MIX CHALLENGE - MC070 October 2020 - Winners announced

[m_tree]

By the way ...
I started mixing 8 years ago and take it a few years very seriously. I've read literature, was practicing a lot and found my own mixing-style. Since a few years I do this semi-professional and focused especially on mixing.
Furthermore I'm a drummer with overall 18 years of practice and 13 years band-experience in over 20 bands so far (now I'm 30).

Even if I master my mixes by myself, I do this in a second step and render the mix exactly in the manner for the challenge here. I.e. with a few dB of headroom and far away from common loudness levels like -14 LUFS (Spotify, Youtube, ...). This leads me to my next point:

@Mikael
This is a mix challenge and usually mixes are delivered to mastering-studios within the given boundaries of this challenge. Every mastering-engineer wants a non-limited mix with enough headroom. So the rules are simply a guide for a professional workflow. Just in my opinion ...
This isn't my first mix-contest in general and at first I was 'amazed' about the detailed and strict rules here, too. But I think this is very reasonable and the whole challenge seems very professional to me. You should appreciate the opportunity ... just in my opinion.

@Oba Ozai
I've heard your mix. For detailed analytics I took the first instrumental refrain and compared it with my mix - after loudness-normalization and with an analyzer ...

Average frequency spectrum (first refrain):
Your mix - https://www.dropbox.com/s/utcogwr5m7q9s ... n.jpg?dl=0

My mix - https://www.dropbox.com/s/wd62vreyftmax ... n.jpg?dl=0

My mix for comparison:
https://www.dropbox.com/s/etdnmrnplg1f4 ... e.wav?dl=0

Voxengo SPAN is a free VST-plugin, by the way:
https://www.voxengo.com/product/span/

The main problem, especially with the low-end, is the negative correlation of your left and right channel (Correlation meter of the analyzer). Negative values means the side-signal is louder than the mid-signal and vice versa. The mid-signal is the mono-content in the center and exactly the same on both speakers. The side-signal is the difference of the channels and disappears in mono.
But this doesn't only affect the mono-compatibility. The bass-frequencies contain the most energy of the spectrum ... thus they're mostly panned in the center, to spread the bass-energy evenly on both speakers.
The 'industry standard' for panning is since a few decades: kick, snare, bass and lead-vocal in the center - 0% panning.

Usually the correlation of the channels should be between +0.5 and +1 and certainly not in the negative area. You rarely find professional mixes with lower values than +0.5.
Your mix has only negative values, my mix only values over +0.5.

Another important point are the balances - and this often is underestimated by novice. Your mix is very unbalanced. The lead-guitar in this part is too quiet and the rhythm section is too dominant. You can also see this in the frequency spectrum of the analyzer.

The Kickdrum is 'ducking' or 'pumping' your whole mix, by the way. But this isn't an EDM track. Even ducking the bass by the kick via sidechain-compression is an artistic thing only, in my opinion. I want the kickdrum and bass to play simultaneously. But this is also a matter of taste.

@Mister Fox
Why I use dither on 24bit export is a good question. I know about the noisefloor of 24bit and 16bit and the hearing range for music (which is about 60dB). Even the noisefloor of my mixes exceeds the -96dBFS of 16bit, because of my analog-style plugins.
I think it's an innocent habit. But thank you for the hint ... I should restrict dithering to 16bit exports again.

[Mister Fox]

Why I use dither on 24bit export is a good question. I know about the noisefloor of 24bit and 16bit and the hearing range for music (which is about 60dB). Even the noisefloor of my mixes exceeds the -96dBFS of 16bit, because of my analog-style plugins.
I think it's an innocent habit. But thank you for the hint ... I should restrict dithering to 16bit exports again.

There are plenty of heated topics regarding Dithering (pro and contra). I might pull this into a separate conversation at some point as it would otherwise exceed the scope of this thread. You just mentioned it in your documentation, and I was merely curious as to why you went that route.

