[gregorio]

[1] When going from live sound in nature (0) to speakers (1) we lose lots of original spatial information, but new spatial information is generated. The amount of spatial information isn't the issue, but the fact that ORIGINAL spatial information is lost.
[2] Headphones without crossfeed present the spatial information in wrong context ...
[2a] Crossfeed doesn't really change the amount of spatial information but [2b] converts it into a more correct form for our spatial hearing.
[3] When there is no real life sound, what the recording sounds like on speakers kind of becomes the "real thing".
[4] As I have mentioned, fabricated spatiality is ok if each individual sound object in the mix has proper spatial context.

1. What live sound in nature and what ORIGINAL spatial information is lost? The original spatial information of say the snare drum on a typical recording will be the sound + spatial information from 1 inch away from the drum (in mono) AND the sound + spatial information from 6ft away (in stereo) AND the sound + spatial information from say 20ft away (in mono), this is all mixed together, EQ'ed, compressed and artificial reverb is added and then it's mixed with all the other instruments in the band, all of which have different spatial information from different positions in different acoustic spaces and is EQ'ed and compressed differently! Your whole concept of "natural" and "original" spatial information is a fallacy that OBJECTIVELY does NOT exist and pretty much the last thing we would want is natural or original spatial information!

2. Yes they do but so do speakers and headphones without crossfeed because the spatial information is all in the "wrong context" to start with (on the recording). In fact, it's in a bunch of significantly different and conflicting contexts!
2a. Correct.
2b. No it doesn't, it's not only the same amount of spatial information, but exactly the same improper, unnatural, conflicting spatial information in every other respect, just crossfed! There's no complex processing occuring, no "converting it into" a natural, proper or "correct form", the technology to do that does not exist and if it did, it would be pretty much the last thing we would want! This, AGAIN, is the objective facts! Now clearly, you perceive something that sounds "correct" to you (with crossfeed), which is fine but that is NOT the objective fact, it is YOUR perception and therefore NOT applicable to everyone else.

3. Are speakers magic, are they implementing some technology that doesn't exist? Speakers are obviously not the "real thing", they do not make a recording "kind of become the real thing" because there is no real thing and there isn't intended to be. What speakers do is just transduce and replay the recording and the listening environment adds room acoustics and the result is an illusion which is NOT the real thing but hopefully pleasing.

4. That's exactly my point, each individual sound object in the mix does NOT have "proper spatial context"! Again, the typical snare drum "in the mix" does not have a proper spatial context, it has about 4 very different and contradictory spatial contexts. In fact, it's hard to image a spatial context that's LESS "proper"! So, you are contradicting yourself (again)!

Gregorio, my point was that more channels mean more possibilities for sound location and more specific sound fields.

Yes, I realise that and while your statement is often/sometimes true, that's not always the case in practice. For example, if we have standard 2 channel stereo and then add a 3rd (centre) channel we do not have more possibilities for sound location, we have exactly the same possibilities. Additionally, the paradigm of more channels isn't really how Atmos works anyway. Maybe my explanation to bfreedma below will help.

Nice post.
Only addition/correction is that there are a number of 16 channel Atmos processors available (Storm Audio, Emotiva, Bryston Acurus, Monoprice (soon). Storm Audio also has 20 and 32 channel processors, though I'm not sure how the 16 channel Atmos processing works with the additional channels in post processing.

