Using the Maximizer

[Jmax]

Hi guys,

I'm currently working on a project with quite a few samples.  All samples are high quality and don't clip.  Yet when I play the overall track I do get a bit of clipping. I figured I could solve this using the Master Section Maximizer but it still seems  I get clipping with that on.  I wondered if I have it in the correct place in the 'chain' or if I'm missing something. I can always just turn down levels and will do more EQ'in but by default shouldn't the maximizer control the clipping? 

[Benedict]

Hi

Yes in theory the Maximizer will swallow all overs (provided the Look Ahead is on). However it will create distortions of its own if pushed too hard.

NN-XT can and will clip itself if you have layered samples. Often you don't notice till the mixdown. If this is the case then turn the Level of every sample down (not the master) then turn the level up using the Gain on the SSL, a Compressor Output (no compression required) or RE like Selig's Gain.

[Tincture]

Selig did a great tutorial on the MClass Maximiser that might be around somewhere? From what I remember if you have the 4ms on and fast attack you should be safe (certainly if soft-clip is above zero). It might be in your favour if I'm wrong here 'cause I reckon someone might just correct me, and say much more to your favour

[Dan Topic]

It's always better to pan and EQ than limit...

[Tincture]

It's always better to pan and EQ than limit...

True but it can be hard to use much Pan on loud elements like kicks, snares, basses. Also, depending on the genre of music you may need a certain amount of bass (loud) on certain elements.

My current strategy is to use slow attack comp to bring up the punch then use fast attack at as higher threshold as possible to cut the peaks.

[Dan Topic]

Well, kicks and basses should stay mostly centered - low frequencies are difficult to determine directon they come from anyway.

I have listened to couple of your track and mostly they sound like many instruments are in center. I would make e.g. for piano 3 parallel channels and split 2 to the right (let's say 30 and 45) and 2 to the left (-30 and -50). All with slightly different settings to get a rich sound. Then you have enough place in the center for bass and drums (also hi-hats and snare and percussion I would pan a bit left/right but not to collide with piano). If the bass ocupies the same frequencies as kick drum then I would use subtractive EQing to make free room for both of them. Also compressor comes handy at this point.

The point is to make own space for all sounds whether by panning or EQing. If everything is centered the signal sums and clipping can occur. 

[Tincture]

Thanks for listening Dan. Sounds like good advice.

[Dan Topic]

There are some ¨problem children¨ like guitar, piano and sax for example. These instruments can occupy wide EQ spectrum so it's not easy to combine them together, in addition with vocals. So I always parallel them and pan aside. They sound like centered but more wide and there's still enough place for vocals in center then.
 
This is one of the most important things for good mix: http://www.reasonforums.com/showthread. ... 65#post365

Think about overall sound like it's a box with opening toward you and place individual sounds in it not too overlap each other much. Then you won't need limiters that much. Who want's limited sounds, heh?

[Dan Topic]

Check my SC profile and listen to some of my newest songs (the newer the better mixing and mastering I guess ) and hear where individual sounds are placed. 

I'm not a master but trying to do my best. I had been also struggling with good sound - like everybody I think

Take care man!

[Dan Topic]

A note to ad: It's always good to route the main instrument buss and all parallel busses to a new output buss to have overall control of whole instrument and you can see the sum of all signals on New output buss meter.

[Dan Topic]

Here's a image how I managed vocals in a song:
 
From the left:
1) Vocal audio track processed (EQ, compressor etc.) - left panned
2) Backing vocals parallel channel - left panned
3) Vocal audio track processed (EQ, compressor etc.) - right panned
4) Backing vocals parallel channel - right panned
5) Interline echo (inserts) - centered 
6) Echo strip (inserts) - centered
7) VOCALS - New output buss

[Dan Topic]

Here you can see it when playing. Even though some channels are reaching 0 db, the new output buss is still holding the volume around - 10db.

 

[selig]

Here you can see it when playing. Even though some channels are reaching 0 db, the new output buss is still holding the volume around - 10db.

 

That's only because you lowered the level of the bus - nothing to do with panning. In fact, in Reason, panning ADDS gain (3 dB when hard panned) and will contribute even MORE to clipping the output.

[submonsterz]

Here you can see it when playing. Even though some channels are reaching 0 db, the new output buss is still holding the volume around - 10db.

 

That's only because you lowered the level of the bus - nothing to do with panning. In fact, in Reason, panning ADDS gain (3 dB when hard panned) and will contribute even MORE to clipping the output.

lol that was exactly what was in my mind when I read that and see the pics. Lift the buss back to 0 it would without a maximiser be clipping like a batch.

