Sync'ing double system sound w/ timecode

[Rippie]

Is it possible to sync double system (separate) sound in FCP
with time code stamped wav's and a time code slate?


In telecine, they can just freeze the picture on any frame of film with the slate
in it, punch in the time code number on the screen, it goes and finds the audio file with that time code..ya move on.


Ive been told a few times there is no way to do such a thing in FCP
and you have no choice but to sync every take by hand using the sticks.

Is this true?
If the regular way to do this is impossible are there any work arounds?

Im not an editor nor do I know FCP but it would help to sell double system
audio to clients if it wasn't such a time consuming process by hand.

I assume it would be easier if the audio recorder was jammed to the camera timecode but im interested
more in "film style". Recorder and slate have the same timecode and no timecode from camera or to camera at all.

Thanks guys

 

[DCSensui]

I just got an Edirol R-44 recorder and it does embed time-of-day timecode into the WAV files (this particular type of WAV file is called a BWF file).

The Sony EX1 also embeds time-of-day timecode into its files.

I can use the timecode to determine relative position of video to audio. But it won't provide me a frame-accurate sync because the two devices aren't locked together at any time.

To get frame-accurate sync I'd have to "slate" each shot with a flash or hand-clap. If there's nothing like that, I can sometimes use the sound waveforms to determine a precise alignment of the recorder's audio.

In my case I would have an audio recorder constantly running. The camera runs only when something starts to happen. In post I'll use the time-of-day timecode to locate the audio.

It's possible to get a frame-accurate match between a recorder and a camera, but you'd have to find equipment that are designed to do that. Otherwise something as simple as a slate or handclap will provide the necessary sync point. That's how it was done in the "old days" of film cameras with tape recorders.

Even when doing multi-camera jobs I'll have the talent clap their hands once which provides a perfect sync point.

Slating shots may or may not work with your project. It depends on the nature of your production.

 

[Brad S.]

Hey Rippie,
I've been looking into FCP lately and I believe the way to do this is:
1. Load your video and your audio separately into the bin.
2. Select the video and the audio in your bin, and from the Modify Menu select Merge Clips.
3. Select the timecode button and click OK.

That should create one clip with both video and audio. Both will last the full length of the media and start with whatever started recording sooner.

I'm a sound guy and I've never done this myself, but I just happened to me researching FCP when you asked this question and this is what the book I have says to do.

 

[Rippie]

Hey Rippie,
I've been looking into FCP lately and I believe the way to do this is:
1. Load your video and your audio separately into the bin.
2. Select the video and the audio in your bin, and from the Modify Menu select Merge Clips.
3. Select the timecode button and click OK.

That should create one clip with both video and audio. Both will last the full length of the media and start with whatever started recording sooner.

I'm a sound guy and I've never done this myself, but I just happened to me researching FCP when you asked this question and this is what the book I have says to do.

Hey Brad,
does the way you mention lock the "audio file timecode" and the "video timecode"?

Im looking to do the sync'ing using the manual slate (visual numbers on screen)

Think actual film. (no timecode on film)

 

[Rippie]

I just got an Edirol R-44 recorder and it does embed time-of-day timecode into the WAV files (this particular type of WAV file is called a BWF file).

The Sony EX1 also embeds time-of-day timecode into its files.

I can use the timecode to determine relative position of video to audio. But it won't provide me a frame-accurate sync because the two devices aren't locked together at any time.

To get frame-accurate sync I'd have to "slate" each shot with a flash or hand-clap. If there's nothing like that, I can sometimes use the sound waveforms to determine a precise alignment of the recorder's audio.

In my case I would have an audio recorder constantly running. The camera runs only when something starts to happen. In post I'll use the time-of-day timecode to locate the audio.

It's possible to get a frame-accurate match between a recorder and a camera, but you'd have to find equipment that are designed to do that. Otherwise something as simple as a slate or handclap will provide the necessary sync point. That's how it was done in the "old days" of film cameras with tape recorders.

Even when doing multi-camera jobs I'll have the talent clap their hands once which provides a perfect sync point.

Slating shots may or may not work with your project. It depends on the nature of your production.

huh?
I do appreciate your input did you read my question i was posing?

 

[DCSensui]

huh?
I do appreciate your input did you read my question i was posing?

I read your question.

The answer is: it depends on what equipment you're using.

If you can get camera and recorder to jam-sync their timecodes then you can use timecode to line up audio and video tracks, and get them in perfect sync.

