BLOG

Audio Post Production Techniques

By Jake Harley posted Jan 26, 2017

Philosophy

Video is an excellent communication tool; churches everywhere have started realizing this and heavily utilizing video during weekend services. Something that is often overlooked however, is the quality of the audio that goes along with these videos. Even if a video looks great, it can still be ruined by muffled dialog or music that is too loud or too soft. 

For this reason, the audio for every video that is played in a weekend service is mixed in a controlled environment and tested multiple times in different rooms. Room acoustics are arguably the most important aspect of how audio playback will sound, along with the quality of speakers and speaker placement. Bob Katz, an award-winning mastering engineer, puts it this way: “Good speakers placed poorly in a bad-sounding room does nobody any good”.

Note that every room is different, so even if a mix sounds good through studio monitors, it can still sound muddy coming through a PA in a large room. For this reason, we actually export our mixes as a stereo file with dialog on the left and music on the right. This way it can be adjusted on a pair of faders depending on how it sounds in different venues.




Session Organization

I won’t spend too much time here, but session organization is important! When working with multiple speakers and lots of dialog and music transitions, you want to be able to navigate your session quickly and accurately. For this reason I do things like place markers (Memory Locations in Pro Tools), add a 2 pop and tail pop two seconds before the beginning and at the end of the video reference, and Color all of my tracks so I can identify things at a glance. 

It is worth noting that I also route all the dialog via an auxiliary bus to one stereo Aux track. I do the same with all the music in the session. I then route these two auxes to a stereo Mix Bus aux, and assign this mix to the Master Fader as the final output destination. Routing things this way has several advantages: I can see exactly where my music and dialog are metering, I can process them as a group if necessary, and it is very easy to set up a Left-Right mix like I mentioned earlier, as well as separate predubs for music and dialog to provide as backups to our video editors.

Editing, Cleaning, and Gain Staging

After listening to all of the dialog edits from our video editors, I ensure there are no clicks and pops and fix edits if necessary. This can sometimes be tricky if the edit is ‘baked into’ an audio file. When given an OMF, this is easier. 

After editing I will often import the dialog files into iZotope RX and do things like some preliminary dialog denoising, spectral editing, and render FX like EQ to clean up boominess/mud. 


If you’ve been working in audio long, you probably know how dynamic the human voice can be. It can get very loud and very soft very quickly. For this reason (and a multitude of others) Gain staging is very important in the post production workflow. You can see from the edits above that I use clip gain fairly liberally to achieve a more consistent level before I use any plugins. There are other options to get dialog to sound more consistent, but I find a combination of clip gain, compression, and volume automation generally sounds the best. 
There may be times in your video when there is no dialog and you want the music to ramp up or down. There are several ways to do this, such as volume automation or clip gain - I find my chosen method depends on what files I receive from our video editors, and in this case I was able to accomplish this with a simple crossfade as seen here.

Plugins and Processing

EQs are the most important tone-shaping tool in the audio engineer’s arsenal. Everyone’s voice is different, and thus no two EQ curves will ever be the same. However, there is a general mentality that I find helpful when EQing dialog to sit well with music. 

Hi-pass to a point where you aren’t cutting too much of the fundamental out of the voice, but you clean up the low end. You don’t want the voice to sound thin, just clean. Either boost or cut the fundamental frequency of the speaker’s voice, depending on how the dialog was recorded. Then deal with any midrange weirdness (often in the 2-4kHz area). Boost or cut top end to taste, and make sure your dialog isn’t sibilant, meaning the ’S’ sounds aren’t annoying or strident. This can often be accomplished by a small cut in the 5-8kHz range. If it is too harsh for EQ, use a De-esser in this range. 


As I mentioned before, the human voice is pretty dynamic. For this reason it is important to use compression to even out dialog, especially when it has to be heard clearly over a music bed. When compressing dialog, I find the most success with a low ratio and a low threshold, moderately fast attack and release. I typically use a 1.5:1 ratio and time the attack and release so that the compressor reacts quickly, and doesn’t hold over into the next word.



I find that compression often sounds the most natural when the compressor is activated all the time. This is where the low ratio/low threshold theory comes in. This way you don’t hear the compressor kick in suddenly when it passes the threshold like it would if you were using a higher threshold and a higher ratio.

