The end really catched my ear: Both ambisonics and Atmos can be used also in game design, but in other ways than wwise etc. How are they different and how can they still be used for game audio what are their limitations and why? Thank you for these awesome videos! 💯
I accidentally deleted a comment that was quite interesting. So sorry for that! The question was if you can convert an Ambisonics file into a Dolby Atmos Masterfile and how you can do that. The answer is, yes, this is quite straightforward to do if you have Nuendo. In Nuendo you can just setup an Ambisonice track and route it into the Atmos bed.
If you are on a Mac you can do this with the external Dolby Atmos Renderer and Envelop for Live. If you are on Windows you are out of luck, i‘m afraid.
Excellent overview, and important point at the end about fixed location- really important for people making interactive experiences. Thanks for the great work
Great video and explanation of both, though I'm kinda missing the real comparison of both. Which one is easier to set up and learn? Which one is more open and adaptable to more environments? Which one might benefit which use-case more? (Like Binaural) Which one takes overall less CPU to compute? Would be Ambisonic B Format as widely usable as Atmos or even more, if manufacturers would give a damn? Which one is more likely to be widely adopted or avaliable for longer? Or why is Atmos present in so many things even though I personally don't see a real advantage other than corporate push in all directions? (Yes I'm beginning to sound bitter, but I see so many people complaining about Atmos, without even looking at Ambisonics even though it could be the better alternative for many)
These are very different technologies that are used is different use cases. Atmos’advantage is the channel based component that makes it gracefully downgradable to older systems.
I've had ambisonic CDs for over 30 years by Nimbus. I am glad that I can use my Meridian DSP processor to decode it, although I suspect you're talking about a complete different technology, not the UHJ analogue format 😊
One thing to note regarding ambisonics decoders: there is an objective standard for decoding ambisonics, even if there are different decoders. One thing to note about either of these, is that there is a benefit to spatial audio even on stereo devices, because there is a difference between ideal headphone decoding and ideal front stereo speaker decoding.
As for listener position in ambisonics, there is a method for ambisonics which accounts for listener position, and I'm working on a prototype for producing it and decoding it.
Wow, this is great technical in-depth explanation of Atmos, thank you! I came to you from a consumer side, trying to figure out why my new Q990c soundbar is limited to 7.1.4 channels output with Atmos input in it's "purist" standard mode, despite having drivers for 11.1.4 channels and up mixing to all of those 11 in surround modes... I understand now it must have to do with underlying "bed" but still cannot answer the question: why does Samsung limit it to 7.1.4 output when all 11 drivers could've been used to render objects more precisely in space derived from Atmos metadata? 🤔
thanks for this video, just wanted to mention, that with Davinci Resolve 19, there is full Ambisonics support as well, next to Atmos support... maybe you could you post a video on that?
Hi Michael, thanks so much for this video it's really helpful. I notice at the end you said that to create an immersive audio experience that enables freedom of movement you don't recommend either ambisonics, or dolby atmos? I want to create an immersive, geolocated sound field sound walk experience but it needs to be accessible for people using normal headphones...for that reason I was thinking ambisonics might be best...but now I'm not sure....do you mind if I ask what you'd advise?
Ambisonics (as well as Atmos) is only advisable if you don’t move too far away from the sweetspot. For installations that require full 6 degrees of freedom movement, the best way to go is through a game engine. That is a bit more complex though because you also need to track the position of the listener. What area is your sound field experience supposed to cover?
If I understand correctly, Atmos objects can only be the output of mixing where someone specifies the position and size of each object? Conversion from mid-side microphones goes into a regular surround bed. What methods does JOC use to cut an object out of the 5.1? That seems to be impossible like unscrambling an egg. I heard that the sum of object channels has only around 192 kbit/s, which is not enough for saving a copy for direct subtraction.
You could also use objects for the conversation. I am not sure how JOC encodes Dolby Atmos objects but it is certainly lossy in terms of spatial properties.
I'm not sure why you say "there's no such thing as a negative microphone." Just invert the polarity. That said, it's more flexible to account for that with the transfer matrix to spherical harmonics (B format.) That can handle arbitrary microphone layouts in addition to necessary inversions.
@@michaelgwagner yes, you'd need bidirectional mics to pick up the opposite side. You'd still want to convert with a transfer matrix, so there's no practical benefit. Nonetheless, I had to point out that you can have a negative microphone.
I've just spent the morning watching your very helpful videos and experimenting on my Cubase 12 pro. In the past, I have used ambisonics via ATK in Supercollider and later Reaper; I've used an number of the free open source ambisonics plugins, e.g., Sparta, IEM, in Reaper and Cubse. Now I'm trying to get my head around Dolby Atmos. Given that my main output target in stereo/binaural via headphones, I have to say, "Why bother with Atmos?" It's easy-peasy to create a binaural export (even with the Steinberg ambi environment) on the other hand, I'm scratching my head after an hour of 'this and that' trying to get a binaural or stereo file out of a rudimentary Atmos project. Or even find information about doing so. I'm sure it's me; but really... ! Not a criticism I hasten to say; just a kind of 'if it ain't broke, don't fix it' comment. Any suggestions? (I'm a masochist at heart.) And, yes, thanks for all the work.
It is generally easier to get good results with Atmos because of its close relationship with channel based formats. In Ambisonics you do depend a bit on the quality of the decoder. But if you are happy with your workflow and results and produce stereo only then Atmos is likely overkill.
