What's new

SSO mics and mixing

OP
Virtual Virgin

Virtual Virgin

Active Member
You also realize that the tree mic configuration is composed of 3 separate mics, and that's merged down to stereo in the samples, so you're already dealing with crosstalk and time delay and mic overlap in the tree recordings?
Yes, and those do present more issues for realism.

What I am looking for is the conductor's perspective; having the closest thing to a one-point perspective as the starting point in an orchestration template. I am willing to use any device with any level of "artificial" manipulation as is required by the material or at the behest of the narrative into the beyond as each work develops, but I desire a clean slate at the onset.

Honestly, what I would likely prefer is binaural recordings from the podium with my own custom HRTF. I don't think Spitfire offers that yet :P
 

Joël Dollié

New Member
Yes, and those do present more issues for realism.

What I am looking for is the conductor's perspective; having the closest thing to a one-point perspective as the starting point in an orchestration template. I am willing to use any device with any level of "artificial" manipulation as is required by the material or at the behest of the narrative into the beyond as each work develops, but I desire a clean slate at the onset.

Honestly, what I would likely prefer is binaural recordings from the podium with my own custom HRTF. I don't think Spitfire offers that yet :P
One single binaural head in the middle would sound pretty terrible. The overall tone of some instruments would be dreadful.

It wouldn't work very well on speakers, and create all kinds of issues.
 

Paul Cardon

Ninja Otter Music
Yes, and those do present more issues for realism.

What I am looking for is the conductor's perspective; having the closest thing to a one-point perspective as the starting point in an orchestration template. I am willing to use any device with any level of "artificial" manipulation as is required by the material or at the behest of the narrative into the beyond as each work develops, but I desire a clean slate at the onset.

Honestly, what I would likely prefer is binaural recordings from the podium with my own custom HRTF. I don't think Spitfire offers that yet :P
I'm getting into the weeds here, but you're very likely never going to be able to do this with almost any sample libraries available, at least with any degree of realism. There just aren't any decent tools for properly simulating and emulating realistic binaural audio from multi-mic'd audio sources, especially stereo sources. There are just too many factors.

But honestly, if you want to find some better forums where people will have ideas for this, you should try joining some Facebook groups on binaural audio mixing and 360 audio. There's some good groups you can probably do searches in, but again don't expect much success with the audio sources you're going about this with. Practically all sample libraries are not recorded for anything but standard stereo mixing practices.
 
Last edited:

Jdiggity1

Stroking The Frog
Moderator
I've been working on a template recently using the SSO and Joby Burgess percussion.

I'm having a dilemma though:
I quite like the the distance imaging for the tree mics, but I find the width of the stereo imaging to be much too thin. Almost the entire orchestra is very close to center.

I don't want phantom images, so I'm not into mixing close mics in a position other than the in situ locations.

Has anyone had success with widening the tree image without damaging the sound of the room?
It seem just panning will yield distortions and smudge the space around, damaging the unity of the image.
If you're not wanting any processing or plugins active to begin with, your best bet is just to pan the kontakt instrument. It's not doing any overlapping or smearing of the stereo image as it works by simply lowering the volume of the other side.
For the violins, keep the tree mic active and try panning 15-20 L. All it's doing is lowering the volume of the right signal a little bit, which is causing the problems with the imaging. The violins (and other instruments) ARE in fact recorded quite wide, but the reverb and reflections coming from the other side are so strong that it creates this centered image.

If you don't want to pan anything, it might be worth exploring Mid-Side processing to scoop out some of the center, leaving the sides. But I still think the problem with SSO is the strength of the opposite side signal, not the center.
 
OP
Virtual Virgin

Virtual Virgin

Active Member
One single binaural head in the middle would sound pretty terrible. The overall tone of some instruments would be dreadful.

It wouldn't work very well on speakers, and create all kinds of issues.
On the former point, I beg to differ.
On the latter, of course. Binaural is meant for headphones.
 
OP
Virtual Virgin

Virtual Virgin

Active Member
I'm getting into the weeds here, but you're very likely never going to be able to do this with almost any sample libraries available, at least with any degree of realism. There just aren't any decent tools for properly simulating and emulating realistic binaural audio from multi-mic'd audio sources, especially stereo sources. There are just too many factors.

But honestly, if you want to find some better forums where people will have ideas for this, you should try joining some Facebook groups on binaural audio mixing and 360 audio. There's some good groups you can probably do searches in, but again don't expect much success with the audio sources you're going about this with. Practically all sample libraries are not recorded for anything but standard stereo mixing practices.
That's why I included the :P at the end of that statement, making light of my expectations.

I've already suggested 2nd order or higher Ambisonics to Spitfire. At some point libraries will be adding those to the mix.
 

Paul Cardon

Ninja Otter Music
On the former point, I beg to differ.
On the latter, of course. Binaural is meant for headphones.
The conductor's location is not a great point of balance for an orchestra. There's a reason why audiences are always out in front, not inside, the orchestra.

That's why I included the :P at the end of that statement, making light of my expectations.

