Hallo,

i wanted to ask you what's the best workflow to capture the audio of an interview when recording trough a zoom h4n or TAscam while the video is captured by an D7100 Nikon dslr.
I used to record the good audio trough the external recorder and then i used to synchronize in post with the footage from my dslr.
My question is: If i connect the external recorder to my DSLR is there any chance to avoid in post the synch process? I mean if i connect the DSLR to my recorder i'll have a compressed audio or it will be the same good quality of audio that i drag in the timeline before synch?
Many thanks and sorry if my question sounds obvious

Most DSLRs do not sound very good no matter where the audio originated from. In lieu of timecode, many folks feed the audio from an external recorder and use 'PluralEyes' to sync the second system audio. Many NLEs support 'PluralEyes' and some even have a native sync process. Vegas Pro 17 and 18 has a native 'PluralEyes' like audio sync option, but reports indicate it does not work very well. Same with Premiere.

Sync is NOT a problem, and you certainly don't need software. My system is always the same. Three cameras in my studio and the main mic into the interface - so that is four tracks to align in Premiere.

I hit record on the four devices and then just clap my hands loudly. As long as each camera can hear the clap then the process is so simple in the edit. Slap the four clips onto the timeline. Move them all maybe a minute into the timeline to give slip room and look at the audio waveforms. Low level fuzz then a big spike. As long as you had silence before and after the clap you can roughly align them, then zoom in and move them frame by frame till they all align. on a 25p timeline, often one will by bad luck be half a frame from where it should be. If I have to make the choice, and one of the four is a half and half, I bring it early. I cannot hear half a frame. That is it. It takes less time than using software. If I am syncing my theatre work I clearly don't have a clap - but there is often visual sync markers. Now we use LED, lights going from full to off usually do it instantly, not a quick dim out as we get with tungsten, so you can use a lighting snap as a cue point - and this is easy to do. You can use moving objects - things flying in and you can see them touch the floor. That works too.

I constantly read people having sync issues and with tape we'd often lose an odd frame but with digits we don't. A sync point can be anywhere where something happens that every source includes. In fact - the waveform of the entire clip is often really easy to use.

One thing I've found is that the alignment of audio tracks to the video in my cameras is generally off (early), so any automatic method that uses the audio tracks would only be approximate. I would end up fine tuning the sync manually, which is my normal workflow anyway.

The attached photo shows how much I've offset the keeper audio (bottom) from the camera audio (top). It's only 12 milliseconds, but that was enough to look odd to me. It might not matter much with dialog, but with music 12ms makes a big difference.

I hit record on the four devices and then just clap my hands loudly. As long as each camera can hear the clap then the process is so simple in the edit. Slap the four clips onto the timeline. Move them all maybe a minute into the timeline to give slip room and look at the audio waveforms. Low level fuzz then a big spike. As long as you had silence before and after the clap you can roughly align them, then zoom in and move them frame by frame till they all align. on a 25p timeline, often one will by bad luck be half a frame from where it should be. If I have to make the choice, and one of the four is a half and half, I bring it early. I cannot hear half a frame. That is it. It takes less time than using software.

Paul, I used to do something like that too, but that method is severely outdated now. In Premiere, all you have to do is highlight the clips you want to align . . . right click on one of them . . . choose Synchronize . . . choose Audio . . . click OK . . . and the clips will be lined up almost instantly. I have never had it fail to get it perfect the first time and it takes less time than it took to describe the steps here.

Doug - please tell me you're joking. How did I miss that feature? It goes to show that no matter how well you think you know, you don't!

I will try it and have a play - and kick myself!!

Cheers Paul

Synching in FCP X for a multicam for instance is trivially easy. The same as Doug described for Premiere essentially. I haven't used Plural Eyes for years.

I just tried it and it had a think - and did it perfectly. Why did I not spot this. Annoyed (but very appreciative).

I have, however, just spilled a cup of coffee into my MacBook which is now hopefully drying out on top of my central heating boiler - I am not having a good evening.

Doug - please tell me you're joking. How did I miss that feature? It goes to show that no matter how well you think you know, you don't!

I will try it and have a play - and kick myself!!

Cheers Paul

Ha, ha. I felt the same way when I discovered it about a year ago. I have no idea if it was a recent addition or has been part of Premiere for years. Either way, it is a great tool to have.

I just tried it and it had a think - and did it perfectly. Why did I not spot this. Annoyed (but very appreciative).

