2 Techniques for Setting Sound Effects Library Mastering Levels

2017/05/10

Last week I saw an interesting conversation appear on Facebook. Field recordist and sound designer Charles Maynes created a post in the Independent Sound Effects group to discuss establishing a standard for audio levels in sound effect libraries. It’s a valuable idea. In fact, I’ve received a handful of messages from readers about this in the last three months alone.

After all, new sound library publishers often wonder how loud their tracks should be. Are room tones best at -15, -20, or -35 dBFS from peak? Should sound design impacts be provided at -5 dBFS, or is a less aggressive -12 dBFS? What’s correct, what’s easiest, and what do sound fx library fans want?

The post generated good discussion. Some people passionately believed current sound fx libraries are mastered with levels set too high. Others felt that the current trend of “hot” levels was appropriate.

Which technique is best? Is it better to master sound effect libraries loudly, with presence? Or, is the best idea to provide sound fx with reduced levels perfectly suited to slide right into editing projects and mix sessions?

Today’s post will present both perspectives and allow you to decide. In the future I’ll share my own choice, and some community-sourced standards.

Please note: I explore this idea in detail. This article should take you about 10 minutes to read. Click the button below to email the article to yourself to read later.


Two Types of Field Recording Mastering Techniques

I’ve spent a long time thinking about mastering sound effects. I’ve been cleaning and publishing field recordings since 1998. After a lot of careful thought, I began to see mastering sound clips fall into two schools of thought.

What are the broad strokes? Well, some people prefer tracks levels perfectly prepared to a broadcast standard for editing and mix sessions. Others like to have levels set to help preview them easily.

I call these two approaches to mastering sound fx levels for projects or mastering them for selection, respectively.

Let’s take a look.

The “Project” Sound FX Levels Mastering Style

Imagine you are auditioning sound effects in Basehead or Soundminer or another metadata app. You need wailing winds for a mountainside scene. The first clips in your search results sound perfect. They’re at ideal levels. They’re suitably faint with a slightly louder, piercing wail, and you can imagine dropping them into the editing session you were listing to just a moment before. And that’s precisely what you do. The tracks mesh with the others without touching a single fader.

One more wind sweetener track would give the mix options. You return to your search results and play through the list. Suddenly, you are blasted with a wall of noise. Another library had mastered a wind track louder than the last. The track’s not bad though, but needs to be diminished to work well. You drop it in and lower the level. Then you return to auditioning, timidly riding the level for every track you play.

It’s tedious to constantly adjust levels while listening to sound effects and using them in projects. Wouldn’t it be much more efficient to simply drop effects into projects already levelled to what the mix requires?

The Approach

Most supporters of this approach work with projects that require a broadcast standard, such as the American and Australian post standard of -20 dBFS (i.e., 0 VU or +4 dBu). Most specific sound effects range from -12 dBFS up to peak level. Ambiences can span from -20 to -8 dBFS. Really, though, levels are all over the map. Currently, there isn’t a guideline.

The upshot? Supporters of the “project” mastering style find themselves reducing the level of these sound effects by a substantial amount to even begin using them in their projects. Quiet atmospheric tracks such as winds, air tones, and room tones cause the most hassle since they typically need to be diminished dramatically. It holds true for most specifics as well, with gunshots, crashes, and impacts being provided in sound libraries as loud as peak level. Those levels will overpower the soundtrack, and need to be reduced, too.

The Goal of the Project Style of Mastering

  • Master sound effects at “lower” levels suitable for broadcast guidelines.
  • Levels will be appropriate for immediate placement in editing sessions and mixes.
  • Emphasize using sound effects easily.

Pros

  • Editors and mixers do not have to lower sound effect levels as dramatically (or at all) before using them.
  • Levels are not overwhelming in calibrated workspaces.

Cons

  • Sound effects mastered at lower levels may not support proper auditioning. In other words, quiet tracks may obscure vital details.
  • Lower-level sound effects may have their noise floor obscured, but become prominent when the track gain is raised.


When You Should Use The Project Mastering Style

  • Your customers work in post or are restricted by specific guidelines.
  • You are using sound effects in a project, or combined with other sound clips.
  • You are mastering for an in-house sound library that will not be shared or published.

The “Selection” Sound FX Levels Mastering Style

What’s the other perspective?

Imagine you are listening to traffic sounds from a sound fx library. It sounds just as if you were present when it was recorded. You can hear the tire rubber groan as it passes over a grate. You hear the character from each engine. A passing siren is much louder, just as it would be if the ambulance were to pass beside you. The track is rich and detailed. You hear every nuance in the sound effect. You drop it into your YouTube video editing timeline.