@@michaelgwagner Thanks for the quick response. I *do* like the quality of the binaural monitoring in the project that I set up as per your tutorials and the docs -- as 7.1.2/4. But when it comes to Rendering/Exporting the results via the Export function... no audio at all. There is something missing in my workflow (and understanding) at this point and I can't find the information to fill in the blank. Suggestions?
@@michaelgwagner OK. Wave off. Confusion reigns but there is now stereo ouput. But on another matter: Is it the case that the Mix Console does not display levels for tracks sent to the Atmos Group while the faders *do* change the levels? This may have something to do with where the ADM utility installs the rendering Insert?
@@michaelgwagner Much obliged. Sorry for the nits. There's a lot to cover and with a great deal of detail. I will watch all of all you vids on this topic now. :)
Hey Michael, awesome video as always. Im currenty going through your unity/wwise tutorials in order to be able to use it in my master thesis. Im not quite sure what the correct approach is though. Ive built a 2D game with a fixed listener/player in the center and objects that have indiviudal sound sources attached to them that spawn at a radius around the listener. They take different paths towards the listener position and once they are close enough you can interact with them. This is an audio game without visual aid so you have to able to track the objects movement very precicely. Somehow i struggle to wrap my head around whether 3D spatial audio is the only approach i can take? As i understand it ambisonics and dolby atmos are primarily used for immersive sound scenes with a global sound object and not so much for moving sound objects? I guess i could simulate all the paths and then bounce the audio with sth like Envelop in Ableton but id like to be able to adjust the sound objects movement freely without having to create new audio files. What would you suggest? Thanks!
In game design you would render the audio in real time, so you would simply attach the audio to the object (in Unity 2d games are just flat 3d games). A game designer would touch spatial audio only for ambiances. Everything else are usually just mono objects.
@@michaelgwagner Thanks for the answer! By spatial audio you mean ambisonice/dolbyatmos or even the 3D spatial audio in unity? Can you think of a way to make tracking of the soundobject with attached sound audiobjects easier? For now i have a hard time even following the object circling around the center/player and it doesnt seem to matter whether i use the 2D or 3D option unless i use reverb zones which makes tracking of the objects a little easier.
I know Im just a bachelor student in music technologies, but I am writting my bachelor project on mixing in ambisonics and quite honestly, while you have said many things correctly about ambisonics - I would object some that you seem to be extremely sure of, and the only explanation I would have for your strong beliefs is faulty loyalty to Nuendo and to Atmos. First, going into higher order ambisonics is encouraged by everyone who works with it and every book or article Ive read about it - some of the most magically sounding spatial things Ive heard are based on Higher Order, at least 5th or 7th order. Second, I was enraged when I heard you say that the similarity Atmos and Ambisonics share is the sweetspot. Excuse me, but WHAT THE F? Yes, while recording you have the center of the soundfield, but when it comes to monitoring and mixing, there are academic papers written on ambisonics and how using higher order ambisonics enlarges the listening area. Let me explain. Even in 1st order ambisonics, the listening area is not a spot, but an actual area, in which you can move around without losing the spatial experience, and only by walking too close to the speakers you lose that feeling and the experience drops down to listening to individual speakers. But the higher order you use - the larger that area of spatial experience becomes. Even at 5th order, you already have most of the area between the speakers covered, and you have to go pretty close to the individual speakers to lose that spatial experience. And no, I did not only read about this, but felt it and heard it myself in a spherical system. And no, I would not consider Cubase implementation to be great. I tried it out and I hated it. Its automated to the level of confusion. The processing chain is too rigid for me to do what I want to, and the detailed tailoring of some of the sound parameters is dumbed down to a single knob, while in Reaper I could freely adjust my processing chain and use advanced plugins that actually let me do what I need to. So no, I dont like ambisonics in Cubase, they are somewhat ok to mix for binaural purposes or 5.1/7.1, but I wouldnt trust it with any advanced ambisonic systems. And I have heard of game engines that rely heavily on ambisonics to render most of the sound in their games. So, to sum it all up - I find your knowledge on ambisonics either insincere or lacking. Yes, ambisonics is hard to understand, yes, its hard to market successfully, which is mostly why it has been in the academic community for over 50 years and only gets utilized and shared with the consumers in its simplest forms, but by god, dont downplay it, because it might bite you in the rear end in the future.
Well, first of all, thank you for your comment. It appears that you are confusing the quality of the spatial experience with its representational accuracy. Let's say you are in the field recording with an ambisonics microphone and you position the mic next to a tree while recording the ambience around you. That tree will influence the way sound is received at the mic. If you later want to recreate that experience with an ambisonics speaker system, that experience will only be accurate at the sweet spot. As soon as you move away from it, you will no longer perceive the correct auditory impression of that tree at that location. This is not a question of filling the sound between the speakers, it is a question of accurately representing the sound space around you. In that sense, ambisonics and Dolby Atmos are similar. They are only mathematically accurate at the sweetspot. Having said that, with a real life ambisonics speaker system, you can probably move around a lot before this becomes an issue. The human brain can compensate a lot. However, that changes immediately as soon as you are talking about the use of ambisonics in cinematic VR or game design. In both cases you are not dealing with physical speaker systems. Rather, you are positioning the ambisonics audio in 3d space directly. Therefore, understanding that the auditory impression is no longer accurate as soon as you move too far away from the sweetspot is absolutely critical in these fields. The order of the ambisonics can bee seen as its resolution. The higher the order, the more resolution you get from the higher order spherical harmonics. What you are talking about seems to indicate that in ambisonics reproduction, the resolution of the speaker system (i.e. the number of speakers) might be less important than the resolution of the recording (i.e. the ambisonics order). That is interesting by itself but only applies to our understanding of ambisonics reproduction systems and not so much ambisonics itself.
