Mix Challenge - Gossip and Discussion

How to do this, that and the other. Share, learn, teach. How did X do that? How can I sound like Y?
User avatar
KVRAF

Topic Starter

12344 posts since 22 Nov, 2000 from Southern California

Post Tue Jun 10, 2014 11:33 pm

Dean Aka Nekro wrote:Really good idea to get off the ground. I am certain I will be able to offer material for a multi-track mixing contest, Not just my stuff but that of my close friend and studio 'wife' :hihi: His material is a lot more varied than my own. If you want me to get in touch about this at some point Eric and Compy my friends then just give me a shout. A lot of it is on soundcloud (not so much of my stuff but the person I am talking about) and I am pretty certain that he would like to hear different takes on his own mixes.
Great music and great mixes! I hope the music's authors will consider using some of the mixes that get turned in.

User avatar
KVRAF
2883 posts since 26 Oct, 2007 from Kent, UK

Post Wed Jun 11, 2014 12:42 am

Uncle E wrote:Also, thoughts on main buss compression? Are we allowing it?
I say yes. I actually use it more often than not. It's nearly always very low ratio compression for glue and I level match...

User avatar
KVRAF
2883 posts since 26 Oct, 2007 from Kent, UK

Post Wed Jun 11, 2014 12:46 am

Uncle E wrote:How would everyone feel about 0dBFS digital peak max? I'm concerned that we'll all be turning in low mixes and having people judge them against commercial mixes.
I don't personally see the point in deliberately lowering the peak level to something arbitrary like -3 or -6. When I finish a mix (that's not intended to be a pre-master) I always normalize to -0.3.

User avatar
KVRAF
6179 posts since 4 Oct, 2007 from Escaped At Last

Post Wed Jun 11, 2014 1:54 am

That is the way I'm set in myself, I have no problem adhering to rules if it will make the challenge a test/a switch up of what is normal/I'm stuck in my ways doing since there is always something to be learned from such. So yeah I will happily stick to whatever is decided upon.

Perhaps if I may suggest, If the rule for a lower level mixdown is put in place when for example there is a multi-tracked mix challenge which is also decided to be further worked on in the form of the odd spit and polish "home brew" mastering challenge (which would work nicely to change things up from time to time), Then delivering -6dBFS mixdowns would be an ideal time to put that into action? Anyway just throwing that one out just to see where it lands

It is just the reality that most people aren't interested in doing their own level matching between a reference set of tracks they are familiar/use as a yardstick with which to measure the qualities of other material against. So I do agree that a lower level mixdown will probably suffer due to that. I know also that I am guilty of doing it myself to be honest. Still as said I'll go with whatever is decided upon by democracy as obviously it is paramount we are all always on the same page in that regard :)

Edit; It just popped into my under developed brain which works pretty slowly when I would like it to not do so that Mike Senior whom has done many of the SOS magazine 'Mix Rescue' articles along with releasing a really very good book which he has a link to whole truck load of mulit-tracked recordings that are updated when more are added that might be a viable option to use? I wouldn't know what would be involved getting the ok from Mike to utilize those resources, Here is the information and resources http://www.cambridge-mt.com/MixingSecrets.htm & http://www.cambridge-mt.com/ms-mtk.htm
Again obviously it would be preferable to keep everything with the community here on KVR, Still thought it was worth throwing in the mix just in case the well ever dries up for whatever reason/reasons

Eric, I'm sure Paul would definitely do so if he liked the mix, It is mostly just him on his own as a one man band a lot like how I do things, Except way more varied and often times way more ear friendly too than I could/can ever muster. So it is only one writer and musician in his case. Unfortunately he is banned from here, Still that doesn't matter as we see each other usually once a week so getting the rendered multi-tracks from him won't be an problem

Dean :)

KVRian
505 posts since 2 May, 2014

Post Wed Jun 11, 2014 2:30 am

Dean Aka Nekro wrote:I wouldn't know what would be involved getting the ok from Mike to utilize those resources
...http://www.cambridge-mt.com/ms-mtk_UsageFAQs

