DYohn
Emo VIPs 
Posts: 18,493
|
Post by DYohn on Jun 10, 2014 19:59:15 GMT -5
While you all debate, I'm enjoying my new Atmos soundbar!   |
|
|
|
Post by Dark Ranger on Jun 10, 2014 20:02:56 GMT -5
^ Now that's a man with his priorities straight.  |
|
|
|
Post by srrndhound on Jun 10, 2014 20:05:56 GMT -5
For those who haven't heard this before, Atmos treats each "track" as an "entity"... This is really useful for the mixing engineer, since you can independently move that object around, or adjust it, and Atmos supports a lot of objects. Today's film mixers operate in a Pro Tools world, where sounds are already treated as objects, segments of audio with associated metadata such as panning, level, etc, with no shortage of channels. What Atmos (and other "immersive object audio" formats) bring to the table is the height dimension, and an ability to export the audio without throwing away the individuality of the objects and the metadata. Forwarding some of that through to the playback system is where the adaptability comes from. There will indeed be extra work because consumer formats will not support as many objects as cinemas. But there is no reason to expect additional licensing fees, as to date Dolby has never charged licensing fees for encoded content. This is part of the extra work just discussed for mastering the home release. If the content is to fulfill the promise of improved dialog intelligibility under difficult circumstances, the dialog stems will need to be properly tagged and encoded. I share your concern that the ball may be dropped... But if not, it should be easy for the end user to get the benefit. Your speakers would have to be in rather radically incorrect positions before the normal mix would fail noticeably. And if they are radically wrong, like the surrounds are in the front, there's no fixing it with signal processing. The main thing is to have a second layer of speakers, 2 or 4 typically, and the benefits of the height signals will flow forth. But for the home, unlike the cinema, many of the benefits of object audio are other than the 3D spatial effect. The main benefit at home is not the ability to correct misplaced speakers, but as Andrew alluded, to allow the end user to decide how many speakers to use, not the content maker. No more up/down mixing (except legacy content :  I think you'll find Atmos will be delivered as an extension to TrueHD, not instead of it (at least in Blu-ray and its high bandwidth successors). That will ensure the bitstream will be compatible with current hardware. It's expensive to upgrade a cinema to Atmos. According to Film Journal: >>Built into the CP850’s list price of $33,750, even though “nobody ever pays that,” is Dolby’s Commissioning Service.<< That's rather more than the reported $6k for a Datasat AP-20. $100k for the whole conversion is not unheard of, especially if structural reinforcement is needed. That's why we tend to see 1 or 2 of a plex's 16 screens getting the upgrade, and the rest do not. Other audio solutions will come in at lower cost, so many more screens will get some form of object audio as time goes on. The reason DTS-HD MA overtook TrueHD was for very real differences. The main one is that while both lossless codecs also carry a "lossy" 5.1 version for player compatibility, in DTS the lossless is built as an extension to the lossy core, whereas in TrueHD they are separate components. The TrueHD streams are easier to decode, but HD MA is easier to encode and QC (1 pass vs. 2). Lower cost won out. |
|
|
|
Post by srrndhound on Jun 10, 2014 20:22:38 GMT -5
Not to leave you with the impression that the only home user benefit is the number of speakers, the following lists some of the ways object-based audio can bring benefits to a 5.1 speaker system.
ETA: 5.1 speakers driven by an object-based audio processor.
Case 1.
Let's say your TV screen is small, and that it would improve the enjoyment of the presentation if the onscreen sounds (sounds associated with a visible source) were to remain better tied to the screen size, thus avoiding exaggerated lateral cues for sources near the sides of the screen. This can be achieved with object rendering that is informed about the screen size. Then it will know how to map "screen edge" sounds to the phantom locations accordingly. As with any phantom image, it can shift depending on the location of the viewer. However, these phantoms are not built between L and R, but with L/C and C/R, so the error is cut in half. Furthermore, the error is never worse (and usually better) than the displacement error with standard L/C/R playback around a small screen.
