Dolby Atmos walkthrough

[Ben@Harrison]

A nice summary of Mixbus's Atmos implementation:

https://www.youtube.com/watch?v=-Skjxc6ofBg

[calimerox]

Hi Ben,

this is a great summary, and it adresses also the questions I had about the Beds and Objects. As I work at the intersection of sound and film I was thinking if I could use Mixbus10 for Atmos Postproduction for film productions?
As the beds refer more to the traditional setups of movie theatres (like 5.1 etc) it still should not be a problem working totally object based rendering to film? I could use the Mixbusses for example to hard-pan the dialog to the center speaker, the ambiences to left and right and for foley just use Objects for example? What would you guys suggest? Till now I work in Davinci Resolve and although it is very complete I am more used to Mixbus and would like to come back to Mixbus for Film Post. Or at least for the part of the film post that is not Dialog and Ambience. Any ideas assigning of the Mixbusses etcetera?

[hodger]

Hi calimerox.

First, the capabilities.
MB10 can easily accommodate those things you wish to do. The post industry has embraced objects for a long time, and if the MB10 "snap" feature works as advertised, you'll still have a way to ensure a particular "speaker" gets your intended signals from an object panner.

Second, the approvals.
You'll need to confirm with whomever is receiving your deliverable. You'll find challenges trying to get these files into a cinematic release, maybe not for technical reasons but for licensing (think FEES) reasons. You may also encounter some particular requirements for bed usage. The primary pursuit of music-streaming for MB10 makes this all go away, but those other property owners have distinctly different deliveries and workflows. I recommend submitting a small sample for them to evaluate, and then get their response. Other than loudness specs, there may be other expectations simple based on how things have worked historically.

h

[calimerox]

Hi calimerox.

First, the capabilities.
MB10 can easily accommodate those things you wish to do.  The post industry has embraced objects for a long time, and if the MB10 "snap" feature works as advertised, you'll still have a way to ensure a particular "speaker" gets your intended signals from an object panner.

Second, the approvals.
You'll need to confirm with whomever is receiving your deliverable.  You'll find challenges trying to get these files into a cinematic release, maybe not for technical reasons but for licensing (think FEES) reasons.  You may also encounter some particular requirements for bed usage.  The primary pursuit of music-streaming for MB10 makes this all go away, but those other property owners have distinctly different deliveries and workflows.  I recommend submitting a small sample for them to evaluate, and then get their response.  Other than loudness specs, there may be other expectations simple based on how things have worked historically.

h

Thank you hodger. 

great that the snap feature gets audio in to the right place. I will play around with it! 

Could you elaborate on the licensing topic? 

I understand that maybe a platform , just to pick an example like netflix, would have a spec sheet with an obligation to use the Bed for example. 


How does routing through a Mixbus affect the Atmos panner of the tracks? I was thinking for example as a workflow routing all the dialog tracks to a Mixbus, all the ambience tracks to another Mixbus, all the Music tracks to a third, etc. 

When I do that, does the panner in the Mixbus overrides the actual panning in the tracks? so if I have set one ambience just on the rear speakers on one track, on another track to the front speakers, and I route them both to the same Mixbus, does these get altered I guess? I am trying things around but do not have a 7.1.4 setup yet in the studio where I could just test on my own..

[johne53]

There was HUGE excitement here when the new Dolby feature first got announced but it's hard to deny the disappointment that users felt when it eventually materialised. And the reason was because of the very dramatic effect it's having on Mixbus's DSP figures - in many cases, increasing DSP beyond 100 percent and producing nothing but xruns 

It's very encouraging to see in the video some reasonably low DSP readings and it's renewed discussions about how to optimize your computer - but in the absence of a small Atmos test session, it's not going to be easy to judge which (if any) optimizations will help. So the sooner someone releases a small Atmos session for Mixbus (maybe just a simple binaural one) the sooner we'll start getting some decent feedback. And I agree with hodger and calimerox that the licensing situation is quite confusing. Would it even be feasible to make a MB Atmos session generally available without running into licensing issues? 

