FCP OMF sync problems with only some regions?

[Max Holland]

I just recieved an OMF and quicktime from a 50 minute doc 23.98 with some odd quirks.


Only some audio regions in the OMF were out of sync, some by 1 second, some by 3 and a half. its not drift because the music over the credit bed is phase perfect with the quicktime but the dialog is not.


The OMF reported some 'Parsing errors' with the fades files, about 20 of them, is it possible that bad fades are 'pushing' the audio regions around and out of sync?


I am having the editor double check to make sure his audio is all of the same sample rate and bit rate, also he will strip the fades etc.


Has anyone seen anything like this?

[eoats]

Were there different formats & were they all inputed the same? I've had this recently where there was some Sony EXHD, DV & Flip Camera all co-mingling inthe timeline.

[smurfyou]

Honestly I would ask them to close their project, re-open it, then export the OMF. Or at least duplicate their sequence before exporting. I swear that has "fixed" slippery regions on more than one occasion.

Also check with the editor to make sure those video clips weren't moved without the audio.

As far as the fade error, theres no telling. Doesn't FCP have a checkbox for including fades in an OMF export? You could have them export without fades to see if that works any better.

[georgia]

FCP is the landfill of editorial systems... It accepts just about anything, any codex, any sample rate into the time line and piles them all together..... then just try to get an OMF out with a struggle.

If you are very very careful to assure that you maintain appropriate management and stick to one sample rate in the time line you actually stand a chance of getting the materials out without a problem. I just exported 18 tracks of a 30 minute broadcast show with Pan info, Volume info and fades... no problems at all. But I edited the project and made damn sure everything was conformed to a standard format and sample rate prior to importing into the project.

cheers
geo

[Max Holland]

I am going to have the editor make sure everything is the same sample rate, and make a copy of the sequence and export that.


I might also bring the OMF into and Avid on my end and re export it, see if Pro Tools is just reading something wrong.

[Sonsey@mac.com]

Geo and Will are correct...

1. FCP's realtime sample rate conversion is one step below useless - plus multiple codecs on a timeline, while sounding like a good idea, cause no end of problems, so have them check that.

2. There is a repeatable bug, that I've mentioned before I think, that if you open an FCP project and make changes, then output an OMF, the changes MAY OR MAY NOT be reflected in the OMF. Saving, closing, then re-opening the project before exporting will insure the OMF properly reflects the timeline. I have confirmed this on V6 of FCP... haven't had a chance to test V7 yet...

[musikwerks]

Originally Posted by Sonsey@mac.com

Geo and Will are correct...

1. FCP's realtime sample rate conversion is one step below useless - plus multiple codecs on a timeline, while sounding like a good idea, cause no end of problems, so have them check that

...and there began the dumbing down of editorial services. Used to be you had to know what the different standards were. Now you just chuck it into the timeline and let the computer sort it out. God, help us.

[georgia]

one of the really "fun" problems in final cut is exporting OMFs with various sample rates in one time line. What happens is that, although FCP allows these and does sample rate conversion on the fly, OMF can't. OMF is a very strong and very stupid tool. It looks at all the database info regarding the audio samples and builds it's database/files for transfer and doesn't care what the actual sample rate is. It just assumes that everything is the right sample rate.
As an example, if you have a 48Khz timeline in FCP but you have a couple 44.1 sample rate files, everything seems to be ok in the editorial FCP session, sounds right, plays correctly. But... OMF sees a file that it thinks is 48Khz, puts it at the correct sample point data in its database time line, but the file is "missing" samples so it's duration is shortened by the difference in samples due to the 48Khz/44.1Khz differential. Then it really starts.... the next file is places early on the track point within the OMF dataset the number of samples "missing" from the previous file. every file on that track is now early xxx number of samples. and it's cumulative. It just gets worse and worse, and its on a track by track basis.

I can't tell you how painful this can get if you've got an editor that isn't paying attention and is using 44.1, 48, Mp3, 32k files from multipe assets and is just dropping them on the time line instead of format concerting them to a standard for the session.

