Single or multiple reverb instances in a mix?

[andyjh]

 I have always mixed on the basis that reverb is a room (be it concert hall or studio) simulator, and you can only be in one place in one room at any time, so I have always made a point of only using one reverb instances and everything routed to this. Different depths of reverb giving the depth of space. Now with convolution reverbs, it seems now that you need many instances of a reverb and then mixing these multiple reverb instances together, my concern here is that phase cancelling and issues will result from using more than one reverb instance, or do we now have to look at the reverb belonging to the instrument rather than the listening position?

If two different reverb types are used, then that's got to be wrong, as you cannot be in two rooms at the same time, but maybe two different settings of the same convolution reverb would be OK, so is a reverb instance for each instrument the ultimate set up? Current processing power may require "per section" rather than "per instrument".

I've been mixing "one reverb" approach for 30 years now, so reluctant to change my approach, but maybe technology is allowing this now, so which approach is the best nowadays?

[Dietz]

Whew ... that's food for a book rather than a forum message ... 8-)

In principle, you're right: A typical orchestra should be located in one room at any time (if you're aiming for absolute realism, that is ... most pop- or rock mixes _live_ from breaking this paradigm - but that's a different topic).

But  "one room" doesn't necessarily mean "one reverb" (in our understanding of the word) - quite on the contrary! In fact one single room will sound totally different  depending on the sound source itself, its radiation pattern, its pitch, its position, and of course depending on the listener's position, and (in case of a recording) of the polarity pattern of the microphone(s). IOW - yes, every source will create its own unique reflection- and reverb-pattern. Using just one single reverb signature for all instruments of the whole orchestra is not unlike sampling the middle C of a Grand Piano in one velocity, stretching it across 88 keys and trying to play Chopin etudes with it. 8-)

As a matter of fact all of these thoughts (and some more) led us to the development of MIR - our Multi Impulse Response reverbation and mixing engine. Here, the reverb "belongs to the source" _and_ the listening position at the same time, just like in real life. Of course, phase cancellation and other interactions between the resulting individual reverb trails _will_ happen, but that's part of the game, as similar phenomenons occur in a real hall, too. Actually, these interactions are one of the reasons that the blend between the single instruments happens at all, which allows us to achieve that typical "orchestral" sound in the virtual world.

... that said, there has always been a blurry line between "pure realism" and "perceived realism": Many great orchestal performances from the best halls on this planet have been enhanced on behalf of additional reverb, delayed microphones, even modulation and so on ... so in the end, we come back to the archaic First Law of Sound-Engineering: "If it sounds right, it IS right". There is no "best" solution, just several valid ones. The "one reverb" approach has worked for you for so many years - do you feel you're missing anything nowadays ...? Or did it just work due to technical and/or financial restraints ...? Food for thought ... :-)

Kind regards,

[andyjh]

 Yes, it probably was financial/practical reasons for using a single reverb, as hardware reverbs cost thousands, so using many of these for a mix, was (at the time) unrealistic, but sample libaries never used to be so realstic that the reverb was an issue, now my VSL library is maybe showing up the room simulation (reverb) as the weakest link, for that final believability of realness.

So I thought I may have to  rethink my approach to room simulation, I thought my "one reverb" approach was OK, but I read so many times in this forum about using multiple reverb instances, I was questioning my own methods.

 It would be good to hear other users approach to this. 

[Beat Kaufmann]

@andyjh said:

... It would be good to hear other users approach to this.

Hello Andyjh

Take into account that listening to a CD (a mix) could be quite another thing than sitting in a live concert behind a pillar in the 18th row.

Over the years we are used to get CDs which are sounding as a CD. In other words: So mixing music can also be a possibility of creating

a spectacular virtual stage or concert hall. So the dimenson "far" and "close" can be one element of "try to get the best result with only two speakers".

As an example for this, listen to Demo 4 at http://www.beat-kaufmann.com/vitutorials/vienna-choir/index.php

The soloist sings close(r) to the listener (with reverb as well) while the choir sings behind him (in the depth).

I like to use the dimension depth. You can nicely bring out soloists or other important thing.

So using different reverbs means probably too: "using different depths" .

Another nice example with different depths (reverb instances) is this Sympohnic Rag

• Depth 1: Marimba, Solo Violin, Flute, Honk
  
• Depth 2: Strings
• Depth 3:Trumpet
• Depth 4: Xylophone, Horns

You asked for another approach. This is one.

It is another possibility as you have with MIR which trys to simulate and get a more live-concert situation.

I believe, that both have their advantages. It probably depends on the music and of course: It is always a matter of tast.

All the best

Beat

[William]

One thing almost no one ever talks about here is the main approach of recording/mixing that used to be done in nearly all recordings.  (Of course I realize that Dietz and others know about this but I am pontificating.)   

It was not creating an imitation of being in a concert hall at all.  It was turning your living room into a concert hall.   So there was none of this rather artificial approach of making strings close, winds a little farther back, percussion way in the back with ERs, etc. 

