Help! My sound is horrid because I have no idea what I'm doing

[jrolsan]

Hello!

I'm a rather new filmmaker and I'm just finishing up my 3rd short. However, I notice that my sound seems to suffer in all of my shoots. This makes sense since I know slightly more than nothing regarding sound.

I have an Azden SGM-1X shotgun mic attached to my makeshift boom pole that feeds into one channel of my Zoom H4 recorder.

When I record, the sound is left channel only (it's in Input 1 which is left channel for stereo) and incredibly quiet. I end up having to go into SoundForge and move the left channel to center and then up the volume on the recorded WAV file. The raising of the volume has a tendency to cause a lot of background static which lowers the quality of the audio.

Any tips or missing equipment recommended?

Jason

 

[clive]

I don't want to give tips on fixing bad sound in post, because there are very few good fixes on bad audio that aren't worse than reshooting or redubbing the audio in post.

However, with that said... here are a few tips on recording audio for film.

1) Never make a film without monitoring the audio... ie listening to what you're recording as you're recording it

2) Never record audio without having a way of monitoring the audio levels (because you can't judge levels by ear)

3) Get the mic as close to the action as physically possible and pointed directly at the actor who is delivering the line.

4) When your sound is bad, either reshoot or bring your actors back in to redub the lines (something you can only do if you wildtracked the ambience of each location)

 

[Will Vincent]

5) If your input levels aren't high enough, get yourself a nice preamp. It will accomplish a few things: 1, give you more headroom to work with and 2, increase the signal to noise ratio so you'll end up with not only louder audio, but cleaner audio. Most of them will also help shield against the horrid buzz that is often picked up from light fixtures and other "dirty" power sources.

 

[clive]

Amen to 5)

and

6) know the oddities of your camera's audio system -- for instance, you really want to control levels manually on pretty much every camera and avoid auto settings on audio, but for instance VX200's, VX9000's and PD150's all have this incredibly hissy inbuilt pre-amp which ramps up the noise when you go in manually, but ramps down on auto.

Basically it means pouring over the fine print in the audio section of the owners manual and then trawling the forums for that particular camera until you've got all the quirks nailed.

It's easy to mess up audio, even when you know what you're doing! LOL

plus

7) Your audio is only as good as your mic (there's no such thing as a good, cheap mic)

 

[Will Vincent]

If all else fails, record to an external device that's explicitly designed for audio recording and sync in post...

 

[jrolsan]

Currently, I'm recording on an external device (the Zoom H4) and am using an Azden SGM-1X shotgun mic.

I'm definitely seeing that point mentioned about getting it to work in production and not trying to deal with it in post.

Would the pre-amp put the audio into the center channel? It sounds like I need a pre-amp definitely for the raising volume.

How close is 'close enough' when using a boom? Usually the boom operator is keeping the mic about a foot distance between the actors and the mic. Is that too far?

Thanks again for the help so far.

 

[Will Vincent]

General rule of thumb, you want to be JUST outside the edge of the frame.. need to make sure it's pointed in the right place too.

You shouldn't worry about whether it's in the center or on the left or right so much... In fact by not having it centered it forces you to at least do something to your audio in post. TOO MANY people don't spend nearly enough time on sound in their films.

This is a good (albeit lengthy) read: http://www.filmsound.org/articles/designing_for_sound.htm

 

[Beeblebrox]

One problem might be your combining the left channel and right channel into one. I would try dumping the right channel entirely and using just your left channel as a mono dialogue track. It might not make any difference, but it's worth a shot.

If your volume is recording that low on set, then the problem is most likely incorrect level settings or a bad mic. It doesn't seem like an incorrectly aimed mic would produce such consistent bad results. I mostly feed my audio directly into my camera via XLR cable from a shotgun mic or a lav or both. Works just fine.

 

[clive]

I've had a quick look at the kit you were using and the most likely explanation for what went wrong is you didn't set your levels properly on the digital recorder.

In terms of the panning: most shot gun mics are mono, hence the single output. So you would need to change the settings in your recorder from stereo to mono, that ought to place the mono input across both channels, for future recordings.

This isn't vital, because you can copy and double up your mono track in post.

What I can't tell from what you've written is why you decided to record your sound to a separate unit -- were you shooting on film?

Except in fairly rare cases it's almost always better to record your audio into the camera. Even cheap domestic camcorders have excellent audio recording as standard. In fact, I've know people shooting on film who have used a DV camcorder to record the sound for their film.

In terms of your post production problem, low recording levels are almost unrecoverable -- however, with that said you might want to track down someone who has Premiere Pro, as the audio software that comes with that package is capable of unpicking some fairly nasty noise issues.

I was at the Adobe roadshow last year and watched a guy get pretty clean sound from a recording I would have considered beyond the pail.

The issues are fairly basic... in order to boost the signal you'll have to boost the noise, so to fix it you have to be able to isolate the frequencies that you want discarding the ones you don't. Premiere Pro's audio software is set up to do exactly that. However, I still think marginal sound is not worth the time in post and in your situation I'd sort out my recording problems and reshoot if I possibly could

 

[jrolsan]

Thanks, Clive!

That made a ton of sense. I don't think that the recorder has a mono mode. It seems to have 'stereo' and 'mixer' mode. I've been using stereo (hence the left channel), but I'm completely idiotic when it comes to 'mixer' mode. It's probably the right mode, I just need to figure that out.

I'll dig into the levels settings on the recorder tonight and do some tests.

The main reason I wasn't using the DV recording was that it was too ambient and echoey (is that a word?). It picked up the background as loud as the voices. Also, you get a little hiss from the camcorder itself (a weakness of the Sony HDR-HC7).

