There’s no magic wand waving when it comes to audio post production – The golden rule is to ‘get right at the source’, that is, try to record your source audio as best you can during the shoot. Sometimes though, different takes from different days need to be monkey-hooked together, or stock audio is needed, low budget ‘Foley‘ must be used, combined with found sounds such as from freesound.org, for that matter. For most of us re-recording is simply not an option and so we look to post production to clean, smooth and enhance our scene’s sonic landscape.
In It’s Right Place
Don’t use a tonne of different plugins in any random order. Audio degrades rapidly when pushed and pulled to extremes, much more so than video. Under-process in post and you’ll discover that with just a few plugins in your chain you can fix messy audio very quickly.
Crop Your Audio
Your recording, sound effect, ADR or Foley can feel bloated when frequencies below 20Hz (imperceptible to the human ear) are left present in encoded audio information. Crop your audio inwards in order to expand what’s pleasant to your ears.
- Insert an EQ on your audio channel, or set up your in built channel EQ
- Cut all frequencies below 20Hz by engaging the ‘High Pass‘ control. During playback slowly sweep the control upwards to 100Hz.
- Listen for a lighter sensation – a clarity in the audio due to it being less bloated, and stop. You might stop at 40Hz or even 60Hz.
- If the sound is thin, pull the control back down towards 20Hz a little to reintroduce any welcome body or fullness. Find a sweet spot between these two points.
Know What You’re Dealing With: Use an Audio Analyzer
When the eye aids the ear you can be more sure of what doing, what you’re seeing aligns with what you’re feeling. What frequencies make your audio messy? What clicks and pops and handling noise occupy what sonic space? What constant average level do our special effects need to be at in order for us to properly mix them into our project?
- Bringing an Audio Analyser into your project and applying it to both your master output and individual channels can be a real time saver. Insert the analyzer on your audio channel and during playback of your low quality clip or sequence watch the graphical display
- Lead with your ears and follow with your eyes. When a certain harsh passage, a grainy sequence, a cracked moment, a rough, loud or distracting sound occurs try to catch it’s graphical movement.
- Many Analysers have a ‘Freeze’ function. This button, once pressed, will freeze the graphical display in place and show you a snapshot of the frequencies present at that instance in time. You can then home in on the sound, even if it is brief, and correct it.
- During playback very quickly click ‘freeze’ when you hear a distracting unpleasant sound. Now you have a snapshot of that particular moment. If you see a very sharp peak at 3,100Hz (3.1kHz) then this is your harsh, sheering metal grinding sound. If you see a large peak at 250Hz, this is your boxy, guttural, boomy sound.
- The handy ‘Freeze’ function also renders the output level of that instant in time. Look at the output meters on the your Analyser to see how loud that particular phrase was, and also how intense that particular frequency was. With this information you can begin audio correction
Low / Low Mid / High Mid / High
It’s unlikely you’ll go into each waveform and individually EQ out each undesirable frequency by hand. In fact, this wouldn’t be advised as it would fall into the realms of over processing. Before you zone in on the most obvious clips, or offending parts of a longer waveform then the rule of thumb is often this: use broad, subtle sweeps and peaks to generalise and not scrutinise your audio correction. EQ across the entire audio channel (and not the individual clip) gives us this initial general remedy.
- Insert a parametric EQ onto your audio channel or clip
- Think of your audio in 4 quarters, rather than a whole. The first quarter being called ‘Low’, the second ‘Low-Mid’, the third ‘High-Mid’ and the fourth ‘High’
- Don’t forget to apply the frequencies you discovered in the audio analysis here. The frequencies below are just examples to get your started:
- Select a band and set a -3dB cut. e.g. Cut -3dB @ 60Hz to correct too much bass, bloated sub and muffled low end.
- Select another band and set a -3dB cut @ 250Hz. This is where boomy, boxy, low room echo resides, can bunch up and be problematic
- Select a third band and set a -3dB cut @ 3,000Hz (3kHz) . This is where harsh, cutting, sharp sounds often become prominent
- Finally set up your fourth band with a -3db cut @ 16,000Hz. Here the very high frequencies can sound hissy, digitally bright and burned
- You might not need all four EQ bands engaged together. Use your ears and your audio analysis to select EQ bands to process
- The last step is to use the output gain on the EQ plugin to give back what’s been taken away (in terms of level/volume). Set it from somewhere between +1.5dB and +3dB to replenish your reduced audio and in turn lift those pleasant frequencies that were not cut by EQ
Audio Correction: Attack the Clip
With some subtle Eq’ing now applied to your audio channel you can focus your attention on individual clips, sliced parts of sequences or how two clips in combination end up feel messy and dissonant. To successfully fix messy audio video editing techniques tend to parallel to audio techniques. Techniques such as deNoise, deClick, Noise Gate but also ambience, reverb, width control and other mastering tools can be used.
Here are a few techniques to fixing problem clips
- Use the audio analyzer to home in on errors. Utilise the ‘freeze’ function, as mentioned earlier to get a snapshot of an instance of time when the problem sounds out. Use your eyes too, zoom into the waveform and look for sharp spikes and peaks – these are tell tale signs of overly dynamic sonic problems
- If the audio clip is too dynamic (too loud and too quiet – inconsistent across time) then insert a compressor onto your clip. Set your compressor’s ‘attack’ to fast – somewhere from 0.01 to 0.30ms. Set your compressors ‘release’ to fast – somewhere from 10 to 50ms. Set your compressors ratio to 6:1 (for every 6dB going in the compressor will only let 1dB out). Set your output gain to +3dB. Now, whilst playing back, slowly reduce the ‘threshold’ control down from 0dB to -25dB. Watch the meters and listen for audio to ‘even out’. Quieter sounds are now closer in level to louder sounds
- Surgically remove clicks and pops by zooming into your audio waveform and slicing offending sections out completely. It can be wise to use third party audio editing software such as Audacity, Sound Forge or Wavelab. Alternatively, if the clicks are too frequent, consider using dePop and deClick plugins that can mask and reduce clicking sounds
- Noise gating can help to dynamically reduce the ambient room sound of an audio recording. Most noise gates are different from noise reduction in so much as they reduce level, as apposed to high frequency hiss and noise. Insert a noise gate plugin before EQ and before compression in order for the noise gate to work more effectively.
If you’re set on damage control and are determined to bring your recordings up to an improved quality then experimentation, as well as trial and error, is recommended. EQ and Compression are your main tools, while audio analyzers, noise reduction and noise gates are allied to those tools. For me, audio analyzers allow for a great amount of accuracy when pinpointing sonic deficiencies that are not obvious to the ear right away.
I hope this article brings you closer to fixing messy audio and lays out some of the tools that make approaching audio with analytical ears easier and more intuitive – Let us know how it works out for you!
We’re doing some cool stuff with audio presets for video editors
find out more White Rabbit Audio Presets