Making tracks louder

[selig]

Thanks, Selig. Re: Boggg1's post, I meant only the reference to the flow diagram. Agreed, his post is very interesting.

Anyone have any thoughts on my questions above (post dated Tue Sep 29, 2015 1:10 am)? Any answers much appreciated. This gain versus volume issue is melting my brain.

The best word to use IMO is audio LEVEL, since it is a less ambiguous term. Level is traditionally expressed in decibels in the audio engineer world.

Gain was traditionally used for amplifiers to specify the ability to increase the level, but is now pretty much synonymous with level.

Volume is a term that is similar to gain and loudness, and has also become fairly synonymous with level. It is typically a dimensionless value often expressed as a percentage.

Loudness is a more specific term (while also being a more ambiguous value), as it is the human perception of level/volume. As such it is more difficult to quantify. Like temperature, it is difficult to get all listeners to agree on what is half as loud or twice as loud, as it is difficult to get folks to agree on what is twice as hot, or half as cold. There are attempts to define loudness over the years, but as humans tend to average what we hear when perceiving loudness and therefore there are many variables involved. A steady tone would be the easiest to measure, but things get more complex with music as the level may change drastically from bar to bar - if you measure one bar you get one value, measure the next you'll get another. Since humans average what we hear, loudness measuring tools also must do this, but there has always been some disagreement over exactly how to do this. Even the current standards have various ways to interpret the reading of "loudness" depending on what you are attempting to do - are you measuring a single song, a broadcast stream, a TV commercial, etc. As one example, imagine an extreme example: a TV commercial with a loud sound at the beginning and then silence for the remainder of the commercial while there is text on the screen - how loud is it?

To make matters more confusing, let's just focus on how audio level is measured. There are different metering systems to measure level, including VU and Peak which can show widely different values for the same audio signal. This is why it's so important to understand this issue and to use the correct metering tool for the job. VU was an early attempt to measure loudness, and so it's an "average" value. As such it will always read lower than Peak levels. In analog systems this system was more important since there wasn't a well defined clipping point. But in digital systems the clipping point is very specific, and as such peak levels are widely regarded as the best way to measure signal level in a digital system. After all, it's the peaks that will clip first!

The difference between Peak and VU levels isn't fixed and can range anywhere from 6- 30 dB or more! Because of this discrepancy you can have a VU level that is WELL below clipping while the actual signal level is well above clipping! This becomes frustrating in Reason because the channel meters show VU, and well intentioned folks will tell you to set all channel levels to a specific point such as - 10 dBFS. The problem here as I've already mentioned is that at that VU level you are most likely already clipping the mix. Additionally, with all tracks at the same VU level, some will be wildly clipping and others may only just clip in a few places if at all! It is for this reason it is suggested to use PEAK levels when working with digital audio in regards to setting levels, and VU when judging relative loudness. Folks like Paul White at Sound on Sound have even gone so far as to suggest that in digital systems, VU stands for "Virtually Useless" !

[selig]

Further to my post above:

1) If I've never used the blue Gain knob on any channel on any of my tracks before, why do I need to start using it now? Is it not possible to create perfectly acceptable mixes in Reason by simply turning Volume faders up and down?

Yes. But if your input level from audio tracks or instruments is too high, compression and other non-linear effects may be more difficult to set effectively (though it's probably rare if you are even half-way paying attention to your levels).

2) The Gain knobs, as per the Volume faders, make signals louder or quieter. What's the advantage of using the Gain knobs instead of the Volume ones? And if both of these things are performing the same function, why do the Gain knobs exist at all?

Because it was on the original SSL and gain staging is important in the analog domain (due to limited dynamic range at each gain stage compared to digital). This is also where the term "gain stage" comes from, because an analog console is in fact made up of many individual analog circuits, each with their own amount of headroom. Therefore, to gain stage means to set all levels at every gain stage to their NOMINAL LEVEL. In contrast, in digital systems such as Reason that employ floating point audio levels, you have SO much headroom there is no need for traditional gain staging as it is impossible to ever clip a gain stage. Folks still cling to the term "gain staging" today to simply mean "Set Levels" overall. I prefer to suggest an overall "reference level" and avoid the term "gain staging" since it has a different meaning in the analog world.

3) If the blue Gain knobs are so important, why are they hidden out of the way at the very top of the channel strip?

Again, that's the way the original was laid out, mainly because hardware controls must be close to the circuit they control and the input circuit is at the top of the channel module. This is also an indication of how likely you would be to actually need to use this control, which even on an SSL (which I've used since the early 80s) this isn't a control that is often required.