I assume they must be consumer Atmos processors? Theatrical Atmos processors are 64 channel, although in both cases they're not really "channels", they're outputs. Before we consider Atmos, we need to get away from the consumer concept of a "channel" and a corresponding physical output/speaker for it. In a consumer stereo system we have a left channel and a right channel and a sound hard panned to say the left channel will exist only in the left audio channel and will be output only to the left speaker. Same with a consumer 5.1 system, we have a left front channel, a corresponding left front speaker and a sound exclusively panned to the front left channel will be output exclusively to the front left speaker. And the same is true with all the other channels/speakers. However, that's not the case with a theatrical 5.1 system. We still have a front left channel and a front left speaker and we also still have a surround (rear) left channel but we don't have a corresponding left surround speaker, instead we have an array of (diffuser) speakers all around the walls. If we pan a sound exclusively to the front left channel of a theatrical 5.1 system, it isn't output to only the front left speaker, some of it is routed to the foremost (nearest the screen) diffuser speaker (and what we would actually hear is a phantom position somewhat further left than the actual left front speaker). A cinema 5.1 system has the same 5.1 audio channels but 30 or more individual outputs/speakers, each of which is a different signal made up of different combinations of the 5 main audio channels in fixed ratios. So for example, a diffuser speaker on the left wall exactly between the front left and surround left positions would receive a signal that is a 50/50 combination of the front left channel and the rear left channel, the next speaker along that wall might receive a signal that is a 60/40 split and so on. This is all setup during installation and our same 5.1 audio channels can therefore feed a small cinema with say 20 speakers or a large cinema with 60. Now we come to Atmos and the first thing to bare in mind is that it's exactly the same, same array of speakers and same fixed ratio combination of what the speakers are fed as with a 5.1/7.1 system. It's important to understand this because the thing that makes Atmos different (audio objects) is in addition to this traditional 5.1/7.1 setup, not a replacement for it! When we're mixing we can simply mix a sound as we always would or, we can assign it as an "audio object", in which case it effectively bypasses the fixed ratio combination of channels that would feed a particular speaker and allows us to address all the speakers individually. In other words, our speaker in the middle of our left wall would output a fixed 50/50 split of the front left channel and surround left channel, plus a completely variable (0% - 100%) amount of an audio object. Atmos also provides us with two additional arrays of (ceiling) speakers, which are accessible as audio objects. In terms of physical audio channels, Atmos has 8 channels (configured as traditional 7.1) but it also has an additional bunch of data, which defines the audio objects. It's a more complex and "intelligent" system because in the case of a smaller cinema with far fewer than 64 speakers, the processor will effectively work out a phantom position for a sound using the available speakers, if a physical speaker doesn't exist in that (panned) location. The consumer version is effectively an extension of this, obviously using even fewer speakers.

I'm not sure how well this explanation helps? Hopefully you can see that the paradigm of an audio channel and it's corresponding physical output isn't really applicable to Atmos?

G

 

[bigshot]

I realize there are always little exceptions and rare situations, but for the everyday purposes of all of us listening to music in our homes, those details rarely matter. Feel free to mention all that because it's interesting, but I'll focus on the broad strokes that are most important and most effective and solving the real world problems of achieving good sound in a home situation. I'm always try to keep information in context. Too often the unimportant details get more discussion around here than the things that really matter. Lack of context is what I believe gets people worrying about jitter ratings and high bitrates and all of that audiophoolery that sends people down pointless rabbit holes. I went to design school where they taught the importance of KISS.

 

[Davesrose]

I realize there are always little exceptions and rare situations, but for the everyday purposes of all of us listening to music in our homes, those details rarely matter. Feel free to mention all that because it's interesting, but I'll focus on the broad strokes that are most important and most effective and solving the real world problems of achieving good sound in a home situation. I'm always try to keep information in context. Too often the unimportant details get more discussion around here than the things that really matter. Lack of context is what I believe gets people worrying about jitter ratings and high bitrates and all of that audiophoolery that sends people down pointless rabbit holes. I went to design school where they taught the importance of KISS.

I'm now listening to a concert BD that's DTS MA 92khz (Crossroads 2010), and I do keep thinking these detailed discussions are banal. So my new Dolby Atmos/ 3D surround receiver is a Denon AVR-X6500H. Before, I was a diehard Harman Kardon fan: my previous main receiver is an older HK TrueHD/DTS MA 7.1 receiver. My dad gave me a more modest entry tier Denon 7.1 surround receiver when he upgraded to an Atmos enabled one. I wired it and tried a few different blu-rays and found it wasn't as enganging (music sounded flatter as there was less mid-bass and DR seemed a bit more reserved). I would like to have seen what Harman Kardon would have come up with for current Atmos/DTS:X/Auro-3D receivers, but unfortunately they're out of the market. I have been impressed with my new Denon for how it processes 5.1 to 3-D in Auro-3D mode. Interestingly, I'm finding the EQ and levels can change quite drastically with selection of surround mode (let alone settings for surround parameters and crossover). When it comes to producing for Atmos, I have read that movie productions do tend to mix *at most* 7.1.4 core audio. And to try to confirm what gregorio has written, 3D surround formats are based on tracks instead of "channels": Atmos allowing 128 "tracks" (combination of objects and "core channels"). IMO the main advantage of the new systems are greater flexibility for spacial reproduction in a cinema with many multi-channels to resampling to 5.1.4 home cinema. But what matters most with what you're listening to is your own processor. With the BD I'm listening to now, I'm hearing better dynamics and bass response by switching my surround mode to Dolby Surround (Dolby's new Dolby Atmos matrix surround). When I'm critical about another track, who knows...I may prefer Auro-3D or DTS: Neural. When it comes to native 3D surround sound: there's more instances of brands mandating an exclusive sound mode (I've found that's true with Atmos for DD streaming but not TrueHD BDs, and DTS:X BDs).