Case 2. (reverse of case 1)
Let's say you are using your TV's internal speakers or a soundbar. The front sound stage will be no wider than the screen. If the device has some sort of stereo widener, then it widens everything, even the sounds that should have stayed near the screen. Object rendering allows the widening to be applied selectively, only to the ambience or music, while leaving the onscreen cues tied to the image.
Case 3.
It's night time, and turning down the volume obscures the dialog. DRC can help, but with object audio, DRC is a new ballgame because the dialog can be elevated independently of the rest of the mix. The mind boggles at the number of opportunities this presents for all manner of use cases; high ambient noise, hearing impairments, sleeping kids.
Case 4.
The surrounds are too far to the back of the room. Objects can be remapped by the renderer to adapt. So can the channel beds. If the speaker to which a bed channel is is assigned is not in the correct location, the renderer will have the option to revert to using that speaker's nominal position, treating it same as an object.
Case 5.
A 5.1 system has no height speakers. Yet there are ways to manipulate sounds to give an impression of height. With object audio, the system knows which sounds are elevated and which are not, and can funnel the elevated sounds to a "virtual height" processor without altering the other sounds.
The end user decides whether to use any of the above features or not.
|
|
|
|
Post by Jean Genie on Jun 10, 2014 21:12:41 GMT -5
This seems like awesome tech, almost like a precursor to holographic video. Will it be available in our automatically driven flying cars?  ?? Man, I love living here in Futureworld! |
|
|
|
Post by ÈlTwo on Jun 10, 2014 22:11:15 GMT -5
I thought what Keith said in his first post in this thread is what Keith said the first time he wrote about Atmos, but . . . The only thing about using Atmos to mix for 5.1 and 7.1 or 7.2 is that, when mixing, the editor must assume the exact positions of the speakers in the home theater. We all know that many a home theater does not have its speakers in the correct positions, and thus the Atmos mix may be anywhere from partially to significantly off. I hate to assume, but I will take the plunge here and assume that an Atmos processor will have the ability to be programmed for specific (dare I say nearly exact) speaker positions, either manually or automatically. Thus the Atmos mix will be decoded properly, despite speaker placement. No that is not the case. The Atmos-enabled AVR will 'know' where the speakers are in the the room and it will also know where the sound is meant to be spatially located. It will use a physical speaker where it finds one and it will phantom-image the sound to the correct place where there is no physical speaker in the right location. If you have your speakers in theoretically 'wrong' places, as many people do due to room constraints, WAF etc, then your Atmos-enabled unit will 'know' this and it will use the speakers you have to in order to place the sound object in the proper position - the position the mixer intended it to be in. emocustomer, Re-read my answer, in the first paragraph I'm referring to Atmos mastered material played on a non-Atmos AVR (as referenced in Keith L's post), that is why I refer to an editor. My second paragraph refers to a unit containing an Atmos processor, be it in a theater, AVR or other processor. |
|
|
|
Post by FilmMixer on Jun 11, 2014 1:03:26 GMT -5
Not to leave you with the impression that the only home user benefit is the number of speakers, the following lists some of the ways object-based audio can bring benefits to a 5.1 speaker system. Case 1. Let's say your TV screen is small, and that it would improve the enjoyment of the presentation if the onscreen sounds (sounds associated with a visible source) were to remain better tied to the screen size, thus avoiding exaggerated lateral cues for sources near the sides of the screen. This can be achieved with object rendering that is informed about the screen size. Then it will know how to map "screen edge" sounds to the phantom locations accordingly. As with any phantom image, it can shift depending on the location of the viewer. However, these phantoms are not built between L and R, but with L/C and C/R, so the error is cut in half. Furthermore, the error is never worse (and usually better) than the displacement error with standard L/C/R playback around a small screen. Case 2. (reverse of case 1) Let's say you are using your TV's internal speakers or a soundbar. The front sound stage will be no wider than the screen. If the device has some sort of stereo widener, then it widens everything, even the sounds that should have stayed near the screen. Object rendering allows the widening to be applied selectively, only to the ambience or music, while leaving the onscreen cues tied to the image. Case 3. It's night time, and turning down the volume obscures the dialog. DRC can help, but with object audio, DRC is a new ballgame because the dialog can be elevated independently of the rest of the mix. The mind boggles at the number of opportunities