@@michaelgwagner Well, your assumption of my confusion only applies when talking about mathematical accuracy and actual recordings. But what if we were to talk about the spatial objects we create through sound synthesis or encoding mono or stereo tracks of instruments? We cant quite say we are mathematically accurately representing them, since their spatial information has not been recorded - nothing to actually represent. And yes, the ambisonic order we use as an input makes a fair difference compared to what order system we are using. Its a simple matter of whether or not we are supplying the system with lacking, abundant or just enough level of information. To say that the input resolution is unimportant is to say that it doesnt matter what we supply the system with - the sound quality will stay the same. And its obvious that its not the case. Yes, the resolution of the system is very important, binaural or surround reproduction will not be as accurate as spherical systems, but if we supply spherical systems with only 1st order or 0 order information - we wont get the spatial experience we desire, which then defeats the whole point of using such a system. To me, it makes sense to supply a high quality system with abundant level of information. And when it comes down to virtual 3d space - its mostly binaural applications, which dont quite care if you move or not. Well, unless you are using the head tracker. And yes, in real ambisonics systems you can move quite a lot even with 3rd and 5th order recordings - I could walk around 2 tables freely without feeling like Im listening to individual speakers, and that is the real ambisonics experience - to live in the soundfield we are producing. While mathematical accuracy is important to us engineers, but if we want to create the right experience for the listener we need to find a way to connect with them. And that, so far, has been the ability to walk around virtually freely in said soundfield.
Greetings from Peru, a query I want to remix a production made in Protools, this audio has to be exported from ProTools in WAV Stereo or has to be exported in another way to be binaural, and this audio I want to adhere to a normal RUclips video, you have to encode in a special way the audio and video for binaural? or add any metadata? Thank you for your reply.
Super helpful video! 11:40 I'm just starting to mess around with Ambisonics using ProTools. I'm wondering why PT didn't make this list. It seems to have all the implementation, 1st, 2nd, 3rd order ambisonics, plus TBE plugins. Is it just a personal pref, or are there good reasons for me to consider one of these DAWS if I want to get serious about Ambisonics. Thx!
Very good question! At the time I recorded this you needed the Pro Tools Ultimate version to gain access to Ambisonics. And that was prohibitively expensive. The inclusion of Ambisonics in the base version of Pro Tools is a very recent development.
@@michaelgwagner Thanks. Well, I have discovered one problem with Pro Tools. The TBE Spatializer video overlay function is broken on my M1, and the components needed to make it work won't install. It looks like some ppl have gotten it to work, but it's complicated. Thanks again!
Hello, Mr. Wagner. Thank you for providing and sharing great information. I really appreciate it. I've been using both of Ambisonics and Dolby Atmos for a while and I was wondering if Ambisonics and 'Atmos objects' are using the same technology. If you set up speaker layouts in your room simillar to 7.1.4 in Ambisonic using reaper, would it be same as Atmos objects? Since both are using object-based audio technology... I just assumed 'Atmos objects' borrowed and using the Ambisonic tech and just added surround tech for 'Atmos beds' which is channel based.
Ambisonics is not object based. But you can use an Atmos speaker system to play Ambisonics. You would need a proper encoder to do that. Some Ambisonics plugin suites include a decoder with such a speaker setup.
@@mjmj7170 You could use IEM's AllRA Decoder, write in your speaker setup, balance it out with imaginary speakers (usually one at the bottom below the listener), press to calculate the decoder and then add distance compensator plugin. Voila, your mix is decoded for the exact system you are using.
Just to make sure, I understand this correctly: If one uses three figure of eight mics (one for x,y,z axis) and an omni, he can order them as w,y,z,x. Isn't that AmbiX (B-format)?
No, because some of the pickup patterns are negative, which is physically impossible to realize with an actual microphone. That’s why they are called “virtual microphones”.
Hey man, good video, learn a lot. Thanks. BTW, do you know how to record the Atmos audio while playing Blue-Ray Movies (whose audio track is Atmos) and AAA Games (whose audio setup supports ouput Atmos)? Thanks in advance!
@@michaelgwagner Nuendo can import the ADM file (*.wav) exported by itself,and recover all VST automation to allow modifications . And this is what I mean the JOC means. If it doesn't match your understanding, I'm sorry for not describing correctly before.
@@michaelgwagner However, I did find that Nuendo imports the exported Atmos file and restored the audio assets with the exact same VST automation as I had exported (by Nuendo).
I have a question, is that possible to transform ambisonic audio captured with the H3-VR into a mono track without choosing a specific directionality? The reason is that I am trying to capture audio in all directions but want to make it as fast as possible to deliver, hence in mono. I basically capture videos with a 360 camera which I transform into normal videos by reframing them. But the native audio quality of the camera is pretty bad so I would like to capture with an external mic without losing the spatial capture. ▼ vidIQ badge
@@michaelgwagner @Michael G Wagner Oh ok it's that simple! Thank you!! I am actually doing some tests. When I import inside DaVinci Resolve the raw audio file from the Zoom H3-VR and drag it to the timeline I get the 4 mono tracks of each mic. Since I basically have the 360 camera on a hand-held stick I could only put the microphone at the bottom of the stick which is a little far from the mouth of the speakers which are at camera level. So what I am trying to do is to for example cut or reduce by half the audio level of the two microphones that point downwards to the floor from the stick so I have more of the upward audio. I feel this is getting somewhere! :) It's kind of as if my 360 video was reframed but my audio was left in equirectangular format if I do a metaphor on the image. It's a good quick solution for now. Ideally, I would find a way in the future to apply the image reframing directionality automatically to the audio. (I use the KartaVR plugin in Davinci Resolve for Reframing)
@@michaelgwagner Oh I just learn that thanks I am been using both dear VR Pro in Logic but I donot know how to set it up in Reaper for binaural 13.1 surround sound this would be for film and Video games redesign I learning maybe when time allows You could do tutorial on this Thanks Sir
I was wondering about the potential in game engines in particular. You would need to deal with 0 mics but that would have issues I assume? And then playback would be an issue? Unless there is a compatible receiver. Or audiochip. But Cmedia is not playing, XMOS could be done but I am not a star ar this at all. I think using a 4.1.2 setup, maybe using 2 lower placed speakers would be wonderful, pinpointing sound like nothing else. I wonder why this (as far as I could find) as a no longer patent encumbered technology has not caught on at all besides a few repo's by Google in the shape of Resonance Audio (with few if any updates to any of them) and even then aimed at VR only.