4) Tutorials such as the one recommended by Boggg1 above, and likewise some of the posts above, discuss 'recommended' levels for channel signals. How can this make sense? In most mixing situations you will want some instruments to be louder than others, not all at the same volume. So what's the logic behind using the Gain knob (in conjunction with the Volume fader? Or not?) to drag all signals to the same level? For instance, in the aforementioned tutorial the level of –10 to –7 ("pick one", he says, apparently arbitrarily) is suggested.

The more I look at this, the more complicated – and unnecessary – the Gain issue seems to be …

Yes, it is being made far more complex than it really needs to be IMO. Still, there are great reasons to adopt a reference level for all tracks, but not by the approach suggested in those videos IMO. As far as mixing, that's what the faders are for! You can start with all tracks at the same peak level (if chosen wisely) and still have plenty of room to use the faders to mix. The faders in Reason have PLENTY of resolution even when down at 10-20% of their travel, so nothing 'bad' will happen if you lower or raise the faders far from their 0 dB default setting.

Here are the advantages of adopting a peak reference level (with headroom) for all audio entering the SSL mixer in Reason.

1: For external audio signals recorded into reason you want to leave headroom for many reasons. First, all the analog components have a nominal level where they give the best results. Recording too hot can affect the quality of these circuits, and recording too low can add unnecessary noise (though you REALLY have to record low levels before this becomes a problem in practice). Second, by leaving headroom you will essentially ensure you won't ever clip when recording!

2: Consistency. You can save presets for compressors, amps, saturators etc. and they will work "as is" for all your tracks, audio and instruments alike. This is only true when all channels have a consistent peak level.

3: Headroom on all tracks leaves room in the final mix to avoid clipping without chasing the master fader every time you add a track to the mix.

4: More Consistency: most of the FSB patches are already set to peak around -12 dBFS when playing them, so adopting that level for ALL tracks simply makes the mixing process simpler which allows you to focus on the creative side rather than the technical.

BTW, I totally practice what I preach here, and find that is has greatly simplified the production process for me!
Hope this makes sense, and if not feel free to ask further questions because this IMO is an essential concept to grasp when working with digital audio.

[avasopht]

One big difference between the gain knob at the top and the fader is the level going into your inserts (/sends when set to pre-fader) and channel compressors.

[skie]

OK so two silly questions: The master fader is post-master inserts, yes? and two - is there a difference between the master fader level meter and the Big Meter? Thanks!

[skie]

I was watching a video of tutorial of the producer !llmind making a beat from scratch (he's heavy in the game) and he said he always shoots for his snare to hit at -6 and kick at -3 db (I assume he meant Full Scale). This is somewhat far off from the -12 dbFS people are recommending, no?

[selig]

OK so two silly questions: The master fader is post-master inserts, yes? and two - is there a difference between the master fader level meter and the Big Meter? Thanks!

Master fader is post everything! The way to test this for yourself is to put any device with meters in the insert and pull down the master fader - note that the levels do not follow the master fader.

The master meter and big meter are typically the same, with a few exceptions. One would be if the click track is playing, which will only show on the Big Meter. Another case would be if you patch any device between the master output and the hardware device input. In this second case clipping will no longer be displayed on the master meter due to the fact that it is no longer the last device in the signal path (and could therefore show inaccurate data).

[skie]

OK so two silly questions: The master fader is post-master inserts, yes? and two - is there a difference between the master fader level meter and the Big Meter? Thanks!

Master fader is post everything! The way to test this for yourself is to put any device with meters in the insert and pull down the master fader - note that the levels do not follow the master fader.

The master meter and big meter are typically the same, with a few exceptions. One would be if the click track is playing, which will only show on the Big Meter. Another case would be if you patch any device between the master output and the hardware device input. In this second case clipping will no longer be displayed on the master meter due to the fact that it is no longer the last device in the signal path (and could therefore show inaccurate data).

Cool! That's what I thought, just making sure

[selig]

I was watching a video of tutorial of the producer !llmind making a beat from scratch (he's heavy in the game) and he said he always shoots for his snare to hit at -6 and kick at -3 db (I assume he meant Full Scale). This is somewhat far off from the -12 dbFS people are recommending, no?

No, these are two different things. He is speaking about how hot to make the kick/snare in the mix, and the -12 dBFS level is for individual tracks. For final mix levels it is suggested by many mastering engineers to keep the un-mastered mix to peaks between -6 and -3 dBFS. What !llmind suggests wouldn't work in this case, and besides this would come down to a personal preference. If you used different kick/snare samples than he does you may choose different levels, or if you had different tastes than him you may choose a different relationship between the kick and snare levels. In any event, putting the kick and snare that hot would probably only work if you ALSO do everything else he does (and want to sound like him to some degree), which you have not explained.

