Help, cannot reverse telecine from DVCAM on XPHD5.5 for Mac

[Michael Phillips]

Hey there…

The slow down is a different issue that the missed cadence. A couple of things to check:

Are you sure the down convert was done with “A” frame on 0 and 5 of a non drop 30 frame timecode?

Is the setting in Film/24p correct for the “A” frame relationship?

Are you properly set for NORMAL cadence (this is not self detecting).

Try logging the same clip (less than 30 seconds of motion) 5 times with the foll wing relationship:

:00 = A
:00 = B
:00 = C
:00 = D
:29 = B (minus 1 frame from the other clips)

And see which one has the proper cadence. Make sure you are watching them back in green/green mode to get the full 2:3 interlace signal out over FireWire. If you are not in green/green mode then you are seeing 4+1 on the monitor due to the nature of the quarter frame display and its affect on 2:3 cadence.

Once you get that straight, we can work on the slow down issue…

Michael

anything 24fps

[Bill Russell]

Hi Michael!

Yes, on 4.8 problem was clear, Xpro only allowed Advance mode. We could see that tapes were correct, A frame landed on 0, and C frame juddered because reverse needed to be standard instead. Everything was in sync. Upgrading to 5.5, now we have the new problem of the slow video.

An important note is, for all we know, the reverse would be correct now. BUT, video seems to be playing back 25% too slow compared to audio, one video frame to one film frame. However, XPRO seems unaware of it because the 30 TC of the clip is skipping numbers 03, 13, and the ABCD are counting correctly. (We do not have a window in order to compare numbers – sigh.)

Anyway, to answer your questions:

– Logging at difference cadences. I’m not in the room today, but I will ask them to do it and they can report back here. However, based on experience in 4.8, I’m 98% confident that A=00 is correct and worked first time fine. Again, for all we know, reverse is correct now. Just stupid video is playing 25% slow.

– Ah ha! Green-Green mode, we’ve been completely ignoring that. (Since we were seeing a 2-field judder, does that mean we were in green-green aleady?) What is 4+1 display?

Thanks as usual for your thorough and always on-it reply. You’ve saved the day more than once…..

“THE LOST SKELETON OF CADAVRA” –

And more…

[Assistandres]

Hi Michael,

I’m assisting on the doc Bill has posted about. I was able to capture the 5 clips, with A, B, X, C, and D frames at the :00 timecode and performed the “Cmnd + D” function on the Start column.

None of the cliips has the proper cadence because they’re all 30 frames playing in a 24 frame timeline. Sound drifts because audio is playing at normal speed. The drift is 1.25 frames/sec^2.

The monitor contains a timecode viewer, and here’s where it gets interesting:
When I bring the clips in, it defaults to counting frames from 0 – 720. That’s 30 seconds * 24 frames. I can’t figure out where the remaining 180 frames of video goes, because the clip follows the parameters of audio duration. The TC30 extracts the appropriate frame number from the timecode depending on the “A” relationship, just not the actual frame.

NOW…here’s some more for the puzzle:

I tried bringing in some of the footage using a 14:1 res instead of DV25P 411 (curiosity). With :00 at “A”, the footage is synched. The frames are pulled out IN RELATIONSHIP TO THE AUDIO. HOWEVER, the “C” frame is a TRIPLE INTERLACE. The top half of the frame (literally the top 212 lines) are a C and D combination and the bottom half is a B and C combination. It looks like it’s trying quite hard to pull that C frame out, but that’s the result.

Any thoughts? Help is much appreciated as Bill said before.

Thanks,

Andres

[Michael Phillips]

I will send your posting to testing in QA to see if they can replicate this. Just a point to be clear on…

When you say 30fps – the sources are truly 24fps (23.976) with 2:3 puldown correct?

Michael

anything 24fps

[Kchunk]

[MichaelP] “When you say 30fps – the sources are truly 24fps (23.976) with 2:3 puldown correct?”

Yes. According to the telecine house, that’s the specs for our D5 masters.

Thannks,
k.

[Michael Phillips]

OK, the miscadence capture was replicated in house and will be fixed in an upcoming point release. Any other information the slow down issue?

Michael

anything 24fps

[Bill Russell]

Hi there, Andy & Michael —

I was in the edit room only on monday, following this since then from afar.

Michael – So was it the triple field C frame anomaly in 14:1 capture that y’all replicated? Or what was it? On 5.5?

What I know so far is that the edit room has downgraded all the way to 4.5 now, and are testing bringing video in over component analog into the Mojo, controlling a DSR-40 over RS422. Now they have a +/- 1 frame accuracy problem — (Andy, does that mean reverse telecine is working now, at least when you are lucky enough to capture on the right frame?)

(P.S. — yes, definately 23.976 HD sources)

“THE LOST SKELETON OF CADAVRA” –

And more…

[Bill Russell]

P.S. I want to clarify my post just above — the reason they are trying analog component capture now is, as I understand it, the said limitation in 4.x that allows only Advanced mode reversal applies only to video from *firewire*. But when doing analog capture using Mojo, Advance *and Standard mode* are allowed in 4.x. Right?

As far as a possible return to 5.5 — I would try: a) installing 5.5 from scratch on a different clean start-up disk, OS 10.6, to rule out any impossible install/device/software/Quicktime bugaboos hiding in your current environment; and b) try firewire capture direct through the computer rather than the Mojo, in case the Mojo is messing it up.

“THE LOST SKELETON OF CADAVRA” –

And more…

[Assistandres]

we actually have not yet downgraded to 4.5

the problem in 4.x with advanced pulldown may only occur over Firewire, but the warning is about compatability with the DV25P resolution. Our res choices (in 5.5 at least) coming in component are DV25P, DV50P, or 28:1.

And yes, the reverse telecine works when we hit the right frame, which happened 4 out of 7 capture attempts.

[Bill Russell]

Oh, okay then, so as I understand it, this all applies to 5.5 on OS 10.6 — for the record here in this post:

– Firewire capture to DV25 = slow video (seems to be frame for frame playback at 5/4th speed).

– Firewire capture to 14:1 = the triple-field C frame.

– Analog component caputure = viola! Good reversal. Only problem is +/- 1 frame timecode (may be the DSR-40, a problem you are troubleshooting).

So I’m still curious to know what cadence problem the Avid lab discovered?

Also, I still think it might be worth installing 5.5 on a different 10.6 boot just to see if it works… and/or try a firewire capture from deck direct into the machine, instead via Mojo. If firewire works, then would be nice if you could get direct digital capture of the audio to eventually go to sound finishing. Also, timecode info should be glued to picture when capturing over firewire, worth checking.

“THE LOST SKELETON OF CADAVRA” –

And more…