this presents for all manner of use cases; high ambient noise, hearing impairments, sleeping kids. Case 4. The surrounds are too far to the back of the room. Objects can be remapped by the renderer to adapt. So can the channel beds. If the speaker to which a bed channel is is assigned is not in the correct location, the renderer will have the option to revert to using that speaker's nominal position, treating it same as an object. Case 5. A 5.1 system has no height speakers. Yet there are ways to manipulate sounds to give an impression of height. With object audio, the system knows which sounds are elevated and which are not, and can funnel the elevated sounds to a "virtual height" processor without altering the other sounds. The end user decides whether to use any of the above features or not. Next time why don't you link to the original post that copied from... Original post by Roger Dressler at AVS |
|
|
|
Post by FilmMixer on Jun 11, 2014 1:26:41 GMT -5
For those who haven't heard this before, Atmos treats each "track" as an "entity"... This is really useful for the mixing engineer, since you can independently move that object around, or adjust it, and Atmos supports a lot of objects. Today's film mixers operate in a Pro Tools world, where sounds are already treated as objects, segments of audio with associated metadata such as panning, level, etc, with no shortage of channels. What Atmos (and other "immersive object audio" formats) bring to the table is the height dimension, and an ability to export the audio without throwing away the individuality of the objects and the metadata. Forwarding some of that through to the playback system is where the adaptability comes from. Pro Tools doesn't quite work as you described... Sounds don't contain metadata in regards to panning, eq, volume, etc.. broadcast wave files do contain time stamps. All of the other information that is created in the PT session is overlaid over the audio that is playing on a given track. And we don't operate in a Pro Tools world.. while it is certainly the industry standard for editing sound for film, it is still not ubiquitous in regards to mixing.. that is slowly changing, but a majority of Atmos films have been final mixed on a traditional console with Pro Tools feeding the audio as a source unit.. I believe that only 6 films so far were mixed "in the box..." Of the 12 Atmos stages in Los Angeles, only 2 are Pro Tools only Icon stages... all of the others have large format digital mixing consoles. Where Pro Tools is integral in regards to Atmos is the organization of what audio will be an object and the associated panning information that accompanies said audio. Any sound that is designated as an object still needs to be recorded with applied processing (eq, dynamics, reverberation, etc) as a "baked in" audio file with it's associated panning data.. Atmos also brings some other key differences besides object audio.. It adds two extra channels behind the screen. It uses full range surrounds.. in addition to the maximum 62 (+1 for traditional LFE with a twist) channel output, it also adds another sub output if desired, which can be used for bass management (the twist of the original .1 output is you can bass manage to that sub if you don't add additional subs in the theater proper) The rendering engine scales to the measured capabilities of the auditorium. Adds in automated room measuring and eq. MDA also includes many other advantages over what has come before us... |
|
|
|
Post by cwt on Jun 11, 2014 2:46:08 GMT -5
I have been reading a lot of posts in this new thread and there are some misconceptions about how an Atmos track is created and how it can benefit users if it comes home. Time is short now but over the next couple of days I will try and clarify some posts from others. In the mean time, I'm happy to answer any questions I may know the answers to. The Dolby site offers very easy to understand whilst papers about the format. A good place to start. I have a bit of experience with the format.  Hi Filmmixer ; something I read elsewhere And something emocustomer just said ; Are these statements referring to the same thing I assume so needing a proper atmos decoder to get the proper localization ? Will the atmos room eq be a replacement for what we would use now such as audyssey; ARC; emoq et al or complement it ? As an aside I read that ceiling speakers are localized to a high degree and wall speakers are much more apropos for atmos ; I can always get a smaller projector screen if that's the case |
|
|
|
Post by sme on Jun 11, 2014 3:55:03 GMT -5
Atmos mixes almost certainly require substantially higher data throughput than 5.1 and 7.1 mixes. I don't know what the actual figure is, but I doubt it can be transported over HDMI 2.0. (If anyone has this information who wants to volunteer it, I'm all ears!) HDMI 2.0 maxes out at 32 channels at 16-bits and 48 kHz. If you want higher resolution, then you have to give up channels. Other bandwidth limitations occur also with the Blu-ray format. Naturally, these formats can evolve to support it, but I imagine there are technical and political hurdles to address along the way.
