Has Atmos Mixing Changed? (@Stephen W Tayler)

[OldAsMono]

Who's Next was the first time I heard a 'different' Atmos mix from Steven Wilson that continued to Dweezil's Machine Head and now others too. You have to turn these up to get a complete balance. Who's Next seems to me to have the thunderous drums mixed low until turned up. Machine Head needs volume to also kick the super mix into gear. Worse yet are the many front heavy Atmos mixes that are being outshined surround wise by their Quad counterparts (as mentioned by @edisonbaggins in another thread.)

I have to ask Stephen ((@Stephen W Tayler) or any other Atmos mixers for streaming services IF they are being directed to turn back the surround at all by the labels or artists? I ask as I'm curious if our beloved Atmos mixes for discs are being dialed back for the masses on streaming services and then released on disc?

Quad sort of forced the mixing of good surround to fill only four channels (speakers) and it seems with Atmos many mixers are shooting for big stereo. The new Joni discs are amazing! I find the variety of stereo, quad, 5.1 and Atmos to be great examples. I rate the surround mixes as follows myself:

A) Atmos - Hand's down the best distinct sound. So fresh and about perfect. The only thing is "I' would have made the surround include more rear channel activities. These sound so crystal clear perfect.

B) Quad - As mentioned above, these quad mixes are pure all-around surround. A better surround experience then Atmos but lacking the crisp, perfect quality of Atmos.

C) 5.1 (Atmos downmix) - Alternate mixing choices and unfortunately sometimes the vocals include a slight echo that detracts from Joni's great voice.

Lastly I have to acknowledge that we here are all still tuning our Atmos systems to some extent or we wouldn't have such drastic different listening experiences. Besides my selfish desire for Atmos to become the de facto music standard, I also know these Atmos mixes need to impress the masses that we have a new (better) way to complement our stereo versions.

 

[jimfisheye]

I can tell you that the down mixing of Atmos when you point it at smaller speaker arrays and finally stereo goes "volume war" if your Atmos mix is above -14 LUFS. That if nothing else has forced people's hands in not delivering volume war loud masters. If there was one single thing Dolby did really really really right in this whole thing it's that!

The dial norm parameter was new to some of us and a few releases went out not set to -31 (the nominal setting for music). And those are quiet. Up to 24db quieter, I believe.

After that I think it's just mix vs mix. Two things that force you to turn up the volume: Super dynamic mix (think classical music) or a murky mix. The latter happens sometimes.

A main feature Dolby has going with Atmos is the encoding. They're using this to push hardware sales with hardware based decoding. To the point that they are still not licensing their decoder codec to software based media players beyond their own reference player app (not sold commercially). So there are maybe going to end up being some faux Atmos mixes done in stereo or even with headphones monitoring off the Atmos renderer in binaural mode. Maybe some of the front heavy ones on streaming are this? The extension to surround sound part of this is certainly more niche!

I'm not in the mainstream. Very small and independent and unknown. No clue what people getting corporate gigs are being told. I don't suspect anyone is requesting faux surround specifically though. Might be happening in a pinch here and there but probably not actually requested. Any surround hating producer would be listening in stereo anyway and none the wiser. And they know most streaming listeners will be in stereo ear buds too. I suppose weird things happen though. Like throwing away fully produced Jimi Hendrix bluray box sets!

So far the worst unimaginative front heavy Atmos mix I've heard wipes the floor with a stereotypical volume war mastered release. (Yep the pun works too.) Again, Dolby putting the limiter and set level on their downmix engine to specifically screw over anyone going for volume war is one of the best and unexpected moves in the industry in a long time! I still have enormous hate for the decoder business. That's really the only main negative for me.

Remixes are always contentious too! Some mixes have some really fiddly nuanced mix moves. Miss something or change something and someone is going to be upset!

What else...
Channels are channels. Full range audio is full range audio. 4 channels or 12. No difference between an Atmos mix and a quad mix there. It's really all about the mix! There are examples of crisp perfect quad mixes next to murky half-assed Atmos mixes too.

 

[sjcorne]

Who's Next seems to me to have the thunderous drums mixed low until turned up

The original stereo mix isn't like this too?