In my case the timecode from the audio recorder and the camera are not perfectly matched. But they're accurate to within a second or so. That allows me to easily locate the approximate relative position of video and audio tracks. From there I can manually get a perfect sync if I have a slate point. It's a little more work than having everything automatically synced but getting four additional audio tracks is worth the effort.

You can use "merge clips" in FCP to line up all the tracks but that's only valid if the timecode of your video and audio tracks are a perfect match. Otherwise your sync will be off.

That's where a slate comes in. A hand clap or standard filmmaker's slate works well. Double system sound has been around since "talkies" first appeared on the silver screen. The same techniques of slating shots are still valid today.

 

[Rippie]

hah dean,
you really gottta read my question a little better.
Im not asking about manually sync'ing via the sticks,
nor am i asking for approximate TOD clock reference,
nor am i asking about sync'ing timecode from an audio file WITH "video timecode"

I do appreciate you taking the time to respond but you're not really even in the ballpark of what
im asking. I'll have to assume you really dont know what I mean.
Dont worry about it.

I read your question.

The answer is: it depends on what equipment you're using.

If you can get camera and recorder to jam-sync their timecodes then you can use timecode to line up audio and video tracks, and get them in perfect sync.

In my case the timecode from the audio recorder and the camera are not perfectly matched. But they're accurate to within a second or so. That allows me to easily locate the approximate relative position of video and audio tracks. From there I can manually get a perfect sync if I have a slate point. It's a little more work than having everything automatically synced but getting four additional audio tracks is worth the effort.

You can use "merge clips" in FCP to line up all the tracks but that's only valid if the timecode of your video and audio tracks are a perfect match. Otherwise your sync will be off.

That's where a slate comes in. A hand clap or standard filmmaker's slate works well. Double system sound has been around since "talkies" first appeared on the silver screen. The same techniques of slating shots are still valid today.

 

[DCSensui]

Rippie...

You're not comprehending any of this, are you?

As an very experienced FCP user and someone who makes use of timecode, I'm telling you exactly what you need to know in order to make an informed decision.

If you're going to sync audio timecode to video timecode, BOTH MUST SHARE THE EXACT SAME TIMECODE right down to the precise frame.

Doesn't matter if it's TOD or free run. BOTH MUST MATCH EXACTLY. Frame-for-frame. Or it's not going to work. That means the camera and recorder must be talking to each other all the time, or be accurate enough to not drift at all during the entire time they're operating if you jam-sync them at the start of the day.

My message says it clearly. "If you can get camera and recorder to jam-sync their timecodes then you can use timecode to line up audio and video tracks, and get them in perfect sync."

I use TOD TC because it suits my workflow. Other people might use free run TC. Either way works just fine if all equipment are exactly in sync.

Furthermore, equipment that CAN synchronize either wired or wirelesly is not cheap. That's why I'm suggesting you consider the equipment you're using and weigh that against your budget. And THAT is why I'm suggesting a manual slate. If a jam-sync or similar system is outside of what your company can afford, manually slating each shot provides a very viable cost-effective option.

Hope this clears up some of the confusion.

 

[Alex H.]

Dean,

What we're talking about here is postproduction workflow for non-sync, double-system sound.Rippie 's already well-versed in how to provide sync points during production, with and without TC jam-sync. The question was seeking a way to batch process video and audio clips to merge them together.

Unfortunately, there's not really a good way to do this except manually and clip by clip, as you have to match the frame of picture where the sticks come together to the spike in the wave form where the clap hits, then lock the two together.

The original question was not of how to sync sound to picture, but, again, how to batch-process the merge in post-production.

Rippie...

You're not comprehending any of this, are you?

As an very experienced FCP user and someone who makes use of timecode, I'm telling you exactly what you need to know in order to make an informed decision.

If you're going to sync audio timecode to video timecode, BOTH MUST SHARE THE EXACT SAME TIMECODE right down to the precise frame.

Doesn't matter if it's TOD or free run. BOTH MUST MATCH EXACTLY. Frame-for-frame. Or it's not going to work. That means the camera and recorder must be talking to each other all the time, or be accurate enough to not drift at all during the entire time they're operating if you jam-sync them at the start of the day.

My message says it clearly. "If you can get camera and recorder to jam-sync their timecodes then you can use timecode to line up audio and video tracks, and get them in perfect sync."

I use TOD TC because it suits my workflow. Other people might use free run TC. Either way works just fine if all equipment are exactly in sync.

Furthermore, equipment that CAN synchronize either wired or wirelesly is not cheap. That's why I'm suggesting you consider the equipment you're using and weigh that against your budget. And THAT is why I'm suggesting a manual slate. If a jam-sync or similar system is outside of what your company can afford, manually slating each shot provides a very viable cost-effective option.