User avatar
KVRAF
6179 posts since 4 Oct, 2007 from Escaped At Last

Post Wed Jun 11, 2014 3:30 am

Nice one 1wob2many, I really ought to perhaps read the details fully on websites which I go to. Time is often my enemy, Appreciate that as it answers all questions and covers what the guys would need to know :tu:

All the best :)

Dean

KVRAF
4592 posts since 2 Nov, 2006

Post Wed Jun 11, 2014 3:46 am

Submitted my mix ;)
Image

KVRAF
2450 posts since 12 Sep, 2004

Post Wed Jun 11, 2014 6:25 am

Dean, tell your friend he's got some real quality stuff there... esp. "Suspended in the Dark"... ooofaah... you just don't hear those types of songs and productions anymore... what a fantastic piece... that's got indie film soundtrack written all over it...
You need to limit that rez, bro.

KVRAF
14476 posts since 19 Oct, 2003 from Berlin, Germany

Post Wed Jun 11, 2014 6:27 am

Before I get to the technical questions, just a couple of infos:

I updated the Submission thread in terms of rule refinements and donations. satYatunes informed be that it should be mandatory to all participants to also vote - else no chance to win anything. I kind of agree with that - full commitment, or not commitment at all.

Since it's the first challenge, we will cut a lot of slack. But with MC02 - those that do not stick to the rules will be instantly disqualified. We made these rules for a reason, and I hope they are as understandable as possible.


Uncle E wrote:How would everyone feel about 0dBFS digital peak max? I'm concerned that we'll all be
turning in low mixes and having people judge them against commercial mixes.
I speak against that. Though let me explain why.

The idea/concept behind this mix challenge is for people to learn something. And people can't learn something if everything is pressed to it's limits, or overcompressed to "hide" certain things. I recently listened to a lot of tracks with a very capable declipper to see how well an overcompressed track would perform compared to a production from the late 1990's to early 2000's - and most of these tracks fell apart on listening. The overcompression hid the main issues with the song - the mix wasn't great.


Which brings me to the concept to use -15dB RMS/+3VU or a reference level of -18dBFS - which is a rule for this competition:

Mix engineers up until this day (at least those that care and do not give a sh*t otherwise) still swear upon outboard consoles - or use excellent ITB tools these days. And those use a reference level at a certain volume/voltage (the hotspot of the hardware) - which usually resolves around the magical (superficial) -18dBFS. They do not(!) apply mastering effects during mixing. They leave that for a second process step. Especially if they brand the material on to 2-track tape or re-record into the DAW. So they have to stick to certain rules/levels.

Not to mention do we have a lot of plugins these days, that use an internal worklevel of -24dBFS to -15dBFS. So we are inevitably forced to go that route. Which is IMO a good thing.


To bring up the quote by "Android":
do_androids_dream wrote:I don't personally see the point in deliberately lowering the peak level to something arbitrary like -3 or -6. When I finish a mix (that's not intended to be a pre-master) I always normalize to -0.3.
But this is the point here.
We try to pull off a fun competition, which also offers a learning courve in terms of "how to mix properly" before the so called "Pre-mastering" (the final touches and loudness raise process before the "Glass Master", or the actual "Mastering" if we go by original technical terms).

Normalizing tracks would IMO destroy that purpose. And people still don't learn how to do it "the right way" (IMO!). And this is what is a major issue I'm having for the last decade... audio engineering schools, video tutorials, music magazines.. they all do it wrong (again, IMO). Always mixing at the limits, always having a very crap fader resolution, no correct explanation of using metering tools. Just peak limiting, peak limiting, peak limiting.

Not okay - and no learning factor. Because in this case, Joe Normalguy can go to YT and hack in "Mixing Tutorials" and get a drop of like 30-50 videos in how to "overcompress sh*t". Heck even people on GearSlutz of all places go that route. Because that's obviously the way to do. I don't think so.



