How would you cue audio

[Dave Baker]

Up to now I have recorded audio directly to a camcorder, but I have just bought a Zoom H4n to allow me some leeway in recording video in some circumstances. At times I will use it on the camcorder shoe and link the output into the camcorder, the plan then is to keep the recorder running, starting and stopping the camcorder as required.

As it's not a clapper board situation, I was wondering how you would approach cueing the camcorder starts in the recorder audio so I don't spent forever searching for the clips.

In testing I have simply spoken and it worked, but do you have any other, maybe better ideas?

Dave

[Edward Carlson]

How to Automatically Sync Dual-System Audio in Premiere Pro (With & Without PluralEyes)

[Dave Baker]

Thanks, but that's not what I am asking and in any case, I don't have Premier Pro.

Dave

[Richard Crowley]

You have created a situation for yourself that is typically solved using timecode. It is difficult to attempt to respond without such information as what camera and what editing software you are using.

Are you aware of the "PluralEyes" software which claims to sync audio and video automatically? Of course, since you have not disclosed what software you are using, we don't know whether this is a useful suggestion or not.

[Gary Nattrass]

Why use a dual recording system when straight into the camera will save you a lot of time and problems?

I never use a dual system if I am on my own and have only ever really done it when I was recording a music concert so just left the recorder and camera running for the whole duration and did a sync up before the edit.

[Robert Benda]

Dual system is more useful for us at a wedding if I, for instance, run a pastor's wireless mic to the recorder, along with a shotgun mic, then both into the camera. We use a 5d Mark ii so there is one or two frames dropped every 12 minutes. Having continuous audio is useful.

Here, you'll have to find a handy way to mark the files for yourself.

For me, I'll have the subject say the name and the date, always, when I first hit record. But then, I'm referring to weddings, again. If you're talking about frequent starts and stops, like in interview on the street situations, simply saying 'interview 1', 'interview 2' would work, if its just as much about identifying which audio file needs to be matched up as it is about syncing.

[Dave Baker]

Thanks for your replies, however we seem to be going off on a tangent. I am asking simply if you have any better ideas than mine for cueing camcorder starts, not about syncing audio in post, hence the reason I didn't mention what NLE I am using.

The idea is for shooting travelogues, e. g. when on a guided tour and the guide gives lots of interesting information in amongst the dross and I miss it because I am not recording video at that moment. If I leave the audio recorder running, I won't miss any useful titbits and I can add them in as part of the voiceover, only shooting video when the guide points something out that is worth watching. That bit will already be in sync. because it will be recorded by the camcorder.

Robert, your solution is what I was trying during my (short) tests, if that's the best idea then fine, I'll do it. As long as I can identify which video clip goes where in the audio stream, it will do.

Dave

[Greg Miller]

Are you using the Zoom's internal mics, or are you feeding some other signal into the Zoom's audio input?

*IF* you are feeding in an external signal, then I suggest you record that continuous audio on just one input channel of the Zoom. Feed your camera's headphone output into the Zoom's second channel input. You may need to devise some sort of pad to get the levels right. (Presumably the camera's headphone output will be silent when you are not filming.)

When you look at the audio in your NLE, the portions with audio on both tracks will correspond to the various camera takes. That will make it easy to see where each take starts and ends. (But of course you will NOT use the "second channel" audio in your final mix, because it will have some added noise and distortion after going through the camera. You will only ever use the "first channel" audio in your final mix.)

- - - - -

Otherwise, before you start the day, make sure the internal date/time clocks in both devices are pretty closely in sync ... at least within a second or two. Then by subtracting the timestamp on the audio file from the timestamp on each video file, you will come up with some "offset" numbers for each video file. Those numbers, in turn, will tell you roughly where to look on the unedited audio track.

e.g. if the audio track starts at 09:11:22 and a given video file starts at 09:23:45, then your "offset" would be (09:23:45) - (09:11:22) = 12:23. So if you listen at around 12 min 23 sec after the beginning of the [unedited] audio track, you should find the sound that belongs to that video take. After that, you already know how to get them in sync.

[Seth Bloombaum]

Originally Posted by Dave Baker

...I am asking simply if you have any better ideas than mine for cueing camcorder starts, not about syncing audio in post, hence the reason I didn't mention what NLE I am using...

The thing is, you're starting a workflow that ends in post. That's why you're getting these questions and proposed solutions. There is no "I just need to know how to shoot it, post doesn't matter" workflow.