[selig]

Hi guys,

I'm currently working on a project with quite a few samples.  All samples are high quality and don't clip.  Yet when I play the overall track I do get a bit of clipping. I figured I could solve this using the Master Section Maximizer but it still seems  I get clipping with that on.  I wondered if I have it in the correct place in the 'chain' or if I'm missing something. I can always just turn down levels and will do more EQ'in but by default shouldn't the maximizer control the clipping? 

This is a VERY simple issue that no one so far has addressed (forgive me if I missed something here). It's PURELY a level issue, nothing to do with EQ, panning, or limiting/maximizing. 

IMPORTANT - this can't be emphasized enough since many (including you) don't seem to understand it. The individual instruments (samples in your case) CANNOT clip on their own. It is ONLY the hardware output (Digital to Analog converter in your sound card) that can clip. Putting a limiter/maximizer on the mix to "fix" clipping is like a tall person cutting off their head so it doesn't hit the ceiling!

It's very simple - to prevent clipping, LOWER LEVELS. To increase loudness, use a limiter BUT as other have said, be aware there is NO free lunch - a limiter will eventually impart a sound that is often not desirable when pushed too far. 

My suggestion is to use lower levels on each track in the future, but for now simply lower the master fader. IMO it's best to keep things as simple as possible.

[Dan Topic]

Here you can see it when playing. Even though some channels are reaching 0 db, the new output buss is still holding the volume around - 10db.

 

That's only because you lowered the level of the bus - nothing to do with panning. In fact, in Reason, panning ADDS gain (3 dB when hard panned) and will contribute even MORE to clipping the output.

Yes but that's the level I wanted. And I wrote that routing to the new output bus is good that you can see the sum of all signals on New output buss meter. Without it you could think all levels are OK and clipping cannot occur. And when all channels are centered the signal strenght increases. Panning always helped me to solve this. 
As I said I'm not the master, actually I'm a newbie - started 1 and 1/2 year ago without any previous production background. This technique helped me to get better sound and I just wanted to help

Here's the song with these vocals. You can listen how it sounds:

Perhaps not perfect but I'm still learning. Any feedback appreciated
Here's the resulting track:
 

[Dan Topic]

Here you can see it when playing. Even though some channels are reaching 0 db, the new output buss is still holding the volume around - 10db.

 

That's only because you lowered the level of the bus - nothing to do with panning. In fact, in Reason, panning ADDS gain (3 dB when hard panned) and will contribute even MORE to clipping the output.

lol that was exactly what was in my mind when I read that and see the pics. Lift the buss back to 0 it would without a maximiser be clipping like a batch.

I don't understand why would I lift it to 0db. There are other instruments in the song and, of course the signals would be sumed with vocals, causing clipping. I always keep individual main tracks about -10db to have room for the others. Master bus shows you sum of all tracks and it can be hard to find which one is to blame. Grouping the channels and parallels makes it easier. That's all I wanted to say. And there's no need for limiting if you stage all carefuly. I use limiter when mastering but very seldom when gain staging.

[submonsterz]

Here you can see it when playing. Even though some channels are reaching 0 db, the new output buss is still holding the volume around - 10db.

 

That's only because you lowered the level of the bus - nothing to do with panning. In fact, in Reason, panning ADDS gain (3 dB when hard panned) and will contribute even MORE to clipping the output.

lol that was exactly what was in my mind when I read that and see the pics. Lift the buss back to 0 it would without a maximiser be clipping like a batch.

I don't understand why would I lift it to 0db. There are other instruments in the song and, of course the signals would be sumed with vocals, causing clipping. I always keep individual main tracks about -10db to have room for the others. Master bus shows you sum of all tracks and it can be hard to find which one is to blame. Grouping the channels and parallels makes it easier. That's all I wanted to say. And there's no need for limiting if you stage all carefuly. I use limiter when mastering but very seldom when gain staging.

my point was all you had done was turn down all the volumes equally by lowering the bus fader to stop the clipping . It's just the same thing as you'd be doing with a grouped fader on all them channels and pulling thier volumes down equally the same after you had set you're balance for the parallels You have. My point was not to turn it to 0 db but the point all you had done via the bus was turn it all down . And the point was if you hadn't turned it down or had it you'd have had to turn them all down to get same result with out limiting .. and I find it no problems finding a signal that's too hot in a mix to cause the clips just by muting individuals and using the peak meter it's pretty easy to see what channel or sounds is the culprit that way to me .

[Tincture]

Ok... I see you got a bit excited Dan! Nothing wrong with that! Good to see Selig chip in... he really does know what he's talking about

11/2 years is nothing really so I'm sure you are doing really well

Here are some mistakes I have made (relevant or not):

Panning the same material (i.e. no changes) left and right doesn't widen it at all. It just pans it by the difference in the pan amounts. To make parallel material spread across the stereo field you must add some tuning (pitch) or timing (delay) change. Only then will you get the separation expected.

