Help with Audio Correction/Mix Strategy

Hi -
I've had a pretty steep learning curve with some tough audio obstacles in my video clips.  Wind noise, background traffic noise, audio spikes from me banging around metal objects (automotive repair content), and the toughest part it seems...audio that is too quiet, with some sections where I turn my head away from the camera and it gets quieter!  I even have some audio of other people in the background that is very quiet and I want to boost those levels to be understood, with the rest of my voice.

I was going along fine (I thought), using the FCPX Voice Isolation to cut out a lot of the unwanted background noise (this works well).  I found a plugin called CrumplePop for removing wind noise (expensive model to get just this one plugin, but I figured I'm still learning - I'll deal with it for now), and it works pretty well. 

I managed the audio spikes by cutting the video clip and reducing the volume of that section, using the "Volume" adjustment in FCPX, while adding an Alex4D sound transition to clean up the audio on each side of the cut.  In the remainder of the clip, I raised the audio using the "Volume" adjustment (all the way to +12dB), and in some cases...even adding more than the +12dB to the clip using the Gain effect.

That all worked pretty well..considering the hours it was taking to do all that on each 10 minute video.  And I have hours and hours of footage like this.  Additionally, I was having some issue getting the audio "reduced" by Youtube, because I was using my "ears", the Macbook speakers, and the FCPX audio meter to adjust volume levels, and when the projects are exported, they must have been too loud.

I also wanted to save some time, so I posted a few messages on the FCPX FB group to get some help.  One idea was to use the Compressor effect, rather than the Volume adjustment in my clips, because the Volume adjustment was considered a "post effect", and would be applied after other effects were already applied.  Apparently, the Compressor effect would be applied along with other effects, rather than after.

So, I went through the learning curves of compressors (and Limiters) so get a reasonable output, but it really isn't saving any time.  And, frankly, it really isn't better sounding (maybe I still stink with Compressor settings).  In some cases, I think it's worse.  Plus, I'm a bit confused now, since I've learned that Youtube monitors LUFS, not dB as represented in the FCPX audio meter.  Also, appears there is some controversy as to whether versions of FCPX after 10.6.5 represented the audio meter differently (the built-in meter vs. the Compressor effect meter vs. the Limiter meter, etc.). 

So, I upgraded to 10.6.8 and it seemed to break my CrumplePop plugins (turns out the the Apple Compressor is ignoring the CrumplePop plugins in 10.6.5 also, so that wasn't the primary issue).  I figured out how to use the Time Machine and restore the old version of FCPX 10.6.5 to get back to where I was, without those complications.  I think I'm back where I started now, and I'd like to learn a better strategy for fixing audio and preparing a final sound mix.  I have a good handle on mixing the SFX and music levels, but the dialogue audio correction is hurting me.

What am I doing wrong?  It is more productive to use Compressor effects on the audio portion of each individual video clip, instead of the Volume adjustment?  I noticed that neither the Volume adjustment, nor the Gain effect change the actual LUFS level in the MultiMeter effect either. I could have sworn it was affecting those levels before I went back to FCPX 10.6.5, so maybe there is a bug here?  Or, perhaps there was a major change in how the effects are displaying?  Maybe that speaks to those people in the FB group who commented about the audio meter in FCPX 10.6.5 vs 10.6.6 and beyond, showing a -6dB difference now, for the same audio level??

Any help would be greatly appreciated!

mattj wrote:

<snipped>

Or, perhaps there was a major change in how the effects are displaying? Maybe that speaks to those people in the FB group who commented about the audio meter in FCPX 10.6.5 vs 10.6.6 and beyond, showing a -6dB difference now, for the same audio level??

Any help would be greatly appreciated!

The pan law as implemented in FCP results in a 6 dB difference in levels when changing between mono and stereo. That might explain some of what you're experiencing.

Typically, compressors and limiters are among the last things added to audio in a final mix. However, if individual clips are problematic, it may be useful to use one or both.

When you select a clip, you can then (with the audio inspector open) apply clip level volume changes, as well as audio component level changes. That's a total of 24 dB if you use both.

The hard lesson in this is probably to learn how to better/properly mic things when filming, so you start off with good audio, rather than "fixing it in post". As you've noticed, there are some amazing tools available that can almost work miracles with audio issues, but they're supposed to be used more for the exceptional cases, rather than the majority of cases...

And, yeah, dialogue editing, sound cleanup, and mixing are very laborious, as you now know.

Another caveat... various meters can show different levels in FCP, especially depending on what's selected in the timeline. There are some quite good LUFs meter plugins available for FCP, including some that will change levels to meet a specified required LUFs value. Here's a free LUFs meter (there are others):

youlean.co/download-youlean-loudness-meter/


Using audio roles may also help separate different types of audio, making it easier to mix things. You might check out the FCP user guide or this white paper for more info:

Audio Roles White Paper

DaveM wrote:

mattj wrote:

The pan law as implemented in FCP results in a 6 dB difference in levels when changing between mono and stereo. That might explain some of what you're experiencing.

Typically, compressors and limiters are among the last things added to audio in a final mix. However, if individual clips are problematic, it may be useful to use one or both.

When you select a clip, you can then (with the audio inspector open) apply clip level volume changes, as well as audio component level changes. That's a total of 24 dB if you use both.

The hard lesson in this is probably to learn how to better/properly mic things when filming, so you start off with good audio, rather than "fixing it in post". As you've noticed, there are some amazing tools available that can almost work miracles with audio issues, but they're supposed to be used more for the exceptional cases, rather than the majority of cases...

And, yeah, dialogue editing, sound cleanup, and mixing are very laborious, as you now know.

Another caveat... various meters can show different levels in FCP, especially depending on what's selected in the timeline. There are some quite good LUFs meter plugins available for FCP, including some that will change levels to meet a specified required LUFs value. Here's a free LUFs meter (there are others):

youlean.co/download-youlean-loudness-meter/


Using audio roles may also help separate different types of audio, making it easier to mix things. You might check out the FCP user guide or this white paper for more info:

Audio Roles White Paper

Thank you!  
Ah, that pan law may definitely be part of what is happening here!  Does it get 6dB louder when it is switched to Stereo?  

Yeah, I'm definitely learning that the audio recording must be better.  I am feeling that now. . 

Thanks for the thought on the LUFS meter.  I did download that one today, and it doesn't work with FCPX 10.6.5 for some reason.  I'm working with their support team to fix it.  

Thanks for the thoughts on audio roles.  I am using that and that has improved my workflow immensely!

I'll read through the white paper too, and see if there is more to learn! 

Thank you!!