Potential syncing organizational nightmare in the works

[Stephen Hall]

My general practice is to organize and rename files after the shoot and before delivery.

Here's what I typically do to get audio to the editor:

1) use Sound Devices Wave Agent software to bind the individual WAV files from a take on my R44 Pro to a single poly-wav file.*

2) rename the poly-wav file per scene and take number -- this is by listening until I hear the slate. E.g., S041_T03

3) any wild sound takes, location effects I captured, room tone, are labeled appropriately, as: WILD_Feet_on_Carpet_Office_Scene

4) everything is bundled together with a text file containing my contact info, scene-notes if required ("car horn in the middle of Joe's line", "lav clothing rustle when Sue sits on the couch"), EQ suggestions if I have time, explaining how to drag a buried lav's sound back to the surface.

Everything is written to a data DVD and handed off within a day or so of the shoot. Extra time, but it's what I feel I need to do to support post.

Stephen H

* the poly-wav files can be dropped onto the timeline of an NLE and they just become a group of linked, synced audio tracks. Fast and easy.

 

[davefisk]

yeah. this is going to be tedious. you are going to need to listen to every file, but even then you won't know which take is the select take until you drag them into the timeline and try to match them up with the picture. Unless it's reality tv, I don't care how "run n' gun" a shoot is. Sound reports are extremely important, and this is why. This is also why timecode can really help. If everything was synced up with timecode, you could have easily done the sync when importing your video clips with audio. Syncing is the very first thing done in post production....before a frame of picture is actually cut.

 

[parallactic-pictures]

Hey everyone,

I wanted to thank you all for contributing to this thread with your advice. I can't tell you how much I appreciate that. Here's an update on the situation: I got in touch with one of the two sound recorders that was on set for the shoot. He encouraged me to download the 1 month fully-working trial of DualEyes and then to use it to scan through all the audio and match up the video. Surprisingly, this worked... sort of. I haven't done it with all the video selects yet, but I tested it with a few big video takes (the raw video files of the take in my timeline) and, while it did generate around 3 finished 'synced clips,' only one of them would obviously be correct. The other two would be misfires-- I guess b/c of their similar wave form audio clips. Still, seeing that this was even possible gave me hope.

Next, the sound recorder revealed that he had purposely recorded all sound low, so as to "get a cleaner take." So he told me to download a program called The Levelator which would audio boost the clips. Well, I did this and I have to say, I wasn't too happy with the results. In fact, I think the camera audio sounded better, and more natural. So at this point, I may try some audio boosting from tools already natively available in Final Cut X. And if that doesn't work, then I'd show it to a sound engineer and get his/her take on it.

My question for you all, is, what's the best way to do an audio online? I'm not even sure if that's the correct name for it, so I'll just explain what I want to do:

- First I'll try exporting an HD Quicktime of my project from Final Cut X.
- Then I'll take that Quicktime and load it into DualEyes.
- Then I'll load in all 400+ of my raw audio files and perform the operation until the Quicktime timeline output is synced up.
*Note: I'm sure this won't be so easy, as DualEyes has proven it misfires a lot.
- Regardless, once I have some sort of synced up file, I'll reimport the newly synced Quicktime file back into FCX.

If this is not doable for some reason, then yes, the more difficult 'audio online' will have to happen. Similar to what an editor does when they re-import the original raw files onto the timeline in place of the optimized / compressed video files... I want to do this for the audio files. Meaning, I may have to individually sync all my raw video files with their appropriate audio files, and then reimport these files into my project. Is there an easy way to do this without overwriting aliases or stepping on the original raw files (the ones without the synced audio?).

I know I'm making this question way more complicated than it needs to be. But hopefully a few of you here know what I'm asking and can lend some advice. Thank you.

 

[GaryNattrass]

I just love it when people go off and shoot stuff without being organised then dump it on the post guys to sort out.

If people took the time to do the research there are tried and tested workflows that have been in existence since Al Jolson sang in the early talkies.

People with DSLR"s in particular seem to think that just buying a camera and a recording device makes them movie makers and they strive to re-invent the wheels and get caught out time after time.

There are always three rules in recording location sound:

1: Keep it simple and don't over complicate things that will cause sync problems later.

2: Always sync the rushes before anything gets edited.

3: Don't try to cut corners thinking you will speed up the edit.

32 years ago one of my main jobs was to work in the transfer bay creating 16mm sep mag reels from the nagra 1/4 inch tapes that came in as dalies, everything was logged on report sheets and the whole syncing ran on a 50hz pulse sync system.

Once the rushes were transferred to 16mm mag an assistant editor then sync-ed them up to the pictures on a shot by shot basis, with timecode or any other modern separate recording device the same applies today and there are some processes that new media/dslr people think may be out of date but they are still valid and have always existed for a reason.

 

[parallactic-pictures]

I just love it when people go off and shoot stuff without being organised then dump it on the post guys to sort out.

