Anyone here working a ScreenX theater?

[Marcel Birgelen]

I'm pretty sure, my 3-5 spread mix uses three pro logic decoders and a mixer/leveler to get it all correct. Again, my final test was 5 discrete sounds that had to come out of the right "hole" and they did but if you soloed channels you could hear bleed.

Mine doesn't use the matrix decoder, it simply "averages" L+C and C+R to create two "phantom" channels. Yes, it somewhat diffuses the center channel and both left and right channels, but in the greater scheme of things, it's not really that noticeable. It will fill in the gap between those speakers quite nicely though.

I didn't see the reasoning behind using the matrix decoder on something that shouldn't contain matrix encoded sound to start with. Maybe the matrix decoder does a bit more than just matrix decoding?

As for panned dialog and The Sound of Music...I'm VERY familiar with that mix and if you had 5-screen channels, it should not have been very jarring at all. If you had 3-screen channels, you had holes in your sound stage. I don't know what they did for the 4-track mix-down but the 70mm 6-track mix was one of the best, ever. There was a lot of LC/RC information for dialog. When there were three actors on the screen, they each got a speaker.

The 7.1 DTS-HD MA mix on Blu Ray keeps the panning of the dialog intact, but you'll obviously loose the LC and RC channels. I don't really perceive any real gaps in between, so they must have mixed those channels down into the Center and L+R channels. Still, it's a pity there isn't a modern home format that can simply contain and reproduce the original mix. Maybe if you'd use Atmos and put the dialog into objects, but that would be an expensive remix...
There is also some limited surround sound in the mix, but it's very subtle and mostly well done.

Unfortunately, while Atmos can support 5 front speakers, the format itself isn't really geared towards it, as it's essentially built on 5.1 and 7.1 "stem configurations". I've yet to see a mix where the dialog is consistently put into objects instead of the stems...

[Steve Guttag]

I didn't see the reasoning behind using the matrix decoder on something that shouldn't contain matrix encoded sound to start with. Maybe the matrix decoder does a bit more than just matrix decoding?

Really? The matrix decoder provides steering. You can't accomplish that with a simple mixer. If you put a matrix decoder with Left and Center as its inputs, the output will be are rather discrete Left, Left-Center, Center output rather than a mush of the three. If you were to pan the signal between the two then just the LC speaker will play. With a mixer, you'd get all three speakers playing with an artificially wide, mushy sound. Try it again with the Matrix Decoder and compare the two and see if one doesn't do a much better job.

[Marcel Birgelen]

Really? The matrix decoder provides steering. You can't accomplish that with a simple mixer. If you put a matrix decoder with Left and Center as its inputs, the output will be are rather discrete Left, Left-Center, Center output rather than a mush of the three. If you were to pan the signal between the two then just the LC speaker will play. With a mixer, you'd get all three speakers playing with an artificially wide, mushy sound. Try it again with the Matrix Decoder and compare the two and see if one doesn't do a much better job.

The sound isn't as mushy as you make it out to be. It does certainly sound "wider", but that's not necessarily always a bad thing, especially on large, wide screens.

And, yeah, really, because it all depends how the Matrix Decoder actually works under the hood, if it makes sense for the purpose or not. If you look at the Dolby Pro Logic algorithm, in essence, the Center speaker is derived by "summing" both 0,71L and 0,71R and boosting it by about 3dB. This works reasonably well with 2.0 content that wasn't encoded this way to begin with. But many modern decoders pull off other tricks to increase the sense of channel separation. Modern algorithm like the DTS Neural are actually pretty good in extracting a full surround stage from 2.0 program material, but they also don't support a 5-front speaker configuration.

The problem I have with using a matrix decoder that's wired like the Pro-Logic decoder is that I'm affraid it will over-accentuate the LC and RC channels, as it is designed to do so with the Center channel. But I'll give it a try to see if it's an improvement.

[Steve Guttag]

As mentioned, I have a mixer/leveler to ensure that the balance is not mucked up. You should be able to pan through all 5 channels without the level going up/down. There is more steering in a pro logic decoder than you are alluding to. That is, if L and R are equal in level, there should be almost no Left or Right, just Center.