I think he is doubt the real value of Ambisonics&Atmos in games, actually I wanna ask also......since most player's speaker is headphone or external stereo speaker(maybe 99%?), can player feel the difference between stereo and those two techs??? Does it improve a lot game experience???@@michaelgwagner
@cabab Well, when I talk about spatial audio in games I am usually talking about the development part of it and not so much the actual gameplay. On the development side it isn’t even a question, you need a spatial audio approach, otherwise you are not going to feel immersed. On the gameplay side things are different. It really depends on your listening environment.
Thx for your patient reply!!!!!Actually I'm newer to this industry and I'm working for a new project now, but, for example, UE5 have a lot of feature, such as "Soundscape" plugin for environment's sound building, or Spatialization、Air Absorption.... setting in "Attenuation" module. So if I use a lot of these thing, can it approach the effect of "Ambisonics&Atmos"? Or, UE5's built-in settings can replace those two tech???? Does that mean if I don't use those two tech or other additional tech mean I can't build immersed feeling?? so complex so confused......need some tips....QwQ@@michaelgwagner
Game audio is object based by default. In that sense Dolby Atmos isn’t really something you need to think about. Game Engines can generate Atmos automatically. You would use Ambisonics primarily if you want to bring in real world ambience recordings. For everything else, you would either use the tools of the game engine, or the audio middleware in case you are using one.
Hi. I have hard time to understand what is going on in ambisonic player meant for H3-VR. Are you familiar with this microhpone? After watching (please open the link) this I am confused ruclips.net/video/kIquru2hAH4/видео.html. If I have recorded Ambix audio from this Mic and open in Zoom Ambisonic Player. I can choose Stereo ,Binaural 5.1, surroound and Custom LR for listening and export. What and where is listening position? What does it mean in such reality when I have just simple stereo speaker setup? What these blue arrows really mean and why they are at angle? What is stereo mode? I know binaural stereo is for headphones and it would be like placing binaural microphone in center of sphere. Stereo modes are mystery for me. How software render stereo mix from ambix audio what are priciples? Why I hear all sounds in speakers even recorder below and above in this stereo mode? And big mystery is Custom LR. How it is relflected in stereo speakers?I can adjust sphere angle they say It is changing listening position?? I can adjust L R in random positions on sphere ("you can adjust listening position of the left and right channels independently through out 360 space") , they say these are channels.I dont know what it means ...Maybe we can adjust stereo width or something thatway??? How the thing corresponds to real stereo speakers setup? Thank You Michael if you willing to answer.
Yes, we have one of those. The Zoom player is just helping you to convert the recording into one of the standard formats. Since it records in Ambisonics you can rotate the signal around or convert it into stereo based on custom spesker positions. If you are only interested in the Ambisonics recording (Why would you use an Ambisonics mic if you only want to record stereo?) then you do not need most of the functionality of that player.
dolby labs atmos is junk . ambisonics is the way it should be captured and mixed as other movie mixes in stereophonic to dolby labs is junk rubbish . ambisonics is miles better
The end really catched my ear: Both ambisonics and Atmos can be used also in game design, but in other ways than wwise etc. How are they different and how can they still be used for game audio what are their limitations and why?
Thank you for these awesome videos! 💯
Thanks! Those are big questions. I might do a video.
@@michaelgwagner Thats what I was thinking, these questions might be a good idea/theme for a video! :)
I accidentally deleted a comment that was quite interesting. So sorry for that! The question was if you can convert an Ambisonics file into a Dolby Atmos Masterfile and how you can do that. The answer is, yes, this is quite straightforward to do if you have Nuendo. In Nuendo you can just setup an Ambisonice track and route it into the Atmos bed.
Is there a way to do this if your DAW is Ableton Live? Are there any free tools available?
If you are on a Mac you can do this with the external Dolby Atmos Renderer and Envelop for Live. If you are on Windows you are out of luck, i‘m afraid.
I don't know where you get the time to create all these fantastic videos! Very helpful and well explained. Thank you so much!
I sometimes wonder that myself. Lol. Thanks!
holy shit, that was explained so well! I needed a quick refresh on the topic and that was just perfect, thanks!
Thanks! :)
Excellent overview, and important point at the end about fixed location- really important for people making interactive experiences. Thanks for the great work
Check out our answer above :-)
Great video and explanation of both, though I'm kinda missing the real comparison of both.
Which one is easier to set up and learn?
Which one is more open and adaptable to more environments?
Which one might benefit which use-case more? (Like Binaural)
Which one takes overall less CPU to compute?
Would be Ambisonic B Format as widely usable as Atmos or even more, if manufacturers would give a damn?