Thanks again for the help, all! If there's anything else worth pointing out, please let me know!

 

[knightly]

You'll want the microphone to be about 3-6 feet maximum from the talent and be pointed at their chest...which is the resonating chamber for the voice. Getting the mic closer will increase their volume relative to everything else, thereby reducing the relative volume of everything else...echoes, other actors, on set sounds.

If you're shooting master, single, reverse types of shots, mike the two actors centrally and turn the mike between lines in the master...then in each of the singles, keep the mike trained on the actor in the frame. When you use these to edit, you'll have really tight audio to use for each actor...and these can often be used to dub over the master shots as well to get fuller audio in the shot where the mike has to be farther away from the actors.

 

[clive]

The main reason I wasn't using the DV recording was that it was too ambient and echoey (is that a word?). It picked up the background as loud as the voices. Also, you get a little hiss from the camcorder itself (a weakness of the Sony HDR-HC7).

Ah... I understand now.

OK, the kind of sound you're talking about is caused not by the recording ability of the camcorder, but by the onboard mic. On board mics are an absolute NO NO, when it comes to recording audio for drama.

What you need to do is use your mic on a boom and then put the audio from that into your camcorder.

If you're lucky your camcorder has XLR in for audio... if not, then it probably has a mini-jack audio input somewhere and you'll need something like the Beachtek XLR to minijack adapter (which also has a niffty pre-amp)

I've just looked up the specs of your camera and it does have a audio line in... it's bound to be mini-jack.

Also, you get a little hiss from the camcorder itself (a weakness of the Sony HDR-HC7)

One of these days I'm going to figure out why it is Sony can't make a cheap camcorder without a degree of built in hiss, when their more expensive cameras are noise free. It's not that big a technical challenge and Panasonic manage it effortlessly.

However, the hiss isn't a huge problem in post (unless your levels are too low and you're pushing the signal).. that kind of hiss is fairly easy to filter out with the audio tools in Premiere and in FCP.

 

[Will Vincent]

One of these days I'm going to figure out why it is Sony can't make a cheap camcorder without a degree of built in hiss, when their more expensive cameras are noise free. It's not that big a technical challenge and Panasonic manage it effortlessly.

It's probably due to use of lowest-cost-possible parts for the amplification circuitry. A cheaper opAmp will always be noisy.. but you'd think that they could manage to spend an extra bit of change and get a higher quality one.

 

[clive]

I agree Will... except that the VX2000, VX9000 and PD150's all had notorious hiss issues and they were supposed to be serious professional videographer's cameras.

My guess is Sony has always differentiated between its professional broadcast cameras and its videographer's units.

I think this is also the reason Adobe have put much more work into intelligent, innovative fixing of poor sound, where FCP have lagged. Premiere have always targeted at working videographers over broadcasters, where FCP went all out to knock Avid off it's perch (and have been pretty successful in that)

It's a shame that with the current trends in the digital revolution Sony have continued to treat their "consumer" cameras differently from their broadcast cameras.

They fixed the preamp problems with the PD170, because it suddenly became a serious broadcast camera.

At the low budget end of cameras it has always been a trade off. Panasonic always have rock solid sound on their cameras (even the incredibly cheap ones), but their cheap camcorders have always had less in terms of shooting options. Sony has always offered remarkable picture acquisition options.

I guess most indie are waiting for someone to figure out that cheap $900 cameras are being used to make "real" films and throw in some of the functionality of the higher end cameras. How cool would it be if you could get a 24p, native 16:9 camcorder with XLR inputs, manual overides and an interchangeable len mount for under $1000?

 

[Will Vincent]

Agreed that would be mighty cool. (Though my new purchase gets me a goodly chunk of it: 16x9 native, 24p... and it's HDV -- will do full 4:2:2 1080 HD with a capture card via the HDMI port too) Can't wait to get my little HV20.. 2 more days.

 

[knightly]

Apple fixed their lagging sound capabilities with the bundling of everything into studio...now you do your sound work in SoundTrack Pro which has great audio scrubbing/filtering tools in it.

 

[Beeblebrox]

Is the Soundtrack Pro in FCS 2 any better than the one in FCS? Because I find it to be a potentially great but utterly buggy piece of crap that crashes so constantly that they really should take the word "pro" off of it. And that's using different copies on different machines.

 

[clive]

In the one area of noise reduction, Adobe's audio software is significantly better, simply because they went lateral with the problem.

Apple's system is set up with exactly the tools a studio engineer would use to solve the problem, "Low band pass" "high band pass" and "multi channel eq" which are great if audio is your area of expertise. The Adobe lads went at it differently, basically they converted the signal into a complex visual waveform... to extract the unwanted hiss, all you have to do is sample a few moments of "hiss only" from your waveform, this sound is then subtracted from the whole waveform, leaving clean, but slightly thin sound.

The thing that sets it apart is by making the manipulation issues visual rather than aural, they could then use all the tools they developed in Photoshop for the selection and manipulation of those waveforms and therefore the sound.

In terms of end results I wouldn't say the Adobe software was better overall at post production sound, however at noise reduction I think it is.

Of course I do have to add the slight caveat that my own version of FCP isn't the absolutely latest, so maybe I'm talking out of my hat! (No change there then)

 

[knightly]

I've got FCS 1, so if you've got soundtrack pro (which admittedly does crash every once in a while for me too), you probably do have the same as me. I've found it annoying that after sending a project to STP, making a change to a chunk of dialog operates on the whole damn DV file's audio (duplicate thereof)...huge waste of space and processing time. I understand that I don't import the way they intend for us to do, but that's kind of silly. It also won't allow you to grab nested sequences' audio, which is specifically how I work. My bad, but I'm not changing I have work arounds figured out now.