[Marcel Birgelen]

I fiddled around a bit with the Active Matrix Decoder and the thing does more than the original ProLogic implementation. It actively removes the summed Center signal from the Left and Right, something vanilla ProLogic doesn't do. It does boost the center speaker too much though, so some extra leveling is required. The end result is pretty convincing, although I just ran some limited test material through it.

[Marcel Birgelen]

Getting back a bit on topic again, the video below shows the conversion of a "regular" screen to ScreenX:

You can see the two by two short-throw projectors being installed inside the auditorium and also how they simply project stuff over the surround speakers...

[Martin Brooks]

I also still clearly remember Children of Men (2006 and also an Alfonso Cuarón movie), which also put effects exactly where they happened, including the dialog... unfortunately, that movie never got an 8-channel SDDS release, I guess because SDDS was already considered a failed format by 2006.

I could never perceive the difference when I went to see an 8-channel SDDS release. I always got the feeling they were doing the spread and not sending discrete audio to the extra channels. I could also never perceive a difference for Dolby-EX and DTS-ES, where they added the derived center rear "channel".

I agree that many sound mixes are not very good. I've heard some great 5.1/7.1 mixes (The French Dispatch and Belfast come to mind), but as much as I like the idea of Dolby Atmos, I'm usually disappointed with the mix in which I think opportunities are lost. I remember seeing one film that took place in a jungle or rain forest and expected to hear leaves rustling, rain, wind and animal sounds coming from individual surrounds with wind, rain and thunder coming in part from the overheads, but there wasn't much. And I'm guessing this happens because the mixers get the movie last, usually late, but they still have to make release dates, so they have to mix fast. There's also a difference if they mix in Atmos first and derive the 5.1 and 7.1 (and then fix as necessary) or whether they mix in 5.1/7.1 and then sweeten with Atmos objects. I think in the latter case, they don't take full advantage of the capabilities.

[Steve Guttag]

By and large, the SDDS-8 used LC/RC as a means to keep music and effects out of the dialog. As such, one isn't going to perceive them as much but one may notice an improved sound. You'll perceive LC/RC when dialog or objects pan around the image. As for "spread mixes"...If they were from Columbia/Tri-Star...it wouldn't surprise me. Those were the house-studios and while SDDS was a thing for Sony, Columbia/Tri-Star had a lot of SDDS-8 releases.

As for Dolby Atmos, I suspect that the mixes start as 5.1/7.1 as Atmos had their "Bed" channels which are the 5.1/7.1 channels. Most sounds are NOT objects (which get to move around individually). Another strike on some Dolby Atmos installations is that, in order to lower the cost, Dolby allows for "culling" (where two rows of top surrounds can be reduced to one row) and pairing (where two adjacent surrounds are turned into a single channel and aimed to cover the room more. So, just because you might see 11 speakers going down a side wall, that might be just 6-channels (the front "zone" is not allowed to be culled/paired). You can visually see if the top surrounds were culled, pretty easily. If you see a side surround...move your eyes up and see if there are top surrounds in line with the side surrounds...if not, the top surrounds were culled.

Thus far, none of the Dolby Atmos systems I've designed have either culled or paired the surrounds...they are all individual speakers with separate amplifier channels. They all have 5-screen channels too. If the people that mix the movie put the sounds there, those systems will reproduce them. If one looks like the individual meters on an Atmos mix, you can pretty quickly see which ones they had "fun" with and those that barely pass for Atmos.

As for Surround-EX/ ES...it too depended on the mix. I'd say the Toy Story movies that were in EX would give a good representation of moving audio around the back of the room. Gary Rydstrom mixes would use them well...he is the one that wanted the format for Star Wars Episode 1 <sigh>. I could have definitely passed on surround EX. It was rather miserably implemented too with the SA-TEN. With the CP650, it came along rather painlessly.