I used to do a lot of things like this, expecting the changes to be as I thought, without really listening whether that was the case. Since learning a LOT from experts here (and they really are) I now challenge my expectations with my ears!! Sounds daft but I used to "hear" so many differences just because I was expecting them. When actually nothing was happening!

My biggest mistake of all was not to realise that there is a difference between mixing and mastering. I was always aiming to get things limiting by a few db to make sure my songs were loud. I even used to mix with a limiter in place! I thought I had to have things hitting 0db to be loud and modern.

Once I realised (thanks to Selig mainly) that it's best to leave lots of headroom when mixing and go for around -12db peaks (I tend to exceed that by quite a bit for beats) my songs started getting much better. If the final mix is peaking at -6 or -4db so what! That's what limiters or other gain units are for.

Hope this doesn't sound condescending... maybe it'll clear something in someone's head somewhere? I was :t0152: before

[eox]

Ok... I see you got a bit excited Dan! Nothing wrong with that! Good to see Selig chip in... he really does know what he's talking about

11/2 years is nothing really so I'm sure you are doing really well

Here are some mistakes I have made (relevant or not):

Panning the same material (i.e. no changes) left and right doesn't widen it at all. It just pans it by the difference in the pan amounts. To make parallel material spread across the stereo field you must add some tuning (pitch) or timing (delay) change. Only then will you get the separation expected.

I used to do a lot of things like this, expecting the changes to be as I thought, without really listening whether that was the case. Since learning a LOT from experts here (and they really are) I now challenge my expectations with my ears!! Sounds daft but I used to "hear" so many differences just because I was expecting them. When actually nothing was happening!

My biggest mistake of all was not to realise that there is a difference between mixing and mastering. I was always aiming to get things limiting by a few db to make sure my songs were loud. I even used to mix with a limiter in place! I thought I had to have things hitting 0db to be loud and modern.

Once I realised (thanks to Selig mainly) that it's best to leave lots of headroom when mixing and go for around -12db peaks (I tend to exceed that by quite a bit for beats) my songs started getting much better. If the final mix is peaking at -6 or -4db so what! That's what limiters or other gain units are for.

Hope this doesn't sound condescending... maybe it'll clear something in someone's head somewhere? I was :t0152: before

Okay this leads me to a question I've had!

For the last year or so I've taken Selig and others advice on dropping my channels to keep the peak around -12dbfs -10dbfs which has helped quality majorly as if you can get your synths, drums, etc. to found full already around those levels then the mastering stage seems to make it pristine. Yet, even though during the mastering stage (I really have no idea what I'm doing but I try) I keep my levels low where they were during the mixing stage (as I usually don't bounce to wavs) and though my master bus begins to peak out at -.01 dbfs or so, it sounds very quiet. Like my mix is just..I'm not sure, low I suppose. But when I listen to my track everything just sounds right even if it's quiet. So am I needing to starting bumping the levels up on each channel during the master stage? Or am I just mastering it all wrong? I hope I made sense..haha

[Tincture]

Makes perfect sense!

I know nothing about mastering either but have picked up certain things from posters here and the PUF.

If you "only just" limit your mix, it will sound quiet. I used to aim for an average complete song level (available in programs such as Har-Bal etc) in the -14dB region but have adjusted that so that I measure my peaks using Flower Audio Loudness Meter at RMS 5sec in Reason and make sure they don't go over -9dB for long (the odd -8.9 or 8.8 I'll tolerate) and this has tended to give me louder songs that average around -13.5 to -12.5db. When I look at Ozone I rarely see it doing more than 2 dB reduction. I tend to have fast settings.

In other words, once you have a mix peaking well below 0dB don't be afraid to add gain before the limiter (or with the limiter) to see where it sounds nice and loud but not too crushed (beats especially).

[CharlyCharlzz]

what about that new RE with a soft-knee GClip ? anybody tryed it ? 

https://shop.propellerheads.se/browse/?q=gc

[Tincture]

what about that new RE with a soft-knee GClip ? anybody tryed it ? 

https://shop.propellerheads.se/browse/?q=gc

yeah it's a great re but I'd only use it to get certain kinds of distortion... Not for comp on the master bus! Just maybe on certain instruments...

[selig]

Here you can see it when playing. Even though some channels are reaching 0 db, the new output buss is still holding the volume around - 10db.

 

That's only because you lowered the level of the bus - nothing to do with panning. In fact, in Reason, panning ADDS gain (3 dB when hard panned) and will contribute even MORE to clipping the output.

lol that was exactly what was in my mind when I read that and see the pics. Lift the buss back to 0 it would without a maximiser be clipping like a batch.

