Mix Challenge - Addendum: Summing Bus Treatment
Posted: Sat May 27, 2017 15:04 CEST
The following questions pop up over and over.
These are all valid questions that I'd like to address with this post. In fact, there are Rules/Guidelines about this given already, and I posted several times how to perform best in this situation. But it seems to be a recurring question, so let us please discuss/tackle this topic more in-depth.
THE CONCEPT OF THE MIX CHALLENGE...
The "Mix Challenge" is as the name implies. It is a mixing competition. The focus is on properly setting up a project, edit it (cutting tracks "clean", setting proper fade in/out), mix it within giving specifications and bonus-rules, then export it (re-record for hardware-only users). During the mixing process, you can use whatever tools you feel like (within given limitations), or any form of mix techniques you know (vintage mix techniques, Brauerize method, mix-in-mono first then spread it out to stereo, etc).
But the main idea is to understand how mixing works, gather experience (if you're inexperienced) and practice techniques, maybe completely overhaul your up-to-this-point used workflow (some users tend to not gain stage, and constantly exceed the limits of the digital meter, adjust the summing bus fader, etc).
WHAT IS MEANT WITH "TRY NOT TO PRE-MASTER THE TRACK"?
As mentioned in the previous point, your focus should be on mixing - and mixing only(!). These days, mixing and "mastering" (technically pre-mastering) is considered to be one and the same step. But the Mix Challenge tries to get away from that and draw distinct lines. Reason being for possible further challenges(!) and for the real world, giving the mastering engineer way more room to work with of course.
If you'd handle all the EQ fixing, stereo field override, limiting and loudness raising yourself already during/after mixing - then the mastering engineer has no choice but to use(!) this as source for his further edits. Making things just more and more loud. Or trying to fix issues that were messed up during the mix with a global EQ/compressor, etc.
I am fully aware that the lines are thin these days. And there are always the questions:
"how much is too much?", "what is considered tasteful summing bus treatment?", "how to do it right?"
The short answer is, that there is no definite / right solution for this. There are only recommendations.
Ultimately it is up to you, the mix engineer, how much wiggle room you give the next person. Though keep in mind, the audio realm is currently changing. "Loudness Normalization" is a huge topic since early 2010s (and since 2017 now more than ever). Meaning that your drastically over-compressed track is being reduced in loudness to go down to a certain value. Your now "limited" track has to compete with material that still has his transients and dynamic (not to misinterpret with dynamic range!) fully intact. Which production do you think would sound better?
Notable examples to this topic:
Metallica's "Death Magnetic" from 2008 or Imagine Dragons' "Smoke + Mirrors" from 2015
vs Michael Jackson's "HIStory – Past, Present and Future Book I")
WHAT CAN BE CONSIDERED AS GOOD HANDLING OF THE SUMMING BUS?
Let's start with an illustration:
This image shows 4 possible ways how to handle a production on the summing bus (or "master bus" in certain hosts) without overdoing things. From "no treatment" to "considered borderline too much". This is just an example, not a clear rule.
EXAMPLE 1 - Metering only:
In this example, there is of course no further treatment happening. You mix as well as you can, and only check for levels or frequency issues on the sum of your available mix console.
EXAMPLE 2 - "The Glue Effect":
For some productions, there is a need to "gel" things together. Which in turn emphasizes certain instruments and pushes things even further. This was (and still is) a famous technique especially used in the early 90ies for Grunge and Rock productions. EDM productions also use summing bus compression for added "impact" (sometimes even with specific setup multi-band compressors). You might have heard of this in video tutorials as so called "record enable technique". However - don't misinterpret this as "mastering and loudness raising". It is an artistic choice.
The idea here is to have a moderate-to-fast responding compressor with a ratio of 2:1 to 4:1 "compacting" the incoming signal by 1,5dB to 3dB of gain reduction absolute maximum (ideally, not higher than 2dB of Gain Reduction). No output gain compensation, no raising of the volume (and therefore the perceived loudness). The effect is supposed to be subtle, but having enough impact to get the feeling that there is something "larger than life" happening to the mix. Or in other words "glue things even more together".
