VBAP panning in ardour

Does anyone have any information on the VBAP feature in ardour. I tried a demo of Dolby atmos and I do like this new format. I remember ardour always had this feature called VBAP panning. Can it render 3D sound to a stereo file etc

Atmos is a commercialization of license-free technology called Ambisonics, which goes back to the 1970s.

Google and you’ll find articles and stuff on Ardour and ambisonics - we were one of the earlier DAWs to support it.

Ambisonics describes the placement of the sound in 3d space, and then renders that to whatever actual output configuration you are using, including stereo.

VBAP cannot do this (certainly not within ardour).

I came across some binaural plugin that allows panning in a 3D kind of way. That renders to stereo as it’s a plugin. I’m wondering if other methods other then Dolby will be used. Since it’s expensive get started unless people already have software that supports it.

Sorry, but Atmos has absolutely nothing to do with ambisonics: atmos is an object-based spatial representation system which can associate mono/stereo “obects” with spatial coordinates and more traditionnal multichannel “bed”. They don’t use ambisonic to be rendered but a method comparable to LBAP, which can be a layered variation of DBAP or VBAP. I don’t know which one is used but it may be logically DBAP since a Dolby Atmos speakers arrangements are not based on triangulation.
Since Atmos is a proprietary format which needs proprietary tools, you cannot actually produce an atmos compatible file without using these tools (in Protools, Pyramix, Logic …)

But you can of course spatialize what you want in a 3d space based using or not on the same spatial arrangement, either object based with other tools/formats (MPEG-H, Sony Reality, ADM) or more freely with all the powerful multichannel “panners” and tools that exist.
Ardour panner being 2D you cannot use it here.

You can also use ambisonics with the appropriated plugins like the excellent and free from IEM and SPARTA.

… but if all you need is a binaural output it will be simpler to use binaural panners :wink:

I probably spoke a bit sloppily. I did not mean that Atmos is Ambisonics. I meant that there is fundamentally nothing about the Atmos model that isn’t covered by Ambisonics, and that the only reason Atmos exists as a distinct technology is the lack of licensing possibilities for Ambisonics, which over the years has led a variety of companies to try to develop their own proprietary version of the same concept.

Ambisonic and Atmos or other OBA techniques are indeed very different in the manner in which they process the sound.
As a render technique, ambisonic can be used to render a spherical object based mixing, like will also do multichannel techniques like VBAP or DBAP, binaural or even WFS. Each one will have its domain where its shines and other ones where it simply don’t work.

But the “spatial coding” in ambisonic is something very particular, and for instance is not adapted to the geometry of theaters that are far from being spherical. It is also totally incapable to create other spaces than periphonic, like for example does 4D-Sound, or will need an infinite order to make the voice sound only on the center speaker ;-).
Despite its qualities in some applications, there is a lot of things that are not covered by ambisonic, be it with Atmos or other more original and less proprietary approaches that use different and also interesting concepts …

You are mixing up several different things. VBAP is a mathematical description of how you distribute audio power between multiple channels.
Original VBAP paper by Pulkki

Dolby Atmos is a method of distributing audio to theaters and home theaters that has base audio channels (like the traditional Dolby 5.1 style distribution), and additional channels that have descriptions of where those audio streams should be located, and the playback processor determines based on the speaker configuration in use how to best distribute those additional channels to the available speakers. I think in principle a Dolby Atmos decoder could use VBAP for placing the additional audio channels, or could use a different panning law. I don’t know the details of Atmos licensing, I am sure there is some standardization involved, but conceptually I don’t think there would be anything preventing using an Ambisonics panner to place the Atmos streams, and then rendering speaker feeds using Ambisonics decoding.

In other words, Atmos is a delivery format, and that is an orthogonal concept to panning power laws, which would be the general category VBAP falls under.

Rendering 3D sound to a stereo file is yet something different, and depending on what exactly you mean could involve matrix encoding, like the old UH-J encoding of horizontal ambisonics to two channel, or the original Dolby Pro-Logic encoding of front stereo plus surround channel into a two channel file that was decoded by home theater processors back into left, right, and surround channels on playback.

It could also mean encoding a multichannel file in a way that when you listen right in the middle to a set of stereo speakers you get the impression of sounds panned wider than typically possible with stereo, or even behind you, like the old QSound system used on some CD’s in the 1990’s.

It could also mean using a generic head related transfer function (HRTF) or even a customized-for-you HRTF to render a surround mix to a headphone feed. Dolby did have a version of that technology marketed as Dolby Headphone, and seem to have an improved version marketed as Dolby Atmos for Headphone. According to the wikipedia article on Atmos there are Atmos for Headphones implementations for several phones, and the XBox One ships with an implementation for game use.

With that as background, you would need to decide what you would actually like to do, and what workflow is practical. Using an Ambisonics workflow can be very flexible, but Ambisonics decoding works best with regular arrays of speakers (i.e. each speaker is identical, placed the same distance from the single listening location, and the angle between each speaker is identical). There are some decoders which attempt to compensate for the irregularity of the typical 5 channel home theater style placement with varying degrees of success.

If you only want to target 5 channel Dolby style layouts, you can use a 5 (or 7) channel output bus, and use VBAP for panning sources to the output bus. That is conceptually the most straight forward, and most applicable for home theater delivery, but least flexible, since you end up with 5 or 7 channels, and then have to decide what to do for stereo release, and if you targeted 7 channel release how to handle 7 channel playback on a 5 channel system.

No matter which workflow you choose for the basic surround output, converting to a headphone equivalent would typically be a separate step at the end where you give the surround mix to a binaural encoder and it produces a two channel file for use with headphones. How well that works depends on how well the HRTF matches the size and shape of your head. There are decades of papers around how to try to make something generic work, how to measure your own specific HRTF, how to estimate which of an existing HRTF would be closest to yours, etc. Depending on how picky you are it could be a complete separate learning experience in addition to figuring out surround mixing workflow, or could just be run an audio file through a processing program and call it done.

1 Like

This topic was automatically closed 91 days after the last reply. New replies are no longer allowed.