|
|
|
|
Post by emocustomer on Jun 11, 2014 4:24:46 GMT -5
No that is not the case. The Atmos-enabled AVR will 'know' where the speakers are in the the room and it will also know where the sound is meant to be spatially located. It will use a physical speaker where it finds one and it will phantom-image the sound to the correct place where there is no physical speaker in the right location. If you have your speakers in theoretically 'wrong' places, as many people do due to room constraints, WAF etc, then your Atmos-enabled unit will 'know' this and it will use the speakers you have to in order to place the sound object in the proper position - the position the mixer intended it to be in. emocustomer, Re-read my answer, in the first paragraph I'm referring to Atmos mastered material played on a non-Atmos AVR (as referenced in Keith L's post), that is why I refer to an editor. My second paragraph refers to a unit containing an Atmos processor, be it in a theater, AVR or other processor. Misunderstood you. In a system without an Atmos-enabled AVR the sound will simply be the regular 5.1/7.1 as it has been designed to be backwards compatible. |
|
|
|
Post by emocustomer on Jun 11, 2014 4:27:17 GMT -5
Next time why don't you link to the original post that copied from... FM, I suspect there is a very good reason for the similarity (copy) of the posts made by Roger over at AVS which srrndhound himself will have to reveal, but which is known to many of us |
|
|
|
Post by emocustomer on Jun 11, 2014 4:30:00 GMT -5
Atmos mixes almost certainly require substantially higher data throughput than 5.1 and 7.1 mixes. I don't know what the actual figure is, but I doubt it can be transported over HDMI 2.0. (If anyone has this information who wants to volunteer it, I'm all ears!) HDMI 2.0 maxes out at 32 channels at 16-bits and 48 kHz. If you want higher resolution, then you have to give up channels. Other bandwidth limitations occur also with the Blu-ray format. Naturally, these formats can evolve to support it, but I imagine there are technical and political hurdles to address along the way. Atmos can be compressed and even Netflix etc will be able to use Atmos mixes, simply by compressing it down to DD+. No higher bandwidth capability needed for the bitstream. Unless I am misunderstanding your point.. Remember, Atmos isn't some half-mythical thing promised for the future - it is here now and on demo, using existing systems with no changes other than the AVR and (if desired) additional speakers. For me, one of the brilliant things about it is the way it is backwards-compatible with legacy gear, the way it can be carried on a standard Bluray disc, the way it will work with a soundbar or a full-blown multi-speaker setup, the way Netflix will be able to stream it etc etc. It is a genuine, major, revolutionary step forward in HT audio. It will eventually become as 'normal' as Dolby TrueHD or DTS-HD MA is today. |
|
|
|
Post by cwt on Jun 11, 2014 4:58:23 GMT -5
Next time why don't you link to the original post that copied from... FM, I suspect there is a very good reason for the similarity (copy) of the posts made by Roger over at AVS which srrndhound himself will have to reveal, but which is known to many of us Don't worry mate most here know what surroundhounds claim to fame with pl2x is  |
|
|
|
Post by sme on Jun 11, 2014 5:32:51 GMT -5
On the subject of Atmos mix-downs to Blu-ray: In The Hobbit: The Unexpected Journey, in the scene with Golem in the caves, the sound is convincingly 3 dimensional, including height, on my 5.1 system. Another film with an Atmos mix, Oz the Great and Powerful (hi filmmixer!) had even more compelling 3D imaging throughout, but if I understand right, the mixes on the Blu-ray were not derived from the Atmos mix.