Hope this clears up some of the confusion.

 

[Rippie]

Dean,

What we're talking about here is postproduction workflow for non-sync, double-system sound.Rippie 's already well-versed in how to provide sync points during production, with and without TC jam-sync. The question was seeking a way to batch process video and audio clips to merge them together.

Unfortunately, there's not really a good way to do this except manually and clip by clip, as you have to match the frame of picture where the sticks come together to the spike in the wave form where the clap hits, then lock the two together.

The original question was not of how to sync sound to picture, but, again, how to batch-process the merge in post-production.

thanks C2V.
I dont understand why multiple posts telling him he wasn't reading my question
correctly that he insisted that he was answering it.

 

[DCSensui]

The original question was not of how to sync sound to picture, but, again, how to batch-process the merge in post-production.

The point is that in order to get double-system sound to conveniently sync with the click of a mouse, you need to consider the entire system and provide the computer and software with accurate time data. You can't isolate one process from the rest and expect technology to provide a magical result. Camera, recorder and matching timecode are part of the system.

You can use "in" points to quickly align audio with video using the "merge clips" command in FCP. But it's a whole lot faster if you have matching timecode.

However, even roughly matching timecode -- accurate to within a second or two, can makes life a whole lot easier when it comes to correlating which audio clip goes with which video clip -- or flim strip.

TOD timecode is easily logged by an assistant: call "action" and the assistant just logs the time on his watch. The clips are easily found during the edit later.

But if you have no timecode to match video with audio, there's no way to log the material. And whoever has to sort out that mess by painstakingly scrubbing through hours of material will hate you.

And that's the point I was trying to make.

Maybe this might be helpful to someone out there.

 

[Rippie]

wow, just wow.

 

[Alex H.]

But if you have no timecode to match video with audio, there's no way to log the material. And whoever has to sort out that mess by painstakingly scrubbing through hours of material will hate you.

Then how did I end up with a log sheet from my non-time-code Zoom H4 from the last film I worked on, if there's no way to log that material?

Furthermore, any non-TC source that comes into FCP is automatically assigned TC that starts at 00:00:00:00. And the audio clips are logged by the file names assigned to them by the recorder, which have been logged in the field to correspond with their scene/take numbers. No "scrubbing through hours of material" required.

You're telling us that we cannot possibly be doing exactly what we have been doing. And what about folks who used recorders like the Tascam DA-P1 DAT for years? It has no time code, yet it was a very popular machine for double-system sound.

As long as the recordist has a proper log sheet, postproduction will not have any major issues...

... except, possibly, the desire to batch-merge.

 

[DCSensui]

Sure FCP will give the material an arbitrary timecode that starts at zero, but how does that correlate to the footage from the camera?

Here's my workflow as a one-man band on a small boat (no room for a soundman):

-- The audio recorder's clock is set to match the clock in the camera. It's set to record and left running the whole day. There's no soundman to start/stop the unit and this will capture everything the whole time.

-- At the end of the day there are two to three hours of coverage with a few hundred shots. Each have timecode that is a close, but not perfect, match to the audio track. However, all the shots are consistently off by less than a second.

-- After editing the material from the camera I can find the additional audio by using the timecode from the video clip to locate the identical timecode on the audio clip. The alignment isn't perfect, but if all the video clips are consistently off by 19 frames, it's a simple matter of shifting all the clips by 19 frames. Bingo. Clip found. Sync is achieved.

Unfortunately FCP can't mass-merge a lot of video clips. Not yet. But we can all send in that suggestion to Apple and hope for the best with a future version.

Meanwhile, aligning each piece of video with audio is still a piece-by-piece task. What I am trying to point out is that with TOD timecode, the tedious process of finding specific audio is made tremendously easier. In my case, since there's no PA or soundman to log anything, I have to be able to find dozens of specific chunks of audio. And TOD TC is the most efficient way to do that.

If the clocks don't drift a great deal during the day, then TOD TC will let you nail the sync even if all the shots aren't slated. Again, this is also very useful for documentary work where multiple cameras and a recorder may be operating in conditions where a slate isn't always practical.

If the producer hasn't yet purchased equipment then I'm suggesting they look at audio gear that have the ability to embed TOD TC: The Tascam P2 and the Edirol R-44 are among them. I'm specifically mentioning the Edirol R-44 because it's not noted anywhere that it's capable of doing that. I had to find that out on my own.

Anyway, I'm not trying to fire up anyone. Just providing possible solutions that have helped me and might help others. That's what this forum is supposed to be for.

 

[Alex H.]