To put it another way, if your mastering engineers requests your mix not exceed -3 dB, this technique won't work if the kick and snare hit at the same time or if there are any other tracks playing at the same time which might push the overall level over 0 dBFS. Much of these techniques are down to factors such as how many other tracks you use etc. For example, if you record solo piano, then sticking to the -12 dBFS peaks reference level could be too low, and conversely if you typically record/mix 48-96 tracks or more the -12 dB level could actually be too high!

[boggg1]

In my post I was offering a strategy to an OP who was asking for help, not advising experts who know more than me. Apparently I did not make that clear.

I still argue that setting the levels of significant (not quiet) sounds in the region of -12 using the blue knob and with the sliders set at ~0 is a good strategy for those who do not claim to be expert. Not everyone has been using real SSL desks for 10 years, has self mixed 12 number 1's or can program DSPs while peeling an orange and quoting Shakespeare. We amateurs need simple strategies which maximise our chances to create lovely music without needing a degree. So my post was my suggestion which I believe is factually correct even if it is not the only way to proceed. By the way set quieter sounds at whatever is suitable to broadly be right for the mix, not everything can be at -12, of course. That purely technical technique is what is being called setting a reference level in this thread - sounds as good a name as gain staging to me. It normally takes a few minutes to do. When that stage is finished, then sit back, relax and use the sliders to do the mix, forgetting technical and thinking music, music, music, music and glorious music.

Deal with loudness and bus components (such as bus compressors) in the mastering process as a separate issue.

If you have the knowledge you can break every rule in the book. Invent you're own strategy, use VU meters to set peak levels, overload you're REs, sit in the noise floor, have red lights flashing everywhere, alter the mix when you're mastering, mix with a bus compressor turned on, I don't care. If it sounds good, I'll buy it. I'm just offering a strategy for those self confessed beginners in their Reason journey. Experts look away, I can't help you. And thanks for teaching me what you know.

As for the diagram, I know it is version 2 and I know it states that it does not include every configuration and I know it is all covered in the manual. But it is still the best diagram of the Reason signal path I have seen. If you know any different then help me improve by pointing me at something better or explaining its faults.

[selig]

In my post I was offering a strategy to an OP who was asking for help, not advising experts who know more than me. Apparently I did not make that clear.

I still argue that setting the levels of significant (not quiet) sounds in the region of -12 using the blue knob and with the sliders set at ~0 is a good strategy for those who do not claim to be expert. Not everyone has been using real SSL desks for 10 years, has self mixed 12 number 1's or can program DSPs while peeling an orange and quoting Shakespeare. We amateurs need simple strategies which maximise our chances to create lovely music without needing a degree. So my post was my suggestion which I believe is factually correct even if it is not the only way to proceed. By the way set quieter sounds at whatever is suitable to broadly be right for the mix, not everything can be at -12, of course. That purely technical technique is what is being called setting a reference level in this thread - sounds as good a name as gain staging to me. It normally takes a few minutes to do. When that stage is finished, then sit back, relax and use the sliders to do the mix, forgetting technical and thinking music, music, music, music and glorious music.

Deal with loudness and bus components (such as bus compressors) in the mastering process as a separate issue.

If you have the knowledge you can break every rule in the book. Invent you're own strategy, use VU meters to set peak levels, overload you're REs, sit in the noise floor, have red lights flashing everywhere, alter the mix when you're mastering, mix with a bus compressor turned on, I don't care. If it sounds good, I'll buy it. I'm just offering a strategy for those self confessed beginners in their Reason journey. Experts look away, I can't help you. And thanks for teaching me what you know.

As for the diagram, I know it is version 2 and I know it states that it does not include every configuration and I know it is all covered in the manual. But it is still the best diagram of the Reason signal path I have seen. If you know any different then help me improve by pointing me at something better or explaining its faults.

If this is addressed to me, I'm not sure I was being clear - you sound like an expert to me! Heck, I even said your post was "excellent"!!!
Apologies if my post sounded like I was correcting anything you said! Carry on, you're doing great!
BTW, here is an old signal flow diagram I created to show detailed routing of the channels in the SSL mixer in Reason:


Here is the signal flow chart I saw, and my corrections intended to help the casual user to not be confused or misled - don't know if any of these were ever adopted fwiw…

[Thousand Ways]

Selig, Boggg et al, thank you so much for all your help.