Then you listen to cars from another library. They’re quite faint. Even when crank the gain, you can’t make out details clearly, but it sounds OK. You drop it into your project, only to discover while mixing later that there is a muted conversation on the track. It was quiet in the metadata app, even after raising gain. You didn’t hear that before you soloed the track and pushed up your session faders. Annoyed, you realize you just wasted time because you mistakenly chose a clip with surprising, unwanted audio. Not only that, but when you raised the gain, the noise floor increased substantially, introducing too much hiss in the track. You will need to begin your search all over again, and replace it.

You realize that to use anything from that library you’ll need to either choose tracks based on guesswork or raise levels while listening in your metadata app. Guesswork is a waste of time. Wouldn’t it be ideal to hear all details without any more work when flipping through your sound fx search results?

The Approach

The supporters of this type of sound fx mastering approach listening to clips differently. They listen and use sound fx in isolation, unattached to a specific project. Because of this, it’s acceptable for the sound to be louder. In fact, each clip needs to be prominent and present. That helps these people hear all details in the sound, and use it in isolation for multimedia and other projects where a specific broadcast level standard is irrelevant. They also may not listen to the sound in calibrated rooms. So, they don’t need to adjust the sound fx levels as dramatically as their project-mastering style brothers and sisters.

This approach also considers the arc of a sound effect’s lifespan. In short, for a sound effect to be selected and used, it must be heard and evaluated first. Quieter sound effects are more difficult to audition on the Web or in metadata apps. The nuances of the tracks are more difficult to perceive, especially when a customer’s listening environment is unknown. So, selecting quietly mastered tracks may introduce surprises when levels are adjusted, such unwanted audio or even substantial noise floor that was hidden by reduced level. Overall, the idea is that if a sound can’t be heard, it won’t ever be used. So, the approach attempts to present the sound in a way it will be selected before others: with all its features – good and bad – present and clearly audible in even consumer listening environments.

This approach also grows out of a common field recording practice. When recording, it’s often best to maximize recording level to ensure a good signal-to-noise ratio. In other words, sound effects are captured at higher levels to ensure there’s as much quality sound present in the recording (the “signal”) and as little component hiss from the field recording equipment (the “noise”). To over-simply, higher recording levels are less likely to introduce noise when they’re adjusted, as compared to effects recorded more quietly. The result? Sound effects are mastered at the same level they’re recorded: loudly and prominently. After all, when mastering these files, it isn’t intuitive to diminish the levels solely for broadcast guidelines. In a way, reducing the level of a “loud” field recording feels like crippling it unnecessarily.

The Goal of Selection Sound FX Mastering

  • Present a sound effect with all important details immediately discernible.
  • Master sound clips to match field recording levels that maintain a healthy signal-to-noise ratio.
  • Represent sound effects closer to the levels they were actually recorded at.
  • Support low-end listening environments.
  • Emphasize selecting sound files easily.

Pros

  • Can hear all details of the sound.
  • Less chance of noise floor in the sound clip.
  • Less chance of surprising audio being present.
  • A wider range of listeners can enjoy the sound.

Cons

  • Overwhelmingly loud when heard in calibrated environments.
  • Level must be modified to be used in certain projects.
  • Contributes to the loudness war.


When You Should Use Selection Sound FX Mastering

  • You expect sound effects will be primarily auditioned (repeatedly), especially with users with lo-fi monitoring.
  • The selection process is essential.
  • The sound effects will be used in isolation.

Finding Common Ground

Now, it appears that most of the time these techniques are at odds. Put simply, the “selection” people prefer tracks louder, and the “project” people prefer them quieter.

However, there are a few things that all sound effect mastering techs agree upon:

  • Master proportionally. Avoid normalizing clips to the same level. Wind ambiences should not be as loud gunshots. Why? Well, editors know what to expect when auditioning effects that are mastered proportionally. They’ll see a gunshot clip and brace themselves for the loud blast. They’ll safely increase the gain when auditioning quieter winds. If the winds are as loud as gunshots, auditioning them will be deafening. Not good.
  • Represent reality. A rough guideline for either school of thought is to master sound effects to represent their levels in the real world. A ticking clock will be far quieter than a car pass. A sniffle may be a bit louder, but still much softer than the car. The level of the sound effects should approximate how they are heard around us.
  • Do not master to peak level. It’s best to leave a bit of headroom between the highest peak of the sound effect and 0 dBFS. This is true both when field recording and mastering the sound effect later. On the field recording side of things, this allows a bit of space to accommodate for surprising leaps in audio, and avoid peaks. When mastering, it allows editors room to play with the clip levels. Moreover, it ensures no one will be blasted by overpowering audio.
  • Be consistent. Choose one method and stick to it. Your sound library fans will grow accustomed to the style you choose. Alternating between them creates confusion when selecting or editing them.