[White Punk OD]

Perhaps we want to continue the peer to peer help and solidarity in one of the general threads, as the hints and techniques are also general.

A word about relationship, which is mentioned not too often. We could easily say, all these technicalities should not matter, because we can change that fast, and level can be changed by everyone after the fact etc. etc.,
but there is a very different point here to observe:
The client, or song provider, or executive producer, the one who pays for our work if it is IRL, has to know that we ourselves are absolutely serious, and respect their requirements and whole business. They would themselves fail, if they do not enforce requirements, or have another and another roundtrip, until everything is finally ok. A lot of business, if you want to make a living, is happening with media, radio, advertising, PR etc., and they are very strict industries, where every little detail sums up, every penny. Industry is a about a gazillion pennies, adding up to a billion bucks.

If I do only "underground", and personally not very interested in buck, I can play funny games, and have to gladly take the blowback, but at a certain point I will lose nice and convenient opportunities, and not make friends.

[TomImmon]

While you do not need to "fill out" the full 24bit dynamic range with your edit, a drop in signal strength can lead to a loss of bits (see canese, dadomachado, eladnaimi, etc). As White Punk OD explained, various tools give various readouts. I barely used Bitter, and I don't know of MusicScope. But "realtime readout" compared to "offline analysis" can be a night and day situation. Wavelab is also very picky (always has been).

It is just a warning that something might be wrong. That your loudness dropped off, that you didn't utilize the full dynamic range at your disposal (if you want to go that route), that there is another underlying issue (offsets, etc). Don't worry too much about it. Anything between 22bit and 24bit is what we can still consider "fine". However, if you're exporting in 24bit and the read-out bitrate is capped at 20bit, then there is something wrong.

On the other end of the spectrum... if you read "25bit", then that means you've overloaded the dynamic range (aka clipping).

Thanks a lot for the detailed infos, Mr. Fox!

[letsmix]

Have any of you guys watched the YouTube channel called "Mixing Night with Ken Lewis"? He has "credits on 99 gold & platinum records and 28+ years experience". His shows are super long and he answers a lot of questions. One thing he does that fascinated me is he does these sprint mixes where he has tracks loaded into ProTools, all faders down, and usually zero plugins and he does a pro mix in as little as 10 minutes! For the skeptical among us he keeps showing the timer on his iPhone and it's one continuous shot. Here's a sprint mix where Pro Tools crashed and he posted it anyway, it's enlightening! https://www.youtube.com/watch?v=gAY8c7QbSSw&t=320s (This link should start at roughly where he explains the Sprint mix)

I found MC070 about 4 hours before the deadline and I initially said no. No way, not enough time, forget it. So I went off to do other things. At the 3 hour mark before the deadline I said well, maybe I'll download the tracks and see if I can do a "Ken Lewis" thing. In my mixing notes I said I did the mix in 2 hours, which is lightening fast for me. I failed to mention I spent 30 minutes or so getting the tracks into my template and labelled so I can fly through the mix and then time at the end to do mix notes, bounce, label and upload. It was exhilarating to just make fast decisions and go with it. Once I had phase right, it was mainly level and pan. It really helped having everything in my template so there was, for example, some of my favorite drum reverbs, vocal effects, and my 2 bus chain already there.... Just push up the faders until you like it.

Is it my best mix ever? No? But I learned something new. I meant to test a new plugin but I forgot. Since this new plugin wasn't in my template I forgot to test it under the time constraints. Ha! Fingers crossed I make it to the next round, but I did make some bold choices in tone that they may hate! I was feeling it at the time so I turned it in. Listening now I would make changes for sure. This is why I always "sleep on" a mix and send it the next day to my local bands, but what is done is done on this one.

p.s. After watching a lot of pros mix on Mix With The Masters, PureMix and Youtube I've learned that having a template with plugins and routing you like and a 2 Bus chain you mix into is so important. Then you really can focus on how everything is interacting with everything else and not be stopping playback or hitting solo so much. Food for thought but I really enjoy this process.