Sometimes the only way to fix a completely screwed session is to output the audio from FCP as AIF files that are the duration of the session. One per track.

cheers
geo

[Max Holland]

Originally Posted by georgia

one of the really "fun" problems in final cut is exporting OMFs with various sample rates in one time line. What happens is that, although FCP allows these and does sample rate conversion on the fly, OMF can't. OMF is a very strong and very stupid tool. It looks at all the database info regarding the audio samples and builds it's database/files for transfer and doesn't care what the actual sample rate is. It just assumes that everything is the right sample rate.
As an example, if you have a 48Khz timeline in FCP but you have a couple 44.1 sample rate files, everything seems to be ok in the editorial FCP session, sounds right, plays correctly. But... OMF sees a file that it thinks is 48Khz, puts it at the correct sample point data in its database time line, but the file is "missing" samples so it's duration is shortened by the difference in samples due to the 48Khz/44.1Khz differential. Then it really starts.... the next file is places early on the track point within the OMF dataset the number of samples "missing" from the previous file. every file on that track is now early xxx number of samples. and it's cumulative. It just gets worse and worse, and its on a track by track basis.

I can't tell you how painful this can get if you've got an editor that isn't paying attention and is using 44.1, 48, Mp3, 32k files from multipe assets and is just dropping them on the time line instead of format concerting them to a standard for the session.

Sometimes the only way to fix a completely screwed session is to output the audio from FCP as AIF files that are the duration of the session. One per track.

cheers
geo

you hit the nail on the head! that sounds exactly like whats going on, by the end of the program the pop at the tail is 4 seconds off on the tracks one and 2, unfortunately those are dialog tracks so fairly important!


I am hoping that the editor can easily convert the audio to 48khz, because there is a lot of dialog editing and denoising I have to do, I cant deal with continuous aif files of each track! especially not the dialog


thanks everyone, I will report how this works out on thursday (when the editor returns from vacation)

[musikwerks]

Originally Posted by georgia

I can't tell you how painful this can get if you've got an editor that isn't paying attention and is using 44.1, 48, Mp3, 32k files from multipe assets and is just dropping them on the time line instead of format concerting them to a standard for the session.

So, so, so true.

I see it all the time when I'm onlining projects. Always from FCP "editors". Never Avid... because up until now you couldn't mix resolutions/standards in Avid. That will all change now that Avid has opened up the timeline though. Yay... more things to fix in an already time/budget squeezed session.

I don't know that it's a matter of not paying attention. Generally speaking (please note "generally") I see it as a completely different school of editor.... those that grew up in a linear world and absolutely had to understand why you must keep your luminace at certain levels (for example) and those that simply bought themselves a mac bundled with FCP and thought editing might be fun. Editors who don't know what correct field dominance is and who only monitor on an LCD screen and can't see/don't even know there's interlacing errors. Ahhhhh... sorry..... I'm just tired of fixing amateur errors and having to explain to the client why their bill is higher than first quoted. I don't like pointing out the flaws of others in a professional situation but I'm forced to these days in order to protect my own business. It's like no one pays attention to SMPTE/EBU standards anymore.

Exciting days with the new technology that's always emerging, sad days for quality and excellence.

[cananball]

Is the bug described here isolated to a specific version of FPC or all versions?

[Jfriah]

Originally Posted by Sonsey@mac.com

Geo and Will are correct...

1. FCP's realtime sample rate conversion is one step below useless - plus multiple codecs on a timeline, while sounding like a good idea, cause no end of problems, so have them check that.

2. There is a repeatable bug, that I've mentioned before I think, that if you open an FCP project and make changes, then output an OMF, the changes MAY OR MAY NOT be reflected in the OMF. Saving, closing, then re-opening the project before exporting will insure the OMF properly reflects the timeline. I have confirmed this on V6 of FCP... haven't had a chance to test V7 yet...

...don't forget she said FC belongs in the landfill... oh, wait, I misread.
(chuckle)


And THANK YOUR EDITOR for putting sync pops. You don't know how difficult that is to drive home for some of the ones I work with. THIS IS EXACTLY WHY----so things can be traced. Good one.

Jeff

[smurfyou]

Originally Posted by cananball

Is the bug described here isolated to a specific version of FPC or all versions?

Pretty much all versions that I know of.