If people took the time to do the research there are tried and tested workflows that have been in existence since Al Jolson sang in the early talkies.

People with DSLR"s in particular seem to think that just buying a camera and a recording device makes them movie makers and they strive to re-invent the wheels and get caught out time after time.

There are always three rules in recording location sound:

1: Keep it simple and don't over complicate things that will cause sync problems later.

2: Always sync the rushes before anything gets edited.

3: Don't try to cut corners thinking you will speed up the edit.

32 years ago one of my main jobs was to work in the transfer bay creating 16mm sep mag reels from the nagra 1/4 inch tapes that came in as dalies, everything was logged on report sheets and the whole syncing ran on a 50hz pulse sync system.

Once the rushes were transferred to 16mm mag an assistant editor then sync-ed them up to the pictures on a shot by shot basis, with timecode or any other modern separate recording device the same applies today and there are some processes that new media/dslr people think may be out of date but they are still valid and have always existed for a reason.

Thank you for this, Gary. A few other people said the same thing, and I agree with it all. But can you help me with the specific question I asked above this? At this point, I am where I am, but I think I've found a solution. I'm just not exactly sure how to do this solution.

Thanks!

 

[Alex H.]

Next, the sound recorder revealed that he had purposely recorded all sound low, so as to "get a cleaner take." So he told me to download a program called The Levelator which would audio boost the clips. Well, I did this and I have to say, I wasn't too happy with the results. In fact, I think the camera audio sounded better, and more natural.

I don't even know what to say to this. It's a cop-out on the recordists part. There's no excuse for audio levels that are way too low, and if there's a risk of overmodulation due to loud screams or other sudden peaks in the script, there's always the safety track (a backup recording that's set lower, while the primary record tracks are set higher). Getting "a cleaner take" is about actively working with recording levels throughout production, be it through mixing or through mic placement. With digital audio, recording too low uses too few of the recording bits, and bringing up the level will leave the audio sounding like a very bad Internet download.

- First I'll try exporting an HD Quicktime of my project from Final Cut X.
- Then I'll take that Quicktime and load it into DualEyes.
- Then I'll load in all 400+ of my raw audio files and perform the operation until the Quicktime timeline output is synced up.
*Note: I'm sure this won't be so easy, as DualEyes has proven it misfires a lot.
- Regardless, once I have some sort of synced up file, I'll reimport the newly synced Quicktime file back into FCX.

Is this going to give you access to the individual audio clips once you're back in FCP? Also, I'm not sure that DualEyes is going to perform any better that way than with any other attempt you've made.

Unfortunately, I think that you're in a situation that's way too convoluted for relying on automated software to fix the problem. It really is going to take listening to each file and finding the takes you've used, one-by-one, and syncing them manually in the time line. I know that's not the answer you want, but it seems to be the reality at this point.

 

[henryqiu]

I'll just explain what I want to do:

- First I'll try exporting an HD Quicktime of my project from Final Cut X.
- Then I'll take that Quicktime and load it into DualEyes.
- Then I'll load in all 400+ of my raw audio files and perform the operation until the Quicktime timeline output is synced up.
*Note: I'm sure this won't be so easy, as DualEyes has proven it misfires a lot.
- Regardless, once I have some sort of synced up file, I'll reimport the newly synced Quicktime file back into FCX.

If this is not doable for some reason, then yes, the more difficult 'audio online' will have to happen. Similar to what an editor does when they re-import the original raw files onto the timeline in place of the optimized / compressed video files... I want to do this for the audio files. Meaning, I may have to individually sync all my raw video files with their appropriate audio files, and then reimport these files into my project. Is there an easy way to do this without overwriting aliases or stepping on the original raw files (the ones without the synced audio?).

If by "an HD Quicktime" you meant a single continuous movie clip (or one with all edits burned in), I doubt DualEyes would save you more time than it would cost you.

BTW, I use PluralEyes and Premiere Pro. By my experience PluralEyes works best when it's fed with original files, not ones split or trimmed. And the audio part of each video clip to be synced had better be kept intact. A level-boosting program, if all it does is to raise audio volume, does not change audio signature. So IMO it offers no help to the syncing program. In fact, one of the features that PluralEyes offers is to boost audio level. So it obviously knows how to handle the audio level, even if you may think it's very low.

To me, it does not matter whether to sync before edit or to sync after edit. That's because I don't use PluralEyes to help me create the final sequence. To me, all that it does and that it does really well is to create standalone synced sequences for reference purposes only. I have never tried to turn any synced sequences created by PluralEyes into my final sequences. Whenever I need certain better/different audio from the audio recorder for a particular visual, I reference the synced sequence(s) to find/cut out the audio clip with which to swap out the existing audio on my edit sequence.

