How to Proceed from Here

graywolf

graywolf

Well-known member
OK, now I have the F4, mic's, camcorders, etc.

I understand how to use the camcorders (Sony PDX10's), mic's, and even how to record through the F4 to the camcorders. But, what I need to learn now is how to link the higher quality tracks in the F4 (24bit, lower noise floor) to my video. I know of no way to lock the camcorders to the recorder to sync the sound in post (Premiere Pro CS5).

Suggestions? Pointers to tutorials? Whatever? Would be appreciated. Please remember that I am a hobbyist with limited resources.
 
The PDX10 has two XLR audio inputs. The two will be synchronised to the video. Why use an F4 and separate audio? I rarely use separate audio for the sort of interview/talking head videos you are looking at.

Try the 16 bit audio I think you will find it is good enough. Start simple. You are not doing a Hollywood feature film. I think you will find the microphones and their placement will be more important than the difference between 16/24 bit sound especially as most will be listening on PC speakers :-)
 
Last edited:
Create a new Audio Track below the one from the camcorder. Put the file from the F4 on the new track. Expand the tracks in the timeline so you can see all the details in the sound. Drag the F4 file to match the original file. If you hear an echo in the playback you're not quite there yet. Once you got it matched - and saved - you can delete the original sound. Lock the new file to the video clip and you can edit the video all you want, the sound will stay synced.

In real life, you probably didn't stop and start the F4 in sync with the camcorder so the file from the F4 will probably be bigger and contain stuff not on the video. You will, of course, have to trim this prior to the above procedure. To my knowledge, this is the most common way of doing this. Unless you can do timecode stuff.
 
Networker said:
Unless you can do timecode stuff.
Click to expand...

The RDX10 mentions it has a time code synch but uses the word "approximate" in the description :-)
The F4 also seems to have a time code sync.

that said: Given this is a talking head/interview I would try with the audio channels in the PDX10 first.
 
clapperboard-acrylic-black-white-clappers.jpg


This is the easiest way to handle non-sync sound. Sending audio from the recorder's mix out to the camcorders will give you a better "scratch" track, but the slate gives you a hard reference point from take to take. Line up the snap in the audio waveforms with the frame of video where the sticks come together. There are programs out there, like PluralEyes, that automate this process, but I find it just as efficient the way I work to go through and manually link the files in my NLE.
 
I was going to suggest a capper slate but I seem to remember that Graywolf is both camera/sound man and interviewee. It is probably just as easy for him to clap his hands. However I come back to the fact that the camera has 2 audio channels anyway.
 
If it's like the F8, the F4 has two main L/R outputs via TA3 connectors. Plug one or both of these into the camera's XLR inputs set to line. The cam's rec. volume will likely need to be cranked, since the Zoom's outputs are less than nominal +4 (another Zoom f__k up).
For docu inteviews and such, the camera's audio tracks are adequate for most clients, though I record ISOs if they want them.
I think the F4/8 has unbalanced mic level sub-outs as well for consumer type gear.
+1 Alex, even with TC and a sync audio trk., I usually clap slate as well.
 
jagraphics said:
I was going to suggest a capper slate but I seem to remember that Graywolf is both camera/sound man and interviewee. It is probably just as easy for him to clap his hands.
Click to expand...

Or have the talent clap their hands. As long as all cameras (I typically roll two) can see the hands.
 
Plluraleyes is probably the easiest way to get tight sync in a NLE like Premiere. Resolve has a similar feature built in. The problem with manual syncing in a NLE is that if it's even possible it usually takes some tricky moves to nudge the sound by less than a full frame. In a DAW like protools it is very quick to sync the waveforms visually and you can nudge by a sample.
 
Yes, many folks use PluralEyes. I recall the latest versions can compensate for drift .. (which TC cant), but it's usually not an issue with short takes
Magix Vegas (formally Sony Vegas.. prior to that, Sonic Foundry) also has the ability to easily move audio down to the sample level. It has a very good integrated DAW ... one of the best IMO.
 
I still think that at this stage separate audio is over complicating things. I would start with the audio channels the camera has though the XLR. Try that.

Remember most of your viewers will be on PC's tablet's and phones. It is not as though it is going to get amplified for a theatre or be complex audio. You will probably find that the room setting and microphone placement is more important that 16 vs 24 bit sound. It will also save a lot of messing about in post.
 
jagraphics said:
I still think that at this stage separate audio is over complicating things. I would start with the audio channels the camera has though the XLR. Try that.

Remember most of your viewers will be on PC's tablet's and phones. It is not as though it is going to get amplified for a theatre or be complex audio. You will probably find that the room setting and microphone placement is more important that 16 vs 24 bit sound. It will also save a lot of messing about in post.
Click to expand...

Since you have followed my posts in the Documentary thread, I think I will heed your advice, at least for that. As I said, I figured I was getting ahead of myself anyway.

On the other hand, learning how to do double system seems worthwhile in itself. If no one noticed, in that first video in the link I posted above, it tells you how you can sync double system sound with Premiere Pro itself. It looks fairly simple to do, although I have not tried it yet.
 
Having done a lot of post I will say that in my experience post syncing is not a big deal, providing the takes are not long and the clocks on the gear are running the same speed. If they are actually locked then the duration doesn't matter.

However having shot some tests and such as a OMB I will say that doing more than one job on a shoot even if it's just a test and there is no time constraints or money issues is a big PITA. I preached that "film was a team sport", but I was just doing some personal learning junk so... It was a PITA. It would have been MUCH more fun/ better to have at least one more person helping.

So adding dual system into that mix would be even a bigger PITA. I shot single system, but I wasn't planning on using the production audio. These were learning excursuses for shooting and editing. From the variety of stuff you talked about doing some seemed like they might be worth double system, the concerts for instance. Interviews I would probably go single because if you, the interviewer, are dealing with a bunch of tech you are going to strain your relationship with the person you are interviewing. You could always set up a set and forget second recorder as backup and a way to play with post syncing, but if it were me I would use the camera as your A roll.

Having been my one man crew and working on a few two man crews I can say that the extra person makes it 1/3 the work. My next round of excursuses I'm going to look for a minion. Even a teenager who is interested and trainable would make it better.

My hats off to the folks on this list that actually make a living doing OMB projects.
 
Compressed video/audio interleaved file formats like AVC do not have any inherent TC lock between audio and video. The only audio recording format that is absolutely timecode locked is broadcast wav PCM (.BWF). In an NLE the decoded video and audio are separate files on the timeline. Even if the cameras audio and video clocks were synchronized, often the computer's GPU and sound card clocks are not, with error rates in the .008 fps range common.
This is not enough to notice on most production length shots. It is enough for there to be significant sync drift on long continuous takes like interviews or event recordings.
The chances of any external recorder's clocks being any more precise than your PC's sound card are nil absent a positive timecode lock with an external master clock source for both camera and audio recorder.
Pluraleyes will match an external recorder's audio file to the camera's audio for sync by matching waveform timing more precisely than one can do by ear or eye. Even if you manually sync audio with video at the head of a shot, long shots may drift out of sync slightly by the end. Video and audio will show different total clip lengths by a few frames. Pluraleyes can correct these slight clock speed discrepancies.
I don't know how one would do this manually with any efficiency in Premiere, but in Lightworks I use a command called "fit to fill" to precisely match video and audio track lengths to the frame. lt involves marking I/O points at the beginning and end frames of video, then inserting a similarly marked audio clip using the command to correct the sync drift. It amounts to a subtle correction in playout speed so the audio sample rate clock matches the video.
 
You must log in or register to reply here.
Back
Top