Deciding Which Technique is Best For You

So, which technique is best?

Like field recording, the best method depends on your audience.

Are you mastering for facility mixers that work to a broadcast standard? Will the tracks be used by in-house editors cutting projects with multiple tracks? Do you expect to share clips with post-production and game audio pros? If so, the project standard may be best for you.

Editors and mixers tend to choose the project method because they were editing audio before they began making sound libraries. It’s natural for them to think about how they themselves would use the sound and master it for that reason.

Do you host sound effects on an a la carte sound library website? Will the clips first be heard in isolation, or used that way in projects? Is emphasizing sound selection most important? Are you unsure about your audience, and want to provide for all possibilities? Then the selection method is best.

Dedicated field recordists may gravitate to this method. They are more involved in capturing a sound and not as invested in using it afterwards. Because of this, they’ll record sounds with as great as possible a signal-to-noise ratio, which contributes to a louder file when mastering techs polish it.

Sound library publishers may also prefer this technique. After all, if people cannot hear a sound, they won’t select it and use it. And as we know, if no one is using a sound, there’s no point in having it. So, they master levels higher to ensure the sounds are selected.

Searching for a Standard

Charles’ original Facebook post was created in part to establish a standard for mastering sound effects. This is an excellent idea for many reasons. First, it makes sound effects easier to use for everyone. It also provides guidelines for inexperienced people creating their first sound fx libraries. These ideas, when combined with adopting common suggestions such as avoiding peak levels, proportional mastering, crafting realistic levels, and a maintaining a consistent practice create something everyone can agree upon: helping guide the field recordings we love to capture to place them wisely into the ears of sound library fans.

Do you have your own ideas about sound fx mastering levels? Share them in the comments below.





To stay in touch, receive free updates by email newsletter or RSS feed.  |  Follow on Twitter, Facebook, YouTube, or SoundCloud.

4 responses to 2 Techniques for Setting Sound Effects Library Mastering Levels

  1. Charles Maynes 2017/05/10 at 14:03

    very nice article Paul-

    the only thing I would mention as a sound effects library author, is that when mastering for either method you defined, removing unwanted things like errant conversations and such would be a requirement of best practice mastering- the other point that is worthy of comment is that when recording effects, if the recordist is elevating the preamp levels to a point where they are contribution significantly (noise wise) is NOT a best practice- the recording chain should hopefully be as transparent to the end product as possible-

    when doing things like BGs- it is usually very helpful to spend time trying to calibrate your input levels to your headset gain as well, which is why production mixers, and broadcast engineers set their levels using tones to get to the optimal settings for their recording chain, then they can simply adjust their mic pre amp input gain to get the incoming signal into a nominally efficient range for the recording chain. pushing the signals into the edges of the electronics capabilities- either on the low, or high ends will result in more distortion and non-linear signal integrity than if the signal is in the optimal performance window at the pre-amp and buss junctions- as well as the A to D’s and headphone amps.

    • Thank you for the comment Charles, and for your ideas and advice. I especially like the idea to calibrate the headset – something that I am consistently guilty of forgetting in the field – and I don’t think is mentioned enough.

    • Thanks Paul, really thought provoking article.

      Charles, really interesting reply also, I wonder if you might be able to elaborate on the process of calibrating input levels?

  2. Justin Macleod 2017/05/18 at 01:27

    Really liked the article. I totally agree with everything you say about proportional mastering.

    while I’m more a fan of selection mastering, when I have to decrease clock sounds by as much as 20-30 DB to use them in a project, it does make me wonder why the publishers mastered them quite so loudly in the first place.

    In the past, I’ve always worked to Ric Viers’ advice on sound effects mastering, i.e. that beeps shouldn’t be normalized because of their piercing qualities, spot effects should be mastered to around -0.5 db, as long as this doesn’t make noise audible, and ambiences should sit anywhere from around -18 to -6 db.

    In response to Charles’ comment about how the recording chain should be set up, I completely agree, though sadly I own sound effects libraries from big names that contain sounds with unacceptable noise floors. I’d rather know about them right away during the selection process than have to find out after importing them.