While I'm at it, I'll mention that I have visited my local Atmos theater (a Regal RPX cinema) a few times. I don't visit often (and I stay away from the other theaters these days), but I always try to sit in the center, 2/3s back. I must say, the sound is better than any movie theater in my memory. The sound is very clean at high levels, and the subs put out plenty of bass under 30 Hz even. With that said, I prefer the sound of my home system. While the cinema can pan sounds above my head with a lot more precision than my system at home, my system at home images a lot better in the front 180 degrees, and in my opinion, that's where it matters most. I especially notice the difference when I listen after coming home from the cinema. The locations of sounds and instruments are spatially more distinct. If I could add a pair each of wides, heights, and rears, I can only imagine how much better it would sound. Maybe if I traveled to another state, I could find an Atmos cinema with better sound.
What I really hope to see with consumer Atmos is smarter handling of dynamic range (DR) and compression in such a way that "near-field" mixes would be made to retain the DR of the theatrical track. I don't like the fact that most Blu-ray releases have tracks mixed with reduced dynamic range. When the compression is done skillfully the result can be okay, but many major Blu-ray releases have mixes that are very loud and have very obvious clipping. It also frustrates me that there is no standard level for Blu-ray mixes. It is distracting to have to subjectively gauge the correct playback level at the start of a film. I typically adjust the level for subjectively natural sounding dialog, and find that my preferred level for Blu-rays typically ranges between -2 and -12 dB (where 0 dB is theatrical reference).
The sound tracks I play at closer to -2 dB typically sound a lot cleaner and more natural than the tracks I play at closer to -12 dB, despite my system having to produce peaks 10 dB higher. I realize most home listeners do not have a system with the headroom mine has, but I also know that many if not most home listeners find even the lowest level DVD and Blu-ray mixes to be too dynamic for their equipment or their listening environment. In my view, the typical DR reduction done for DVD and Blu-ray releases is a compromise that doesn't really serve anyone well. My hope is that Atmos objects and metadata improve the performance of DR compression (DRC) on consumer devices to the point that there is no longer justification for doing Blu-ray mixes at reduced level and everyone just mixes at theatrical reference level.
|
|
|
|
Post by sme on Jun 11, 2014 5:54:09 GMT -5
Atmos mixes almost certainly require substantially higher data throughput than 5.1 and 7.1 mixes. I don't know what the actual figure is, but I doubt it can be transported over HDMI 2.0. (If anyone has this information who wants to volunteer it, I'm all ears!) HDMI 2.0 maxes out at 32 channels at 16-bits and 48 kHz. If you want higher resolution, then you have to give up channels. Other bandwidth limitations occur also with the Blu-ray format. Naturally, these formats can evolve to support it, but I imagine there are technical and political hurdles to address along the way. Atmos can be compressed and even Netflix etc will be able to use Atmos mixes, simply by compressing it down to DD+. No higher bandwidth capability needed for the bitstream. Unless I am misunderstanding your point.. Remember, Atmos isn't some half-mythical thing promised for the future - it is here now and on demo, using existing systems with no changes other than the AVR and (if desired) additional speakers. For me, one of the brilliant things about it is the way it is backwards-compatible with legacy gear, the way it can be carried on a standard Bluray disc, the way it will work with a soundbar or a full-blown multi-speaker setup, the way Netflix will be able to stream it etc etc. It is a genuine, major, revolutionary step forward in HT audio. It will eventually become as 'normal' as Dolby TrueHD or DTS-HD MA is today. As I see it, an Atmos mix, down-mixed to 5.1 isn't an Atmos mix anymore. It's a 5.1 mix. I imagine that a clever decoder could up-mix the 5.1 channels mix to additional speakers and restore some of the spatial fidelity that