 

[ar surround]

...So far the worst unimaginative front heavy Atmos mix I've heard wipes the floor with a stereotypical volume war mastered release. (Yep the pun works too.) Again, Dolby putting the limiter and set level on their downmix engine to specifically screw over anyone going for volume war is one of the best and unexpected moves in the industry in a long time!...

Good.

 

[LuvMyQuad]

Good.

I didn't recognize the new avatar. I guess there were no good pics of Rodgers in a Jets helmet eh?

 

[austinsignal]

Well, the tools used for ATMOS mixing have changed - the ability to compress/clip/limit across the whole mix (bed and objects), and incorporating a multitude of tools (all our favorite plugins) to do so is now completely doable. That wasn't true a couple years ago. That's not to say that the mixes are all heavily compressed/limited - it's just that there are more tools available. For indie mixers like me, that's really nice, and I think improves the end product.

As mentioned by jimfisheye, the loudness thing is not a thing with ATMOS, which is a very welcome change - also means that you may have to turn things up a bit to get full effect, as opposed to heavy compression/limiting making things seem energetic even at low volumes.

The stuff I'm hearing coming from the major releases just keeps getting better. Lots and lots of great mixes out there.

Regarding listening to 5.1 downmixes - I think the challenge there is the incredible variability in 5.1 speaker placement in the consumer's homes. Many have the surround speakers in the ceiling (or other weird places), which creates a completely different sound than having speakers in their 'proper' place. I have a monitoring setting that lets me check 5.1 with the surrounds up above and in back..it's actually pretty remarkable that it sounds good at all - Dolby does a nice job with that - but it does sound different for sure. FWIW, when I check mixes in 'normal - not ceiling surrounds - 5.1 against the 7.1.4 toggling back and forth - I really enjoy the 5.1. As least for energetic program material, it has more 'thump' to it. It's less spread out, more focused than 7.1.4.

 

[Stephen W Tayler]

... incidentally I am not sufficiently experienced with Atmos mixing to comment on this thread at this stage - I have been experimenting for some time but have only completed one project that has not yet been released (hopefully soon!). So I will need to see how things actually work out. Hope to talk about it all soon!! SWTx

 

[jimfisheye]

I'm weirded out by the comments that talk about 4.0 vs 5.1 vs 7.1.4 as though they are completely different formats with a different sound. There's something to learn here! There's a correct setup for speaker placement and balanced levels across channels. Do something odd and it can change the mix as much as listening to a Beatles stereo mix with only one speaker connected. The Atmos format isn't aiming to do anything about that either. Nor should it. You can't really deal with variables like that... the system needs to be set up in an expected way. Surround is still this all or nothing engineer's kind of system with that. Exponentially more now with 12 channels!

The downmixing is secondary and always a compromise. So that stuff aside, it's really still all about the mix. Anything with more punch or directness in a 5.1 mix vs a 7.1.4 mix is 100% the mix decisions. The format is 24 bit audio mixed to full range speaker channels for both. Mix elements sound very literally the same in either. The commercially released 7.1.4 mixes I've heard don't have any different sonic aesthetic than 5.1 mixes. Obviously everyone's mixes are unique to their style but the format doesn't have a different basic sound to it. It's still transparent 24 bit digital with no sonic imprint or artifacts.

I suppose it's simply the result of experimenting and trying new things and failing with some of them. I've had that exact experience where a 5.1 mix sounded more focused. I'm going to say that means the 7.1.4 mix is wrong when that happens! Some initial attempts to merely expand a 5.1 mix to 7.1.4 didn't seem right either. The expanded system with those height channels begged to reinvent some things.

One thing I do notice in most 5.1 and now 7.1.4 mixes (and a decent number of 4.0) is the front channels are usually full blooded with kick drum and bass and usually some other main instruments to the mix. Mine included. Even some mixes that don't sound front heavy will look front heavy with that (in the waveforms). So this hasn't changed. I suppose it's a little inevitable. Some quad mixes had more experimental stuff with that though. Not all of it worked, of course! There are some surround mixes with the main 'wall of sound' element sounding pulled apart when that goes wrong.