Which one is more likely to be widely adopted or avaliable for longer?
Or why is Atmos present in so many things even though I personally don't see a real advantage other than corporate push in all directions? (Yes I'm beginning to sound bitter, but I see so many people complaining about Atmos, without even looking at Ambisonics even though it could be the better alternative for many)
These are very different technologies that are used is different use cases. Atmos’advantage is the channel based component that makes it gracefully downgradable to older systems.
I've had ambisonic CDs for over 30 years by Nimbus. I am glad that I can use my Meridian DSP processor to decode it, although I suspect you're talking about a complete different technology, not the UHJ analogue format 😊
Probably… 😄
Coming back for a refresher, Thanks for this Michael. You are genius
Thanks!!
Many thanks for this video, you did a way better job of explaining it than my uni did! =)
Haha, thanks! Much appreciated.
great concise video
Thanks!
One thing to note regarding ambisonics decoders: there is an objective standard for decoding ambisonics, even if there are different decoders.
One thing to note about either of these, is that there is a benefit to spatial audio even on stereo devices, because there is a difference between ideal headphone decoding and ideal front stereo speaker decoding.
As for listener position in ambisonics, there is a method for ambisonics which accounts for listener position, and I'm working on a prototype for producing it and decoding it.
Very interesting!
Great informations Wagner. Thank You!
Thanks!
awesome! subscribed
Very interesting and knowledgeable!
Glad you enjoyed it
Great, clear and concise information. Thank you.
You’re very welcome!
Thank you for making this video , liked and subscribed !
You are very welcome, in every meaning of the word. ;)
Wow, this is great technical in-depth explanation of Atmos, thank you! I came to you from a consumer side, trying to figure out why my new Q990c soundbar is limited to 7.1.4 channels output with Atmos input in it's "purist" standard mode, despite having drivers for 11.1.4 channels and up mixing to all of those 11 in surround modes... I understand now it must have to do with underlying "bed" but still cannot answer the question: why does Samsung limit it to 7.1.4 output when all 11 drivers could've been used to render objects more precisely in space derived from Atmos metadata? 🤔
Thanks! 7.1.4 is sort of the home theater standard layout.
thanks for this video, just wanted to mention, that with Davinci Resolve 19, there is full Ambisonics support as well, next to Atmos support... maybe you could you post a video on that?
Yes, should come up in the not too distant future. 😉
Hi Michael, thanks for this new interesting video. By the way, you forgot to put the links in the description below the video.
Yes, thanks. I wanted to add them today but then got lost in meetings for my day job.
Awesome need more contents like this
Thanks!
Hi Michael, thanks so much for this video it's really helpful. I notice at the end you said that to create an immersive audio experience that enables freedom of movement you don't recommend either ambisonics, or dolby atmos? I want to create an immersive, geolocated sound field sound walk experience but it needs to be accessible for people using normal headphones...for that reason I was thinking ambisonics might be best...but now I'm not sure....do you mind if I ask what you'd advise?
Ambisonics (as well as Atmos) is only advisable if you don’t move too far away from the sweetspot. For installations that require full 6 degrees of freedom movement, the best way to go is through a game engine. That is a bit more complex though because you also need to track the position of the listener. What area is your sound field experience supposed to cover?
If I understand correctly, Atmos objects can only be the output of mixing where someone specifies the position and size of each object? Conversion from mid-side microphones goes into a regular surround bed. What methods does JOC use to cut an object out of the 5.1? That seems to be impossible like unscrambling an egg. I heard that the sum of object channels has only around 192 kbit/s, which is not enough for saving a copy for direct subtraction.
You could also use objects for the conversation. I am not sure how JOC encodes Dolby Atmos objects but it is certainly lossy in terms of spatial properties.
I'm not sure why you say "there's no such thing as a negative microphone." Just invert the polarity. That said, it's more flexible to account for that with the transfer matrix to spherical harmonics (B format.) That can handle arbitrary microphone layouts in addition to necessary inversions.
Interesting point, thanks. The argument is that you would have to invert the polarity for only a part of the pickup pattern.
@@michaelgwagner yes, you'd need bidirectional mics to pick up the opposite side. You'd still want to convert with a transfer matrix, so there's no practical benefit. Nonetheless, I had to point out that you can have a negative microphone.
Yes. thanks. I might have oversimplified.
Would be great to find a comparison of 4k, 8k and 9k recording setups and its benefits…
I've just spent the morning watching your very helpful videos and experimenting on my Cubase 12 pro. In the past, I have used ambisonics via ATK in Supercollider and later Reaper; I've used an number of the free open source ambisonics plugins, e.g., Sparta, IEM, in Reaper and Cubse. Now I'm trying to get my head around Dolby Atmos. Given that my main output target in stereo/binaural via headphones, I have to say, "Why bother with Atmos?" It's easy-peasy to create a binaural export (even with the Steinberg ambi environment) on the other hand, I'm scratching my head after an hour of 'this and that' trying to get a binaural or stereo file out of a rudimentary Atmos project. Or even find information about doing so. I'm sure it's me; but really... ! Not a criticism I hasten to say; just a kind of 'if it ain't broke, don't fix it' comment. Any suggestions? (I'm a masochist at heart.) And, yes, thanks for all the work.
It is generally easier to get good results with Atmos because of its close relationship with channel based formats. In Ambisonics you do depend a bit on the quality of the decoder. But if you are happy with your workflow and results and produce stereo only then Atmos is likely overkill.