was in the original, but this definitely represents a downgrade in quality versus a native Atmos format mix. For a lossless version of the original Atmos mix, how much storage is required, and what is the potential peak data rate? I'm not aware of AVRs or BD players installing lots of RAM to use for buffering, so these devices and the transports that connect them will need to be capable of providing the necessary bandwidth for the peak data rate. I don't think HDMI 2.0 is up to the task, but with buffering, it may be possible. Another issue may be that some Atmos mixes are simple enough to playback on consumer equipment without modification while others exceed bandwidth limitations and must be re-mixed. I imagine down-mixing objects to fewer objects is non-trivial to automate, but mixing down to beds is trivial. So perhaps consumer Atmos tracks will be more like legacy 7.1 soundtracks tracks, except that they'll have more channels and more metadata. Perhaps a few channels will be objects, which could be used to isolate dialog originating from a few different source locations, so it can undergo different DRC from the rest of the track. This approach isn't necessarily a bad one, but it's definitely not the same as theatrical Atmos. Before we all get carried away by the hype, we need to wait and see how this actually gets implemented and how it actually sounds. I'm also skeptical that adding ceiling speakers is a good idea in residential rooms with low ceilings, and I wonder if they are even required for convincing imaging above the head. |
|
|
|
Post by srrndhound on Jun 11, 2014 6:38:12 GMT -5
Next time why don't you link to the original post that copied from... Because I wanted to make some small edits to my original post to remove reference to MDA, and to make it easier for anyone with a question to quote within the Emo forum. Sound good? |
|
|
|
Post by srrndhound on Jun 11, 2014 6:45:56 GMT -5
Pro Tools doesn't quite work as you described... Sounds don't contain metadata in regards to panning, eq, volume, etc.. That's good, because I did not say they did. I said "associated," not "contained." This is a chat forum, not an AES conference. I err on the side of conceptual descriptions where possible so as to not get too deep into minutiae. As one of the inventors of MDA, I think I know how this stuff works. Carry on. |
|
|
|
Post by srrndhound on Jun 11, 2014 7:01:13 GMT -5
According to roger dressler truehd can accommodate atmos in the extension data, which would mean backwards compatibility with all current decoders. The same goes for the dts version of atmos. And something emocustomer just said ; Are these statements referring to the same thing Let's ask him: Yes. Same thing. To get any of the Atmos benefits, yes, you will need an Atmos decoder. If Dolby provides room EQ, it is separate from Atmos, per se. Atmos mixes almost certainly require substantially higher data throughput than 5.1 and 7.1 mixes. I don't know what the actual figure is, but I doubt it can be transported over HDMI 2.0. Since HDMI 1.3 is capable of object audio delivery, I do not see why 2.0 would have any trouble. It is not necessary to decode the audio before passing it over HDMI. In fact it is preferable not to do that so the AVR's speaker setup details can optimally influence the audio rendering. Note that home versions will likely carry fewer objects than the cinema soundtrack does in order to limit bandwidth and DSP processing requirements. |
|
|
|
Post by FilmMixer on Jun 11, 2014 12:53:50 GMT -5
emocustomer, Re-read my answer, in the first paragraph I'm referring to Atmos mastered material played on a non-Atmos AVR (as referenced in Keith L's post), that is why I refer to an editor. My second paragraph refers to a unit containing an Atmos processor, be it in a theater, AVR or other processor. Misunderstood you. In a system without an Atmos-enabled AVR the sound will simply be the regular 5.1/7.1 as it has been designed to be backwards compatible. Which is why I stated that Atmos offers no advantage for those with 5.1 systems on the other thread... where you then told everyone that I was wrong... |
|
) One alternative would be to assume that you have 5.1 or 7.1 speakers "in normal locations" - but, if you do that, then you've thrown away a lot of what differentiates Atmos.