I don't understand why would I lift it to 0db. There are other instruments in the song and, of course the signals would be sumed with vocals, causing clipping. I always keep individual main tracks about -10db to have room for the others. Master bus shows you sum of all tracks and it can be hard to find which one is to blame. Grouping the channels and parallels makes it easier. That's all I wanted to say. And there's no need for limiting if you stage all carefuly. I use limiter when mastering but very seldom when gain staging.

You don't seem to be aware that the channel meters read in VU, and therefore don't show the true peak levels. When you say you keep tracks at -10 dB you don't specify the meter mode which can make a HUGE difference. If you use VU, you can be clipping and the meter won't show it. 

As for your contention the master meter doesn't show "which one is to blame", that is totally irrelevant! If you lower an individual mixer channel you are changing your mix balances, which is NOT what you want to do to prevent clipping. The way to prevent clipping is NOT to change panning, EQ, or individual track levels. Those adjustments are at the heart of mixing - once you have your mix you simply need to use the master fader or a master insert to set the overall level to not clip. 

What you are suggesting is like using less of ONE ingredient in a recipe in order to adjust the total amount of the dish you are preparing - it would RUIN the recipe! The better solution is to use a bigger pot or to adjust ALL ingredients proportionally, NOT to use less or more of one ingredient.

[selig]

Ok... I see you got a bit excited Dan! Nothing wrong with that! Good to see Selig chip in... he really does know what he's talking about

11/2 years is nothing really so I'm sure you are doing really well

Here are some mistakes I have made (relevant or not):

Panning the same material (i.e. no changes) left and right doesn't widen it at all. It just pans it by the difference in the pan amounts. To make parallel material spread across the stereo field you must add some tuning (pitch) or timing (delay) change. Only then will you get the separation expected.

I used to do a lot of things like this, expecting the changes to be as I thought, without really listening whether that was the case. Since learning a LOT from experts here (and they really are) I now challenge my expectations with my ears!! Sounds daft but I used to "hear" so many differences just because I was expecting them. When actually nothing was happening!

My biggest mistake of all was not to realise that there is a difference between mixing and mastering. I was always aiming to get things limiting by a few db to make sure my songs were loud. I even used to mix with a limiter in place! I thought I had to have things hitting 0db to be loud and modern.

Once I realised (thanks to Selig mainly) that it's best to leave lots of headroom when mixing and go for around -12db peaks (I tend to exceed that by quite a bit for beats) my songs started getting much better. If the final mix is peaking at -6 or -4db so what! That's what limiters or other gain units are for.

Hope this doesn't sound condescending... maybe it'll clear something in someone's head somewhere? I was :t0152: before

Okay this leads me to a question I've had! For the last year or so I've taken Selig and others advice on dropping my channels to keep the peak around -12dbfs -10dbfs which has helped quality majorly as if you can get your synths, drums, etc. to found full already around those levels then the mastering stage seems to make it pristine. Yet, even though during the mastering stage (I really have no idea what I'm doing but I try) I keep my levels low where they were during the mixing stage (as I usually don't bounce to wavs) and though my master bus begins to peak out at -.01 dbfs or so, it sounds very quiet. Like my mix is just..I'm not sure, low I suppose. But when I listen to my track everything just sounds right even if it's quiet. So am I needing to starting bumping the levels up on each channel during the master stage? Or am I just mastering it all wrong? I hope I made sense..haha

It's a mastering issue - peak levels do not represent "loudness". Crest Factor, which is the difference between peak and average levels, is a better judge of loudness. 

For example you can have two mixes that both peak at 0.1 dBFS but one sounds a LOT louder than the other. How is this possible? Here's an extreme example: a track that has ONE loud section that hits 0 dBFS vs a track that continually hits 0 dBFS on every beat. Both have a peak level of 0 dBFS, but the second one will obviously sound louder. 

One way you could deal with the first track is to use a limiter that would 'catch' the loud section. But if that section is a great deal louder than the others, you risk distorting the one loud section. If the track is extremely dynamic, you may instead opt to automate the master fader to raise the soft sections, plus use a limiter to get a little extra dynamic range reduction.

As I have previously stated, there is no free lunch. Reducing the dynamic range of a track can make it sound louder but can also introduce unwanted artifacts. 

My "mantra" on this subject is that a loud mix starts at the arrangement stage, not at the mastering stage. By choosing sounds and arranging them to sound as loud as possible at the beginning of the project (possibly using light compression on individual tracks), and then mixing with loudness in mind (riding levels, light bus compression on sub-mixes, light mix compression, etc.), you will not need to do much mastering limiting and your final product will not sound as "compressed" or processed.