Sure FCP will give the material an arbitrary timecode that starts at zero, but how does that correlate to the footage from the camera?

Here's my workflow as a one-man band on a small boat (no room for a soundman):

-- The audio recorder's clock is set to match the clock in the camera. It's set to record and left running the whole day. There's no soundman to start/stop the unit and this will capture everything the whole time.

-- At the end of the day there are two to three hours of coverage with a few hundred shots. Each have timecode that is a close, but not perfect, match to the audio track. However, all the shots are consistently off by less than a second.

-- After editing the material from the camera I can find the additional audio by using the timecode from the video clip to locate the identical timecode on the audio clip. The alignment isn't perfect, but if all the video clips are consistently off by 19 frames, it's a simple matter of shifting all the clips by 19 frames. Bingo. Clip found. Sync is achieved.

You are operating with a very unusual production work-flow with your audio recorder running constantly. That leaves you with one, large audio file covering your whole shoot, and yeah... that's going to demand a really different way of thinking in post. While it's great that you've found something that works for you, it is not they way that most productions operate.

And the OP was not referring to a one-man-band approach that you seem to think. Most of us who use non-sync, double-system stop and start audio recording with each take, which starts a new sound file each time (and each take is usually slated for picture and sound). That means that each take from the camera has a corresponding audio file all to itself. So even without TOD or jam-sync, it only takes the sound report to find the audio file from the same take as it has a file name that imports into FCP with the audio information. And a lot of us use record run on our cameras, rather than TOD/free-run.

Unfortunately FCP can't mass-merge a lot of video clips. Not yet. But we can all send in that suggestion to Apple and hope for the best with a future version.

Those sentences right there are all it takes to answer the original post.

Meanwhile, aligning each piece of video with audio is still a piece-by-piece task. What I am trying to point out is that with TOD timecode, the tedious process of finding specific audio is made tremendously easier. In my case, since there's no PA or soundman to log anything, I have to be able to find dozens of specific chunks of audio. And TOD TC is the most efficient way to do that.

If the clocks don't drift a great deal during the day, then TOD TC will let you nail the sync even if all the shots aren't slated. Again, this is also very useful for documentary work where multiple cameras and a recorder may be operating in conditions where a slate isn't always practical.

If the producer hasn't yet purchased equipment then I'm suggesting they look at audio gear that have the ability to embed TOD TC: The Tascam P2 and the Edirol R-44 are among them. I'm specifically mentioning the Edirol R-44 because it's not noted anywhere that it's capable of doing that. I had to find that out on my own.

That's great to know about the R-44. Strange that it isn't publicized.

But again, you are assuming that your particular work flow is the only work flow, which it isn't. And you're assuming that it's Rippie's work flow, which it isn't.

I'll also say that you have given a long tutorial on double-system sync to some one who is already well-versed in field production audio. I would be lying if I said that I don't get offended or lose patience when some one assumes what I don't know and tries to teach it to me like I'm an amateur, so I can see why Rippie was frustrated. You're a one-man-band producer/director/editor, who tried to school a working production audio professional.

Anyway, I'm not trying to fire up anyone. Just providing possible solutions that have helped me and might help others. That's what this forum is supposed to be for.

You're right. It is. And when an answer drifts from the point of the question, we need to steer it back to help the OP.

 

[DCSensui]

Fate is a funny thing. And at my age I try not to take anything for granted or shoot down other ideas from anyone. Because you just never know when you'll find the answer you're looking for.

Just this afternoon I got a call from someone who I worked with on an independent full-length feature film. Not everything I do is one-man-band, and I'm involved in a lot of workflows that include compositing, color grading, sound editing, etc.

I got the opportunity to work with this crew through a set of unusual circumstances. And that situation led to other life-changing events that just makes you wonder about how life works sometimes.

This producer/director/cameraman/editor just got back from NAB and during our conversation he just happened to tell me about a very interesting app that works with FCP.

It takes double-system material recorded and shot with TOD TC and lines them up for you. It can even be assigned an "offset" value to automatically bring all the clips into perfect sync, assuming they're all off by the same amount throughout.

I just tested it and the results blew me away. It's exactly what I was looking for, exactly at the right time. It took all the separate video clips and lined them up along the audio track, exactly where they should be in relation to the time code. I then grabbed just the video track and shifted them so all the clips were in perfect sync.

And the best part: It's free. For now.

http://www.spherico.de/filmtools/sequenceLiner/index.html

Aloha,

 

[Alex H.]

Fate is a funny thing. And at my age I try not to take anything for granted or shoot down other ideas from anyone. Because you just never know when you'll find the answer you're looking for.