[Marcel Birgelen]

Most Atmos soundtracks are rather quick and dirty mixes, where most of the content is still in the 5.1 or 7.1 bed configuration and only the odd sound effect is being pushed into objects. Dialog is practically never pushed into objects, it's simply mixed into the Center bed. It's a safe bet, because you don't really have to worry about 5.1 and 7.1 downmixes: simply hit the export to 7.1 and 5.1 in your ProTools and your 7.1 and 5.1 mixes are done/done.

I consider Atmos a "first generation" object audio system, where the objects are more like a nice gimmick. I hope in future revisions, they simply get rid of the idea of "beds" and simply put the sounds where the "action' is all the time. Instead of just audio-point surfaces, they should add the possibility to add virtual audio emitting surfaces/primitives too, that would fix many of the limitations of an "object only mix".

If you have ever played a modern computer game in a screening room or cinema auditorium, you realize what kind of immersion can be achieved with a 7.1 configuration, as long as you simply put the sounds there where the action is. Maybe, there is something like too much surround sound, so much that it becomes distracting, but it's clear to me that your average cinematic sound mix is being created from the same template or "templated mindset" of how a cinematic sound mix should sound: Dialog flat center in the middle, music in the L and R and the surrounds are there just for the support. It's like we're stuck in the 1970's forever, where the soundtrack is still supposed to work with a single center speaker, whereas the bare minimum for every DCI compliant room is a working 5.1 configuration...

[Steve Guttag]

My own personal observations are that the Dolby Atmos mixes with significant object interactions is closer to 50% than what Marcel has portrayed above. Again, I look at the meters of all of the channels and you can tell/see when objects are in decent use by how they move about...how active are LC/RC, LW/RW...do the TS channels move objects or do they seem to operate like an "array." You can also quickly see the ones that seem to have had a glancing Atmos mix where it is just about 5.1/7.1 throughout with the occasional Array based Top Surrounds and the VERY rare object in any channel. The title of the movie has little indication as to what type of mix will be in it.

[Marcel Birgelen]

You can see it by looking at the meters, but if you've listened to thousands of movie soundtracks and have done some mixing yourself, then you can also easily spot them by just listening at them.

Let's do a "very objective" statistical analysis of recent Atmos releases I've had the distinctive joy to experience:

Doctor Strange in the Multiverse of Madness => Object heavy mix, especially the "in between universe" scenes...
The Unbearable Weight of Massive Talent => Felt just like an average 5.1/7.1 mix
Fantastic Beasts: The Secrets of Dumbledore => Pretty object heavy, but still rather flat mix
Morbius => Sounded like an average 7.1 mix to me with some objects thrown in to justify the Atmos label
Ambulance => Object heavy, all kinds of stuff flying "all around you". Heck, even the top surrounds got a workout in this one... (thanks to that one helicopter...)
Uncharted => Could just have been a 7.1 mix
The Lost City => Same as above.
The Batman => Clearly some objects, but otherwise rather traditional "7.1-style mix"
Death on the Nile => I didn't even know I've seen this in Atmos...
Moonfall => Pretty object heavy, especially clear in the "space" scenes.
Scream => The ringing telephone was clearly an object... so let's count it as an object heavy mix!

I've marked the "object heavy" mixes in bold... so let's count:

11 total, of which 5 object-heavy, so, that's about 45%... close enough to 50%, so I give you that number, at least for recent mixes.

But even the mixes with heavy object use still just mostly dump the dialog in the center bed and the music mostly in the L + R beds. We now have algorithms that can track people on screen. We track thousands of visual reference points when composing scenes with tools like Nuke. You can animate cartoon character's face with facial tracking, so it should be possible to do 90% automated dialogue tracking.

[Steve Guttag]

I'd say you also substantial my claim that the movie title is no indication of an object heavy mix...like The Batman...not Atmos heavy. Yes on Ambulance being object heavy. VERY heavy. Everything got used.

[Martin McCaffery]

Not to divert away from the tech talk, but why the hell does Downton Abbey: A New Era need an Atmos track?

[Steve Guttag]

Sound is half the picture Martin. Come on...get your Dolby ATMOS now!!!!!

[Mark Gulbrandsen]

Sound is half the picture Martin. Come on...get your Dolby ATMOS now!!!!!

If Atmos had included a "sound excitement meter" it would barely move in most cases...