Is there an agreed upon standard for all audio?

So far my methodology for audio levels is "don't let anything hit or get above 0dB on any track or the master fader. Is there a better methodology to follow where I try to have all dialogue levels at a certain dB then SFX and music at others? Is there a certain normalize option I should select at export?

Audio is the last thing in my way to becoming pretty good at most aspects of production/post (not pre, I am bad at writing) so any feedback would be greatly appreciated.

All audio I ever use is LPCM with at least 48kHz 24-bit when possible, even with a 3.5 mic to camera or a recorder.
Most XLR audio I record is 32-bit float 48kHz.

So far my methodology for audio levels is "don't let anything hit or get above 0dB on any track or the master fader. Is there a better methodology to follow where I try to have all dialogue levels at a certain dB then SFX and music at others? Is there a certain normalize option I should select at export?

Yes, there is more than making sure it does not get over 0 dB you want to understand loudness first and then lookup the different standards and requirements.

The agreed-upon standard is don't just blow sound through the speakers but have consideration for how loud it is perceived. The details are obviously more complicated.

So far my methodology for audio levels is "don't let anything hit or get above 0dB on any track or the master fader. Is there a better methodology to follow where I try to have all dialogue levels at a certain dB then SFX and music at others? Is there a certain normalize option I should select at export?

Yes, there is more than making sure it does not get over 0 dB you want to understand loudness first and then lookup the different standards and requirements.

Is there one you’d recommend working towards? It seems like each company and streaming platform each have their own.

So far my methodology for audio levels is "don't let anything hit or get above 0dB on any track or the master fader. Is there a better methodology to follow where I try to have all dialogue levels at a certain dB then SFX and music at others? Is there a certain normalize option I should select at export?

Yes, there is more than making sure it does not get over 0 dB you want to understand loudness first and then lookup the different standards and requirements.

Is there one you’d recommend working towards? It seems like each company and streaming platform each have their own.

Here is a very rough categorization:

Broadcast Television: Follow the regional broadcast standard (e.g., EBU R128 in Europe, ATSC A/85 in North America).

Streaming and Online Content: Target -14 to -16 LUFS to match the loudness standards of popular platforms like YouTube and Spotify.

Music Production: While not as strictly regulated, aim for around -14 LUFS to ensure consistency across various listening environments.

Film and Cinema: Follow ITU-R BS.1770 guidelines, often around -23 LUFS for consistency with broadcast standards. There are also SMPTE and Dolby guidelines/references.

For broadcast, for example, a -3dBTP maximum True Peak target ensures that any downstream processing such as receiver filtering, codecs, sample rate conversion, et cetera do not result in peak overshoots.

To avoid codec overshoots and inter-sample peaks, for example, at minimum 0.5dB headroom is required (often even more headroom may be required) to avoid exceeding full scale digital.

Deliverables need to adhere to the target platform, and delivery specs for programme and advertising will vary between the various broadcasters, streaming platforms, and cinema - but specs will indicate using an integrated, dialogue-gated or Loudness Range measurement for speech, and other relevant targets.

For an outline of the various standards refer to RTW's Worldwide Loudness Delivery Standards article: https://www.rtw.com/en/blog/worldwide-l ... dards.html

Hi, total duffer here who creates video for his own personal use and is too lazy to run around researching broadcast standards. I read somewhere (perhaps here) that you should try to keep your average sound peaks at around -9dB on the sound level meter. That doesn't mean they can't go above that (i.e., quick percussive hits in music), but generally the loudest music should be bumping along at +/- that level. That seems to work pretty well for my videos - I don't have to adjust the volume between watching them and other professionally produced content. That's my main concern when it comes to loudness levels.

As far as the balance between voice, music and effects, my practice is to get the voice at the level I want and then vary the volume of the other elements as necessary to make sure the voice still comes through clearly. That often means dips as needed in music volume, for example.

If you're a pro, go deeper. But if you're just looking for a quick and dirty guideline, that's what I've found to work for me.

if you're just looking for a quick and dirty guideline...

...right click the clips>Normalize Audio Levels>YouTube preset.

sound.jpg (723.13 KiB) Viewed 1011 times

I usually just stay below -6 for the LOUDEST stuff ( like music crescendo) and actors voices or V.O. around halfway there and SFX wherever it's needed effective. A screen shot of short video and then link to video so you can HEAR IT on you computer, and then TURN VOLUME up or down to suit your self.
So there's the image and then the link to vimeo so you can hear and see what's up...

It's relations, sorta.... just like light when filming.

turn up your volume or turn it down or go to the movie house or go to the theatre.... and put your hands over your ears if it gets too loud ... hehe...

Hi Alexander,

indeed you are right, each platform/target has it's own loudness and peak level standards (or lack of them). That's where you need to start - where is this going, what platforms/services, then work backwards from there. Note that there's loudness as an average measurement, as well as peak levels which is where your peaks reach relative to 0dB Full Scale (i.e. when the red lights all start going off). You want a buffer between Full Scale and your highest peaks, and your average loudness (often BS1440 method of measuring) to wherever it needs to be.