I've already suggested 2nd order or higher Ambisonics to Spitfire. At some point libraries will be adding those to the mix.
Have you worked with ambisonic orchestral recordings in a mix before? They're not as independently awesome as you'd think because most 360 mics don't go above 2nd order so your direction accuracy is pretty limited, and low order ambisonic recordings in binaural spatially smear like CRAZY. For single sound sources, you lose a lot of high-end directional accuracy, but in a reverberant scoring stage/hall with 80 players, it's just a blur. You're not going to be able to pinpoint the violins with that 2nd order recording, only the suggestion of their direction as smeared through an HRTR.

AFAIK, most ambisonic orchestra mixes use a diverse combination of ambisonic mics, stereo pairs, and spot mics to build up a more pleasing and higher order mix. The recorded 360 mics are more often just used as a base room recording.
 

Paul Cardon

Ninja Otter Music
I've been working on a template recently using the SSO and Joby Burgess percussion.

I'm having a dilemma though:
I quite like the the distance imaging for the tree mics, but I find the width of the stereo imaging to be much too thin. Almost the entire orchestra is very close to center.

I don't want phantom images, so I'm not into mixing close mics in a position other than the in situ locations.

Has anyone had success with widening the tree image without damaging the sound of the room?
It seem just panning will yield distortions and smudge the space around, damaging the unity of the image.
To get back to why we're here, do take Joel's recommendation for 2C Audio's Precedence. It does some fun things to simulate stereo mic crosstalk panning (mic frequency patterns, time delays, gain differences) that might let you retain the impression of a room and it's spatial characteristics while still "panning" the bulk of the sound whichever way you like.
 
OP
Virtual Virgin

Virtual Virgin

Active Member
If you're not wanting any processing or plugins active to begin with, your best bet is just to pan the kontakt instrument. It's not doing any overlapping or smearing of the stereo image as it works by simply lowering the volume of the other side.
For the violins, keep the tree mic active and try panning 15-20 L. All it's doing is lowering the volume of the right signal a little bit, which is causing the problems with the imaging. The violins (and other instruments) ARE in fact recorded quite wide, but the reverb and reflections coming from the other side are so strong that it creates this centered image.

If you don't want to pan anything, it might be worth exploring Mid-Side processing to scoop out some of the center, leaving the sides. But I still think the problem with SSO is the strength of the opposite side signal, not the center.
I'm thinking it's less about the reflections and more about the fact that all three mics are omni and so close to center. This means the violins are guaranteed to be almost as loud in the far mic. There is not much decibel reduction that takes place in span of those mics when the direct source is travelling the distance.

What I'm reading here is that the Decca Tree was devised as a supplement to fill the center.

I will try you panning idea, as that may be the closest thing to capturing the depth and stereo placement that I am looking for.

The ambient mics do have some more side imaging, particularly for the Joby Burgess percussion, but the depth is too far out.

What I am starting to realize here is that I am finding the SSO to be somewhat disappointing in it's prized feature: in situ recording. I am on board with having a library with a baked-in room, and I quite like the depth, but the stereo imaging with the included mic arrays does not capture the advantages of in situ recording.
 
OP
Virtual Virgin

Virtual Virgin

Active Member
The conductor's location is not a great point of balance for an orchestra. There's a reason why audiences are always out in front, not inside, the orchestra.
LOL. Not this gem again. It's patently obvious that an orchestra itself can barely fit on the stage, let alone the addition of an audience! Your point is not made using a physical impossibility for the alternative hypothesis.

Have you worked with ambisonic orchestral recordings in a mix before? They're not as independently awesome as you'd think because most 360 mics don't go above 2nd order so your direction accuracy is pretty limited, and low order ambisonic recordings in binaural spatially smear like CRAZY.
Did you miss where I said 2nd order or higher?

AFAIK, most ambisonic orchestra mixes use a diverse combination of ambisonic mics, stereo pairs, and spot mics to build up a more pleasing and higher order mix. The recorded 360 mics are more often just used as a base room recording.
I am suggesting high order ambisonics as a supplement, not a replacement. Libraries seem to be going the way of additional mic options and companies like Toontrack have added height information on top of surround.

Anyway, getting off track as ambisonics has its own set of troubles for realistic capture and rendering.

Looking into that 2CAudio plugin... give me some time to see what it does.
 

Paul Cardon

Ninja Otter Music
I guess I'm still confused at what you're trying to do. You're suggesting it as a supplement, but I thought above you were saying you won't mix mics because of "phantom" images?

LOL. Not this gem again. It's patently obvious that an orchestra itself can barely fit on the stage, let alone the addition of an audience! Your point is not made using a physical impossibility for the alternative hypothesis.
Is this... logic debate class? I'm not trying to construct a perfectly sound argument, just trying to lend more thoughts to consider. Of course that's not possible, but recordings give us the ability to put mics wherever we want and create our own spatial realities so anything is possible.
But even decca trees, one of the primary means of capturing an orchestra, are a few feet or more above the conductors head so they capture a better broad perspective of the orchestra. At the conductors position, the balance turns much more string heavy, bodies start blocking sounds more, etc. It's not entirely ideal.