EXAMPLE 3 - The Console / Tape Effect:
This topic precedes "The Glue Effect". Due to the revolution in recent years with "console modeling", you see this setup more and more as deliberate artistic/sound design choice. With the right tools, you create an ITB (In-the-box) analog mixing console (some hosts have this built in, HARRISON MIXBUS comes to mind), and would maybe even like to print the mix down on to tape. So you see/use a console emulation type plugin on the summing bus (resulting in varying amounts of crosstalk and saturation), but also a tape machine (more saturation, possible sound shaping due to tape head bump effect, "tape" compression). Back in the days, this was the only way to record and process a production. These days, this technique is used to spice up mixes that might sound too digital, or if you want to re-create vintage mixing techniques.
Notice that there isn't an additional compression or EQing in the signal chain. Not only that, the console and tape machine is setup to work at their hotspots of -2VU to +2VU. Like "The Glue Effect", this can add something "larger than life" to a production, while not overdoing it as much compared to "The Glue Effect" (which can already be a precursor to pre-mastering).
EXAMPLE 4 - Console, Tape/Compressor and Fixing-EQ (fix one-EQ-band):
Now we're getting into the territory that I (personally) start to(!!!) consider "borderline too much". We now have the console (crosstalk, saturation), the tape machine (head bump, saturation, compression) and then you realize "whoops, now I have too many mid frequencies". In this case, it is still(!) valid (in my opinion) to add another EQ, and then turn down one frequency band only, maybe more if there is another problematic frequency. It can still considered be "general sound design". But then the lines turn thin - everything after that, clearly goes into "fix the mixdown" / mastering territory
The bigger picture here, is that the mixdown should sound great to begin with. The next step (mastering, porting into the right release format) is then only to emphasize on the mix with so called "Fairy Dust" (if needed!).
THE THIN LINE...
Now we're treading on a very thin line. "Was the summing bus/master bus handling too much?", "What about the so called 'Fairy Dust'?", "but if we add more to the summing bus, while keeping a certain loudness value, the mastering engineer still has a lot to work with, no?"
All valid arguments. However, notice that this setup didn't use either of the following:
I am fully aware that the line is thin. But there is a purpose behind it. These days, we have access to near unlimited tools - but will that be the case in a studio you're not familiar with? Or what about a pure analog large scale hardware console? This is the focus of the Mix Challenge - it's a learning experience.
WHAT ABOUT SETTING A LOUDNESS VALUE FOR THE MIXDOWN THAT EVERYONE HAS TO ADHERE TO?
This has already been addressed in the Rules/Guidelines of the Mix Challenge.
It is recommended to work with a reference level of either -20dBFS = 0VU (SMPTE RP 0155 recommendation) or -18dBFS = 0VU (EBU R68 convention).
Simplified, you use a combination of a VU (without weighting filter) and Digital meter. Then you try to mix in a way so that your maximum signal strength (dBTP or dBFS) does not exceed -1dBFS, while your average signal strength (VU w/o weighting filter / RMS realtime with Dorrough specs) does hover between -21dBFS to -15dBFS (RMS realtime) on average or -3VU to +3VU.
There are plenty of articles about that on the KVR Marks of the CTO of the Mix Challenge, and now even on the Mix Challenge Forum (in Production Techniques)
I do not recommend to go with LUFS during mixing music, as this particular metering tool (ITU-R BS.1770-x specs / EBU R-128 and ATSC A/85 are presets of it) has it's main focus on mastering and broadcasting. Also, there are still many people out there, that don't know the fine details of available metering tools. So one might say "I went for xyz LUFS" but does he/she mean average, maximum, integrated, etc?
DR-Metering ("Dynamic Range" metering, or in reality "crest factor measurement", first created by Friedemann Tischmeyer/Algorithmix and ported by Brainworx) is also not suitable for the task IMO, as the value varies heavily depending on the density (how dense the production was mixed) of the source material.
Your focus should really be on the basics (in this case: digital meter and VU meter). This might sound absolutely limiting at first. But I (Mister Fox) have been doing this for 10+ years, and I feel not limited at all. In fact, it's the other way around. As I can incorporate hardware or modeled devices in software form as needed/in combination at will, without adjusting anything.
The question is now...
How to simplify this/make the rule set more apparent?
Or is this thread enough to cover the fine details?
- What do you mean when you say "try to not(!) pre-master the track (please no overuse of mix bus treatment)"?
- shouldn't we just use fixed loudness values on the final render and the participant can decide how much treatment he uses on the summing bus?