I have, however, just spilled a cup of coffee into my MacBook which is now hopefully drying out on top of my central heating boiler - I am not having a good evening.
Sort of reminds me. Years ago I remember someone put their keyboard in a dishwasher to clean it.
Where I worked at this time I could often pick up a good keyboard that was being thrown out. So I took one home put it in the dishwasher then put it out on the washing line to dry. Once all nice and dry I plugged it in and hey presto every button worked as it should and it looked brand spanking new.

My doubt is:
if i connect straight out of the audio recorder a cable to the Input of my reflex, the quality of the file that my reflex is recording will be the same as the file recorded in the SD of the external Audio recorder?

DSLRs in general have very poor audio capability. Over the years I've discovered one thing that always seems certain - if you see a camera of any kind with XLR inputs with gain controls in addition to any auto features or undefeatable limiters, it has better audio performance than anything fitted with 3.5mm jacks. Electronically they should both be potentially as good, but it's a guide to the effort put into audio design.

My studio is rigged with cameras and a totally separate audio system with digital mixers that can output to the system. Better preamps and signal paths - in use, the best I can get. Or, I can simply plug a mic into an input on an interface and go into the computer via that path. There is a very small difference in quality as the mixer preamps are slightly better. Or, I can plug into one of the cameras XLR input, and the audio goes to the card with the picture. For convenience this is the most frequent route for quick in time, or less critical recordings - it's also the lowest quality I can accept. The camera method is the noisiest. With the two totally digital systems the dynamic range is so wide that noise isn't an issue, even when the track is recorded a little lower than optimum. On the video camera too much is distortion, but too little reveal hiss. One recorded last week hit -7dB on the peaks, and averagely most of the time lower and bringing that up cost me dearly. My DSLR always has very poor audio - often usable, but that's not good enough.

Re-recording your audio into the camera will compromise it 100% - no matter how clean it was, it will be worse by the extra analogue to digital conversion. Put it back in the edit in the digital domain. Many DSLRs also struggle with higher level input signals so you will need to knock the level down even more for the camera to then crunch even further when it brings it back up.

Keep in mind that light moves at about 186242 miles per second. Sound moves at about 1125 feet per second. That translates to about 37.5 feet per frame (for NTSC). Sound just naturally reaches our ears (and mic) later than the light reaches our eyes (or camcorder). So if shifting sound, the more distant view (camcorder) should have the later sound to better match in-person perceptions. And sound leading the visual is unnatural.

A clap, clapboard, rim shit, flash are all useful for syncing video and sound. Use what works for you and is available.

Many factors impact the quality of sound, not the least of which are ambient noise levels (both acoustic and electrical), the mics being used, the audio recorder (and factors such as whether or not audio AGC, high compression formats, low sampling rates, and limiting are involved). The way to be sure with your kit is to run some tests to see what flow gives you the best results. In high noise environments like typical receptions it probably doesn't make much difference much as long as there is no clipping.

I work with bad drummers to - plenty of rim shit. I think I'm going to steal this term and use it on a music forum - can anyone help me with a rim shit sound? Brilliant! Typos are always best when they make sense.

That translates to about 37.5 feet per frame (for NTSC.

I used to shoot lots of opera performance video with a camera in the rear of the house about 100 feet from the stage and had to sync audio from the camera with mics on stage. What I found easy to remember is that it takes about 1 millisecond for sound to travel one foot. So, for example, at a distance of 100 feet it would take about 1/10 of a second which is roughly equivalent to 3 NTSC frames.

Thank you guys!You are helping me in this audio world

I don't want to divert the topic too much but regarding sensitivity to time shifts, my last good Alpine car deck had 4 channel time correction. The dealer has known me for life and puts me on to things like that. I'm far from a sound snob but because I'd know how to apply it and appreciate it as more than a gimmick he presented that model.

I can't remember the timing range in milliseconds but even at 60 years of age with the onset of tinnitus I could definitely hear improvements and alternatively, degraded listening just in the confines of a sub compact with two door speakers and two drivers on the back shelf. Dial in individual delays for front left, front right, and rear left to "wait" to arrive at my ears at the same time as the furthest away, rear right and the music was golden. It is perceptible even if you can't really put your finger on what you just tweaked.

I haven't concerned myself with it in video editing or live music recording - yet.

Years ago I sold quadraphonic hifi and remember the effect of the signal suddenly locking in. We would do the demo and watch people suddenly smile. The got out their wallets almost every time. What few realised that the width and depth of the time aligned zone was tiny, localised to about one chair’s width. If you had two people on a sofa, it was not possible to get both in the same zone, so we recommended a prime listening chair, a nice comfy seat for just that one person’s listening. Everyone else got something pretty average, with the time alignment working against them.

It’s dangerous, because your adjustments and tweaks may not sound the same everywhere, which is why my belief is edit suits and recording studios should try to replicate average listening environments.