 

[gregorio]

[1] When it comes to producing for Atmos, I have read that movie productions do tend to mix *at most* 7.1.4 core audio.
[2] And to try to confirm what gregorio has written, 3D surround formats are based on tracks instead of "channels": Atmos allowing 128 "tracks" (combination of objects and "core channels").
[3] IMO the main advantage of the new systems are greater flexibility for spacial reproduction in a cinema with many multi-channels to resampling to 5.1.4 home cinema.
[4] But what matters most with what you're listening to is your own processor.

1. To be honest, that's not really correct, there isn't a 7.1.4 theatrical movie format and Atmos doesn't really have any height channels as such, they are effectively audio objects. The mixing of movies, the formats and routing is very complex, it makes music recordings/mixes look childishly simple. A higher budget film will be mixed in several formats and numerous versions. Typically there will be an Atmos mix and a separate 7.1 mix, there might also be an Auro 3D mix which is based more on the traditional concept of channels and was 11.2 (although there were/are various versions, 13.2 and then AuroMax with 22.2), it was effectively two vertical layers of 5.1 and a single centre, even higher positioned, speaker (often referred to as "the voice of god" channel). There's also likely to be an IMAX version, which is 6 channels (but not in the 5.1 layout) and again requires a separate, dedicated mix. In total, a high budget feature film may have as many as 70 different sound mixes, spread across 5 or more different audio formats. Regardless of what other formats are used though, there's always a 5.1/7.1 mix. It's been this way with moves for over 30 years.

2. Mmmm, not really. I completely understand your confusion though, because the terms "tracks" and "channels" are often used interchangeably which confuses even quite advanced students and occasionally even pros. A "channel" is an audio path, while a "track" is a physical location where the audio is recorded, originally a horizontal region on a tape (or film), hence the term "multi-track recorder". Obviously, to record something on a "track" we need an audio path (channel) routed to that track and in this case the "track" and the "channel" is effectively the same thing. The physical act of recording a musician/sound is therefore often called (in the business) "tracking". Let's take a drumkit as an example to explain when/why they're different: We record a drumkit with multiple mics; a kick mic, a couple of snare mics, one or more mics for the toms, another for the hi-hats, a stereo "overhead pair" and a distant "room mic". Each of these mics requires it's own audio path (channel), so that's about 9 or more audio channels. However, it would be usual to submix that drumkit (those 9 channels) to stereo (another two channels). If we record that submix then we need 2 tracks (which are the same as our submixed stereo audio channels) but the drumkit has actually used 11 audio channels. A blockbuster film will commonly require over 1,000 audio channels, which in the case of a 7.1 mix will be recorded down to 8 tracks, which then require 8 audio paths (channels) routed to the 8 speakers or, as explained previously, 30 or more channels (derived from those 8 tracks) routed to each of the speakers in a theatrical system. Atmos is effectively 8 tracks/channels (traditional 7.1), plus up to 128 audio objects which are not related to tracks at all and are not tied to any specific output channel/s.

3. I'm not sure I would say "greater flexibility for spatial reproduction in a cinema", we don't actually ever record anything even in 7.1, let alone in Atmos. Maybe a greater flexibility for sound positioning would be a better description. Also, it's not really "resampling", nothing is being resampled it's being up/down mixed. You can theoretically up or down mix anything to pretty much anything else even without Atmos. 5.1.4 is effectively 7.1 down-mixed to 5.1 and then 4 ceiling audio channels approximated/calculated from the audio objects. This would typically give a far more accurate representation than upmixing a traditional 5.1 mix and effectively guessing what to put in the ".4" speakers.

4. BTW, the processor is identical in whichever AVR you're using. If your AVR has Dolby Atmos, then the manufacturer of the AVR must have purchased Dolby's Atmos processor, the same Dolby Atmos processor that every other AVR manufacturer has to purchase/licence. Same goes for the other copyrighted formats (DTS and Auro for example). As mentioned, when upmixing (say stereo to 5.1, 5.1 to Atmos, or whatever) the processors are effectively guessing, although not entirely blindly, the processor will try to identify cues (such as phase and various other indicators) but each different processor (Dolby, DTS, Auro) will "guess" at least somewhat differently and which is "better" will vary depending on the individual film/mix, your speaker setup and your personal preference.

