Analyzing your music

[cityle]

So I posted a cringe on how a song was badly mastered, and well it got me thinking about making an intorductory post about music mastering, for your general knowledge. Before I start, just to let you know, I’m using a software named MusicScope, which is really fun to play with. It’s a paid software, but you can use in trial version to test the first 30 seconds of a file if you want to play a little bit around and see how is your music.

The culprit
So that’s the song that started it: Ref: rain, the ED of After the rain.

I rarr it in order to obtain it (thanks mora.jp for now refusing foreign credit card) and when I listened to it, noticed a clear loss in clarity when the music ramped up. Well it appears that the masterer decide that, instead of lowering the overall volume, decided to literally clip the sections where the music ramped up, indicated by the 3 arrows. See how flat it becomes.

Clipping refers to a loss of information when the audio level go above a threshold (the two red circles in the circle graph), and are therefore not recorded, hence the clip, a loss of information.

Pop mastering
sh0ut, 2nd OP of Re: Creators
Unfortunately, clipping is somewhat recurrent because of nowaday practices. With how pop and music in general are recorded, masterer try to get the highest overall volume level. This a two consequences. First, masterer often compresses the music, reducing the dynamic range (the S-Mode number) so there is less variation between the lowest and the highest sound, flattening the music. Second, the music while be on the edge of the threshold in order to get a higher volume. From time to time, it will cliped, resulting in a loss of information.

As you can see in the picture below, music nowadays are master in such a way that the green line, the volume level, stays almost always near the threshold, and sometime clipping.

Proper mastering
Otozureta Henka, OST of Usagi Drop
A proper mastering of music should show a big dynamic range (S-mode). This indicate a file where the masterer didn’t seek to get the highest overall volume of the file. This allow for much richer expression of the music and instrument, and more clarity, vs the compressed song of pop mastering.

Notice how the green line is way more spiky and travels more distance up and down.

Fake FLAC vs True Flac
There is a reason, theme song of No Game No Life ZERO
A last point I want to make you aware with FLAC files is how the quality can vary widely. Often files will be labeled as FLACs, so you’re expecting better audio. But some FLAC files are MP3 that we’re converted to FLAC. This is totally useless as all the information have been already lost with the lossy formating of MP3, hence giving you a fake FLAC. here is an example where we can see the frquency spectrum of a song ends abruptly, indicating a MP3 file that have been converted to FLAC.

If you want more info, lachlanlikesathing did a great video about mastering right here. most of the information i displayed here come from him.

[MazeFrame]

I don´t understand the “crank it to max” when it comes to audio. There is nothing to gain from riding the edge (only audio to loose, as you mentioned).

[BGL]

I’ll be honest, I know very little about music production. What I do know is that some of my earliest CD’s from the early 90’s sound better than one bought 10 years ago, and generally anything I stream from Spotify or Google Play these days. I guess this helps explain it

Still all things are relative, my Dad used to work in a record and CD factory and used to bemoan compromises and slipping standards. He doesn’t do it anymore, he’s now 74 and his ears are not what they used to be, I digress…

[anon25377527]

Inspecting

Same song with things

Absolutely no idea what the acronyms mean

[cityle]

I think it’s because masterers do have to take into account also who will listen to it. By example, with pop music, there will be a lot of people listening with crappy earbuds that would not be able to represent all the dynamic range. So masterer would be more incline to compress the range and raise the noise level so everybody can be able to listen to the music. Vs symphonic music who have a different audience that seek the richness of the music, and thus the masterer will take care to have a better dynamic range.

Also, this technic can be used for aesthetic choice, like a filter that appeal to a certain audience. Lachlan video touches that aspect.

And also in that video which is really interesting

[cityle]

Don’t worry, I don’t know also. Pretty much the S-Mode, the History. and the spectrum frequency is pretty much what is mostly useful and informative to us.

[MazeFrame]

Or the pop music has to operate on tighter budget therefor nobody puts more care than necessary into it.

Also, this technic can be used for aesthetic choice, like a filter that appeal to a certain audience.

“Lets overdrive so it sounds crappy for everyone.”

[cityle]

“Lets overdrive so it sounds crappy for everyone.”

Well, that can be literally it. Like mimicking music that would came out from crappy 60’s speakers for a danceclub by example.

[Positron]

I think it’s because masterers do have to take into account also who will listen to it.

Good informative post!

I was listening to ‘Let Me Ride’ by Dr. Dre the other day. The song has 2, 2, 2 mixes in one!
Most of the song is mixed to be listened to on a boom box.
When Snoop sings, he is mixed in full hi-fidelity.

Once I got in trouble for listening to classical at work.
Because it had a full dynamic range.
The volume ramped up slowly and I didn’t realize my music had become too loud for the office.

[anon25377527]

I think you could create tool like this for Wallpaper Engine with HTML or Unity as already existing EQ’s work like that, though, I dont know which one of the two

Basically it’d need
EQ wiggling
and its target ball
Adjustable time window for that history ghosting
Possibly history averages for unknown acronyms

?

[Karmek]

I only did some audio mastering as a hobby but I know a few things here and there. So much for the disclaimer…

Things like Normalization, Hard Limiter and Single-/Multiband Compressing are steps you do fairly late in the mastering pipeline as it will result in quality loss, that’s right. However, I wouldn’t suspect audio files which received this treatment to be inferior. It’s common practice to increase volume for certain frequencies and in a sensible way in order to make quiet instruments or sounds audible whithout turning the already loud ones even louder.
Some genres, and pop music is a very suiting one here, benefit from proper compression as the track will sound more dense and lively although it’s easy to mess up (happened to me a few times xD).

tl;dr: compressing audio tracks does make sense at times. It depends on genre, source quality and other factors.

[IQCubed]

That’s a fantastic looking program and I’d like to see if it could be used for streaming services. As a subscriber to Tidal Hi-Fi, I’m really enjoying the great sounds from the Jazz and Classical playlists, but the Hip-Hop/Rap and Pop really isn’t for me, per se. It would be reassuring to verify the quality of the stream on each service.

I did have a thought that I’ve been wondering about for some time, and I’d like your feedback. Currently, drive storage, bandwidth throughput and processing power all seem to be breaking some threshold that would have previously held lossless compression streaming back. Do any of you think that the ongoing conversation about, say “which lossy codec sounds best” is losing meaning as time and tech have progressed?

This may be presumptuous, but I think we’re getting to the point where the default or “Normal” setting should be a FLAC 44.1 kHz stream and instead of stepping up in quality, just step down as needed. Would this move have unforeseen consequences?