Two weeks ago I had an opportunity to sit down with Rafael Duyos and test his DSpatial software.

It is still quite difficult for me to explain myself the whole experience, as there was so much information in just one day - it definitely requires more thinking about it than I had available so far.

A brief description would be - DSpatial is a mixing engine that treats sound clips as objects, and it can be configured for any number of speakers you want (or have available in your studio). The number of these objects is limited by the number of channels your DAW can play back.
So far, so good, but the interesting part begins now.
These objects are positioned inside a virtual space that has 3 zones: the first one is inside the speakers' zone which is pretty much self-explaining - it can be your mixing room and the limits are defined by the speakers position.
The second zone is outside the speakers - you can move your object(s) past the limits of your room!
The third zone is the faraway zone, where sounds fade and disappear.
This being said, we come to the next level of complexity: a powerful, multi-channel reverb engine represents the space you're working in - it can be any space you want, interior or exterior small or large, and it is constantly present on all speakers, fronts, surrounds, tops, whatever you have in your studio. The great thing about this reverb is that it is so natural that you never hear it as if it is coming out of all speakers at the same time, just like natural spaces, where reflexions come from all directions.

In classical mixing process we add space to objects - in this DSpatial's one, we actually mix sounds which are inside the space all the time.

We had a multi-touch screen to control the objects - you touch/grab an object or a group of objects and move or rotate them on the screen where you can see the room, the outside space and the fade zone.
Here comes another layer of interesting things: moving objects react to movement as any moving objects in the real world - there is a Doppler effect that depends on the speed, when further away they are more dipped into reverb /it is really hard to call it reverb, I think "space" would be a better name), when they come closer, they sound dry. Another nice touch: when objects are closer, their frequency range is full, when further away, a HF and a LF rolloff is automatically applied in a very natural way.
Every movement is fully automated and reverb/space changes are frame-accurate. There is a whole Reverb designer section where you can tweak the reverb parameters.
In my studio, we ran it on top of Atmos setup, and assigned each virtual speaker output to physical speaker outputs, 32 in my case. Inside the plugin, there is a Room designer where you can fine tune the speaker position according to your room layout.

The plug in was running in Pro Tools (I'm not sure it exists for other platforms), on a very modest machine - New Mac Pro, 6core with only 12GB ram. Pro Tools HD Native Thunderbolt v11 or 12 I can't remember, HDMadi and HDsync.
With inserts on about 40 channels, the processor usage bar was steadily at 40%.

I am sure I forgot many details, but I'll try to bring them in my next posts.
DSpatial has a website, but the info is limited.

so can you do panning across X, Y, and Z planes??? this looks awesome!

Yes, fully 3-dimensional. Like Atmos panner, but with a lot more functions, all of them interactive.
One of the things I don't like in Atmos is non-existent objects grouping, like when you have a sound that's built of several layers and you need to pan them together. In DSpatial you can group any number of objects, move them through the sequence them, then ungroup, and create a different group...

I don't quite understand, is it like an add-on function of atmos? So the studio should have Atmos setup for running this?When the mixing is finished with this system, the mixing should be played back through Atmos setup?

No, it is not an add-on for Atmos. You can work with it in any environment. However, as my studio is equipped with Atmos and the easiest way of accessing speakers is through Dolby RMU, that was how I set it up for this test.
Also, it seems to me it is most practical to use it as an Atmos (or Auro 3D, or whatever may come) add-on, as that would guarantee correct rendering on playback in any Atmos-equipped room.

Reminds me of the Iosono surround panner, except 3d.

Looks similar? Or behaves similarly?

Behavior. You can pan "outside" of your space, and it also allows for automatic filtering / level attenuation based on distance, if I recall correctly.

Makes sense. It (Iosono) is based on wave field synthesis, so you can recreate any source position if you have sufficient loudspeaker "density".
I don't know if DSpatial uses WFS technology, but the definition increases if the number of speakers is greater.

I got a full day demo at "The Dub Stage" today with Rafael from "dspatial".
Really liked what I saw and especially heard using it. It's well thought out and I am looking forward to outfitting both my mix systems with it.
I see lot's of potential with it.

There are / were several processors that allow panning outside speakers' boundaries. However, this software (DSpatial) is unique for other reason - mixing inside the reverb, as I mentioned it in my previous post. The interaction of objects and the space they're placed in is stunning. Also the interaction of objects and the space around the listener is something impossible to achieve by simply adding reverb to sounds.
In this one, you can design your desired acoustical space, and then place and mix sounds inside or outside it.
In my first post, I forgot to mention the possibility of modifying the created space by building a wall around it. You can decide the size of the wall, so it will or won't circle the listener completely. You can add a door and make it as big as you want it to be. You can add furniture to your space and that will change the decay behaviour of the reverb, change materials your walls are made of, and so on...

How is this played back in the theatres? Is this yet another new format or is this rendered to the "common" formats (Atmos, Auro)?

I think it would work best "sitting" on top of already established formats - rendered to Atmos or Auro3D.
What we did during the demo session was to assign an Atmos object to each of the 32 outputs the DSpatial system was configured for. By doing so, you can keep the same position of outputs during the playback in any other Atmos-equipped room.
It is also worth mentioning that Rafael has a proposal for "Universal" speaker layout that could accommodate both Atmos and Auro and/or combine these two into a new, hybrid system.

I am curious about this software specifically for creating headphone-targeted binaural mixes for virtual surround. Has anyone here tried this software on a two speaker system and compared it against the usual stereo mixing "tricks"? I am intrigued...

Also, I'd like to try it for 5.1. Any new thoughts since it gas been a year since this thread landed?

Hey Branko and Marti,

Thanks for the reviews.

Do you think the multitouch is a must (I notice that you purchase the monitors separately)?

Do you know if this can be run on an existing mix system (i.e. As a plugin on the fx mixers side) or would you really consider this an additional mix system due to hardware requirements?

Thanks,
Matt

Hello Matt,
Multi-touch is not a must, but certain movements (like rotation) really benefit from it. I operate it with a mouse, which is sometimes very limiting, but I have limited space on my console and can't find space for two stacked monitors. It is not easy to get used to work with DSpatial, and I'd say the learning curve is a bit steep, but once you master the basics, it becomes real fun. Mastering the reverb settings takes time.
In my studio, I run it on top of existing ProTools FX session, on a separate machine, and I treat it as a plug-in (which it really is). In that session, there are many channels that aren't being processed through DSpatial. I admit I haven't used it for a great number of channels, only for those that need convincing panning, with natural doppler, reverberation and frequency shaping. I'm running it on a 6-core trashcan Mac Pro with 32 GB ram, and it works flawlessly.
hope this helps

Binaural for loudspeakers is called transaural. It doesnt work terribly well because of crosstalk. SPAT is meant to have the best system. I have never heard any magic tricks that really worked with two speakers - some interesting accidents though.