@@michaelgwagner Thanks for the quick response. I *do* like the quality of the binaural monitoring in the project that I set up as per your tutorials and the docs -- as 7.1.2/4. But when it comes to Rendering/Exporting the results via the Export function... no audio at all. There is something missing in my workflow (and understanding) at this point and I can't find the information to fill in the blank. Suggestions?
@@michaelgwagner OK. Wave off. Confusion reigns but there is now stereo ouput. But on another matter: Is it the case that the Mix Console does not display levels for tracks sent to the Atmos Group while the faders *do* change the levels? This may have something to do with where the ADM utility installs the rendering Insert?
The Atmos panner sends the audio to the renderer before it would hit the meter. You need to set the global meter settings to “pre-panner”.
@@michaelgwagner Much obliged. Sorry for the nits. There's a lot to cover and with a great deal of detail. I will watch all of all you vids on this topic now. :)
Hey Michael, awesome video as always. Im currenty going through your unity/wwise tutorials in order to be able to use it in my master thesis. Im not quite sure what the correct approach is though. Ive built a 2D game with a fixed listener/player in the center and objects that have indiviudal sound sources attached to them that spawn at a radius around the listener. They take different paths towards the listener position and once they are close enough you can interact with them. This is an audio game without visual aid so you have to able to track the objects movement very precicely. Somehow i struggle to wrap my head around whether 3D spatial audio is the only approach i can take? As i understand it ambisonics and dolby atmos are primarily used for immersive sound scenes with a global sound object and not so much for moving sound objects? I guess i could simulate all the paths and then bounce the audio with sth like Envelop in Ableton but id like to be able to adjust the sound objects movement freely without having to create new audio files. What would you suggest? Thanks!
In game design you would render the audio in real time, so you would simply attach the audio to the object (in Unity 2d games are just flat 3d games). A game designer would touch spatial audio only for ambiances. Everything else are usually just mono objects.
@@michaelgwagner Thanks for the answer! By spatial audio you mean ambisonice/dolbyatmos or even the 3D spatial audio in unity? Can you think of a way to make tracking of the soundobject with attached sound audiobjects easier? For now i have a hard time even following the object circling around the center/player and it doesnt seem to matter whether i use the 2D or 3D option unless i use reverb zones which makes tracking of the objects a little easier.
Not sure what you mean by “tracking”.
@@michaelgwagner By "tracking" i mean the ability to determine the sound objects position with your ears while they travel around you / towards you.
That probably depends a lot on your audio setup and not so much on how you implement it in Unity.
Hey there, You're talking about Dolby Atmos decoder... but is it hardware or an app ?
Software
Do you read minds? Just yesterday I was wondering what would be the real difference between Dolby Atmos and Ambisonics? As always, thank you very much
By “efficient” @5:00 or so, you mean data- and processing- efficient?
Correct.
I know Im just a bachelor student in music technologies, but I am writting my bachelor project on mixing in ambisonics and quite honestly, while you have said many things correctly about ambisonics - I would object some that you seem to be extremely sure of, and the only explanation I would have for your strong beliefs is faulty loyalty to Nuendo and to Atmos. First, going into higher order ambisonics is encouraged by everyone who works with it and every book or article Ive read about it - some of the most magically sounding spatial things Ive heard are based on Higher Order, at least 5th or 7th order. Second, I was enraged when I heard you say that the similarity Atmos and Ambisonics share is the sweetspot. Excuse me, but WHAT THE F? Yes, while recording you have the center of the soundfield, but when it comes to monitoring and mixing, there are academic papers written on ambisonics and how using higher order ambisonics enlarges the listening area. Let me explain. Even in 1st order ambisonics, the listening area is not a spot, but an actual area, in which you can move around without losing the spatial experience, and only by walking too close to the speakers you lose that feeling and the experience drops down to listening to individual speakers. But the higher order you use - the larger that area of spatial experience becomes. Even at 5th order, you already have most of the area between the speakers covered, and you have to go pretty close to the individual speakers to lose that spatial experience. And no, I did not only read about this, but felt it and heard it myself in a spherical system. And no, I would not consider Cubase implementation to be great. I tried it out and I hated it. Its automated to the level of confusion. The processing chain is too rigid for me to do what I want to, and the detailed tailoring of some of the sound parameters is dumbed down to a single knob, while in Reaper I could freely adjust my processing chain and use advanced plugins that actually let me do what I need to. So no, I dont like ambisonics in Cubase, they are somewhat ok to mix for binaural purposes or 5.1/7.1, but I wouldnt trust it with any advanced ambisonic systems. And I have heard of game engines that rely heavily on ambisonics to render most of the sound in their games. So, to sum it all up - I find your knowledge on ambisonics either insincere or lacking. Yes, ambisonics is hard to understand, yes, its hard to market successfully, which is mostly why it has been in the academic community for over 50 years and only gets utilized and shared with the consumers in its simplest forms, but by god, dont downplay it, because it might bite you in the rear end in the future.
Well, first of all, thank you for your comment. It appears that you are confusing the quality of the spatial experience with its representational accuracy.
Let's say you are in the field recording with an ambisonics microphone and you position the mic next to a tree while recording the ambience around you. That tree will influence the way sound is received at the mic. If you later want to recreate that experience with an ambisonics speaker system, that experience will only be accurate at the sweet spot. As soon as you move away from it, you will no longer perceive the correct auditory impression of that tree at that location.
This is not a question of filling the sound between the speakers, it is a question of accurately representing the sound space around you. In that sense, ambisonics and Dolby Atmos are similar. They are only mathematically accurate at the sweetspot.