There are about a million different choices when it comes to processing dialog - just remember that you should never put a processor on something unless you have a specific reason for it. Not every piece of dialog needs multi band compression! It is easy to over-process and mangle your audio, so be selective with your processors.

All this talk about processing, and I seem to have skipped over the most important aspect of a good mix - levels! 

Intelligibility is king in the world of dialog, so making sure that you can hear the dialog above the music bed is very important. After all, the entire message of the video is carried through the voice (unless there are subtitles).

This we have a general guideline for metering music and dialog here at Northview. We make sure dialog peaks around -3dBFS, while RMS levels on music hover around -18dBFS. This is another area where setting up subgroup busses comes in very handy. If all your dialog tracks are too quiet, but they are sent to a bus, you can simply raise the fader on the bus!




When Exporting, it is often a good idea to build in safeties for yourself and the other members of your team. After you preview the project and ensure that nothing is clipping, export multiple deliverables, such as dialog and music combined and solo’d. This ensures flexibility for whoever is exporting the final video, as well as a backup if there is a problem with a different mix you sent. 

Always remember, even if it sounds great in your studio it could sound wonky in a different room! 

Hopefully these tips help you out when mixing audio for your own videos. If you have any questions feel free to contact me at jake.harley@northviewproduction.us. 




Happy Mixing!

Christmas 2016: Christmas is For-Giving

By Evan Rodecap

The initial concept for the 2016 Christmas design was a ArtDeco or Great Gatsby look. Our immediate thought was gold, gold, gold. We were very intentional in creating a versatile set that would work well with all of our Christmas events and MultiSite locations with very little modification. With the ability to trim lighting at 38ft from the stage deck at our Carmel campus we also wanted to be intentional in creating a very even and big lighting look.


Our gear list is attached below. If you have any questions reach out to me at evan.rodecap@northviewproduction.us!


Happy designing!

Carmel Campus - Full Staging


Lighting Control:

Jands Vista L5 


Video Servers:

Coolux Pandoras Box, Casper


Video Playback:

ProPresenter, ProVideoPlayer


Video: 

Christie S+16k Roadsters

Christie HD12K Roadsters


Lighting:

6 - Martin Mac Viper Profile

7 - VL2500 Profile

8 - Elation 5r Extreme Professional Beam

6 - Martin Rush MH3 Beam

14 - American DJ 412 Professional Wash

26 - American DJ Kling Bar 54 LED

12 - ETC Selador Classic Lustor LED

14 - PixelRange RGBW PixelPar LED

6 - (In-House Fabrication) Straight Line 4lite Blinder

4 - (In-House Fabrication) Straight Line 2lite Blinder


Special Effects:

4 - Artistry in Motion Dual BigSHOT Blasters 

1 - City Theatrical Aqua Fogger 3300


LED Products:

640 feet of SuperNight 12v RGB Tape 

4 - 30AMP 12v 4 Circuit Power Supplies

4 - 8 Circuit 3 Channel 12v RGB Dimmers


Staging:

24 - 4x4 IntelliStage Variable Height Platforms


Set:

10 - PreLit Christmas Trees

4 - 10' CNC Cut Foam Trees

2 - 6' CNC Cut Foam Trees

1 - 10' CNC Cut Grid w/ 6' Included LED Tree

Custom Printed & Cut Coroplast Panels 


Supplies:

90ft 18AWG 4 Conductor Alarm Wire

50ft 12AWG 3 Conductor Power Wire

100 18AWG Crimp Butt Connectors

Multisite Campus - Compact Staging


Control:

Jands Vista S3


Video Playback:

ProPresenter, ProVideoPlayer


Video: 

2 Barco RLM-W14

1 Barco HDX-W20


Lighting:

4 - Rogue R2 Spots

12 - Rogue R2 Wash Zoom

4 - American DJ Kling Bar 54 LED

4 - Generic RGBW LED FlatPars 


LED Products:

140 feet of SuperNight 12v RGB Tape

3 - 30AMP 12v 4 Circuit Power Supplies

3 - 1 Circuit 3 Channel 12v RGB Dimmers


Set:

4 - PreLit Christmas Trees

2 - 6' CNC Cut Foam Trees

1 - 8' CNC Cut Grid w/ 6' Included LED Tree


Supplies:

25ft 18AWG 4 Conductor Alarm Wire

15ft 12AWG 3 Conductor Power Wire

50 18AWG Crimp Butt Connectors