Here's the version of that flow diagram that is currently available at LearnReason.com. It looks as though he did indeed make the changes suggested by Selig.

attachment.jpg (79.35 KiB) Viewed 1925 times

[Exowildebeest]

Boggles the mind why he goes through all the trouble of making an illustrated diagram, then fails to check for huge typos... Makes me cringe.

[skie]

I was watching a video of tutorial of the producer !llmind making a beat from scratch (he's heavy in the game) and he said he always shoots for his snare to hit at -6 and kick at -3 db (I assume he meant Full Scale). This is somewhat far off from the -12 dbFS people are recommending, no?

No, these are two different things. He is speaking about how hot to make the kick/snare in the mix, and the -12 dBFS level is for individual tracks. For final mix levels it is suggested by many mastering engineers to keep the un-mastered mix to peaks between -6 and -3 dBFS. What !llmind suggests wouldn't work in this case, and besides this would come down to a personal preference. If you used different kick/snare samples than he does you may choose different levels, or if you had different tastes than him you may choose a different relationship between the kick and snare levels. In any event, putting the kick and snare that hot would probably only work if you ALSO do everything else he does (and want to sound like him to some degree), which you have not explained.

To put it another way, if your mastering engineers requests your mix not exceed -3 dB, this technique won't work if the kick and snare hit at the same time or if there are any other tracks playing at the same time which might push the overall level over 0 dBFS. Much of these techniques are down to factors such as how many other tracks you use etc. For example, if you record solo piano, then sticking to the -12 dBFS peaks reference level could be too low, and conversely if you typically record/mix 48-96 tracks or more the -12 dB level could actually be too high!

Got it, makes total sense. Thank you. I usually aim for -4dbFS before I hit my mastering goodies. I usually mix on feeling, common sense, and the engineering theory I do know, but I need to have these kinds of discussions to enhance my overall understanding of the process.

[skie]

Lately, in order to ensure I'm around -4db FS before master inserts, I've been adjusting the make-up gain of the master bus compressor. Do you think it would be more transparent if I grouped all the tracks in my mix into one channel on the SSL and used it's fader to get to where I want to be, level-wise?

[selig]

Selig, Boggg et al, thank you so much for all your help.

Here's the version of that flow diagram that is currently available at LearnReason.com. It looks as though he did indeed make the changes suggested by Selig.

attachment.jpg

Though still confusing and not 100% accurate. For example, the output of the sends do NOT come back into the channel fader as shown in the current diagram - they are returned in the master section. In the name of simplicity it would have been better to leave that part out completely. To be fair, it's a difficult concept to visualize with a static drawing IMO. Oh well…

[Thousand Ways]

Boggles the mind why he goes through all the trouble of making an illustrated diagram, then fails to check for huge typos... Makes me cringe.

Writing 'effect' when he means 'affect' is especially misleading.

At the same source there is a downloadable stereo imaging diagram. It looks like a squash court in perspective. I've never tried using a visual aid when placing signals in a stereo spectrum; maybe other people find that helpful.

[gak]

But let's be honest for a moment....... :

http://dictionary.reference.com/help/fa ... e/d70.html

http://www.grammar-monster.com/easily_c ... effect.htm

Vague in the understanding is.... vague.

You could minutia the difference but in reality they could be interchanged in the form of a conversation.

I understand where you are coming from, but gaging language, the necessity for "spell checks" (which are not necessarily grammatically correct) and a general understanding, it's not completely unacceptable. Obviously, I'm not a grammar nazi

[gak]

Now, onto "loudness"

Here's a fact that people get way to technical on that is often misunderstood: The END result is more important than the "mix" result.

In example, If you are "mixing" to get the loudest possible track, your doing it wrong. In the end, you should mix with a mindset of equal value and then apply "loudness" afterwards. Example again, OZONE.

I'm not the leading expert, but I have enough knowledge to understand that getting the "mix" right and then applying "loudness" after the fact is by FAR the most desirable result. It's what the "experts" do, and far be it for me to argue with them.

f... all with getting it loud to start. Do as the manual says, get the mix awesome w/o worrying about loudness and then apply that later.

[avasopht]

What gak said is spot on.

Develop your mix style. Maybe you like tape compressors on every track, or feed a compressor GR to the input control of your distortion and saturation effects that are beefing up your bass.

Learn how each distortion effect affects the signal in both the perceptive and signal domain. Eg, does it sound beefier without increasing the peak level of the signal?

Sometimes you'll find that an effect makes it harder to compress and limit in a way that maximises loudness.

In any case, of you're just making a good mix, eliminating frequency components that don't add to the sound (filtering and eq), and managing the sound space, it'll be much easier to sound loud.

[Thousand Ways]

You could minutia the difference but in reality they could be interchanged in the form of a conversation.

