Mysterious 6 frames out of sync live capture issue

dvxStephen

Active member
I've recently completed a two week green screen studio shoot using 3 HVX200s, all capturing live direct to disk via 24p. We captured about 30 hours of interviews.

All Macs had a fresh install of the latest OS, fresh install of latest FCP, with all updates in place.

All went smoothly with my intel Mac G5. Perfect sync. No significant problems, other than a bug in the intel version of FCP that prevents the waveform monitor and vectorscope from working in the capture module of FCP.

The other two cameras were being captured by powerbooks. Both powerbooks worked fine, but in post I'm finding an interesting glitch -- they're both out of sync by 7 frames or so. Advancing the sound 6 or 7 frames fixes it, and I'm ok. But it's an odd glitch that appears to be a bug with the current version of FCP on non-intel macs, or at least FCP on powerbooks.

Anyone else seen this?

Here's the details:

Capturing 24p using stock easy setup called DVCPRO HD - 720p24, which does automatic pulldown removal during 24p capture, making for very efficient 24p captures, resulting in the same data rate being recorded on my external capture drives as if I was using 24pn via P2 card. Works great, and resultant data rate of 5.7 mb/sec consistent on all 3 cameras. (That's with just 2 audio tracks turned on; it's a bit higher if more audio tracks are engaged during recording.)

Using the non-controllable device setting in the firewire devices popup.

I was using mirrored drives for each computer, fed on a separate bus from the capture firewire input.

By swapping cameras around, I've since confirmed that this was not a camera-specific problem, which was unlikely anyway because I set up all the cameras identically using a SD card.

I'd be interested to hear if anyone else has seen this, or knows of a fix.

Thanks,

Stephen Gagne
Westwood Studio
 
Sorry to dig up this old thread, but in searching the archive I could not find any evolutions on this problem.

I experienced this tonight for the first time (I'm a new HVX user, but I'm a pro sound mixer, so I know what I'm dealing with when it comes to sound sync).

Difference being is that I'm doing live capture to an Intel Mac mini through FCP (5.1.4) at 720p/60fps. My sync is exactly 5 frames off. This is not sync drift or mismatched sample rates or anything like that... it's capture latency. I'm using an external mic, but it is not misaligned with the camera mic being recorded to channel 2, so the entire sound architecture is off by 5 frames. Obviously, this is not a big deal to correct, however I wonder if all of the films that I'm going to start receiving with this popular camera are going to have this inherent sync issue in the raw footage.

It does not seem to be happening on P2 transfers.

-Jamey
 
It could be something to do with the way you are monitoring the picture and sound. Explain your EXACT set up in regards to monitoring picture and sound then also have a look at the sync offset in the FCP preferences and see where it's set.
 
This is an odd one. I was shooting this way quite a bit about 8 months ago for music concerts. My own laptop - a 15" G4 aluminum 1.5 GHz Powerbook has never had the sync issue but 2 friends with 17" 1.67GHz powerbooks did have it. They may have figured out what was causing the problem but then again maybe it came back again. i'll try to find out.

You are definitely not alone though, and its not a monitoring issue - its on the file.

- Lenny levy
 
Same problem here. Doesn't seem to be any rhyme or reason either. One 10 minute clip might be in sync, the next shot, out by seven frames. I had all recent version with an HVX and a 15" Powerbook with an external drive (slow, 5400 rpm) I was capturing 720/24p to. The offset in FCP wouldn't come into play as all other clips captured before this method is all in perfect sync. It's out of sync when viewed in QT as well. I'm stumped.
 
ozduc said:
It could be something to do with the way you are monitoring the picture and sound. Explain your EXACT set up in regards to monitoring picture and sound then also have a look at the sync offset in the FCP preferences and see where it's set.

Hi.
Thanks for the response. Unfortunately, I don't think it has anything to do with my playback settings as it doesn't seem to be affecting my footage that I bring in via P2 import. For monitoring, I'm just using a Dell monitor and the audio outputs on the mac mini... no extra hardware. There are no sync offsets in FCP. Also, when I output an OMF and bring it into ProTools on an entirely different system, it's still out of sync, so it clearly has nothing to do with my monitoring/playback setup.

Lenilenapi said:
You are definitely not alone though, and its not a monitoring issue - its on the file.

Thanks for confirming. I would be very interested to hear of a solution (or at least a cause) if you do find out.

Geno said:
Same problem here. Doesn't seem to be any rhyme or reason either. One 10 minute clip might be in sync, the next shot, out by seven frames. I had all recent version with an HVX and a 15" Powerbook with an external drive (slow, 5400 rpm) I was capturing 720/24p to. The offset in FCP wouldn't come into play as all other clips captured before this method is all in perfect sync. It's out of sync when viewed in QT as well. I'm stumped.

That makes me think that throughput could be a possible target. If your 5400rpm drive is the cause, then perhaps for me it's the fact that I'm using the same firewire bus for both my HVX and my 7200rpm firewire drive....both would be bandwidth bottlenecks. At least it's something to do some tests with.

Cheers,
Jamey
 
iocomposer said:
That makes me think that throughput could be a possible target. If your 5400rpm drive is the cause, then perhaps for me it's the fact that I'm using the same firewire bus for both my HVX and my 7200rpm firewire drive....both would be bandwidth bottlenecks. At least it's something to do some tests with.

Cheers,
Jamey

This could definitely be the problem. To this day I still dont think Apple "officially" supports the use of firewire drives for video capture. Yes everybody does it but it can cause problems. Having the camera and drive on one bus can cause problems to. YOu will need to get a card to add another FW bus to your system. Even though you may have several FW ports (400. 800), they all share the same bus.
 