Having said that, with a real life ambisonics speaker system, you can probably move around a lot before this becomes an issue. The human brain can compensate a lot. However, that changes immediately as soon as you are talking about the use of ambisonics in cinematic VR or game design. In both cases you are not dealing with physical speaker systems. Rather, you are positioning the ambisonics audio in 3d space directly. Therefore, understanding that the auditory impression is no longer accurate as soon as you move too far away from the sweetspot is absolutely critical in these fields.
The order of the ambisonics can bee seen as its resolution. The higher the order, the more resolution you get from the higher order spherical harmonics. What you are talking about seems to indicate that in ambisonics reproduction, the resolution of the speaker system (i.e. the number of speakers) might be less important than the resolution of the recording (i.e. the ambisonics order). That is interesting by itself but only applies to our understanding of ambisonics reproduction systems and not so much ambisonics itself.
@@michaelgwagner Well, your assumption of my confusion only applies when talking about mathematical accuracy and actual recordings. But what if we were to talk about the spatial objects we create through sound synthesis or encoding mono or stereo tracks of instruments? We cant quite say we are mathematically accurately representing them, since their spatial information has not been recorded - nothing to actually represent. And yes, the ambisonic order we use as an input makes a fair difference compared to what order system we are using. Its a simple matter of whether or not we are supplying the system with lacking, abundant or just enough level of information. To say that the input resolution is unimportant is to say that it doesnt matter what we supply the system with - the sound quality will stay the same. And its obvious that its not the case. Yes, the resolution of the system is very important, binaural or surround reproduction will not be as accurate as spherical systems, but if we supply spherical systems with only 1st order or 0 order information - we wont get the spatial experience we desire, which then defeats the whole point of using such a system. To me, it makes sense to supply a high quality system with abundant level of information. And when it comes down to virtual 3d space - its mostly binaural applications, which dont quite care if you move or not. Well, unless you are using the head tracker. And yes, in real ambisonics systems you can move quite a lot even with 3rd and 5th order recordings - I could walk around 2 tables freely without feeling like Im listening to individual speakers, and that is the real ambisonics experience - to live in the soundfield we are producing. While mathematical accuracy is important to us engineers, but if we want to create the right experience for the listener we need to find a way to connect with them. And that, so far, has been the ability to walk around virtually freely in said soundfield.
Greetings from Peru, a query I want to remix a production made in Protools, this audio has to be exported from ProTools in WAV Stereo or has to be exported in another way to be binaural, and this audio I want to adhere to a normal RUclips video, you have to encode in a special way the audio and video for binaural? or add any metadata? Thank you for your reply.
Binaural audio is encoded in stereo. You don’t need any metadat. It’s just a stereo signal.
Super helpful video! 11:40 I'm just starting to mess around with Ambisonics using ProTools. I'm wondering why PT didn't make this list. It seems to have all the implementation, 1st, 2nd, 3rd order ambisonics, plus TBE plugins. Is it just a personal pref, or are there good reasons for me to consider one of these DAWS if I want to get serious about Ambisonics. Thx!
Very good question! At the time I recorded this you needed the Pro Tools Ultimate version to gain access to Ambisonics. And that was prohibitively expensive. The inclusion of Ambisonics in the base version of Pro Tools is a very recent development.
@@michaelgwagner Thanks. Well, I have discovered one problem with Pro Tools. The TBE Spatializer video overlay function is broken on my M1, and the components needed to make it work won't install. It looks like some ppl have gotten it to work, but it's complicated. Thanks again!
Hello, Mr. Wagner. Thank you for providing and sharing great information. I really appreciate it. I've been using both of Ambisonics and Dolby Atmos for a while and I was wondering if Ambisonics and 'Atmos objects' are using the same technology. If you set up speaker layouts in your room simillar to 7.1.4 in Ambisonic using reaper, would it be same as Atmos objects? Since both are using object-based audio technology... I just assumed 'Atmos objects' borrowed and using the Ambisonic tech and just added surround tech for 'Atmos beds' which is channel based.
Ambisonics is not object based. But you can use an Atmos speaker system to play Ambisonics. You would need a proper encoder to do that. Some Ambisonics plugin suites include a decoder with such a speaker setup.
@@michaelgwagner Thank you for the answer. You’re very kind.
@@mjmj7170 You could use IEM's AllRA Decoder, write in your speaker setup, balance it out with imaginary speakers (usually one at the bottom below the listener), press to calculate the decoder and then add distance compensator plugin. Voila, your mix is decoded for the exact system you are using.
Just to make sure, I understand this correctly: If one uses three figure of eight mics (one for x,y,z axis) and an omni, he can order them as w,y,z,x. Isn't that AmbiX (B-format)?
No, because some of the pickup patterns are negative, which is physically impossible to realize with an actual microphone. That’s why they are called “virtual microphones”.
great video, I have a problem that when I render in 5.1 it comes with the volume lower, but in the project it is normal
Hey man, good video, learn a lot. Thanks.
BTW, do you know how to record the Atmos audio while playing Blue-Ray Movies (whose audio track is Atmos) and AAA Games (whose audio setup supports ouput Atmos)?
Thanks in advance!
I mean the recorded audio should containing the JOC data that can be imported to Nuendo and restore the object movements info.
I honestly don’t know if that is even possible. Can Nuendo read JOC?
@@michaelgwagner Nuendo can import the ADM file (*.wav) exported by itself,and recover all VST automation to allow modifications . And this is what I mean the JOC means. If it doesn't match your understanding, I'm sorry for not describing correctly before.
I don’t think you can reconstruct the ADM from a DD+JOC file. There is information lost when the DD+JOC is created.