I haven't heard anything in 7.1.4 that's way out in left field experimental yet. I heard an absurd one where they had the kick drum in every channel. Wonder what the Flaming Lips are up to?

PS. I don't think I'm really qualified to be talking about this yet either! But these are my initial experiences and I posted anyway.

 

[OldAsMono]

I'm weirded out by the comments that talk about 4.0 vs 5.1 vs 7.1.4 as though they are completely different formats with a different sound. There's something to learn here! There's a correct setup for speaker placement and balanced levels across channels.

I appreciate the comments as I wrestle with my own search for "audio perfection" in my system. I know I was way to general when I categorized my listening experience with current 4.0 vs 5.1 vs 7.1.4 mixes. I do feel they can all be unique, and they are controlled by the mixing choices. I love the variety. Each format has its merits. (I'm going to try disabling my heights when I play a 5.1 mix derived from Atmos as mentioned above. This shows I do use sound processing to fill the ceiling speakers on 4.0 & 5.1 & even 2.0 and create my own unique listening experience. ah, the variables.)

My goal was to spark a conversation to help me create a solid listening experience across all formats. Honestly it seems our oldest obstacle with surround is hoping for all mixes to make all channels active participants. I never understand comments of staying mixing choices were to stay true to the stereo mix. We get (have) the stereo mixes, create us something new channel wise as the formats have made possible. I'm not saying to make the actual song different as such, just the mixing choices.

 

[jimfisheye]

For me it's not so much wanting someone to stay true to the stereo mix but more just not phoning in a so-so surround mix after a more accomplished stereo mix. Hate when that happens!

I'll throw this out there for the "difference" between Atmos vs others.
There can be some wild level differences all of a sudden maybe leading to some of this.

The industry has been pushing for -16 LUFS to be nominal and Dolby is playing along here. Streaming has been -14 to -12 LUFS for a while. Apple has been -16 recently though. Youtube is whatever the heck you want.
Volume war CDs have been pushed up to -8 to -6 LUFS for the extreme ones. Usually -10 to -8.

People going from years of volume war CDs to streaming could make the leap. That's still a significant 6 to 10 db change though. Put a loud CD on after governed streaming (the streaming sites have a loud mode too) and you'll be rocked but that's about it.

So Atmos now. More strict -16 LUFS to begin with. Or close.
Dialnorm parameter set to -31 for nominal music.
Wait... what's that, you say? Yeah, so leave that at default -18 and now your Atmos program is playing 13db quieter.

Let's add up the extremes. Quietest Atmos mix with -18 dialnorm vs loudest volume war CD can easily be 24 db difference. 3db is twice as loud sounding. Playing the CD after the Atmos would do more than just rock your world. Probably lose a speaker. Or some hearing.

Audio equipment mostly all performs well down the middle. Nicer gear does better in the extremes. Modest devices can maybe only turn the volume up so far and in the extreme the system will be struggling. Some of this consumer gear sold after the volume war thing started was only designed to handle a more reduced dynamic range. Pushing such a thing with a quiet source could easily sound subdued.

 

[austinsignal]

For me it's not so much wanting someone to stay true to the stereo mix but more just not phoning in a so-so surround mix after a more accomplished stereo mix. Hate when that happens!

I'll throw this out there for the "difference" between Atmos vs others.
There can be some wild level differences all of a sudden maybe leading to some of this.

The industry has been pushing for -16 LUFS to be nominal and Dolby is playing along here. Streaming has been -14 to -12 LUFS for a while. Apple has been -16 recently though. Youtube is whatever the heck you want.
Volume war CDs have been pushed up to -8 to -6 LUFS for the extreme ones. Usually -10 to -8.

People going from years of volume war CDs to streaming could make the leap. That's still a significant 6 to 10 db change though. Put a loud CD on after governed streaming (the streaming sites have a loud mode too) and you'll be rocked but that's about it.

So Atmos now. More strict -16 LUFS to begin with. Or close.
Dialnorm parameter set to -31 for nominal music.
Wait... what's that, you say? Yeah, so leave that at default -18 and now your Atmos program is playing 13db quieter.