Percussion was very dry so it would be clear and crisp.  Strings were slathered with reverb so they would be big and espressivo and more reverb does that psychologically.  Brass were midway so that they were rich and powerful but not too muddy.  Basses were almost bone-dry so there was clarity audible beneath the mass of the other sounds on top of them.  Then, this sound was presented in your listening area not as an IMAGE of a concert hall you went to, but instead as if those instruments were actually playing in your room.  They would not have reverb of some other place, but only an idealized reverb to make them sound better psychologcally/musically within YOUR place.

An example of this applied to orchestral music in truly great recordings is the London FFFR LPs created in the 70s of Mahler's symphonies.  All of the solo instruments would be crystal clear, the percussion almost dry, the strings huge and reverberant.  This is exactly opposite to what people seem to assume you must do with samples in their attempts at reproducing some imagined concert hall.

Also, even with the approach of reproducing a concert hall -  I think that there is almost no audible difference in almost any concert hall between the sound of  a player sitting in the third row of an orchestra compared to one sitting in the second row or the back or wherever. You do NOT  hear the difference from the audience! Whatever difference in sound exists is incredibly tiny in a good concert hall  because your brain totally erases it.  Go to a symphony concert and see if I am wrong!  I was listening recently at several live orchestral concerts with this very thing in mind.  As hard as I tried I could detect ABSOLUTELY NO DIFFERENCE  in sound between trumpets sitting in one spot and flutes in another.   They all basically sounded like the concert hall's overall reverb.  To reproduce the sound of this experience - which I was trying very hard to discern in a real-world setting - you would simply run all  your dry tracks through ONE overall reverb.  And yet  people using samples are adding more wet or adjusting the convolution or whatever between such different imaginary placements, and it is totally artificial in this sense.  I would not be surprised if many or most of the detailed things that they are doing are utterly inaudible because the overall sound and the listener's perception obliterates the difference.  

[Beat Kaufmann]

@William said:

...Go to a symphony concert and see if I am wrong!  I was listening recently at several live orchestral concerts with this very thing in mind.  As hard as I tried I could detect ABSOLUTELY NO DIFFERENCE  in sound between trumpets sitting in one spot and flutes in another.   They all basically sounded like the concert hall's overall reverb.  To reproduce the sound of this experience - which I was trying very hard to discern in a real-world setting - you would simply run all  your dry tracks through ONE overall reverb...

Hello William

This far-away-listener-experience I had as well as I was sitting on the balcony in a concert here in switzerland. But in December 2010 I visited the Chrsitmas Concert in our town church. Because I went one hour bevor the concert for getting a good seat I got a place in the 2nd row!

3m away the 2 singers - behind them the orchestra and far away (20m?)  - up in the height - the choir on a rostrum. The singers have been so close and choir ...

So I had an exeperience with different depths.

Whatever the situation may be, one main aim of a proper mix should (could?) be that no instrument covers an other.

Therefore we can use the following 3 dimensions:

If you would use all of them - or just one - depends on the size of the ensemble, of your matter of taste or of...

Another approach could be that you listen to mixes which you like.

Try then to reach the same results, with all the tools we have today.

Maybe you only use one reverb - OK

Best

Beat

[William]

 No actually I use many - such as this mix I am doing now which is as many as each instrument.   HOwever, I do wonder about it myself.

[doubleattack]

@Beat Kaufmann said:

It is another possibility as you have with MIR which trys to simulate and get a more live-concert situation.

 

Hello Beat,

interesting topic and always interesting to read about your workflow. Your hints are often helpful. So thanks for that.

But: Since I'm using MIR I have to dissent about your assumed target of this tool.

I think even though MIR is using samples of real venues, it's not the goal to get a "live-concert situation" or even a "simulation of the reality". That's a big misapprehension in my experience and in my opinion. I will try it to explain why.

Even though I use it sometimes in this way for making mockups to give musicians an audible suggestion of the sheets on their musicstands, in fact - even though it sounds "believable" and in this way, how musician ears know it from rehearsels in a studio room. It sounds good, really incredible for a never done "performance", but for a music-production this mockup sounds quite boring. Even if the mockup would be made with real musicians, recorded in the same studio-room (since MIR imaginabel.) it would have the same boring effect - 'cause it's not produced.

And this producing makes the different.

For me it's a big difference between "make music" and "music-production". A half-thousand years music happens only in unity of space and time. Since the first recording of a musical piece these unity has been lost and a new kind of art (beside "music") was born - I call it here "music-production".

(This differentiation is important not only to avoiding fights endless comparing apple and pears.) 

I think working in the "art of music-productions" we are aiming for the liveliness and the spirit as we know and love it of making music (in the half-thousand years old traditionell way). That liveliness will be not done with real spaces alone, even not with real musicains!

To get an effect with music productions like we have with making music in real life -  n o t  getting  the  s a m e  effect (we have no unity of space and time here!) - we have to do a lot of unorthodoxly and not "reality-like" things to get an effect moving our hearts. That's why we take all the effort to get an analogical effect as we know it from the reality.