@@michaelgwagner However, I did find that Nuendo imports the exported Atmos file and restored the audio assets with the exact same VST automation as I had exported (by Nuendo).
I have a question, is that possible to transform ambisonic audio captured with the H3-VR into a mono track without choosing a specific directionality? The reason is that I am trying to capture audio in all directions but want to make it as fast as possible to deliver, hence in mono. I basically capture videos with a 360 camera which I transform into normal videos by reframing them. But the native audio quality of the camera is pretty bad so I would like to capture with an external mic without losing the spatial capture.
▼
vidIQ badge
The first channel in the Ambisonics B-Format is the mono track. You can just use that one.
@@michaelgwagner @Michael G Wagner Oh ok it's that simple! Thank you!! I am actually doing some tests. When I import inside DaVinci Resolve the raw audio file from the Zoom H3-VR and drag it to the timeline I get the 4 mono tracks of each mic. Since I basically have the 360 camera on a hand-held stick I could only put the microphone at the bottom of the stick which is a little far from the mouth of the speakers which are at camera level. So what I am trying to do is to for example cut or reduce by half the audio level of the two microphones that point downwards to the floor from the stick so I have more of the upward audio. I feel this is getting somewhere! :) It's kind of as if my 360 video was reframed but my audio was left in equirectangular format if I do a metaphor on the image. It's a good quick solution for now. Ideally, I would find a way in the future to apply the image reframing directionality automatically to the audio. (I use the KartaVR plugin in Davinci Resolve for Reframing)
What about dear VR Pro
Not sure if I understand the question. DearVR uses Ambisonics under the hood.
@@michaelgwagner Oh I just learn that thanks I am been using both dear VR Pro in Logic but I donot know how to set it up in Reaper for binaural 13.1 surround sound this would be for film and Video games redesign I learning maybe when time allows You could do tutorial on this Thanks Sir
I was wondering about the potential in game engines in particular. You would need to deal with 0 mics but that would have issues I assume? And then playback would be an issue? Unless there is a compatible receiver. Or audiochip. But Cmedia is not playing, XMOS could be done but I am not a star ar this at all.
I think using a 4.1.2 setup, maybe using 2 lower placed speakers would be wonderful, pinpointing sound like nothing else.
I wonder why this (as far as I could find) as a no longer patent encumbered technology has not caught on at all besides a few repo's by Google in the shape of Resonance Audio (with few if any updates to any of them) and even then aimed at VR only.
I have a hard time understanding your question. Are you asking about the use of Dolby Atmos in game engines?
I think he is doubt the real value of Ambisonics&Atmos in games, actually I wanna ask also......since most player's speaker is headphone or external stereo speaker(maybe 99%?), can player feel the difference between stereo and those two techs??? Does it improve a lot game experience???@@michaelgwagner
@cabab Well, when I talk about spatial audio in games I am usually talking about the development part of it and not so much the actual gameplay. On the development side it isn’t even a question, you need a spatial audio approach, otherwise you are not going to feel immersed. On the gameplay side things are different. It really depends on your listening environment.
Thx for your patient reply!!!!!Actually I'm newer to this industry and I'm working for a new project now, but, for example, UE5 have a lot of feature, such as "Soundscape" plugin for environment's sound building, or Spatialization、Air Absorption.... setting in "Attenuation" module. So if I use a lot of these thing, can it approach the effect of "Ambisonics&Atmos"? Or, UE5's built-in settings can replace those two tech???? Does that mean if I don't use those two tech or other additional tech mean I can't build immersed feeling?? so complex so confused......need some tips....QwQ@@michaelgwagner
Game audio is object based by default. In that sense Dolby Atmos isn’t really something you need to think about. Game Engines can generate Atmos automatically. You would use Ambisonics primarily if you want to bring in real world ambience recordings. For everything else, you would either use the tools of the game engine, or the audio middleware in case you are using one.
Hi. I have hard time to understand what is going on in ambisonic player meant for H3-VR. Are you familiar with this microhpone? After watching (please open the link) this I am confused ruclips.net/video/kIquru2hAH4/видео.html. If I have recorded Ambix audio from this Mic and open in Zoom Ambisonic Player. I can choose Stereo ,Binaural 5.1, surroound and Custom LR for listening and export. What and where is listening position? What does it mean in such reality when I have just simple stereo speaker setup? What these blue arrows really mean and why they are at angle? What is stereo mode? I know binaural stereo is for headphones and it would be like placing binaural microphone in center of sphere. Stereo modes are mystery for me. How software render stereo mix from ambix audio what are priciples? Why I hear all sounds in speakers even recorder below and above in this stereo mode? And big mystery is Custom LR. How it is relflected in stereo speakers?I can adjust sphere angle they say It is changing listening position?? I can adjust L R in random positions on sphere ("you can adjust listening position of the left and right channels independently through out 360 space") , they say these are channels.I dont know what it means ...Maybe we can adjust stereo width or something thatway??? How the thing corresponds to real stereo speakers setup? Thank You Michael if you willing to answer.
Yes, we have one of those. The Zoom player is just helping you to convert the recording into one of the standard formats. Since it records in Ambisonics you can rotate the signal around or convert it into stereo based on custom spesker positions. If you are only interested in the Ambisonics recording (Why would you use an Ambisonics mic if you only want to record stereo?) then you do not need most of the functionality of that player.
dolby labs atmos is junk . ambisonics is the way it should be captured and mixed as other movie mixes in stereophonic to dolby labs is junk rubbish . ambisonics is miles better
Now Studio one 6.5 has Dolby Atmos mixing in binaural on headphones
Video coming up this week. 😉