- how much treatment is too much?
These are all valid questions that I'd like to address with this post. In fact, there are Rules/Guidelines about this given already, and I posted several times how to perform best in this situation. But it seems to be a recurring question, so let us please discuss/tackle this topic more in-depth.
THE CONCEPT OF THE MIX CHALLENGE...The "Mix Challenge" is as the name implies. It is a mixing competition. The focus is on properly setting up a project, edit it (cutting tracks "clean", setting proper fade in/out), mix it within giving specifications and bonus-rules, then export it (re-record for hardware-only users). During the mixing process, you can use whatever tools you feel like (within given limitations), or any form of mix techniques you know (vintage mix techniques, Brauerize method, mix-in-mono first then spread it out to stereo, etc).
But the main idea is to understand how mixing works, gather experience (if you're inexperienced) and practice techniques, maybe completely overhaul your up-to-this-point used workflow (some users tend to not gain stage, and constantly exceed the limits of the digital meter, adjust the summing bus fader, etc).
WHAT IS MEANT WITH "TRY NOT TO PRE-MASTER THE TRACK"?As mentioned in the previous point, your focus should be on mixing - and mixing only(!). These days, mixing and "mastering" (technically pre-mastering) is considered to be one and the same step. But the Mix Challenge tries to get away from that and draw distinct lines. Reason being for possible further challenges(!) and for the real world, giving the mastering engineer way more room to work with of course.
If you'd handle all the EQ fixing, stereo field override, limiting and loudness raising yourself already during/after mixing - then the mastering engineer has no choice but to use(!) this as source for his further edits. Making things just more and more loud. Or trying to fix issues that were messed up during the mix with a global EQ/compressor, etc.
I am fully aware that the lines are thin these days. And there are always the questions:
"how much is too much?", "what is considered tasteful summing bus treatment?", "how to do it right?"
The short answer is, that there is no definite / right solution for this. There are only recommendations.
Ultimately it is up to you, the mix engineer, how much wiggle room you give the next person. Though keep in mind, the audio realm is currently changing. "Loudness Normalization" is a huge topic since early 2010s (and since 2017 now more than ever). Meaning that your drastically over-compressed track is being reduced in loudness to go down to a certain value. Your now "limited" track has to compete with material that still has his transients and dynamic (not to misinterpret with dynamic range!) fully intact. Which production do you think would sound better?
Notable examples to this topic:
Metallica's "Death Magnetic" from 2008 or Imagine Dragons' "Smoke + Mirrors" from 2015
vs Michael Jackson's "HIStory – Past, Present and Future Book I")
WHAT CAN BE CONSIDERED AS GOOD HANDLING OF THE SUMMING BUS?Let's start with an illustration:
This image shows 4 possible ways how to handle a production on the summing bus (or "master bus" in certain hosts) without overdoing things. From "no treatment" to "considered borderline too much". This is just an example, not a clear rule.
EXAMPLE 1 - Metering only:
In this example, there is of course no further treatment happening. You mix as well as you can, and only check for levels or frequency issues on the sum of your available mix console.
EXAMPLE 2 - "The Glue Effect":
For some productions, there is a need to "gel" things together. Which in turn emphasizes certain instruments and pushes things even further. This was (and still is) a famous technique especially used in the early 90ies for Grunge and Rock productions. EDM productions also use summing bus compression for added "impact" (sometimes even with specific setup multi-band compressors). You might have heard of this in video tutorials as so called "record enable technique". However - don't misinterpret this as "mastering and loudness raising". It is an artistic choice.
The idea here is to have a moderate-to-fast responding compressor with a ratio of 2:1 to 4:1 "compacting" the incoming signal by 1,5dB to 3dB of gain reduction absolute maximum (ideally, not higher than 2dB of Gain Reduction). No output gain compensation, no raising of the volume (and therefore the perceived loudness). The effect is supposed to be subtle, but having enough impact to get the feeling that there is something "larger than life" happening to the mix. Or in other words "glue things even more together".
