Orchestral Placement

[Simon Godden]

I'm not sure if a lot of people will agree with the chart that I am
presenting here, because it's very basic.  What I've done is set a few
panning levels to all the instruments of the orchestra, and given an
idea on how to create depth with the equipment that you have.

Now then, we all know that it is not just the amount of reverb that
gives a sample it's depth.  There's a whole load of other things to
consider, if you want to create the impression that some samples are
nearer to the listener than others.  For example, early reflections. 
Early reflections are the first evidence of reverb (I'm using the word
'reverb' in its loose sense here) and it usually first affects
instruments at the back of the orchestra because they are nearer the
back walls, whereas any voice that sits in the middle of the orchestra
has to pass through a series of heads, music stands, chairs, other
instruments etc, so their early reflections are not as noticeable. 
There is also such a thing called pre-delay.  Pre-delay is the amount of
time it takes before there is any evidence of a reverberation (early
reflections).  Therefore, musicians at the front of the orchestra will
have a longer pre-delay than those at the back because the actual
soundwave that issues from the instrument will reach the listener's ear
uninterrupted by reverberation for a longer period of time.  Then, there
is the release trails (or maybe it's tails) themselves (the ACTUAL
reverb).  The trails are the amount of time it takes for the echo to
stop AFTER the actual noise from the instrument has ceased to sound. 
And finally after all that, there is something about high frequencies or
something, but by the time I get to that bit, my brain has started to
dissolve into the same sort of compound one finds in a trifle.

Now all this is extremely complicated and it takes a magician to work it
out so accurately that a full mix will sound in three dimensions
exactly as if the orchestra are correctly placed.  Some of us (and I'm
not one of them) can afford software that will provide the algorithms
necessary for you to create such depth in your music.  All you are
required to do is 'draw' in where each of the instrumentalists are
situated on a nice convenient 3D diagram on your monitor and hey presto,
instant orchestral depth.  The software (called "Altiverb") costs about
a thousand dollars (I'm not exaggerating, it really is about a thousand
dollars).  It also has recorded the impulse responses of certain famous
auditoriums like the Sydney Opera House for instance, and the
Concertbouw in Holland, so maybe it is actually worth a thousand
dollars.

So, how does one try and create a bit of depth to their orchestral music
without either spending the equivalent of a reliable second hand car or
growing a bushy grey moustache and messy grey hair to match, and
standing at a blackboard with a piece of chalk and a German accent. 
Easy!  All you do is just make the ACTUAL reverb time a little longer
the further you go back.  It isn't by any means perfect, but at least it
will prevent your music from sounding like all the musicians are
standing in a bloody, great long line.  Also, as far as panning is
concerned, you should only really extend each side to about halfway
across the actual maximum parameter.  Otherwise it will sound as if you
are standing in the middle of the orchestra, and some instruments will
actually sound as if they are sitting behind you.

So, rather than you having to work it all out, I have taken time out of
my composition/notation activities to draw a nice easy chart from you to
work on.  I can't post it here because Ning are behaving rather
illogically at the moment, so I've attached the file to the reply that
I'm going to write after this long-winded discussion.

I therefore hope most sincerely that it works for you as well as it works for me.

Cheers,

Simon

[Simon Godden]

....and here is the chart as promised.

Good luck now. If you have any questions, please do not hesitate to ask.

[James Semple]

Hi Simon,

Nice article and a good subject to discuss.

Unfortunately, even Altiverb doesn't work quite as easily as you might imagine. Most people using it don't use the stage placement feature. Honestly there still isn't an easy way to do this that doesn't require a fair amount of fiddling around with levels, reverb settings and EQ. I haven't had a chance to check out MIR from VSL yet but that promises to make life incredibly easy for people trying to do this.

Also while some of the ERs are coming from the back wall (which would probably give the strings a pre-delay of around 40-50ms) we are also getting early reflections from both the side walls and the floor which arrive a lot earlier. So what does this mean? Well either you can use two different ERs for the strings or you can just do it by ear which I think it still what most of us do.

At the moment algorithmic reverb (I assume meaning 'with the properties of Al Gore') is coming back into popularity again and convolutions seem to be getting less popular.

... and while Altiverb may seem a tad expensive, remember that a Bricasti M7 is around 3 times that price.

[Simon Godden]

Ray...... If you had read more than a couple of lines, you would have seen that I have not put myself up as an expert. In fact, I have admitted that I find the whole concept of reverberation placement very complicated, and tried to explain in layman's terms the limited knowledge that I have gained from various sources. James has kindly corrected me on a few points, and Jesus, that Bricasti thing IS VERY expensive.

Ray Kemp said:

I've only read a couple of lines here so I'll be brief.

Go learn a lot more about this before putting yourself up as an expert.

[Simon Godden]

Thanks James,

I just looked up the Bricasti M7, and I was a bit gobsmacked. I would have guessed that the right way to do business is to make it easier and more user-friendly for the client, but the Bricasti looks a bit complicated as well.