G

 

[gregorio]

I realize there are always little exceptions and rare situations, but for the everyday purposes of all of us listening to music in our homes, those details rarely matter.

True but this isn't one of those occasions (a little detail or rare exception). The example I gave, having two channel stereo and expanding it to 3 channels (with a centre channel) would appear to be very rare/exceptional and indeed there are no 3.0 mixes available to consumers as far as I'm aware. However, in practice it's not a little exception or detail that doesn't really matter, because that's exactly how the front three speakers of ALL 5.1, 7.1, Atmos and modern multi-channel formats work!

G

 

[Davesrose]

1. To be honest, that's not really correct, there isn't a 7.1.4 theatrical movie format and Atmos doesn't really have any height channels as such, they are effectively audio objects. The mixing of movies, the formats and routing is very complex, it makes music recordings/mixes look childishly simple. A higher budget film will be mixed in several formats and numerous versions. Typically there will be an Atmos mix and a separate 7.1 mix, there might also be an Auro 3D mix which is based more on the traditional concept of channels and was 11.2 (although there were/are various versions, 13.2 and then AuroMax with 22.2), it was effectively two vertical layers of 5.1 and a single centre, even higher positioned, speaker (often referred to as "the voice of god" channel). There's also likely to be an IMAX version, which is 6 channels (but not in the 5.1 layout) and again requires a separate, dedicated mix. In total, a high budget feature film may have as many as 70 different sound mixes, spread across 5 or more different audio formats. Regardless of what other formats are used though, there's always a 5.1/7.1 mix. It's been this way with moves for over 30 years.

2. Mmmm, not really. I completely understand your confusion though, because the terms "tracks" and "channels" are often used interchangeably which confuses even quite advanced students and occasionally even pros. A "channel" is an audio path, while a "track" is a physical location where the audio is recorded, originally a horizontal region on a tape (or film), hence the term "multi-track recorder". Obviously, to record something on a "track" we need an audio path (channel) routed to that track and in this case the "track" and the "channel" is effectively the same thing. The physical act of recording a musician/sound is therefore often called (in the business) "tracking". Let's take a drumkit as an example to explain when/why they're different: We record a drumkit with multiple mics; a kick mic, a couple of snare mics, one or more mics for the toms, another for the hi-hats, a stereo "overhead pair" and a distant "room mic". Each of these mics requires it's own audio path (channel), so that's about 9 or more audio channels. However, it would be usual to submix that drumkit (those 9 channels) to stereo (another two channels). If we record that submix then we need 2 tracks (which are the same as our submixed stereo audio channels) but the drumkit has actually used 11 audio channels. A blockbuster film will commonly require over 1,000 audio channels, which in the case of a 7.1 mix will be recorded down to 8 tracks, which then require 8 audio paths (channels) routed to the 8 speakers or, as explained previously, 30 or more channels (derived from those 8 tracks) routed to each of the speakers in a theatrical system. Atmos is effectively 8 tracks/channels (traditional 7.1), plus up to 128 audio objects which are not related to tracks at all and are not tied to any specific output channel/s.

3. I'm not sure I would say "greater flexibility for spatial reproduction in a cinema", we don't actually ever record anything even in 7.1, let alone in Atmos. Maybe a greater flexibility for sound positioning would be a better description. Also, it's not really "resampling", nothing is being resampled it's being up/down mixed. You can theoretically up or down mix anything to pretty much anything else even without Atmos. 5.1.4 is effectively 7.1 down-mixed to 5.1 and then 4 ceiling audio channels approximated/calculated from the audio objects. This would typically give a far more accurate representation than upmixing a traditional 5.1 mix and effectively guessing what to put in the ".4" speakers.

4. BTW, the processor is identical in whichever AVR you're using. If your AVR has Dolby Atmos, then the manufacturer of the AVR must have purchased Dolby's Atmos processor, the same Dolby Atmos processor that every other AVR manufacturer has to purchase/licence. Same goes for the other copyrighted formats (DTS and Auro for example). As mentioned, when upmixing (say stereo to 5.1, 5.1 to Atmos, or whatever) the processors are effectively guessing, although not entirely blindly, the processor will try to identify cues (such as phase and various other indicators) but each different processor (Dolby, DTS, Auro) will "guess" at least somewhat differently and which is "better" will vary depending on the individual film/mix, your speaker setup and your personal preference.