EXAMPLE 3 - The Console / Tape Effect:
This topic precedes "The Glue Effect". Due to the revolution in recent years with "console modeling", you see this setup more and more as deliberate artistic/sound design choice. With the right tools, you create an ITB (In-the-box) analog mixing console (some hosts have this built in, HARRISON MIXBUS comes to mind), and would maybe even like to print the mix down on to tape. So you see/use a console emulation type plugin on the summing bus (resulting in varying amounts of crosstalk and saturation), but also a tape machine (more saturation, possible sound shaping due to tape head bump effect, "tape" compression). Back in the days, this was the only way to record and process a production. These days, this technique is used to spice up mixes that might sound too digital, or if you want to re-create vintage mixing techniques.
Notice that there isn't an additional compression or EQing in the signal chain. Not only that, the console and tape machine is setup to work at their hotspots of -2VU to +2VU. Like "The Glue Effect", this can add something "larger than life" to a production, while not overdoing it as much compared to "The Glue Effect" (which can already be a precursor to pre-mastering).
EXAMPLE 4 - Console, Tape/Compressor and Fixing-EQ (fix one-EQ-band):
Now we're getting into the territory that I (personally) start to(!!!) consider "borderline too much". We now have the console (crosstalk, saturation), the tape machine (head bump, saturation, compression) and then you realize "whoops, now I have too many mid frequencies". In this case, it is still(!) valid (in my opinion) to add another EQ, and then turn down one frequency band only, maybe more if there is another problematic frequency. It can still considered be "general sound design". But then the lines turn thin - everything after that, clearly goes into "fix the mixdown" / mastering territory
The bigger picture here, is that the mixdown should sound great to begin with. The next step (mastering, porting into the right release format) is then only to emphasize on the mix with so called "Fairy Dust" (if needed!).
THE THIN LINE...Now we're treading on a very thin line. "Was the summing bus/master bus handling too much?", "What about the so called 'Fairy Dust'?", "but if we add more to the summing bus, while keeping a certain loudness value, the mastering engineer still has a lot to work with, no?"
All valid arguments. However, notice that this setup didn't use either of the following:
- mid/side treatment / enhancement (stereo widening techniques, bass narrowing)
- adding "Fairy Dust" (parallel EQ, broad "character EQ", more saturation)
- multiband compression (if needed! Compacting the signal even further, which leads to...)
- limiting
- loudness adjustments
I am fully aware that the line is thin. But there is a purpose behind it. These days, we have access to near unlimited tools - but will that be the case in a studio you're not familiar with? Or what about a pure analog large scale hardware console? This is the focus of the Mix Challenge - it's a learning experience.
WHAT ABOUT SETTING A LOUDNESS VALUE FOR THE MIXDOWN THAT EVERYONE HAS TO ADHERE TO?This has already been addressed in the Rules/Guidelines of the Mix Challenge.
It is recommended to work with a reference level of either -20dBFS = 0VU (SMPTE RP 0155 recommendation) or -18dBFS = 0VU (EBU R68 convention).
Simplified, you use a combination of a VU (without weighting filter) and Digital meter. Then you try to mix in a way so that your maximum signal strength (dBTP or dBFS) does not exceed -1dBFS, while your average signal strength (VU w/o weighting filter / RMS realtime with Dorrough specs) does hover between -21dBFS to -15dBFS (RMS realtime) on average or -3VU to +3VU.
There are plenty of articles about that on the KVR Marks of the CTO of the Mix Challenge, and now even on the Mix Challenge Forum (in Production Techniques)
I do not recommend to go with LUFS during mixing music, as this particular metering tool (ITU-R BS.1770-x specs / EBU R-128 and ATSC A/85 are presets of it) has it's main focus on mastering and broadcasting. Also, there are still many people out there, that don't know the fine details of available metering tools. So one might say "I went for xyz LUFS" but does he/she mean average, maximum, integrated, etc?
DR-Metering ("Dynamic Range" metering, or in reality "crest factor measurement", first created by Friedemann Tischmeyer/Algorithmix and ported by Brainworx) is also not suitable for the task IMO, as the value varies heavily depending on the density (how dense the production was mixed) of the source material.
Your focus should really be on the basics (in this case: digital meter and VU meter). This might sound absolutely limiting at first. But I (Mister Fox) have been doing this for 10+ years, and I feel not limited at all. In fact, it's the other way around. As I can incorporate hardware or modeled devices in software form as needed/in combination at will, without adjusting anything.
The question is now...How to simplify this/make the rule set more apparent?
Or is this thread enough to cover the fine details?
