A few questions follwing the Bob Clearmountain Atmos SOS podcast

[Aled Hughes]

Thank Sam and SOS for the Bob Clearmountain podcast - very interesting.

It did raise a few questions that I'd like to ask.

Firstly, BC mentions that he still uses bus compression with Atmos, and that he had a 12 (or was it 16?) channel bus compressor for it. How does this work? Atmos has no fixed outbus, and can be anything from mono to a 128 channels from what I understand, totally dependent on the playback system at the destination. Even if he has (for example) a 12-channel compressor and a 7.1.4 monitoring system, his final Atmos masters are still not going to be fixed at 12 output channels.
I assume that a bed channel can be compressed as usual with a multi-channel compressor, but he did not mention that he works exclusively with beds.

Secondly, the discussion on using the centre channel for music was interesting. BC was saying that it's silly not to, Others say you shouldn't. What's the opinions here? I guess some of the reasons not to are that the centre channel is often voiced a bit differently/compromised in comparison to the L/R speakers in cheaper domestic surround systems?

By extension to the above, the LFE channel is often seemingly misunderstood. Is there any point in using it for music? With properly integrated bass management in a system I can't really see the benefit of having an LFE channel for music. Yet I remember a discussion a while ago on this forum where a label or client had freaked out when there was nothing on the LFE channel. I guess it's not widely understood that LFE is not a 'subwoofer channel'?

Finally, and Google hasn't been able to help here, is there a 'standard' way to downmix a sourround or Atmos mix to stereo for monitoring? Is it simply 'all left/right channels to L/R at unity and centre to both', or is there some attenuation of rear or centre channels? And is the LFE disregarded in this, or is it added to the sum via a HPF?

Aled

[Sam Inglis]

Glad you enjoyed it -- he definitely has an interesting and different perspective.

On the subject of compression, it seems he uses objects for the height speakers but treats these like additional bed channels, statically panned to the appropriate loudspeakers. So I assume that his multi-channel compression is applied to these channels as well. He has a couple of extra objects that are used for any moving elements. Not sure if those get compressed too, but there's no reason why they couldn't be.

I think in the ideal world, the choice as to whether to use the centre channel for vocals and so on should be an artistic one that is made by the mix engineer. It certainly sounds different from panning something to the phantom centre, and I can well imagine you might want to use both to achieve different effects.

I believe the LFE channel is intended only for special FX in movies and so on. I'd have thought that if general LF content is going to be handled by a sub in a particular playback system, the crossover should be handled locally, because the crossover frequency might be specific to that playback system. Hence you shouldn't pre-empt that at the mix stage.

That's a good question regarding fold-downs to stereo — paging Mr Robjohns...

[Hugh Robjohns]

Atmos supports downmixes for stereo (and 5.1) legacy systems (as well as binaural headphones) and the way the beds and objects are mapped into the available physical channels is determined by embedded metatdata programmed by the content creator. The metadata can be based directly on Dolby's own presets for the legacy formats, or customised to specific requirements.

It's all similar to the downmix metadata concepts in Dolby Digital etc and there are more details here if you want them:

https://learning.dolby.com/hc/en-us/art ... -Controls-

Multichannel compression (with linked gain reduction) is essential if you want to avoid disturbing spatial movement, and its just a case of processing the main bed and height channels, however many you chose to have — normally 5, 7 or 9 horizontal and 4, 6 or 8 height.

It's true that Atmos can accommodate 128 channels but few mixes would use them all, and never all for static channels/beds.

As for your other questions, Sam has given good answers. As he says, sounds placed in the centre speaker have a different character and spatial presence compared to a phantom centre, a difference which can be used creatively.

I'd counsel against putting vocals exclusively in the centre channel. It's too easy to strip out for instrumentals or accapella tracks and record companies don't like that.... and it also sounds weird and strangely unnatural to me... but anything else goes! Just remember that there's no guarantee that the playback system's centre channel will be correctly aligned, so critical mix elements might be more reliable as phantom images between left/right channels!

As all the main channels can carry anything down to 5Hz there's no need to use the LFE for music and I'd recommend you don't because the LFE is generally omitted from downmixes. So if you do use the LFE channel you risk serious mix balance problems on different replay systems.