Mastering in the age of streaming

richardyot

I thought this was a decent video, particularly the second half. You can skip the talking head at the beginning and just jump straight to the presentation:

mulletsaison

Thanks for that. It was really interesting to see how loudness works. Yay for the end of the smashed mix.

Tarekith

Don't get your hopes up yet, most people are still opting for "normal" masters and just letting the streaming service turn them down. We're just too used to the compressed sound of modern masters, and there's still not an easy way for average musicians to submit streaming only masters on most aggregators.

It's a start, but sadly not getting as widely adopted as I would have hoped.

richardyot

@Tarekith said:
Don't get your hopes up yet, most people are still opting for "normal" masters and just letting the streaming service turn them down. We're just too used to the compressed sound of modern masters, and there's still not an easy way for average musicians to submit streaming only masters on most aggregators.

It's a start, but sadly not getting as widely adopted as I would have hoped.

I think there is a trend towards more dynamic and exciting masters, I posted these in the previous thread but I think they're great examples of where the trend was 10-15 years ago compared to where it's going now.

This is from 2007/2008, listen to the ending of the song, especially the final chorus and the buildup just before it, it's compressed and distorted to hell. There's a drum roll at 3.56 where the toms have almost no impact because the transients are completely crushed:

And compare this to a more recent song that is mastered with streaming in mind, even the busy sections are much more punchy and dynamic. If you listen with headphones the sax really jumps out of the mix, and even in the busier sections the drums are still punchy and dynamic, even when they're layered with bass, horns, guitars, and backing vocals. In terms of raw sound quality this is light years ahead of the MGMT track:

Tarekith

I hope it goes that way, don't get me wrong. Just not seeing that here in practice yet, most people still want it loud and proud regardless of what I tell them.

u0421793

I have to let a little secret out.

In the late 90s when I was doing a lot of music with my analogue synth collection in my home studio of the time, I didn’t use computers to sequence the bulk of the work, Digital Audio Workstations weren’t really a thing and anyway, my computer sequencer wasn’t capable of it (I had Trax, AudioTrax which could record audio on my Mac IIsi but I just used that to chop up audio as a collage, then moved on to a package called Freestyle, but most of my sequencing was in hardware sequencers) and I was running each instrument in parallel into a Tascam midi-mute mixer into a two-channel ‘live’ output.

I decided in the end that about the best way of ‘mastering’ was actually for me to use the auto-level control on the Minidisc I was recording to (the first ever Sony one, which was portable). My DAT was mis-aligned by then but it started with using the DAT on auto, finding it solved a lot of problems and then realising the Minidisc was providing even better results on auto than I’d imagined I could get myself. This wasn’t the same crude ‘auto level’ we were all brought up on in the cassette age, it was actually quite good (except for still getting startled by quiet starting songs, but I decided to live with that).

I know, it’s not technically mastering, but I decided it was.

dendy

there is well know rule "master you tracks on -14 LUFS if they are for streaming services"

But.. when i mastered my recently released album, i decided to go around -10 LUFS (even throug i know that most of streaming services normalizes it to -14) - simply when i compared my music with other officialy released stuff which i use as reference, until i was on -14 LUFS, my sonds vere significantly quieter, lot les "pumping" (whis is important in EDM/trance genre)

at the end, now when i'm listening my album on spotify, or apple music, and i compare it with other offical EDM/trance music there, i'm very satisfied how it play - it is very much comparable to rest of stuff ... so i think it is more important to use ears than strictly follow some numbers ...

richardyot

There's a few different things to unpack here. First of all "loud" means something different depending on the context: peak-metering (as in CD mastering) and loudness normalisation (as per streaming services) give a completely different meaning to the concept of "loudness".

In the CD mastering era, you could use peak-metering to make your tracks louder than the competition by using the brickwall limiter more aggressively. This came at the cost of losing dynamic range and adding (sometimes considerable) distortion to the music. In some genres this also led to an arms race of who could make their tracks the loudest (and also the most distorted and least dynamic).

But of course if you make a "loud" CD style master now, and the streaming services simply reduce its volume, what have you gained? Pretty much nothing, since your track won't be any louder than the competition. The concept of peak-metering and brickwall limiting to make a track louder than others is simply not relevant any more, and is in fact self-defeating because really aggressive brickwall limiting also seriously compromises the sound quality of the music. So by creating a "loud" CD-style master in the streaming age seems to be totally self-defeating.

However some genres of music use aggressive limiting or compression as part of their sound (something that's mentioned in the video). So for EDM, heavy metal, hip-hop and some types of rock, you might still want to apply limiting or compression as an artistic choice, to fit in with the conventions of the genre. This won't make the music "louder" but it will make it sound more aggressive and compressed.

But, you don't necessarily need to do this using the same methods as CD mastering engineers, because that whole paradigm is no longer relevant. For example instead of using a brick wall limiter, you could apply a compressor with a slower attack on the master bus, and tune the release to the tempo of the track - that way you would still get the pumping sound that is essential to EDM but without crushing the transients in the drums, so the result would be much more punchy and dynamic, and probably subjectively also sound "louder" thanks to the sharper transients, when compared to a brickwall-limited master.

This situation is reminiscent of the early CD masters (which is something mentioned in the video), these were way too bright because at the time the mastering engineers were still stuck in the vinyl paradigm, so they high-passed the music too aggressively since that is what you need to do for vinyl. It took time for people to realise that CD can handle much more bass than vinyl (due to the physical limitations of using a needle in a groove to reproduce sound).

It sounds like we are in a similar situation now, with artists still caught up in the CD mindset, which is over. The rule shave changed, and we all need to adapt to this.

The good news is that music can be far more dynamic and punchy than it was during the loudness wars, with much less distortion. You can still use compression for artistic reasons, but without actually crushing the music itself.