Let's add up the extremes. Quietest Atmos mix with -18 dialnorm vs loudest volume war CD can easily be 24 db difference. 3db is twice as loud sounding. Playing the CD after the Atmos would do more than just rock your world. Probably lose a speaker. Or some hearing.

Audio equipment mostly all performs well down the middle. Nicer gear does better in the extremes. Modest devices can maybe only turn the volume up so far and in the extreme the system will be struggling. Some of this consumer gear sold after the volume war thing started was only designed to handle a more reduced dynamic range. Pushing such a thing with a quiet source could easily sound subdued.

Maybe I'm just thick - but I don't understand this. The only limits I see for ATMOS mixes are -18 Integrated LUFS, and -1 true peak. Everyone plays by those rules, so all available ATMOS tracks should be relatively in line. When mastering an ATMOS album, the common practice is to set the loudest track to -18, and the rest of the tracks -18 or less, depending on how loud they are compared to that initial loudest track. So, at least for me, there will be 1 song at -18, and the rest are -18, -20, -19, or whatever. This is all done by ear, using the Dolby Album Assembler. I don't know of any way to do this via metering, and besides, perceived loudness can vary a lot based on how compressed, how much the front wall has been pulled apart across 7.1.4., etc...so using your ears seems to be a good method.

For me, Dialnorm only comes into play when delivering mp4's to clients for playback of the Apple Spatial evaluation - those mp4's are decidedly quieter than the Binaural renders for evaluation, thereby confusing the client as to 'loudness' of the master. I've been told the only way to change this is to use the Dolby Encoder, but I've just kind of given up and I just walk the client through the fact that their mp4's will be quieter for the purposes of evaluation, and tell them just to turn the volume up!

 

[jimfisheye]

That's all at odds with my understanding and Dolby's literature. The published spec from Apple is -16 LUFS for the album as a whole. Not song by song. (Spotify is -14, etc etc.) There are some Atmos masters above that out in the wild. Streaming will just turn the lossy copy of any of those down just like with any other content. The Atmos downmix engine sets the limiter where it does and it's fixed. This just results in the downmixes getting brick wall limited, volume war CD style, if you go too far over that level. (Seems to be more like -13 LUFS for that ceiling.)

The dialnorm parameter is baked into the encoded Atmos master. Read the "-31" parameter as "nominal music setting". If you set this (or leave it default set) low, the Dolby decoder/player will be instructed to turn down the volume by that much. It's matter of fact like that.

It's not the end of the world. You'd only lose two bits of resolution listening that way. 22 bit instead of full 24 bit. It's more the low levels with some consumer electronics that were dialed in for compressed mp3 sound and then the off putting volume swing from quiet Atmos to full volume war CD.

The Dolby encoder is genuinely the only app that can produce the encoded Atmos output. The mp4 option from the renderer is lossy only. They really thought this through and nailed it down. You need both the renderer and the encoder to produce full Atmos content. The encoder also can not produce the intermediate file set. And then the reference player to play any of it back. The renderer abilities (with full Atmos output still strictly locked out) are available from this newer Fielder renderer app and then a few select DAWs.

All the roadblocks even when you buy the official software and play along feel like poor man's style copy protection inhibition. We have to run renders in real time again like it's 1991? Network connections between these apps? I haven't sworn at the computer this much in a very long time!

Sorry, digressing into content creator issues! Anyway, I think the levels are throwing some consumers with some of the cheapness devices from the volume war era. And it's scarey because they can hit play on something 24 db louder the next moment.

 

[austinsignal]

That's all at odds with my understanding and Dolby's literature. The published spec from Apple is -16 LUFS for the album as a whole. Not song by song. (Spotify is -14, etc etc.) There are some Atmos masters above that out in the wild. Streaming will just turn the lossy copy of any of those down just like with any other content. The Atmos downmix engine sets the limiter where it does and it's fixed. This just results in the downmixes getting brick wall limited, volume war CD style, if you go too far over that level. (Seems to be more like -13 LUFS for that ceiling.)