I don't know why convolution reverb is going out of fashion (are you talking about the natural convolution reverb that EWQLSO provide, or the recorded sine waves (IRs) from various auditoriums?) because I think that they sound better. I have the demo version of Wizoo (unfortunately, because it's a demo, I can only get one instance from it which is great for master settings, but when using different sections of the orchestra I have to use individual reverb settings provided on the IK sampler. But the convolution settings that they've taken from actual spaces sound better than the algorithmic settings. I also have the VST plug-in of IK Multimedia's Classik Studio Reverb (hall interface, that was provided with Notion 3), but I don't know how to use that yet. I haven't been too impressed with it so far.

It's absolutely true what you say about different ERs. It seems that there are so many things to consider when trying to tweak your way into successfully emulating a realistic orchestral space. That's why I thought I'd try and invent a simple way of trying it out.

I think I'll have another go at trying to understand the CS Reverb from IK Multimedia. After all, it's in my system and it's supposedly won awards so I might get something out of it.

Cheers,

Simon

James Semple said:

Hi Simon,
Nice article and a good subject to discuss.
Unfortunately, even Altiverb doesn't work quite as easily as you might imagine. Most people using it don't use the stage placement feature. Honestly there still isn't an easy way to do this that doesn't require a fair amount of fiddling around with levels, reverb settings and EQ. I haven't had a chance to check out MIR from VSL yet but that promises to make life incredibly easy for people trying to do this.

Also while some of the ERs are coming from the back wall (which would probably give the strings a pre-delay of around 40-50ms) we are also getting early reflections from both the side walls and the floor which arrive a lot earlier. So what does this mean? Well either you can use two different ERs for the strings or you can just do it by ear which I think it still what most of us do.

At the moment algorithmic reverb (I assume meaning 'with the properties of Al Gore') is coming back into popularity again and convolutions seem to be getting less popular.

... and while Altiverb may seem a tad expensive, remember that a Bricasti M7 is around 3 times that price.

[Simon Godden]

I think reverb is pretty much a black art to everyone. I hope it works for you.

Newport Hutchinson said:

Just out of interest, there is a set of Britcasti M7 impulses available (for free! - although you can donate to the developer if you wish) from -
http://www.samplicity.com/bricasti-m7-impulse-responses/
I've been playing around with them for a short time in conjunction with Reverberate, and they sound pretty good. Simon - Thanks for putting the chart up - Reverb to me seems to be a black art, and any extra info is always welcome!

[Simon Godden]

Yeah sure Phillip. Have a listen to this. It's the finale of my symphony. But remember, these samples were dry to start with and they're not top of the range samples by a long shot (Miroslav and HALion).

I hope you can hear what I'm trying to do.

Cheers,

Simon

Phillip Park said:

Hello Simon,
Very interesting concept you have laid out. Do you have any produced examples of this concept in practice? Though making sample-based performance slightly more realistic, I can't imagine this making any more than a negligible difference in the performance. Definitely a solid attempt at reaching towards a live performance.

[Simon Godden]

Hold on, Phillip

I forgot to attach the .mp3

[Simon Godden]

The .mp 3 doesn't seem to want to stick to this discussion. Hold on, I'll try something else.

Simon Godden said:

Hold on, Phillip

I forgot to attach the .mp3

[Simon Godden]

Right! I've had to compress the file to 128bps, so it might not be as clear.

[Ann Rodela]

Darn! My internet is too spotty to enjoy it the sound comes in spurts, but what I can hear is very, very good quality. I will come back later. :)

[Simon Godden]

That's because you're used to EWQLSO and VSL. I'm trying to make the best of what I have, but thanks for listening.

Ray Kemp said:

Simon,

It sounds absolutely dreadful to my ears so if I'm in the minority as in past examples I have proved to be, then let me say...... I give in right here right now. I'll transport myself back into the real world out of this cyberspace hell.

[phil Kelly]

Uhhh Simon ..

Where's the chart ?

Also, I'm assuming you're basically discussing creating the stereo illusion of "depth" using samples in a DAW environment ( Which I'm less familiar with than an acoustic environment ) In the Mezezoic era ( when live fiddle players ate their lunches out of paper bags and the brass players would head for the pub on breaks ), Our basic tools were:

1. Several sets of stereo echo chambers to isolate certain subgroups of the orchestra
for separate processing, EQ, and delay times.

2. Multiple sets of delay ( and pre-delay ) lines to enable the use of crossed returns on certain sounds.

Also, as opposed to a typical live orchestral image, we might tend to pan the higher pitched sounds toward L or R and leaving the bulk of the low sounds in the center.

How all this relates to your discussion here regarding these same issues in the world of samples , I know not, but I DO know that Chris Alpiar did quite a good job on some orchestral projects for me earlier this year, so maybe he can share some of his tricks with y'all!