G

It seems you're mainly arguing about definitions and whether height channels are mixed with Dolby Atmos. My language was based on Wikipedia. Wikipedia states that Atmos is comprised of 128 "tracks" (and each track can be assigned a channel or object). It also says that by default, there's a bed of "7.1.2" channels (leaving 118 tracks available for objects). That doesn't seem confusing to me, so not sure why you do want to be argumentative. It does seem the main point you've brought up that's against the literature is that the 128 "objects" are in addition to other mixed channels (that's fundamentally different than what is in the layman's literature, which specifically does say a channel is in one of the 128 tracks). You're also saying height channels aren't included (so to me, that might be the main point that can be confusing). I've seen an engineer say that Dolby recommends a 7.1.4 speaker configuration for home mixing and 9.1.4 for cinema mixing. But perhaps the difference is that 7.1 channels are mixed and the 4 height speakers are for referencing objects? Dolby's literature on cinema Atmos does say that there's a bed of 9.1 channels (and they also state this leaves 118 available tracks).

As for delivered audio formats, I had assumed cinemas could be similar to home. I've noticed that Atmos for home is delivered via a Dolby Digital core (in streaming) or TrueHD 7.1 core (in blu-rays). I also bought one of the few IMAX Enhanced blu-rays to listen to IMAX format (which my AVR supports). IMAX is trying to compete with Dolby, and their audio format is delivered on DTS:X, with their HDR format delivered on HDR10+. It seems like trends are pointing to Dolby being the dominant format for home UHD standards (unlike DTS MA that dominated HD blu-rays).

RE: Atmos processor. I do realize that the Atmos processor is effectively identical in any AVR. My point was that I'm finding sound can be significantly different on which surround processor is selected, as well as different sound settings (such as enabling room correction, bass management, dynamic compression, and other processing in the receiver's system). I have also seen that each brand's processing of stereo/2D surround to 3D has a different name than their 3D format (signifying it being different processing). With Dolby, it's "Surround". With DTS, it's "Neural:X".

Edit: I've just found a good series on Dolby Atmos from a sound engineer, and he goes through an example of doing an Atmos mix. I've just started watching the series, but seems it might help me understand traditional channels vs audio objects. He has mentioned that previous surround (2D) is comprised of panning (and these "channels" comprise "the bed" in Atmos) and that is still included in the RMU (which handles all channels and objects). So it seems height information is handled as objects, and when one hears such and such movie was mixed in 7.1.4, that might be referencing the speaker configuration that was used for mixing:



Interesting that I've found another video in the series where he shows doing a pan and adding elevation. The referenced axis are different than what is used in 3D animation. In audio engineering Z is the height axis: in animation, Y is the height axis. Every discipline has its own terminology, and it's good that we can learn and get clarification of terminology.

 

[bigshot]

My dad gave me a more modest entry tier Denon 7.1 surround receiver when he upgraded to an Atmos enabled one. I wired it and tried a few different blu-rays and found it wasn't as enganging (music sounded flatter as there was less mid-bass and DR seemed a bit more reserved).

It may be underpowered for the number of speakers you're pushing with it. When you add speakers, you need a more powerful amp, because the power ratings get divided up like a pie between all the speakers. Your description perfectly describes an underpowered amp.

The example I gave, having two channel stereo and expanding it to 3 channels (with a centre channel) would appear to be very rare/exceptional and indeed there are no 3.0 mixes available to consumers as far as I'm aware.

There are a number of 3.0 SACDs in the RCA Living Stereo catalog. The added channel was recorded to allow them to adjust the soundstage in the mix. These SACDs are said to have a better defined soundstage because of the added channel. But that is the exception, not the rule. Most mixes are 4, 5 or 7 channel with or without LFE. I've heard of systems that have a center channel in the rear as well, but I don't know how that is handled. There are a lot of 4.0 recordings that don't have a center at all. I don't care for those as much as 5.1 because they tend to focus on sound coming from the corners of the room, rather than creating a soundstage or sound field.

I think the reason that these particular CDs have very little rear cnannel info and tightly defined center is because of remastering. Something they did in the re-release must have messed with the 90 degree phase, and they either didn't know or didn't care. I just got a copy of a Tomita CD that is titled Tomita in Surround. Interested to hear how the Dolby Surround is supposed to sound.

 

[Davesrose]

It may be underpowered for the number of speakers you're pushing with it. When you add speakers, you need a more powerful amp, because the power ratings get divided up like a pie between all the speakers.