[m_tree]

One thing he does that fascinated me is he does these sprint mixes where he has tracks loaded into ProTools, all faders down, and usually zero plugins and he does a pro mix in as little as 10 minutes!

I remember that Chris Lord-Alge is admired for his ability to make mixes in 3 hours - of course with his common 'industry standard" mix quality.
But there are several requirements to make this possible: at first his assistants, who prepare all the tracks (clean every single track with chirurgical EQ, if needed ... build drum-samples from the track, if needed ... comp some tracks down for his SSL and so on). And of course his huge experience and that he only goes forward and doesn't questioning any of his mix decisions. But furthermore I think 3h-mixes are also for CLA only possible with kind of 'standard' tracks ... nothing 'unique' in the arrangement and so on.
And of course the recording itself ... the sound begins with the recording and usually the 'big guys' receive recordings from very good recording-engineers who work also with assistants and instrument-technicians. Nothing like 'we'll fix it later in the mix'

For a professional result with the maximum quality one can achieve, I think 8 hours are a realistic (or minimum) mark also for an experienced, professional mix-engineer, who can't build on the preparation work of assistants.

So, what I wanted to say ... a 'professional' 10 minutes mix is a farytale to impress naive novice.

[john]

Have any of you guys watched the YouTube channel called "Mixing Night with Ken Lewis"? He has "credits on 99 gold & platinum records and 28+ years experience". His shows are super long and he answers a lot of questions. One thing he does that fascinated me is he does these sprint mixes where he has tracks loaded into ProTools, all faders down, and usually zero plugins and he does a pro mix in as little as 10 minutes! For the skeptical among us he keeps showing the timer on his iPhone and it's one continuous shot. Here's a sprint mix where Pro Tools crashed and he posted it anyway, it's enlightening! https://www.youtube.com/watch?v=gAY8c7QbSSw&t=320s (This link should start at roughly where he explains the Sprint mix)
...

good stuff @letsmix - I'm definitely going to check out that video.

We did a similarly focused challenge over a Audio Mix Club a few months ago, although not as aggressive.
The challenge was to mix the provided tracks in 3 sessions. These were meant to occur on three separate days.

• Session 1: do all the things an assistant would do for the "big guys". Import all the tracks, organize them in the template and gain stage everything ending with a static 0-faders mix. All the plugins necessary to mix the song should be loaded and you should be familiar with the tracks provided, so in the next session you can be focused on mixing.

• Session 2: Mix the entire song, and render.

• Session 3: It's a new day, listen to the session 2 mix with fresh ears, taking detailed notes on what needs to be addressed. Then open the DAW, address the items in the notes, and call it done.

This was an exercise, and no one was saying you have to finish every mix in this fashion, but it was a useful exercise for me. It definitely caused me to hone my template. The idea is, in that second session, not to spend any time hunting around for plugins, just focus on the creative aspects of mixing with the plugins you should already have present in your template.

I've carried this thought process into my mixes since, and it has helped to reduce my turnaround time.

-John

[TURANGA]

This is what im taking about. reverb ? (nah, some people are actually saying how they did it.) I just dont have their skill and equipment.

Turanga -16.7 LUFS TP -2.8 db
https://drive.google.com/file/d/1ypTe_Y ... sp=sharing

Hi Oba,
a couple of comments that may help you to get what happens with my mix version:

The two kicks channels were sended to one buss, so, they were processed like "one sound". This prevent phase problem issues.
This is the chain fx for that buss:
https://drive.google.com/file/d/1Z-36dY ... sp=sharing

Same in bass land. except for the DI channel that I used to get the "chorus sound".
Chain fx for the bass buss:
https://drive.google.com/file/d/1UC7yEO ... sp=sharing

For me, "the trick" for this low frequency response, was the use of NLS channel (EMI TG) with total drive. this console (simulation) is famous for increasing the response in the low frequencies.

[letsmix]

These were meant to occur on three separate days.