Let's take a look at audio engineer Butch Vig, and his IMO excellent work on Nirvana. Unfortunately to some also the turning point of driving stuff to it's limits. Butch Vig was (and still is) a master in terms of slightly overdriving the mix console to get a certain sound. But (pre)mastering (loudness raise, final touches) were still handled at a second step. And this is what I aim at with this limitation rule.

I try to force(!) the users to focus on the mixing aspect again. Work as great as possible right from the start. Only adding slight final touches and maybe even a loudness raise during (pre)mastering, instead of creating a sh*t mix, then trying to fix/hide it in a -5dB RMS squarewave and insist that it's "adding to the sound". This is just a lazy excuse. And the very problem why we have such issues in terms of loudness and crap sound these days.

(I'm not the only one: Alan Parsons, John Field, Richard Dodd, Bob Katz, Skip Burrows on GearSlutz, and many more also think that way).


The limited values also don't come out of nowhere. If you mix at -18dBFS to -15dBFS (average), chances are that your peaks won't even reach -3dBFS (more like -6dBFS), which funny enough... is a limitation of the DIN PPM (actually: -18dBFS reference, -9dBFS max peak). And actually also the (IMO good) limitations of outboard consoles.

Normalizing might be one thing in the end, but I still do not support it. I want people to finally understand "why" reference levels exist and "why" we should go for them. This is not to punish people. This is also not to turn people way since the mixes are not "commercial ready". Such comparisons are nonsense IMO anyway.

I just want the users to focus on the mixing aspect again! And this can't work if we ignore the rules, just because we have a 32bit floating point DAW and levels really don't matter to some. Or because there are a ton of Youtube videos that say "just mix as loud as you can and then pull down the master fader" - this is plain wrong and doesn't work. Especially not at large scale studios with analog consoles.

Digital Mix consoles (read our DAW's) came up due to the fact that it makes life a ton more simple. But it doesn't mean that we can ignore how they were built and work.


To sum things up:
My focus with this challenge and the rules in terms of mix levels is to force the users to focus on "mixing" again, do as much as possible in there, rather than trying to "polish crap" with a quick overprocessed mastering.

It should not be underestimated how much you can shape a sound during mixing compared to a 3-4 hours session and trying to fix things while mastering.

We can still have a mastering challenge at a later state, but even here I recommended to not(!) exceed a certain loudness level either (again, a -5dBFS average signal level squarewave does more than often try to hide things, rather than showing off a good production). If we want to win the fight for a good and not over-compressed/processed sound, we have to relearn the absolute basics again.

Hopefully I'm not working against windmills alone on that behalf. But I'm used to that by now.



Uncle E wrote:Also, thoughts on main buss compression? Are we allowing it?
Master bus compression as in form of "slight mix glue" is a concept from SSL console days that is still popular up until this day. Tried it myself several times to "mix through" a compressor, though I quickly went away from that. It can add to a certain sound.

But it should NOT be used to raise the loudness. This is a mix challenge after all.


Dean Aka Nekro wrote:Perhaps if I may suggest, If the rule for a lower level mixdown is put in place when for example there is a multi-tracked mix challenge which is also decided to be further worked on in the form of the odd spit and polish "home brew" mastering challenge (which would work nicely to change things up from time to time), Then delivering -6dBFS mixdowns would be an ideal time to put that into action? Anyway just throwing that one out just to see where it lands
Which is actually a great idea. A mix challenge with a follow up mastering challenge of the "winner track".

Though regarding the -6dBFS digital max peak signal level...
Please folks... learn how to use measuring tools again. I am gladly of service. Actually, my KVRmarks are full with information on that behalf.

Focus your mixing on an average level of -15dB (RMS meter with 300ms) or +3VU (VU meter, 300ms, -18dBFS reference level). Ignore the peaks. If you mix at that level, your peaks can reach up to -2dBFS if you have really, really strong transients.

Once more - my focus with that rule is to slightly force the users to take note of proper gain staging, and going a great job right from the start rather than being lazy. It is not to annoy people.