Only by those who don't understand English. The URLs you've given prove my point: the two words are so utterly unlike that one's a verb and the other's a noun. Incidentally, 'Grammar Nazi' is an internet term invented by people who can't spell or understand grammar, to make themselves feel a little better about it. Can't do something? Call those who can 'Nazis', and pretend that they're pedants. 'Nazi' is a term so serious that it shouldn't be used at a distance from its original meaning. Sadly, it's thrown around in so many situations now that it's lost its meaning.

And now, back to making tracks louder.

[Thousand Ways]

Having reread all of the above, and watched the video recommended by Boggg1 again (https://www.youtube.com/watch?v=kT5Nw-ZoEnE), is the following a valid method of gainstaging, or setting levels? Please excuse the oversimplified prose; I'm trying to get this as certain as possible:

1) Instruments, audio clips, etc are added as channels. The fader for each is left at 0 dB;

2) In the Dynamics section of each strip, compression is added to even-out the signal if required (usually in cases wherein the signal fluctuates a lot, such as a vocal);

3) In the Input section of each strip, the blue Gain control is turned until the channel, on playback, remains around as close to –12 dB (or whichever nominal db level the user is aiming at) as possible;

4) Now that all channels are playing at roughly the same –12 dB level, their relative volumes are adjusted using the fader on each. In each case this will be a matter of either leaving the fader at 0 dB or lowering it; none of the faders should be raised above 0 dB, as this will make the channel in question louder than –12 dB;

5) Now that the individual channels have each been optimised at around 12 –dB, the master level will be above –12 dB, and is likely to be too loud (Selig's more-than-one-glass problem, described above). The overall level of the track is corrected back to around –12 dB by lowering the master fader, and if necessary using the master compressor.

Does this make sense, or is it completely wrong?

[lowpryo]

you're on the right track, but there are some things that I would clarify...

2) In the Dynamics section of each strip, compression is added to even-out the signal if required (usually in cases wherein the signal fluctuates a lot, such as a vocal);

that is one possibility, but I would generalize this step even further. I would say this step can include ANY effect you want - EQ, compression, distortion, etc. and it can be from the channel strip, or any insert FX you want to use. this step is basically "get the sound how you want it to sound".

3) In the Input section of each strip, the blue Gain control is turned until the channel, on playback, remains around as close to –12 dB (or whichever nominal db level the user is aiming at) as possible;

you don't have to use the blue gain control here. it's what Bogg uses, and that's fine. but it's important to remember - that is the level of the signal before any channel strip processing or insert FX (in step 2), and modifying this level will change any compression or distortion that you applied. I personally like to add Selig's Gain rack extension at the end of my insert FX to aim for the -12dB. but you can do this with pretty much any device that has a built-in gain knob (which almost every device does)

5) Now that the individual channels have each been optimised at around 12 –dB, the master level will be above –12 dB, and is likely to be too loud (Selig's more-than-one-glass problem, described above). The overall level of the track is corrected back to around –12 dB by lowering the master fader, and if necessary using the master compressor.

your master doesn't need to be set back to -12dB. the only hard rule is that your master should not exceed 0dB (selig recommended to peak around -6dB to -3dB, but that's just a rule of thumb). the purpose of setting all of the previous tracks at -12dB was so that when they add up, the overall level doesn't reach 0dB (hopefully). then once your track is mixed how you want it to sound, and is not exceeding 0dB, the rest of the loudness can be achieved through mastering.

[normen]

I just want to throw in here that *loudness* is definitely not something that is only achieved by mastering, its very much depending on your mix, the EQing, arrangement and overall crest factor you generate during mixing. The final *level* of your track should definitely be pulled up only during mastering though and you should leave enough headroom during mixing.

What the meter displays is * level*, the *loudness* of tracks can differ very much even when they have the same *level* (or volume).

[zeebot]

[gak]

You could minutia the difference but in reality they could be interchanged in the form of a conversation.

Only by those who don't understand English. The URLs you've given prove my point: the two words are so utterly unlike that one's a verb and the other's a noun. Incidentally, 'Grammar Nazi' is an internet term invented by people who can't spell or understand grammar, to make themselves feel a little better about it. Can't do something? Call those who can 'Nazis', and pretend that they're pedants. 'Nazi' is a term so serious that it shouldn't be used at a distance from its original meaning. Sadly, it's thrown around in so many situations now that it's lost its meaning.

I dunno. Wat da kidz r doin' these dayz making it really hardz.

Agree that a mistake like that should be corrected*, not going to get overly fussed about.

Edited because I said "correctly" instead of "corrected" which just brings the point home, even though you know better, nobody is perfect.