You are here

Need advice on mastering levels - using Audacity

All about the tools and techniques involved in capturing sound, in the studio or on location.

Need advice on mastering levels - using Audacity

Postby mken » Tue Jun 25, 2019 7:41 pm

Hi, hope someone can give me some tips re procedures to get compatible volumes between individual songs and overall " level" to master at. Hope I'm using the correct terms, I've lots of experience with live sound mixing but was asked to make demo cd for a band - 6 songs. It has been a big learning curve but now I'm at the stage of having the 6 songs mixed to my satisfaction and next step is to produce a cd using the wav files. The CD I have ended up with sounds not too bad, but overall , compared to a commercial cd , the volume is lower. Using Audacity if I use the amplify function then it will increase amplitude until a spike hits the "0 " level. Don't know if this means 0db? Also you get option to allow clipping. I have been thro each track and tried to reduce individual peaks which then allows me to "amplify" perhaps another 0.1 or 0.2 db and then repeat procedure knocking peaks out and reamplifying. However this is a time consuming process and I wonder if what I'm doing is in fact " limiting" the levels but manually. Also I wonder if each peak I reduce is actually detracting from the dynamics of the sound. I have tried both "normalise" and "compress" but feel they both produce a much more "squashed sound". So any advice on the steps you would use to get the appropriate levels onto CD. I realise this is a bit of an open ended question as I believe there have been various standards over the years such that a 70's CD will not sound as loud as a current CD because of newer production techniques. One last question, if I take the wav files to a CD reproduction company is it likely that their process may automatically "burn" the cd at current volume levels?
mken
Poster
Posts: 26
Joined: Tue Sep 12, 2017 1:58 pm

Re: Need advice on mastering levels - using Audacity

Postby James Perrett » Tue Jun 25, 2019 8:26 pm

Yes - it sounds like a limiter is the tool you need. I try to just use one to shave off the highest peaks but sometimes you end up with removing a dB or two from every kick and snare hit. Of course, sometimes having the rest of the track go quieter in volume on every beat is a good way to give a track more apparent energy. I use Reaper and one of my favourite limiters is the JS Master Limiter included with Reaper which offers plenty of control.

Quite a few people seem to like the Limiter No 6 plug-in which is available free from

https://vladgsound.wordpress.com/plugins/limiter6/

As far as duplication or replication companies are concerned they should produce bit accurate copies from your master. For absolute peace of mind most mastering engineers send a DDP fileset to be replicated rather than a CD master as there is less potential for errors to creep in with a DDP.
User avatar
James Perrett
Moderator
Posts: 8539
Joined: Mon Sep 10, 2001 12:00 am
Location: The wilds of Hampshire
JRP Music - Audio Mastering and Restoration. JRP Music Facebook Page

Re: Need advice on mastering levels - using Audacity

Postby DC-Choppah » Wed Jun 26, 2019 2:49 am

Upwards (parallel) compression sounds much better to me for mastering a track if you want to compress the dynamic range to make it read better in a noisy environment like a car, etc.

You don't have to touch the spikes in your song at all. In fact you don't want to because, I mean that is the loudest thing in the song, so if you alter it, that will be easily heard and sound squashed and non-musical.

You can try this in Audacity by following this: https://www.youtube.com/watch?v=ghy9DNg0z48

Basically, make a copy of the track and apply the built in compressor with
a very fast attack and fast release
very high compression ratio
soft knee

and render the audio

Blend that compressed track with the original uncompressed track by ear. The highly compressed track is the 'quiet stuff'. The loud stuff has all been compressed down so becomes inaudible against the main track. But when the main track is quiet, the compressed track is what you hear.

So bring the 'quiet stuff' up so that you can hear it all. This is upward compression. Sounds much better to me for mastering. I like to listen with the volume really low and try to make sure I can hear everything. Or, play it in a noisy car and bring the quiet stuff up until you hear everything in the car. Cool!!

After you add these together, you can use a limiter to finish the job but with just a little bit of limiting that you can't really hear but sets the level at the top of the dynamic range without changing the sound. Compare your level to a commercial track that sounds good to you, but not a squashed song. Something good that you want to emulate. We all have our favorite reference tracks that we use. You should pick one you like.
User avatar
DC-Choppah
Frequent Poster
Posts: 1225
Joined: Fri Jul 20, 2012 12:00 am
Location: MD, USA

Re: Need advice on mastering levels - using Audacity

Postby Hugh Robjohns » Wed Jun 26, 2019 11:21 am

There have already been some excellent answers, but just to try and add some further clarity...

[quote="mken"The CD I have ended up with sounds not too bad, but overall , compared to a commercial cd , the volume is lower.[/quote]

The commercial CDs format has evolved to become what is known as 'peak normalised'. What that means in practice is that a lot of music is mastered in a way that makes it squashed up against the end-stops, with peaks banging on the maximum (0dBFS) and the average levels not too far below that. So there is little dynamic range and it sounds very loud (and fatiguing, usually!)

So, if you compare your mixes with a commercial CD the chances are it will tend to sound quieter and more dynamic. Personally, I think that's a good thing... but I get that you want things to sound more comparable... and I'll come back to that in a moment.

In the meantime, it's worth noting, though, that the peak-normalisation idea is now obsolete and rapidly going out of fashion along with the physical CD format. Most music is now consumed online via streaming, and all of the main streaming services employ loudness-normalisation rather than peak-normalisation.

What that means is that each track is assessed for its average loudness (rather than peak levels), and then the track's overall replay level is turned up or down so that its loudness matches that of every other track on the system. The aim is to give the listener a consistent experience without having to continually adjust the volume between tracks.