Dean Aka Nekro wrote:It is just the reality that most people aren't interested in doing their own level matching between a reference set of tracks they are familiar/use as a yardstick with which to measure the qualities of other material against. So I do agree that a lower level mixdown will probably suffer due to that. I know also that I am guilty of doing it myself to be honest. Still as said I'll go with whatever is decided upon by democracy as obviously it is paramount we are all always on the same page in that regard :)
You can still match a certain mix frequency response to your own mix - but only if you're eager to actually use metering tools properly, and pull down the volume of the source file (reference track) rather than pulling up the mix to the level you desire.

There are tools on the market that even offer modes to pull down the average signal level automatically.




Regarding the SoS track offerings, I we have to dig into that a bit deeper in terms of rights. However, the idea to stick to KVR tracks does make sense. Because we could get tracks from the OSC and SWC, then mix it up and go from there. It would result in a constant learning cycle with new productions, rather than trying to overhaul tracks that are already known to the masses (chart productions, remixes, etc). And it would mean, clearing rights.



Anyway... took me almost 2 hours to do the edits on the challenge and writing this post now.
I hope I could make my ideas a bit more clear and it wasn't too strong. Though those people that know me on KVR, know how hotheaded I can be regarding this topic.
[ Studio Page / Twitter ] | [ KVRmarks (see: metering tools) ] | [ Mix Challenge ]

User avatar
KVRAF
2883 posts since 26 Oct, 2007 from Kent, UK

Post Wed Jun 11, 2014 8:14 am

When I talk about normalizing it has nothing to do with peak limiting. I always mix using the K20 reference level. This means that for an average mix I end up with the loudest peak of my mix sitting somewhere between -5 and -2 dBFS on the master (this is the reference level I'm very used to and my monitor volume level and my actual volume knob is all calibrated together). I then render the mix, normalise it to -0.3 dBFS and archive it as my finished mix. If I'm going to master it or send it to an ME another copy of that mix is made to which the gain will be lowered so that the loudest peak hits -6 dBFS. If you would prefer that the -6 dBFS version of the mix is the one uploaded for the competition then no biggie.

But, at no point have I advocated 'mixing at the limit'. Normalising your finished mix does not advocate mixing at the limit either. If you're actually mixing at the limit DURING the mixing process then that's a different thing entirely. I also think there is a danger in imposing many rules within something that is essentially an art... I like how Mike Senior's forum runs - no hard and fast rules but everyone still learns stuff from each other.

I really don't mind if a certain limit is set (-6 etc.) - I can stick to it if deemed necessary for learning purposes but there is a certain danger in imposing all these rules imo - especially if one is declaring that everyone else is wrong and you are right.
Compyfox wrote:...audio engineering schools, video tutorials, music magazines.. they all do it wrong (again, IMO). Always mixing at the limits, always having a very crap fader resolution, no correct explanation of using metering tools. Just peak limiting, peak limiting, peak limiting."
This is a generalisation that I have personally not seen much evidence of. Ok, perhaps in dance music but not really anywhere else.

I ain't out to cause trouble here :) I just think things should be a bit more slack and casual - folks will still learn... or maybe I should butt out and shut up? :D

User avatar
KVRAF

Topic Starter

12344 posts since 22 Nov, 2000 from Southern California

Post Wed Jun 11, 2014 9:40 am

Compyfox wrote:The idea/concept behind this mix challenge is for people to learn something. And people can't learn something if everything is pressed to it's limits, or overcompressed to "hide" certain things. I recently listened to a lot of tracks with a very capable declipper to see how well an overcompressed track would perform compared to a production from the late 1990's to early 2000's - and most of these tracks fell apart on listening. The overcompression hid the main issues with the song - the mix wasn't great.
OK, I can live with that. I'll lower the level of my mix and upload it today.
A mix challenge with a follow up mastering challenge of the "winner track".
Brilliant! This really ups the ante for winning, now you get your music mixed AND mastered. :)

User avatar
KVRAF
2883 posts since 26 Oct, 2007 from Kent, UK

Post Wed Jun 11, 2014 10:26 am

Ok so I've just re-read (a bit more slowly this time - I tend to read very fast/skim read these days as there's just not enough hours in the day... lol!) your post compy. Yes, I whole-heartedly agree with your principles - absolutely. But why not make it easier and suggest that everyone adopt K20? That would mean that everyone would have their average level at 0 on the scale (so -20 RMS roughly) and, hopefully, that would result in similar level mixes to compare and more than enough headroom for a good mix (most of my mixes end up with an RMS of between -16 and -19 depending on the style).