Just this afternoon I got a call from someone who I worked with on an independent full-length feature film. Not everything I do is one-man-band, and I'm involved in a lot of workflows that include compositing, color grading, sound editing, etc.

I got the opportunity to work with this crew through a set of unusual circumstances. And that situation led to other life-changing events that just makes you wonder about how life works sometimes.

This producer/director/cameraman/editor just got back from NAB and during our conversation he just happened to tell me about a very interesting app that works with FCP.

It takes double-system material recorded and shot with TOD TC and lines them up for you. It can even be assigned an "offset" value to automatically bring all the clips into perfect sync, assuming they're all off by the same amount throughout.

I just tested it and the results blew me away. It's exactly what I was looking for, exactly at the right time. It took all the separate video clips and lined them up along the audio track, exactly where they should be in relation to the time code. I then grabbed just the video track and shifted them so all the clips were in perfect sync.

And the best part: It's free. For now.

http://www.spherico.de/filmtools/sequenceLiner/index.html

Aloha,

Excellent! I'll definitely check that out.

Please know, I wasn't trying to shoot down your idea. Simply trying to find a solution for the OP when the conversation went in a direction that was not helpful (even the OP pointed that out, yet you still argued that you'd answered the question). You have a system that you use in a specific situation... and it works for you. The use of TOD is not what's odd here... it's the fact that the recorder runs all day, even when the camera is down. That is what has fueled this debate, plain and simple. That one fact changes the post-production process drastically, and even changes the production process. If you're going to let the recorder run with and without the camera, then either TOD or free-run are absolutely the way to go and equipment purchasing should be considerate of that (assuming that there's still purchasing to be done, or one can afford to update gear to meet the need).

Non-sync (non-TC) is pretty common, be it with older (but proven) devices such as MiniDisc, non-TC DAT (like the Tascam DA-P1, which only has ABS), or the new generation of hand-held digital recorders. And we face another issue over long takes with non-TC recorders. Audio from these recorders drifts over time from the video as it has no reference clock to keep it in line with the SMPTE rate, so the audio has to be time-compressed (minimally... to about 0.998 percent of the original) so that the outpoint is in sync.

The best process? Probably not. But for smaller budgets, if you know the challenges, it works well. I can see a lot of challenges with letting even a locked recorder run all day, but you have figured out a way that makes it work well for you. And short of the link you just posted, which is great, it doesn't make much time difference in post to line up all the clips with either process. On one hand, you have a giant audio file that covers many, many takes... but you do have the TOD reference to find what you need quickly. On the other hand, you have individual audio files for each take, so it's not any more time consuming to open the corresponding audio and video clips, mark the ins using the clapper, and merge them together.

To be honest, the way we achieve 24p production from miniDV is not the best workflow in the world, and the HDV workflow is just as akward... but we know the challenges, and knowing those, they work perfectly well.

 

[Rippie]

thanks c2v for trying to explain better than I would have.

 

[cordvision]

I didn't read all the posts very carefully but if you have a time code slate it is pretty easy to sync the audio to the video in final cut pro. Pause the video when you see the time code slate (which displays the time code of the audio), then go to modify timecode -> AUX1 and punch in the number you see on the slate. Now, you can select the corresponding audio file and the clip, then merge them using AUX1. Final Cut doesn't look for the corresponding audio file by itself, but it is an easy way to get your video in perfect sync with the audio. If somebody keeps track of which take corresponds to which audio file, this is a pretty easy process....

(just had an idea, if all the audio files time code is unique (you don't have two audio files with the same time code), it might be possible for FCP to pick out the right one if you select all the audio files and the clip before you choose the merge command. Might be worth a try...)

 

[Rippie]

if this is indeed the case, this is excellent.
ive been told time and time again that it cant be done like this.
I'll pass this along to an editor I know and see what happens.

I didn't read all the posts very carefully but if you have a time code slate it is pretty easy to sync the audio to the video in final cut pro. Pause the video when you see the time code slate (which displays the time code of the audio), then go to modify timecode -> AUX1 and punch in the number you see on the slate. Now, you can select the corresponding audio file and the clip, then merge them using AUX1. Final Cut doesn't look for the corresponding audio file by itself, but it is an easy way to get your video in perfect sync with the audio. If somebody keeps track of which take corresponds to which audio file, this is a pretty easy process....

(just had an idea, if all the audio files time code is unique (you don't have two audio files with the same time code), it might be possible for FCP to pick out the right one if you select all the audio files and the clip before you choose the merge command. Might be worth a try...)