Originally Posted by Greg Miller

...before you start the day, make sure the internal date/time clocks in both devices are pretty closely in sync ... at least within a second or two. Then by subtracting the timestamp on the audio file from the timestamp on each video file, you will come up with some "offset" numbers for each video file. Those numbers, in turn, will tell you roughly where to look on the unedited audio track...

^^^^^^^^^
This!

That's the kind of workflow I would choose first. Of course it depends on your editing software showing you timecode.

Generally, for starts and stops on multi-camera-recorder shoots of all kinds, free-running Time-of-Day timecode on all devices is the best first choice. It still leaves you with the fine sync to do by other methods, but, if you can at least get rough sync of takes about 90% of the work is done!

[Greg Miller]

If I gave the impression that time code is needed for my solution, I'm sorry. It is not.

You just go by the timestamp on each file. (This assumes that the recorder and the camera assign an accurate date/time to each file. It might be the time that particular record mode started; or it might be when record mode ended and the file was closed. In either case, you just do a little math to figure out where you are.) You can look at the file's date/time stamp in Windows Explorer, or whatever filesystem/OS you're using.

Of course I'm assuming the NLE shows a timeline (as opposed to timecode) so you can pick out the spot that's xx:yy minutes and seconds after the beginning of the audio file.

This may be a rather unorthodox procedure. But with a Zoom and an unknown camera, I think the OP will be unable to easily implement any more usual solution.

[Seth Bloombaum]

Originally Posted by Greg Miller

If I gave the impression that time code is needed for my solution, I'm sorry. It is not.

You just go by the timestamp on each file. (This assumes that the recorder and the camera assign an accurate date/time to each file. It might be the time that particular record mode started; or it might be when record mode ended and the file was closed...

...This may be a rather unorthodox procedure. But with a Zoom and an unknown camera, I think the OP will be unable to easily implement any more usual solution.

In my hurried reading this morning I did miss that - thanks for the clarification. You're right - a timestamp can be very helpful.

Another form of Time-of-Day. Since the OP is reluctant to share details of gear or post workflows, we're swinging in the dark here, but IMO Greg and I are on roughly the same page, whether implemented as timestamp or timecode. Specifics depend on field gear, OS, and NLE capabilities.

[Gary Nattrass]

As others have said you need to define you post processes before considering what is going on at the front end.

I see this all the time in broadcast and it is all too easy to go shoot and record everything that moves or speaks but if you ain't got the post procedures sorted you will spend an awful lot of time sorting it all out.

Timecode is the way pro systems are used to do this sort of thing as there is a common reference for all constant rolling machines and cameras that may be stop starting during a performance.

That said even with one system rolling and the other stop starting you will need to re-sync for each shot.

If you don't have a timecode reference then all you have to go with is manual syncing or eye matching waveforms so as you say "you will be forever searching for the clips"

That is how we have used the wheels of production and how it is easily done but if you wish to re-define the wheels and therefore ripple that into endless hours in post trying to match random clips with kit that is not up to the job then you may have a long journey ahead!

There may be other software solution such as others have suggested but if you aren't following a defined post workflow that works with them then they may be useless anyway!

[Dave Baker]

It's not that I am reluctant to give details of camera, NLE etc., it's rather that I didn't see the point since I thought I was asking a simple question, expecting answers like clap near the mike, hit it with a hammer or shout Geronimo!:-)

However you have taken me where I didn't expect to go and given me answers I almost certainly would never have thought of. But then, that's why I asked.

The camcorder is a Canon HF G30, the recorder is the ZoomH4nSP. My main NLE is Cinelerra, the original Heroine Virtual fork and I always transcode my footage to DNxHD for editing using Winff.

Both the raw camera files and Zoom files have a time stamp, Mediainfo shows them to the second and the file manager to the minute, but it is lost in the transcoded files. I will need to figure out how to incorporate checking this information into my workflow when necessary.

Thank you.

Dave

[Jay Massengill]

The only suggestion I would add is to state the camera's new file number aloud after you start recording (or just before if you have pre-record activated and you can keep up with the actual file number incrementing when you hit the record button.)

Stating aloud the few digits of the camera's new file number is about the least disruptive (shortest time) and most informative thing I know to do.

This should also survive the transcoding function since I assume the new files would at least have the same filename, but you'd need to check that.

If you speak distinctly it also gives you a reasonably easy sync by hand set of waveforms on the timeline.

You don't state if you're using external mics, but I assume if you are that some mic will be close enough going to both devices to pick up your voice clearly.

[Rick Reineke]

'Movies' pretty much always have verbal slates, regardless of the advent of TC and metadata.. been that way since the first 'talkie'.