KVRAF
14476 posts since 19 Oct, 2003 from Berlin, Germany

Post Wed Jun 11, 2014 11:12 am

I'll have to answer more in-depth later. Sorry...


I'm against the K-System (either version) for mixing, since it's:
a) not made for mixing, but mastering
b) the K-System can be abused. Everyone thinks different of "mezzoforte" different. So a K-20 AZ+0 mix at Mezzoforte might result in a K-20 AZ+2 for others.
c) different ballistics (600ms unweighted for K-System v1, 300ms for RMS and VU meter)
d) not everyone understands it fully
e) a VU meter is more definite


Sorry - ultra short answer. And definitely not to bark up anyone's tree.

Something came up over here for today that makes it hard to give in-depth answers.
[ Studio Page / Twitter ] | [ KVRmarks (see: metering tools) ] | [ Mix Challenge ]

KVRAF
7088 posts since 17 Feb, 2005

Post Wed Jun 11, 2014 12:29 pm

How about having all entries normalized to a specific LUFS integrated loudness level? This should work well since we are all using the same material, and it will make previewing and comparison much easier.

KVRAF
14476 posts since 19 Oct, 2003 from Berlin, Germany

Post Wed Jun 11, 2014 1:04 pm

Question - and I mean that in all honesty - do you know, or do you know many people that know, how to actually loudness normalize anything?


There are plenty of tools on the market (however not really affordable ones yet!), and I wrote a couple of tutorials about that already (manual handling - lot of work!). But it seems like as if people are still like "whatever". And... these tools are also not working unified yet for music yet. One tool could interpret -18LUFS different and tool B due to different math. We're still far from actually gaining use out of it. These meters do respond different, depending on how dense a mix is, how much low end was used, how much emphasis was on high frequencies, etc.

Have you ever analyzed a track with the old DR-Meter (Offline Version)? Several times even? 80% of the time you got similar results. But 20% is still a way too high chance value in terms of giving a wrong analysis.

Granted, it would be a simple solution. But this only works:
a) if there is an affordable offline tool that does this (ReplayGain and R128gain are not recommended - yes, I tested them, no they are not suitable for that task since they use their own made up standards!)
b) people sticked to a suitable headroom while mixing (reference levels, peak headroom) - a too much compressed track (read: peak limiting) would result in a bad result if loudness normalized


Why is it so difficult to use a simple VU, a tool that stood the test of time, and work at a level of -18dBFS = 0VU?

There are so many VU's on the market from free to absolute low budget. VUMT is still only 8EUR (Cross Platform), Sleepy Time Records' "legacy bundle" is free (Windows only), Hornet VU meter is also only 4EUR (Cross Platform). Pretty much every host has a RMS meter bundled (ProTools, Cubase has the new "Metering Section", Logic has an RMS meter, Harrison Mixbus already works with VU's). Why use an EBU R-128 meter - the most difficult one to understand? Especially if most of the users in here still don't know the difference of a QPPM, digital PPM, what a K-System does, what the "guidelines" are, etc.


We're currently not talking about A/Bing here. This might turn interesting while mastering where a set level should not be exceeded. We're still talking about mixing and using proper metering. Saying "let's just mix how we feel like, at a level how we're comfortable with and then Loudness Normalize the result" is another thing that I would call "lazy". No learn value whatsoever.
[ Studio Page / Twitter ] | [ KVRmarks (see: metering tools) ] | [ Mix Challenge ]

Return to “Production Techniques”