This approach also means that there is no longer any need -- or benefit -- to cranking the level and squashing the life out of tracks to make them appear loud. Everything now has the same perceived loudness as everything else. In fact, a typical peak-normalised CD track now tends to sound what it is -- flat, dull and lifeless -- when played on a loudness-normalised platform. And the huge benefit for music lovers is loudness-normalised platforms really encourage the use of dynamics -- and a track with good dynamic range tends to really stand out and sound great!

The minor complication at the moment is that different streaming services use slightly different methods of assessing the average loudness. However, there are now formal standards recommendations and they are converging into a consensus. Already the variation between different services is only a few decibels, so it's not too hard to come up with a master that works perfectly well on all of them.

So... what I'm saying is that while you may feel the need to employ a peak-normalisation approach to try and make the CD sound as loud as typical commercial tracks, it would be worthwhile either compromising the amount of squash so that it will still sound good when played on a streaming service... and/or come up with a more dynamic loudness-normalised version specifically for uploading to the streaming services.

Anyway... back to your CD mastering situation:

Using Audacity if I use the amplify function then it will increase amplitude until a spike hits the "0 " level. Don't know if this means 0db?

I don't use Audacity but it probably means 0dBFS -- digital Full Scale

I have been thro each track and tried to reduce individual peaks which then allows me to "amplify" perhaps another 0.1 or 0.2 db and then repeat procedure knocking peaks out and reamplifying. However this is a time consuming process and I wonder if what I'm doing is in fact " limiting" the levels but manually.

You are doing precisely that! It would be quicker and easier to set up a peak limiter and shave off a decibel or two. However, I would caution against making 0dBFS your target level because that is likely to result in clipping in the reconstruction filter in the replay output. Instead, set the limiter to hold everything a little lower -- say -0.5dBFS as a reasonable compromise.

Also I wonder if each peak I reduce is actually detracting from the dynamics of the sound. I have tried both "normalise" and "compress" but feel they both produce a much more "squashed sound".

Yes, obviously reducing the transients will reduce the overall dynamics. the 'Normalise' function shouldn't affect dynamics at all. It is just a level shift to raise the maximum peak level of the entire track to a predetermined level (often 0dBFS, but possibly you can set you own level). 'Compress' is obviously a dynamic process intended to reduce the dynamic range in some way...

I believe there have been various standards over the years such that a 70's CD will not sound as loud as a current CD because of newer production techniques.

Quite so. Back in the 70s vinyl was the standard consumer format and that couldn't cope with excessively loud and compressed music -- you could have one or the other, but not both! The advent of the CD in the early 80s allowed producers to do anything they liked, and so they seized upon the long-standing idea that louder and brighter always sounds better and more exciting (at least in comparison to quieter and more neutral!)... so the 'loudness war' really took off and the peak normalisation thing got progressively more aggressive and silly.

One last question, if I take the wav files to a CD reproduction company is it likely that their process may automatically "burn" the cd at current volume levels?

No, by default they will burn whatever you give them to burn... but some may offer an in-house mastering service that could process the mixes to achieve a more commercial sounding volume. Whether they do that well or not depends on who is doing it I guess... some mastering engineers are better than others! :-)

Finally, I agree with the post above that 'parallel compression' (also sometimes referred to as London or NewYork compression) can be an effective mastering process.

This reduces the overall dynamic range to make the average level higher and so increase the perceived loudness... but whereas a conventional compressor (and limiter) works by turning down the loud bits ('top down'), parallel compression works by raising the quieter bits (bottom up).

Inherently, top-down compression/limiting does most of its work on the transient spikes, to it tends to audibly affect dynamic transients. In contrast, bottom-up compression tends to leave the transients unaffected, which -- in drum-forward music -- can sounds noticeably better.

H
User avatar
Hugh Robjohns
Moderator
Posts: 24871
Joined: Fri Jul 25, 2003 12:00 am
Location: Worcestershire, UK
Technical Editor, Sound On Sound

Re: Need advice on mastering levels - using Audacity

Postby mken » Wed Jun 26, 2019 1:08 pm

Hi everyone, thanks for the very quick and useful tips you have given. I didn't see this project as a streaming project more a CD project for giving away as demos. I guess its down to the individual , just as live mixing, every engineer will mix slightly differently to get their "ideal" mix. So as far as loudness is concerned then everyone will have their ideal.
I have tried to rationalise the problem so I feel comfortable with it and then try to see what is possible without detracting too much from the original.
So I have come up withe following hypothetical scenario. Supposing i am trying to record a soft female vocalist that has only been able to find one person to accompany her and that person is a cymbal player from a marching band. I only have one mic and try to place in a suitable place to capture both sources. The cymbal players starts off - quite loudly and then the vocalist comes in. Forunately the cymbal player has some common sense and backs off slightly when the vocalist sings. So what I end up with is a track that has very loud peaks and the main body of the vocals somewhat quieter. OK I could compress the track which would hopefully bring the vocals up and cymbals down. But anything i do to the original in terms of amplification will always be governed by the peaks of the cymbals. I could run a limiter which will in fact clip the cymbals but that in itself will alter the dynamics of the track. So I guess there's no magic bullet to automatically give me all I want and I'm going to have to spend some more time experimenting - particularly with the parallel compression effect. Thanks for your input , will post when I've done some experimenting.
mken
Poster
Posts: 26
Joined: Tue Sep 12, 2017 1:58 pm

Re: Need advice on mastering levels - using Audacity

Postby blinddrew » Wed Jun 26, 2019 1:33 pm

Actually your example is a good case of when upward compression might work.
Have a read of this: https://www.soundonsound.com/techniques ... ompression :)
User avatar
blinddrew
Jedi Poster
Posts: 8326
Joined: Sun Jul 05, 2015 12:00 am
Location: York
Ignore the post count, I have no idea what I'm doing...


Who is online

Users browsing this forum: No registered users