Very true that amplification is a factor for sound characteristics. I've also seen that current quoted specs of AVRs show as rants for many A/V channels/forums (as they tend to just rate stereo channels and not a specified impedance). The hand me down Denon 7.1's specs quoted a power rating of 2 channels at 6 ohm. My older HK (that was more expensive at the time) quoted specs at 8 ohm. When I ran the calculation for figuring watts per 8 ohms, the Denon should have been about on par with my HK. But forums say that one of the nice features of Harman Kardon was that they tended to be conservative in their power ratings (so that actual RMS might be greater than other competitors).

There are a number of 3.0 SACDs in the RCA Living Stereo catalog. The added channel was recorded to allow them to adjust the soundstage in the mix. These SACDs are said to have a better defined soundstage because of the added channel. But that is the exception, not the rule. Most mixes are 4, 5 or 7 channel with or without LFE. I've heard of systems that have a center channel in the rear as well, but I don't know how that is handled. There are a lot of 4.0 recordings that don't have a center at all. I don't care for those as much as 5.1 because they tend to focus on sound coming from the corners of the room, rather than creating a soundstage or sound field.

That's interesting. I collected quite a few RCA Living Stereo SACDs. I even got one of Julian Bream that I also got as a used vinyl (to compare with debates about best format/mastering). I thought one of the best aspects of the RCA Living Stereo SACDs is that it should reproduce the studio master tapes. The tapes are also before the quadraphonic formats of the 70s. While looking up info about Dolby Atmos, I saw that cinema stereo could comprise of 3 speakers (wonder if there was some intention of them being in a cinema, or if it's just for mixing purposes of the soundstage). Also, when it comes to quad sound being part of a soundstage: it does depend on source. One surround SACD that I have that is the prime example of being quintessential in 4 channel sound is a re-issue quad vinyl record of Bach toccatas in which there are 4 organs in each corner. An exception for sure. From what I'm gathering about Atmos with an Atmos processor: there is more emphasis on localization based on decoded speaker system. Also, it's funny you mention a CD that has some of the matrixed old "Dolby Surround" (4.0) information. Not sure why the marketing folks at Dolby have decided to label the new 2D to 3D surround format "Surround".

 

[bigshot]

Are you running more speakers with the Dennon than you did with the HK?

 

[Davesrose]

Are you running more speakers with the Dennon than you did with the HK?

At the time, no. This was the same 7 speakers I've had for over 10 years (I've since upgraded my subwoofer with a custom built 12" one). Now I have a higher end Denon that can power 11 speakers (to give me 7.1.4).

 

[71 dB]

Spatial hearing hasn't developped for snare drum recordings with multiple contexts and artistical intent. It was developped to hear the lion roar nearby in the Safari so that you know to which direction to run away to not be eaten alive...

 

[Davesrose]

Spatial hearing hasn't developped for snare drum recordings with multiple contexts and artistical intent. It was developped to hear the lion roar nearby in the Safari so that you know to which direction to run away to not be eaten alive...

So we have to wait for this year's Lion King to hear true headphone 3D?

 

[gregorio]

[1] My language was based on Wikipedia. Wikipedia states that Atmos is comprised of 128 "tracks" (and each track can be assigned a channel or object). It also says that by default, there's a bed of "7.1.2" channels (leaving 118 tracks available for objects). That doesn't seem confusing to me, so not sure why you do want to be argumentative.
[2] It does seem the main point you've brought up that's against the literature is that the 128 "objects" are in addition to other mixed channels ...
[2a] I've seen an engineer say that Dolby recommends a 7.1.4 speaker configuration for home mixing and 9.1.4 for cinema mixing.
[2b] As for delivered audio formats, I had assumed cinemas could be similar to home.
[2c] Edit: I've just found a good series on Dolby Atmos from a sound engineer, and he goes through an example of doing an Atmos mix. I've just started watching the series, but seems it might help me understand traditional channels vs audio objects. He has mentioned that previous surround (2D) is comprised of panning (and these "channels" comprise "the bed" in Atmos) and that is still included in the RMU (which handles all channels and objects). So it seems height information is handled as objects, and when one hears such and such movie was mixed in 7.1.4, that might be referencing the speaker configuration that was used for mixing.