[Dingo]

There was HUGE excitement here when the new Dolby feature first got announced but it's hard to deny the disappointment that users felt when it eventually materialised. And the reason was because of the very dramatic effect it's having on Mixbus's DSP figures - in many cases, increasing DSP beyond 100 percent and producing nothing but xruns 

Well others experiences may differ, but running 32 tracks of a recorded session (just Mixbus no additional plugs) sits at about 20-22% enabling Atmos and that goes to about 38% no x-runs at all .... Mac Intel Ventura.

Running the Mixbus Vapor renderer sucks a little bit less resource than using the Dolby Atmos Renderer (DAR) ..
Mixbus is also drop dead easy compared to the mind boggling setup needed on every other DAW I have tested that supports Dolby either native or with the DAR.

Yes Immersive takes more DSP that is just how things are.

My advice is that if your system is struggling, export stems of tracks and or Mixbuses and create an Atmos version rather than trying to do it all in one session.
The other thing I would say is that Mixbus Pro and the addition of Gates, Comp etc has dramatically decreased my need for third party plugins and the Mixbus elements are very DSP light compared to some VST etc.

I've seen discussions going on now about how to optimize your computer but in the absence of a small Atmos test session, it's not going to be easy to judge which (if any) optmimizations will help. So the sooner someone releases a small Atmos session for Mixbus (maybe just a simple binaural one) the sooner we'll start getting some decent feedback.

What kind of session would you like ? A big difficulty is finding material that does not have copyright issues. How about a session from Telefunken that anyone can import and use as a yardstick ? I will see what I can dig up and post back on this.

And I agree with hodger and calimerox that the licensing situation is quite confusing. Would it even be possible to make a MB Atmos session generally available without running into licensing issues? 

My understanding is that Dolby licensing applies to use in Cinema releases not music streaming ... but I could well be wrong...
dingoes often bark up the wrong tree.

[johne53]

Yes, at the moment I think it's just confusing a lot of users. If a downloadable test session was available (i.e. without falling foul of licensing / copyright issues) it can only be a help.

[hodger]

Could you elaborate on the licensing topic? 

I understand that maybe a platform , just to pick an example like netflix, would have a spec sheet with an obligation to use the Bed for example. 


How does routing through a Mixbus affect the Atmos panner of the tracks? I was thinking for example as a workflow routing all the dialog tracks to a Mixbus, all the ambience tracks to another Mixbus, all the Music tracks to a third, etc. 

When I do that, does the panner in the Mixbus overrides the actual panning in the tracks? so if I have set one ambience just on the rear speakers on one track, on another track to the front speakers, and I route them both to the same Mixbus, does these get altered I guess? I am trying things around but do not have a 7.1.4 setup yet in the studio where I could just test on my own..

Your Music Distributor and the Streaming service mutually handle licensing for streaming music.  Aside from the licensing Harrison undertakes to embed the Dolby SDK, we have no licensing headaches when producing music for streaming. 

Video/Film distributors handle licensing for video content differently, and whoever owns the film has a relationship with Dolby, inclusive of fees and paperwork.  Netflix may be an uncomfortable comparison because they are the film owner in some cases and merely a distributor in others.  Unless you're creating video for your own use, you'll likely already have a client or distributor that can work with you.

With regard to Mixbus groups and mix strategy, Nathan has a nice video on the site explaining the logic that happens when you start the immersive engine.  If you've assigned audio through groups, and de-selected the master, the appropriate object panners are created for you without double-routes.  You are free to change them, so make sure to avoid double-routing.  The audio is common, but the master sections and banners are distinctly separate.

h

[johne53]

Nathan has a nice video on the site explaining the logic that happens when you start the immersive engine.  If you've assigned audio through groups, and de-selected the master, the appropriate object panners are created for you without double-routes.

Could you maybe post a link to that video ?

[Jostein]

https://www.youtube.com/watch?v=nZCeDiA03f4   :-)