If I were you, I would most likely try to have PluralEyes create a synced sequence using the original clips, not clips already trimmed or split. Actually, I might try to create two synced sequences. One using the clips ordered by the time they were taken. One using the clips whose exact creation times cannot be determined. Of course, some clips might have to be put on both sequences being created to feed to PluralEyes for syncing. In PluralEyes I would toggle a check box to indicate which is the sequence of ordered clips, and which is not.

By my experience there always seem to exist clips that PluralEyes does not know how to sync. Some of such clips are actually very easy to manually sync, but for some reason... a tool is just a mindless tool. PluralEyes is very accurate IMO. But it does make mistakes. Many of such mistakes are not hard to find and fix, as long as you have carefully compared the waveforms of both the audio from the camera and the audio from the audio recorder side by side on the timeline. Some mistakes can be hard or even very hard to find/fix: for example, when the cameraman took a photo without stopping video recording. It's also possible that with 7D, the visual and the camera audio are not exactly matched - I don't always know why - this may not always be obvious until the camera audio has bee replaced, so listening to the final clip is always a must.

I have no idea why to chose DualEyes instead of PluralEyes. Anyway, PluralEyes has no problem taking on more than dozens of files. It's a great inexpensive tool. Although having it chew on 100s of files might take quite a while to finish, especially if you toggle a check box to force it to work really hard, a great tool if used properly can save a great deal of time.

 

[Noiz2]

I have no idea why to chose DualEyes instead of PluralEyes. Anyway, PluralEyes has no problem taking on more than dozens of files. It's a great inexpensive tool. Although having it chew on 100s of files might take quite a while to finish, especially if you toggle a check box to force it to work really hard, a great tool if used properly can save a great deal of time.

Well PluralEyes was designed to sync up a bunch of video files based on the sound track. Folks started to use it to sync sound but actually that wasn't what it was designed to do. It works pretty sell on master clips but doesn't work well on an edited sequence.

DualEyes was designed to sync sound. It is also supposed to be able to work on an edited sequence.

 

[stefancolson]

With 24-bit audio you can record pretty low and boost the levels in post for clean, problem-free audio. Not uncommon if the boom op is also the mixer, or if you're recording straight to a recorder without a mixer in the signal chain. In such a situation (very common on extremely low-budget shoots) there probably isn't a backup recording or the equipment to make that easily work. If the audio files are really low and 16-bit then you are in serious trouble. I hope for your sake they recorded at 48/24.

You should always sync your audio before the edit. Then it's a simple matter of lining up the slate clap. Because your footage is all chopped up you're going to have to sync based on individual words or syllables. Not that much more difficult, but much more time consuming. ALWAYS sync your audio before the edit.

As far as listening to all of the audio files goes, that's just the nature of the beast. If you record separate system sound then you'll have to verify the slates on each take to label. In a bigger production you'd have an assistant editor to do that grunt work for you, but someone has to do it. If you have great shot logs then this is an easier process, but still really can't be avoided.

There's an easy way to make this process easier on you though. Simply sort all of your files by "date created". If you do that to you audio and video folders, everything will be in the order it was shot. Now you can much more easily find what you need. This has saved my ass many times.

The key to staying sane in the edit is to organize, label, and sync everything before you start editing. It's a tedious process that never gets more enjoyable, but it will save you a significant amount of hassle and headache in the long run. Be thankful that you're encountering this on a short and not a feature.

 

[henryqiu]

Well PluralEyes was designed to sync up a bunch of video files based on the sound track. Folks started to use it to sync sound but actually that wasn't what it was designed to do. It works pretty sell on master clips but doesn't work well on an edited sequence.

DualEyes was designed to sync sound. It is also supposed to be able to work on an edited sequence.

This is not what the official FAQ states. By official FAQ, DualEyes is designed for those who don't run an editing system supported by PluralEyes, and if possible, PluralEyes is always preferred. Here are the officially stated differences between the two products.

[h=3]What is the difference between DualEyes and PluralEyes?[/h] DualEyes was created for those who can't use PluralEyes because they are using an NLE that it does not support. DualEyes is a standalone application and can be used with any NLE.

PluralEyes and DualEyes are used for similar tasks, with the following differences:

1. DualEyes is mainly for a dual-system audio workflow. PluralEyes is also for multi-camera.
2. DualEyes works on files selected from the file system. With PluralEyes you need to create a sequence in the NLE.
3. DualEyes creates new media files: trimmed audio and (optionally) video clips with replaced audio. PluralEyes creates clips in the NLE that behave in the same way, but no new files are created on disk.
4. DualEyes requires the new files to be imported into the NLE, and possibly matched up. PluralEyes creates a new sequence with all the clips laid out, ready for editing.
5. PluralEyes lets you use markers to give it a hint as to how to sync.
6. (Final Cut Pro) PluralEyes automatically creates a multiclip (when possible).

If you use an NLE that is supported by PluralEyes, we generally recommend that you choose it over DualEyes. It offers similar functionality and has the added convenience of integration with the NLE. But both products are available as free trials that are fully functional for 30 days, so you can experiment with both to see which works best for you.