1. I agree, that doesn't seem confusing. However, an explanation isn't necessarily correct/accurate just because it "doesn't seem confusing", in fact the opposite is commonly the case. The reason I'm being argumentative (disputing the assertions) is because that explanation is NOT correct/accurate! In fact, Wikipedia actually contradicts itself. DCP (Digital Cinema Package) is the "package" of digital data delivered to cinemas which comprises the film. DCP is effectively a set of standardised specifications (defined by DCI/SMPTE) which ensures that a DCP will play in all cinemas. DCP specifies a SINGLE audio track, which must contain an uncompressed, 24bit, 48kHz or 96kHz 16 channel bwav file. Wikipedia states: "The Picture Track File essence is compressed using JPEG 2000 and the Audio Track File carries a 24-bit linear PCM uncompressed multichannel WAV file." - This then is a contradiction, if Dolby Atmos were 128 tracks then it couldn't actually be put into a DCP or distributed to cinemas. Even if we consider the term "tracks" to be interchangeable with the term "channels", that would still be impossible because DCP supports a maximum of 16 channels. BTW, in order to be completely technically accurate, DCP does optionally support multiple audio tracks but only one can be output at a time. This feature, when employed, is typically used for different language versions.

2. No, the main point I was bringing up is that tracks and channels are different things, that audio objects are a different thing again and this is important because the traditional consumer concept of channels and speakers does not apply. Case in point:
2a. Yes but that's speaker configuration, not track or channel configuration! Additionally, 9.1.4 maybe the theoretical minimum speaker configuration for cinema mixing, in practice it's much higher. The lowest speaker configuration I've ever seen was about 12.1.8 but the typical configuration in the dubbing stages used by Hollywood would be at least 24.1.16.
2b. As explained, not really. Cinema format is DCP with one audio track, comprising 16 audio channels. In the case of Dolby Atmos, 8 of those channels are used traditionally, IE. They contain 24/48 uncompressed wav audio data in the traditional 7.1 layout. The other audio channels do not contain audio data, they contain data which define each of the audio objects and their (3D, or really 2.5D) coordinates. Because of this, a cinema processor will not recognise these other channels as audio data and will just play the 7.1 channels, unless the cinema processor is a Dolby Atmos unit (RMU), in which case, it of course knows how to decode and use the data in those other channels.
2c. Good, there's hopefully little point in me continuing, as you seem to be grasping the concept.

I can almost hear bigshot saying something like: "All this info about theatrical systems might be somewhat interesting as an aside but it doesn't really concern me or other consumers because we don't have/use DCP or theatrical systems, we use (significantly different) BD/steaming media and home cinema systems". However, 5.1, 7.1, Atmos etc., were all originally designed as theatrical formats and then adapted for home consumer use and all films are mixed on theatrical systems (although often "re-versioned" - somewhat changed/adapted for home use). So the actual audio you are reproducing is largely (or even entirely) influenced by theatrical systems.

There are a number of 3.0 SACDs in the RCA Living Stereo catalog. The added channel was recorded to allow them to adjust the soundstage in the mix. These SACDs are said to have a better defined soundstage because of the added channel.

I didn't know there were some 3.0 SACDs. 3.0 can be a cinema format but is very rarely ever used, I've only ever seen it used for essentially "stereo" documentaries. However, 3.0 is incorporated within 5.1, 7.1 and other formats, it's effectively the front main speakers of all these formats and behaves in exactly the same way. Your stated "my point was that more channels mean more possibilities for sound location and more specific sound fields" - You are making and reiterating a point that is not correct (or that maybe correct but only in certain circumstances). In 5.1, 7.1, etc., the centre front speaker does not mean more possibilities for sound location, the possibilities for sound location are dictated by the front left and front right speakers and adding a centre speaker/channel does not increase or alter those possible locations. It does though, under certain conditions, make the phantom centre position more defined. This fact becomes even more important/relevant when we consider the Dolby Atmos format.

Spatial hearing hasn't developped for snare drum recordings with multiple contexts and artistical intent.

Exactly! So, as pretty much every song contains a snare drum recording with multiple contexts, pretty much every song should sound "unnatural" regardless of whether you reproduce it with speakers or HPs (with or without crossfeed), because neither speakers nor HPs (with or without crossfeed) can correct that "unnaturalness"!