I mean orchestras are often recorded and mixed in very flexible and unrealistic ways, and if not, they often just stick to a decca tree. So it's NOT a bad thing to instead mix towards the impression of a good orchestra, of something that's pleasing. In fact, that's why orchestras are recorded the ways that they are today. I guess I'm really not sure what you're doing or why you're trying to do it.

Did you miss where I said 2nd order or higher?

...

I am suggesting high order ambisonics as a supplement, not a replacement. Libraries seem to be going the way of additional mic options and companies like Toontrack have added height information on top of surround.
But there aren't any ambisonic microphones that record above 2nd order that aren't plagued by imaging and self-noise issues, so the low order issues remain.

Looking into that 2CAudio plugin... give me some time to see what it does.
Definitely try the demo, though I'm concerned it might not do what you're looking for it to do if you're going for ultimate realism. It's a plugin that can give the impression of direction in a way that standard panning can't accomplish, so worth a shot!

What I am starting to realize here is that I am finding the SSO to be somewhat disappointing in it's prized feature: in situ recording. I am on board with having a library with a baked-in room, and I quite like the depth, but the stereo imaging with the included mic arrays does not capture the advantages of in situ recording.
The omnis in decca trees are often fitted with balls around the capsules on the end that add a bit of directionality to the signal in the top end. Then mic directions are splayed across the width of the orchestra. Again, this is one of the most common ways to capture an orchestra as a body. You've likely heard a decca tree in most orchestra recordings.

Is it possible you're listening to this stuff on headphones that have an exaggerated bass/low-mid response, because the upper end content in SSO's trees definitely points towards the left, but because Lyndhurst hall is so active and reverberant, the violins bounce across the space and their dulled reflections come through on the right as well, bringing the image of the mids more center but with a time delay and blur. So without the high end and ignoring time delay cues, the averaged mids and below image may steer towards the center (albeit a very wide "center")
 
Last edited:
OP
Virtual Virgin

Virtual Virgin

Active Member
I've mentioned pretty clearly what the objective is: make an orchestral template that has a high level of realism from a one point perspective, as close being in the room as possible. I want fidelity of localization of the in situ playing positions with as little artifice as possible.

Why is this thread full of resistance to this objective?

I think too many of you are stuck on the purpose of the template itself.

Yes, I understand that classical recordings often use many microphones from many positions.
Yes, I understand that film music uses this even more and blends various techniques from pop, rock, electronic music etc...

AND, as I stated very clearly I have no qualms about using any type of these combinations in the final product all the way down to bit crushing everything, putting it into a reverse Radioshack reverb blender, bouncing it to tape, microwaving the tape then play it back from an iPhone sitting behind a fan and recording that with a wire microphone. Whatever works.

The purpose of the objective of this template however is not that. It is for a specific stage in a creative process. It is not meant to be the last stage of a film trailer production.
 

Paul Cardon

Ninja Otter Music
I've mentioned pretty clearly what the objective is: make an orchestral template that has a high level of realism from a one point perspective, as close being in the room as possible. I want fidelity of localization of the in situ playing positions with as little artifice as possible.

Why is this thread full of resistance to this objective?

I think too many of you are stuck on the purpose of the template itself.

Yes, I understand that classical recordings often use many microphones from many positions.
Yes, I understand that film music uses this even more and blends various techniques from pop, rock, electronic music etc...

AND, as I stated very clearly I have no qualms about using any type of these combinations in the final product all the way down to bit crushing everything, putting it into a reverse Radioshack reverb blender, bouncing it to tape, microwaving the tape then play it back from an iPhone sitting behind a fan and recording that with a wire microphone. Whatever works.

The purpose of the objective of this template however is not that. It is for a specific stage in a creative process. It is not meant to be the last stage of a film trailer production.
I was curious about your goals with the first post and I’m digging a bit to see if there’s other strategies to uncover. If it seems like I’m rubbing against it, it’s because I have a hunch this won’t be easily doable with most or any samples libraries and you’ll have to use some fun tricks to imitate it, such as using close mics and virtualizing, or manipulating stereo mics in some fun and funky ways to ultimately give the impression of listening to an orchestra from a single realistic perspective. That’s why Joel and I recommended that Precedence plugin.
 
OP
Virtual Virgin

Virtual Virgin

Active Member
I was curious about your goals with the first post and I’m digging a bit to see if there’s other strategies to uncover. If it seems like I’m rubbing against it, it’s because I have a hunch this won’t be easily doable with most or any samples libraries and you’ll have to use some fun tricks to imitate it, such as using close mics and virtualizing, or manipulating stereo mics in some fun and funky ways to ultimately give the impression of listening to an orchestra from a single realistic perspective. That’s why Joel and I recommended that Precedence plugin.
I'm using MIR PRO, but that is for other libraries and not for this particular objective.
My problem with Spitfire is the localization and stereo width.
Cinesamples is even worse for a number of reasons. Even with only one mic mix on (say the room or ambient) there is a lot imaging modulation. Many times the release samples don't match well enough and adjacent keys can make an oboe jump a few meters around the room.
 
Top Bottom