[georgia]

yes. to the best of my knowledge its all versions.

cheers
geo

[davidl466]

Originally Posted by georgia

one of the really "fun" problems in final cut is exporting OMFs with various sample rates in one time line. What happens is that, although FCP allows these and does sample rate conversion on the fly, OMF can't. OMF is a very strong and very stupid tool. It looks at all the database info regarding the audio samples and builds it's database/files for transfer and doesn't care what the actual sample rate is. It just assumes that everything is the right sample rate.
As an example, if you have a 48Khz timeline in FCP but you have a couple 44.1 sample rate files, everything seems to be ok in the editorial FCP session, sounds right, plays correctly. But... OMF sees a file that it thinks is 48Khz, puts it at the correct sample point data in its database time line, but the file is "missing" samples so it's duration is shortened by the difference in samples due to the 48Khz/44.1Khz differential. Then it really starts.... the next file is places early on the track point within the OMF dataset the number of samples "missing" from the previous file. every file on that track is now early xxx number of samples. and it's cumulative. It just gets worse and worse, and its on a track by track basis.

I can't tell you how painful this can get if you've got an editor that isn't paying attention and is using 44.1, 48, Mp3, 32k files from multipe assets and is just dropping them on the time line instead of format concerting them to a standard for the session.

Sometimes the only way to fix a completely screwed session is to output the audio from FCP as AIF files that are the duration of the session. One per track.

cheers
geo

So Georgia, I'm actually having this EXACT same problem right now with my OMF's for a feature film that I wanted to ask you about...I've had several different sound mixers on the film and so am having several different sample rates....and so I've just gone through the entire project and it seems that the bigger majority of the audio was recorded at 44.1k/24-bit which seems sorta odd to me but anyway...I was at first exporting at 48k/16-bit but now it is obvious that since most of the big dialogue scenes etc were recorded at 44.1/24-bit that I should go with that...I was wondering if you see films in this format much or is it pretty much always 48k?

I'm also thinking that with my sync issues, that it would help to put just the 44.1/24-bit stuff on dedicated tracks and put the 48/16-bit stuff on other tracks so that at least maybe I can get away with having only the 48/16-bit stuff out of sync...if it ends up that in fact that is the problem...Anyway, hoping you can shed a little light on this for me :-)

David

[Max Holland]

Funny this thread got brought back up as I am now working on the sequel to this doc which has been cut from the same project apparently, lets hope this time around things go smoothly.

[cinetj]

This bug concerns only different sample rates files in the same FCP session or it is related to different bit depths also?

[georgia]

I get almost every project we work on , both in picture editorial world and in sound, at 48KHZ... once in a blue moon we get a 96KHZ session or a 44.1K. But close to 99% of all out projects are 48KHZ. Since I do a lot of our picture editorial and we work with clients early in the process, we normally AVOID these issues.... but every once in a while...


Putting the 44.1 K on one set of tracks and the 48K on others won't fix it. It will just isolate the issue on the 44.1K clip tracks...


its about sample rates... OMF is stupid and very strong. FCP is very smart and well... some call it a feature, I call it a bug.
Final Cut Pro performs real-time bit depth conversion and sample rate conversion when your audio file settings don’t match your sequence settings. Sequence audio is always mixed using 32-bit floating-point values


if you have multiple sample rates in FCP it plays them accordingly and it all works great, until you try to export using OMF. Then the problem hits. OMF sees a a clip that is 48KHZ with missing samples, instead of a 44.1k or 32K clip, because the session that created the OMF was running at 48KHZ. OMF just "slides everything up in time to fill in the missing samples".... and you, as a sound editor, have a sync nightmare on your hands.


Two ways to fix it.

1. call the picture editor and make them create proper sample rate files and replace ann the incorrect sample rate files in their session. (good luck...)

2. start replacing in correct sample rate files and slipping/sliding files until you get it all fixed... remember, it's accumulative when multiple clips with incorrect sample rates are on one track. And its specific to each clip(s) on each track.

Audio Formats to Avoid in FCP
The following formats and audio data formats should be avoided when editing in Final Cut Pro because they require real-time processing for playback: Any file containing compressed formats such as MP3, AAC, and Apple Lossless Codec, QuickTime movies containing compressed audio, such as MPEG-4 and H.264 files, Multiplexed video and audio streams such as MPEG-2 program streams and DV Stream files.




cheers
geo