My advice is if you are doing this for putting out anywhere others will see it, even if only YouTube, get a copy of the standards and set up your system to allow you to keep an eye on overall levels to make sure you are compliant. All platforms have published standards available (Netflix, Disney, Spotify, YouTube, etc). Resolve (Fairlight) has a number of standards built into it's metering system to help make this easy.

Dialogue typically will activate the yellow on your individual track meters, mix everything else around that (so backgrounds and ambiences will be much lower, music will be closer, and FX will be all over the place depending on their place in the story), just as a very rough guide to help give you a starting point. Everything revolves around the dialogue (unless you're working with Christopher Nolan ). Keep an eye on your overall levels (on your main mix bus and Fairlight Loudness/CR meters) as you go.

If you're just making stuff for yourself or for fun, just make it sound good, avoid slamming the meters, but don't let everything stay too low either.

Hope this helps!

Thor

Dialogue is always king.
You should aim to have the individual tracks peak at around -6 dbfs to allow some headroom if you are going to add some audio processing such as eq and compression.
When mixing with music and sound fx you balance them against the dialogue. This is normally done by ear rather than watching the meters so you take into account perceived level rather than electrical level. There are far too many videos on YT where you have quiet dialogue, so you turn the replay volume up, only to be suddenly blasted by a music track.
If there is no dialogue is the video music led or fx led? This will alter how you balance one against the other.
Unfortunately I can't give an example of a dialogue led mix as all the examples I have are copyright, but here are two examples without dialogue. The first is music led with the fx interjecting when appropriate, and the second where the background fx are as important as the music.

Unfortunately I can't give an example of a dialogue led mix as all the examples I have are copyright, but here are two examples without dialogue. The first is music led with the fx interjecting when appropriate, and the second where the background fx are as important as the music.

Both videos are really beautiful! Concerning the sound, my feelings are as follows:

First Video [The Edge of the Sea]:
The sound of the waves is muffled.
Faced with the ocean, what you normally feel is a caress or a roar, depending on sea conditions, that rises up and transcends you. It's not uncommon, in such moments, to lose your gaze on the horizon and end up raising your eyes to Heaven.
But it's quite a challenge to add music (or it should probably be in the background, like if it was playing softly in your head).

Second Video [Hanningfield in the Spring]:
Oh my! This one brought tears to my eyes.

Following on from the second video, you can imagine playing the music in Nature and recording the overall result. You'll be surprised, believe me!
I do it every day (without recording) with a 432Hz recorded version of Amazing Grace. When all the Birds join in, I give thanks to God for the unique Symphony I've been granted.

The Edge of the Sea video is music led so the effects only come into play later. In the opening sunrise the waves aren't muffled, they are quiet as the tide was out. The fx were added to the music, not the other way round, and it's mixed exactly as intended.
Glad you liked Hanningfield in the Spring.

Further to my earlier post I now have an example of a mix with dialogue. This is a remix for illustration purposes.

Though mixed in Pro Tools this shows the levels in Fairlight with the Limiter on the output bus to control the output levels. The loudness meter shows -14 LUFS.

Mix Levels to -14LUFS.jpg (515.81 KiB) Viewed 681 times

The Stats for Nerds on YT also shows it to be the case, showing 100%/100% (content loudness -0 db).

Broughton_Stats for Nerds.jpg (81.56 KiB) Viewed 681 times

(...)

Do you have any experiences to share with "Audio Ducking" (which preserves the clarity of a message, hence the name perhaps)?

I have never used an audio ducking tool, and that goes back before they existed when you would side chain a hardware compressor to do the job. I've always been a "ride the faders" mixer.

You may be going about your audio mixing process backwards.

Normalizing and setting the absolute peak level should be the very last step in the process, after the mix is done.

Two very important and related aspects of mixing audio are dynamic range and perceived loudness. The lower the dynamic range, the louder the perceived loudness can be. A suitable dynamic range is something you want determine when considering the content, the audience, the distribution platform, and the venue or devices the video will be played for/in. Too wide a DR and some parts can be too loud while other parts can be too soft. It can be a complex relationship between setting the DR of each track and the final mix, but necessary.

Once the mix is finished, then normalization will set the level according to the type you choose.

Respectfully, I have been recording and mixing film and video soundtracks since 1972, and for theatrical special venues since 1994.

Normalizing and setting the absolute peak level should be the very last step in the process, after the mix is done.

I'm a small time producer, but that sounds backwards to me.

Take a stage production, for example. I have a board feed and a nat feed. I Normalize the two feeds to the same standard first, and then mix them for a good balance.

I run a check before delivery, adjusting the Main only so that I hit the Target Loudness, but the mix remains unchanged.