So it's often quite barren setting the instrument-icons in MIR in a "real life-position" thinking now it sounds good, 'cause I have real sampled spaces. (So we could think, these production will be good 'cause I have life musicians, made from flesh and blood ;-))

No, the real advantage of MIR is to have no trouble with all the settings of tons of reverbs and panning tools for different space-depths and "reality-like" reflections - the sampled rooms helps to make this inavoidable producing process to make very easy and simple. And so I have space using my mind to think and care about the depht in the other dimensions (your graphic is very good btw!).

My using of MIR if I make a "music-production" is far away of the "reality". Maybe the first violin sitting next to Beethoven (in the foyer) and the last doublebass upstairs on the wall (I think 50 m or 70 away?) - it's not live-concert-situation! (beside the "unnatural" trimed tail of the reverb). Setting the icons like live concert situation sounds good, maybe real, I don't know, but never mind, it sounds boring. Without the liveliness touching my mind and heart.

The visual "crazy" setting do it.

Maybe the lookout of MIR gives a wrong impression of how to use it. Of course in different ways, and sometimes for simple "simulations of reality". But thinking it's made for having a "faked" live-concert-situation is very abridged.

Yes, I'm in love with MIR!

Best

Frank       

  

[Beat Kaufmann]

@doubleattack said:

...Yes, I'm in love with MIR!

Hello Frank

Thanks for your detailed article. A wonderful and another approach for andijh - as he wished.

Keep this feeling as long as possible (being in love with MIR).[:)]

Beat

[doubleattack]

There is no end of the story in sight! [;)]

[Beat Kaufmann]

@andyjh said:

...I've been mixing "one reverb" approach for 30 years now, so reluctant to change my approach, but maybe technology is allowing this now, so which approach is the best nowadays?

Here is another approach in way.

The question is:  How can I mix Classic Rock? Orchestras with a Band?

Or in general: How can I keep the transparency in complex mixes?

Go to http://www.beat-kaufmann.com/vitutorials/vi-tips--tricks-3/index.php and

click on "27."  Find a transparent arrangement for getting a good mix

Read the text and watch the Image-Show

BTW: This approach is for MIR-users as well [;)]

Now have fun

Beat Kaufmann

PS. This image-show is maybe not new for users who often visit my homepage. Sorry!

[jasensmith]

@andyjh said:

 I have always mixed on the basis that reverb is a room (be it concert hall or studio) simulator, and you can only be in one place in one room at any time, so I have always made a point of only using one reverb instances and everything routed to this. Different depths of reverb giving the depth of space. Now with convolution reverbs, it seems now that you need many instances of a reverb and then mixing these multiple reverb instances together, my concern here is that phase cancelling and issues will result from using more than one reverb instance, or do we now have to look at the reverb belonging to the instrument rather than the listening position?

If two different reverb types are used, then that's got to be wrong, as you cannot be in two rooms at the same time, but maybe two different settings of the same convolution reverb would be OK, so is a reverb instance for each instrument the ultimate set up? Current processing power may require "per section" rather than "per instrument".

I've been mixing "one reverb" approach for 30 years now, so reluctant to change my approach, but maybe technology is allowing this now, so which approach is the best nowadays?

 

I do both for my orchestral works.  I set up each section with their own reverb impulse, basically like Beat points out, then I use one overall reverb to sort of meld everything together.  Generally, the overall reverb is the same impulse as the individual sections.  I think it works pretty well for somebody who doesn't have MIR like me.  The nice thing about these convolution reverbs (converbs) is that you have so much control over their characteristic sounds giving them their own personality if you will.  You can control things like high pass, low pass, EQ, dry signal, wet signal, you can reverse the signal which probably wouldn't be useful for orchestrated works but at least we have options now that we didn't have 30 years ago. 

[andyjh]

 Well, it's beginning to sound like the old saying "if it sounds right, it is right", but I still wonder, that if when using multiple instances of reverb, how each reverb is set. I assume that each reverb is of the same type? but is the decay time set differently for each reverb instance? and the same HF decay?.

I am about to start a new project, so I'm going to give multiple instances a go, I just thought I could learn from others experience of what doesn't work - I am aiming for a concert hall sound (big orchestra)

[Stephen W. Beatty]

I have the impression that MIR has made using multiple reverbs  moot. Why not just change the volume or the wet-dry balance if you don't like the natural sound of the instrument. In the extreme case you can always add a different convolution reverb to each dry instrument if you like. I tend to live by the "KISS" or keep it simple sam.

Regards,

Stephen W. Beatty 

[William]

Well MIR is many, many reverbs - all of them done to represent the space in a unified way from different points.  However you can use it as if it were one, with different levels of wet/dry coming out of each instrument.  I decided on the mix I am currently doing to use the same settings for wet/dry I had previously used with hardware reverb which didn't do any of the realistic placement of MIR.  But it can fucntion as an indicator of how big the sound should be psychologically. The wet/dry becomes with MIR a matter of taste though there must be some of both.  I do notice the wet/.dry with MIR does NOT have the simple effect Beat Kaufman is talking about with making something in back wetter, something in front drier, etc. Because of the nature of the convolution it is much more complicated and something that is drier can easily sound farther away depending on the placement/space. This ends up being a huge advantage because you can use the dry side of the ratio to make something that is actually sounding far away clearer than it would be with a wetter setting.