G

 

[Davesrose]

1. I agree, that doesn't seem confusing. However, an explanation isn't necessarily correct/accurate just because it "doesn't seem confusing", in fact the opposite is commonly the case. The reason I'm being argumentative (disputing the assertions) is because that explanation is NOT correct/accurate! In fact, Wikipedia actually contradicts itself. DCP (Digital Cinema Package) is the "package" of digital data delivered to cinemas which comprises the film. DCP is effectively a set of standardised specifications (defined by DCI/SMPTE) which ensures that a DCP will play in all cinemas. DCP specifies a SINGLE audio track, which must contain an uncompressed, 24bit, 48kHz or 96kHz 16 channel bwav file. Wikipedia states: "The Picture Track File essence is compressed using JPEG 2000 and the Audio Track File carries a 24-bit linear PCM uncompressed multichannel WAV file." - This then is a contradiction, if Dolby Atmos were 128 tracks then it couldn't actually be put into a DCP or distributed to cinemas. Even if we consider the term "tracks" to be interchangeable with the term "channels", that would still be impossible because DCP supports a maximum of 16 channels. BTW, in order to be completely technically accurate, DCP does optionally support multiple audio tracks but only one can be output at a time. This feature, when employed, is typically used for different language versions.

2. No, the main point I was bringing up is that tracks and channels are different things, that audio objects are a different thing again and this is important because the traditional consumer concept of channels and speakers does not apply. Case in point:
2a. Yes but that's speaker configuration, not track or channel configuration! Additionally, 9.1.4 maybe the theoretical minimum speaker configuration for cinema mixing, in practice it's much higher. The lowest speaker configuration I've ever seen was about 12.1.8 but the typical configuration in the dubbing stages used by Hollywood would be at least 24.1.16.
2b. As explained, not really. Cinema format is DCP with one audio track, comprising 16 audio channels. In the case of Dolby Atmos, 8 of those channels are used traditionally, IE. They contain 24/48 uncompressed wav audio data in the traditional 7.1 layout. The other audio channels do not contain audio data, they contain data which define each of the audio objects and their (3D, or really 2.5D) coordinates. Because of this, a cinema processor will not recognise these other channels as audio data and will just play the 7.1 channels, unless the cinema processor is a Dolby Atmos unit (RMU), in which case, it of course knows how to decode and use the data in those other channels.
2c. Good, there's hopefully little point in me continuing, as you seem to be grasping the concept.

I can almost hear bigshot saying something like: "All this info about theatrical systems might be somewhat interesting as an aside but it doesn't really concern me or other consumers because we don't have/use DCP or theatrical systems, we use (significantly different) BD/steaming media and home cinema systems". However, 5.1, 7.1, Atmos etc., were all originally designed as theatrical formats and then adapted for home consumer use and all films are mixed on theatrical systems (although often "re-versioned" - somewhat changed/adapted for home use). So the actual audio you are reproducing is largely (or even entirely) influenced by theatrical systems.



I didn't know there were some 3.0 SACDs. 3.0 can be a cinema format but is very rarely ever used, I've only ever seen it used for essentially "stereo" documentaries. However, 3.0 is incorporated within 5.1, 7.1 and other formats, it's effectively the front main speakers of all these formats and behaves in exactly the same way. Your stated "my point was that more channels mean more possibilities for sound location and more specific sound fields" - You are making and reiterating a point that is not correct (or that maybe correct but only in certain circumstances). In 5.1, 7.1, etc., the centre front speaker does not mean more possibilities for sound location, the possibilities for sound location are dictated by the front left and front right speakers and adding a centre speaker/channel does not increase or alter those possible locations. It does though, under certain conditions, make the phantom centre position more defined. This fact becomes even more important/relevant when we consider the Dolby Atmos format.



Exactly! So, as pretty much every song contains a snare drum recording with multiple contexts, pretty much every song should sound "unnatural" regardless of whether you reproduce it with speakers or HPs (with or without crossfeed), because neither speakers nor HPs (with or without crossfeed) can correct that "unnaturalness"!

G

I’m not sure why you want to continue to assert a track is not an audio channel or object. Dolby’s own white paper on mixing with Atmos says that it comprises 128 tracks (5.1 or 7.1 or 9.1 channels taking up some of those tracks). I understand that this is for mixing, and may not be the same as a “track” on a video file. You say my use of the term “track” would make sense, but it’s incorrect. However, Dolby themselves say that a track can be a channel or audio object (in the context of a session).

https://www.dolby.com/us/en/technol...t-generation-audio-for-cinema-white-paper.pdf

 

[bigshot]

In the case of the Living Stereo 3.0 SACDs, RCA claims that playing them in 3.0 with a center channel more precisely defines the placement of the instruments in the center of the soundstage because the recording was made in three channels with a symphony orchestra with a mike on left, right and center. You could mix it to stereo and just have the phantom center serve, but having a separate channel covering the center provides more interior definition of sound sources in the middle-- woodwinds would be spread better, rather than just in an undifferentiated middle. That's what RCA claims at least...