I've never, ever heard of anything but -18 integrated LUFS and -1 true peak for ATMOS mixes. I've submitted mixes to the distro's directly and to labels that use whatever distro they employ, and ALL are the -18 standard. Louder mixes get rejected. I've also submitted singles (usually in advance of the rest of an album) so I have no idea how an average would be calculated at -16. Here's a section from Distrokid's support page: distrokid atmos loudness requirements All the distro's say the same.

I'm not sure what the ATMOS downmix engine is - but I can tell you that it's generally understood the 5.1 fold down is used to calculate loudness by the Dolby ATMOS renderer... I've used Pro Tools, Logic, Fielder, and Reaper to mix projects. When using Apple's logic, and you're working on estimating loudness prior to rendering the final mix, Apple instructs you to set monitoring to 5.1, play the song, and look at the result on the loudness meter plugin, so I'm pretty sure that's what Dolby is doing in the Renderer when running the loudness utility.

The dialnorm parameter is baked into the encoded Atmos master. Read the "-31" parameter as "nominal music setting". If you set this (or leave it default set) low, the Dolby decoder/player will be instructed to turn down the volume by that much. It's matter of fact like that.

It's not the end of the world. You'd only lose two bits of resolution listening that way. 22 bit instead of full 24 bit. It's more the low levels with some consumer electronics that were dialed in for compressed mp3 sound and then the off putting volume swing from quiet Atmos to full volume war CD.

The Dolby encoder is genuinely the only app that can produce the encoded Atmos output. The mp4 option from the renderer is lossy only. They really thought this through and nailed it down. You need both the renderer and the encoder to produce full Atmos content. The encoder also can not produce the intermediate file set. And then the reference player to play any of it back. The renderer abilities (with full Atmos output still strictly locked out) are available from this newer Fielder renderer app and then a few select DAWs.

No idea what you're talking about here. I mix ATMOS all day, pretty much every day. I mix in a DAW, (generally Reaper these days), render a ADM BWF in the Dolby ATMOS Renderer, and from that generate mp4's and BIN WAV files for client/label headphone evaluation. Once all the mixes are signed off, I use the Dolby ATMOS Assembler to get relative volume set for each track (-18 being the loudest track), generate final ADM BWF's with then are sent to a distro (like Distrokid) by me or the artist or the label, depending on the situation. I don't have the Dolby Encoder installed, and never have.

All the roadblocks even when you buy the official software and play along feel like poor man's style copy protection inhibition. We have to run renders in real time again like it's 1991? Network connections between these apps? I haven't sworn at the computer this much in a very long time!

You can run offline renders with some DAW's, like Logic (although there are issues with doing that), Pro Tools, etc. I choose not to, because of various issues that have arisen (downmix settings, etc.), so I go ahead and run the song, and 'record' it using the Dolby Renderer. At least for now, I use Reaper for a variety of reasons, and it seems the most reliable in combination with the Renderer. It's a little backward, but not that big of a deal.

Sorry, digressing into content creator issues! Anyway, I think the levels are throwing some consumers with some of the cheapness devices from the volume war era. And it's scarey because they can hit play on something 24 db louder the next moment.

 

[jimfisheye]

I don't touch the lossy streaming stuff at present. Only the lossless and for that output you do need the encoder to produce .mlp.
Interesting with the request to meter downmixed to 5.1. I've noticed the meter discrepancy across nearly everything with multichannel mixes in general. The more channels, the more questionable. Sounds like it's time for a shootout again! I use Reaper too.

I don't mean to digress with that or argue. Or quote streaming levels that might be different next time someone reads this.
https://www.rtw.com/en/blog/worldwide-loudness-delivery-standards.html
The comment about a different or "less defined" sound has come up more than a few times now. That's what caught my attention. To me it's the very same sound but now with 4 more channels. Same sound but more of it from the ceiling now. So that's the best theory I can come up with - lower than expected levels for certain devices leading to this.

Maybe it's something even more simple like someone not willing to turn up their system 24 db higher than they ever have before? Not realizing that it should just be OK?***

*** Not responsible for any equipment loss, hearing loss, or spousal loss caused by this comment! Use only as directed.