• Session 1: do all the things an assistant would do for the "big guys". Import all the tracks, organize them in the template and gain stage everything ending with a static 0-faders mix. All the plugins necessary to mix the song should be loaded and you should be familiar with the tracks provided, so in the next session you can be focused on mixing.

• Session 2: Mix the entire song, and render.

• Session 3: It's a new day, listen to the session 2 mix with fresh ears, taking detailed notes on what needs to be addressed. Then open the DAW, address the items in the notes, and call it done.

John,

This is actually pretty close to my normal process except:

1) Between Sessions 2 and 3 I would listen in the car or in my earbuds and take some mental notes. Then when I'm in Session 3 I'd ask myself if I still agree with my mental notes. If not, then why not?
2) I would also start Sessions 2 and 3 with listening to a few minutes of my reference tracks and the client reference material.

What do you guys/gals do the same or different? Sorry, I'm always trying to learn!

Letsmix

[m_tree]

@letsmix
My procedure is similar, but not the same. As I already mentioned:

Regarding how to handle 80+ tracks, I'll give a short insight in my approach:

At first I always make a rough-mix project and sight all the tracks, to get an overview. Maybe I'll sort out some tracks at this point - in this mix 13 tracks.
I'll further sort the tracks and make some considerations (possibly with FX-tests), how to mix the song and how I've to adjust my mixing-template for the particular mix.

In addition I comped, respectively merged some of the tracks in the rough-mix project of this mix. More precisely: I merged the lead-guitar tracks until the second (doubled) lead-guitar starts at the end of the song. I used all the guitar tracks except the DI-racks.

So I ended up with 41 tracks that I copied into my final mix-project from my rough-mix project. Before the masterbus I've 9 main-buses (drums, bass, lead-guitar, backing-guitar, ending-guitar 1+2, lead-vocal, backing-vox 1+2) apart from some additional delay- and reverb tracks.

Regarding the question 'what makes a better mix?' ...
I think you should never forget what the goal is: to transmit the song with it's emotion and musicality to the listener. This is a sort of art!
This goes hand in hand with all the technical and physical stuff, of course. For example: how do I get a balanced mix, that will transmit the song with it's emphasises evenly in miscellaneous listener situations - from the poor phone speaker to the adjusted hi-fi system - ...? The main answer is in my perception, by the way: clean and present midrange frequencies! And good balances (which includes dynamics).

So in general I do detailed mix-preparation, too. But I try to improve my accuracy without 'make improvements with fresh ears (after a while)'. The 'big guys" usually mix a song straight forward without long breaks in between. Currently my mixes still benefit from a break and improvements with fresh ears.

I also mentioned another keypoint: the mix should deliver the song as consistent as possible on different speakers, headphones and environments. To get an impression on what this depends, it's very important to check your mix not only on your studio monitors and headphones, sure.
But this is another point the 'big guys' and experienced professionals don't need to do ... and since a few years I don't need this anymore, too. If you know your tools well and know how the mix will behave on different speakers, it's simple in my opinion: the midrange frequencies are the most important area. Both for the human ears ('fletcher munson curve') and the speakers and headphones. Also for the recorded material - except drums and bass, almost all instruments and vocals mainly consist of midrange-frequencies.
And whether you hear the mix with a poor phone-speaker or a hi-fi system with subwoofers - the midrange frequencies are always hearable.
So, for example - if you mix a funk-song and the bass plays a fundamental role for the music, the bass should have enough presence in the midrange (and treble, maybe) to stay hearable on speakers with poor bass playback ... or simply at low listening levels or loud environments.

Some guys occasionally use a 'midrange-loupe' (EQ) on the masterbus to check the midrange. I listen at (very) low levels instead. Anyway it's the best way to mix at low levels the most of the time - analytical hearing works best at low levels and the ears don't get tired so fast. Sometimes it's good to check the mix at higher levels or very low levels (hearing threshold level).

But these are just my experiences and approaches.