We were using a PCMCIA firewire adapter to have separate buses. I just emailed my friend about this and here's what he said :


. "Yes, i figured out that if I totally erase the hard drive and reinstall everything, then it records properly. Why the on board hard drive affects the off board recording I dont' know but it works."

I think on board means the powerbooks internal HD vs. recording to an external HD
 
In the last few months I've sunk up hundreds of hours of 3 camera HVX footage, all captured direct to disk with the same gear, that is an good test of this issue. It's a 720p thing in my experience. The bug appeared in 90 hours of footage that was all shot last summer 720p/24p with three identical HVS200s == matching menu settings via SD card -- and identical computer setups except as noted in a moment -- same version of FCP, all new FCP installs after fresh OS installs. Computers used ONLY for capture. All clean, new, top quality star-quad firewire cables, with stabalant 22 added. I had consistent results -- an intel iMac (brand new) stayed acceptably in sync, although it was one frame off at times -- which I've seen also with DV, so no surprise there.

However the two G4 laptops, powerbooks that is, with maxed-out memory, etc -- both consistently have sound out of snyc with picture upon capture IF shooting 720p but not SD. Haven't tried 1080, though. Easy enough to fix in the edit, of course. The amount off varies from take to take -- 5 frames on low side, 7 on high side. NEVER happened on the intel iMac.

This issue does not come up, using this identical gear, same crew, same cables, same camera settings -- same everything -- when capturing DVC50 SD. We've shot hundreds of hours on this same project, multi camera, with that codec without the 5 to 7 frame out of sync issue, although all three recording setups do vary a frame here and there from each other in DV50 also, in terms of sound to picture sync matchup to slate.

There's no doubt in my mind that there's a bug in the non-intel version of FCP that I was using at last summer for the 720p work. It was the latest greatest FCP version as of last July, but of course there was a new version that's just come out this month, and as I recall seems like there's been a version between then and the new one as well. This bug may be fixed by now. I'll be testing this in the next few weeks to find out.

Stephen Gagne
 
Forgot to mention - we were recording onto external drives, mostly G-techs, mirrored pairs on each computer for data redundancy.
 
Not sure "bug" is the word though as neither Apple nor Panny recommends this method of capture perhaps for this reason. It does work though.
 
Not sure if this helps, but, in SD tape situations, if your camera is set for 32kHz sound, and your Log & Capture window settings are for 48kHz sound, you'll get drift and be out of sync.

Also, exceeding the bandwidth of the hard drive busses will cause all sorts of very strange problems. I've done a lot of trials capturing live with a laptop, both Mac and PC, and have sworn never to even attempt it again. I don't know anyone who "whole hearteldy" suggests it's a solid and reliable way to capture live.

Also, if you're monitoring on your computer, watching the video on your computer screen, with sound routed to an external device, you will get a dely in sound that doesn't actually exist in the file.

I'd avoid live recording to laptops for long durations. They're graphics cards are not as powerful as what you can put in a tower, their data busses are more limited than what's in a tower, etc. Laptops are made to run as well as they can with minimum heat. It's a HUGE engineering feat to do that. One that tends to choke the data pipeline in the system as a whole.

Probably none of this helps, but, there you have it. I'm on meds for a really badly needed, over due root canal that can't be done for a couple of days, pain, like you wouldn't believe, like P2 on a Mac. And I've a low pain tolerance, just like NTSC-DV and color. And it takes a lot to satisfy my demanded comfort level, like editing HDV on a G3 iMac. So I'll just be ranting aimlessly for the weekend...
 
I posted on this exact problem a few months ago, recording direct to disk on a PowerBook 1.5 GHZ G4 with 1 Meg of RAM using FCP 5.1.2, I think. I taped 4 one and a half hour 720p24 clips over a weekend, with FCP stripping out redundant frames on the fly. Three clips were 6 frames out of sync and one 3 frames out of sync, with the audio ahead of the video on all clips. The audio remained 6 (or 3) frames out of sync for the duration of the clip, so it was an easy fix in post just sliding the audio back the required number of frames. I never figured out why this happened, but it was clearly not an audio delay in playback, as the audio was ahead of the video. The DVD I made after fixing the audio in this manner was fine.

I am really curious to have someone figure out what is causing this.

Kevin
 
Ben, when you say, "Over loading the bandwidth of the limited data pipes of a laptop..." you state your view as fact.

Is this something you verified by calibrated tests?

My own experience, as described earlier in this thread in detail, leads me to believe it is a software bug in FCP.
 
BenB may well be right that a laptop simply cannot keep up, but if this is the cause, I would expect to have the audio and video get farther out of sync as one continues to tape. As I mentioned above, this was not the case when I ran across this issue. Also, there were no dropped frames or audio in the footage--all the data was there, just a consistent 6 frames out of sync.

Kevin
 
Well, my 20+ years as an IT Engineer have nothing to do with the idea I may have a clue, technically. What I state is fact, not opinion. All laptops have smaller everything. Heat is a major issue. Laptops have never to date out performed a tower. The graphics cards have to be lower grade due to heat, etc, etc, etc.

I've yet to see a Mac tower that can't capture live with no hiccups, period. Compromises are made on laptops to keep space down, keep heat down, etc. Simple fact, not my opinion.

As for A/V getting more out of sync, based on what data? Nothing says it HAS to work that way, and I'm just saying what I know technically to be the case. I've done tons of live capture on both, I've done tons of technical work professionally with the systems, I stand by the facts I've stated.
 
Last edited:
Weird thing is my 15" G4 1.5Ghz powerbook doesn't have sync issues on live capture while my friend's 1.67 machine does - all other settings equal.
Also listen to my friend who got his sync issue to go away after reinstalling everything on his internal hard disk.
 
Back
Top