apply magic. I thought this was what PF25 (or PSF?) buys me.

So the question more boils down to: how can I tell FCPX to de-interlace by weaving? What would it do if I didn't? Would it blend and hence mess up resolution? It wouldn't be able to find out on itself prior to avchd 2.0, would it? What happens to the 4:2:0 sampling anyway?

Rainer

Right... explain this!

I followed the workflow suggested by another site regarding PsF to avoid it being de-interlaced and subsequent loss of quality. Basically I forced FCPX to recognise it as 30P.

Here are two screenshots.... both were filmed at 30PsF on a Canon. One was brought in to FCPX using "auto" settings (and it was seen as 60i), the other was forced to be 30P. The "30P" one to me looks like it's only grabbed one field of the two that make up the frame. The 60i looks fine. So, it looks like the two fields were not recombined when forcing it to see it as 30P... or am I going completely stark raving mad?

60i

30P??

PS... I'd like to add that I would like to ask the site that suggested the workflow directly, but its sign up requirements are quite intrusive!

Explain? Sorry.

The only observation I could make: you're presenting two complete images, hence de-interlacing has already taken place.

I don't know how sharp your original material is, but the upper image looks a bit blurry, perhaps an indication that is has been de-interlaced using blending (or top-field-only, or dark magic).

The lower one remains a mystery: it looks like two fields taken at different times. If you shot in PSF, that shouldn't happen. So I conclude I don't understand PSF.

Now, when you de-interlace it matters whether you take top-field-first or bottom-field-first. Weaving top-field-frame-1 into bottom-field-frame-2 would be particularly silly, but since you forced it anyway all bets are off. There are a few knobs in the settings panel of the inspector, BTW...

Mind posting the workflow that you used?

Explain? Sorry.

Er, yes, my comment didn't come over as intended! Apologies...!

The two images are clips from the .mov files created after export. They're blurry because I was filming from a car doing 100Kmh, and the oncoming car was probably doing the same speed. The crops are about 1/4 of the full frame. My shutter speed should have been around 1/60th, but I was in "Program" mode so can't be sure.

I am absolutely certain the clip at the top came from using "first clip" settings, which in this case was thought by FCPX to be 60i. The second clip was made using the manual settings for 29.97P in the project, and changing field dominance to progressive, so that it was treated as 30P. Then both projects were exported to QuickTime "using current settings" and the clips are from the resulting files.

However, I can't check all this because I've since dumped the projects and events (I have to once a "film" is made just to keep hard disc space manageable!!).

I tried to replicate the results this morning, and couldn't. Maybe I made a mistake with export settings before... but they're both MPEG-4 movies, and more or less the same size.

Things are now even more complicated because Tom Wolsky has found another checkbox.... de-interlace. Aargh!

Bring back DV!! 🙂

Hey All - we're back from NAB now and (slightly) caught up on sleep, so I wanted to jump in.

We've got a couple potential flags within the SD800 stream that might reflect PSF tagging. Before we're certain, it'd be great to get a couple more samples - in particular, from an NTSC cam (our only sample is PAL).

Can anyone provide such a thing? If so, drop me a line at support@divergentmedia.com and I'll get you an upload URL.

Thanks!

-Colin

Ian,

I try to check back in once a week or so to this thread to see how it's going. Any luck working with Colin?

Dave

Howdy - we've got an internal test build, but need more footage before we can do a release. See my post from the 22nd for details.

-Colin

Ian, Colin,

Well, that sounds like positive progress. The camera companies need to step up and fix this, but since they seem to not move very fast, this will likely be the necessary workaround to avoid the render bar. I frequent differnet FCPX forums and I see this same issue come up time and time again.

Good luck,

Dave