Recording Impulse Responses for Speaker Virtualization
Jan 26, 2024 at 1:37 PM Post #1,786 of 1,816
@musicreo you have helped me loads and I respect and appreciate what you taught me. But I have to disagree on this one. I’m able to switch between using the Dolby atmos from the lg tv sound options, which is the device being listened to by vbcable recording option IMG_1829.jpeg

IMG_1830.jpeg

And it’s a massive difference from just having the sound around me in a ring to having it around me as natural and 3d as can be almost binaural. I’m also getting height (but that is maybe because of the upmix to height setting in the Dolby access app).

Isn’t it the same as getting object based from a 7.1 speaker system. Where it fools the brain into thinking there’s more speakers than 7.1?
you probably have your hrir in vb cable that's listened on your tv that's also using dolby atmos, you should first disable the emulation from hesuvi

the real dolby atmos is the one you hear in cinemas, i doubt that a tv can replicate atmos, even hesuvi has its atmos hrir but it's just a gimmick

the prpper way to emulate atmos is by having more channels recognizable on windows that emulates the upper speakers, from what i'm aware of the only audio card that does this is the smyth realizer
 
Jan 27, 2024 at 4:16 PM Post #1,787 of 1,816
Sorry if I keep banging on about this but just listening to my hrir with Dolby home theatre v4 installed into my tv driver sounds lifelike. I’ve been to cinema listened to Dolby atmos even the new avatar I watched at the best imax I could find with the best sound. And it didn’t sound as good as this.

Just wondering if I am wrong. What is it that we would need to get atmos to work. Not talking about height I’m talking about object based.

From what I understand for Dolby atmos you need 7.1 speakers, Dolby gets the sound information and makes those 7.1 speakers sound like sound is coming from 3d space. So why can’t we just get Dolby hacked into the drivers and make Dolby think our hrir is actual Dolby 7.1 speakers.

Also when I installed the hacked driver for listened to device (lg cx), it now shows supported formats as Dolby atmos and Dolby mat 2.0/2.1 isn’t that what we need to decode atmos?

IMG_1871.jpeg
 
Last edited:
Jan 27, 2024 at 5:44 PM Post #1,788 of 1,816
Also when I installed the hacked driver for listened to device (lg cx), it now shows supported formats as Dolby atmos and Dolby mat 2.0/2.1 isn’t that what we need to decode atmos?
If you connect your TV with bit stream audio it can encode the meta data of the atmos format. But this way you can't use EQ-APO with your hrir. You have to encode atmos on the PC and that is very difficult as there is no media player that can do this.
 
Feb 2, 2024 at 12:35 PM Post #1,789 of 1,816
I would like to cross-post my result here covering my experience with the https://www.earfish.eu/ HRTF acquisition approach: https://www.head-fi.org/threads/anyone-into-crossfeed.961533/post-17950791 (post #32). I had written more about the process in https://www.reddit.com/r/headphones.../?utm_source=share&utm_medium=web2x&context=3. My approach focused on using windowing to approximate a free-field response to EQ toward, then letting SPARTA AmbiBIN apply the head-tracked HRTF changes. At least "under the microscope" of checking how AmbiBIN transforms the reference 30-degree free-field EQs for other directions or for in-phase combined sound, there are technically still flaws in how the SOFA file is interacting with my EQ (tonality may not be perfect, but all the directional stuff is as good as ever) and the phase response rendering likely skewing the tonality of centered sounds, but it otherwise sounds absolutely wonderful with great recordings, exhibiting beautiful imaging and clarity.
 
Feb 3, 2024 at 12:52 AM Post #1,790 of 1,816
I would like to cross-post my result here covering my experience with the https://www.earfish.eu/ HRTF acquisition approach: https://www.head-fi.org/threads/anyone-into-crossfeed.961533/post-17950791 (post #32). I had written more about the process in https://www.reddit.com/r/headphones.../?utm_source=share&utm_medium=web2x&context=3. My approach focused on using windowing to approximate a free-field response to EQ toward, then letting SPARTA AmbiBIN apply the head-tracked HRTF changes. At least "under the microscope" of checking how AmbiBIN transforms the reference 30-degree free-field EQs for other directions or for in-phase combined sound, there are technically still flaws in how the SOFA file is interacting with my EQ (tonality may not be perfect, but all the directional stuff is as good as ever) and the phase response rendering likely skewing the tonality of centered sounds, but it otherwise sounds absolutely wonderful with great recordings, exhibiting beautiful imaging and clarity.
Sounds good! I find interesting that you use free field HRTF since that's often perceived to be too bright having no tilt. What's your take on that?
 
Feb 3, 2024 at 1:18 AM Post #1,791 of 1,816
Sorry if I keep banging on about this but just listening to my hrir with Dolby home theatre v4 installed into my tv driver sounds lifelike. I’ve been to cinema listened to Dolby atmos even the new avatar I watched at the best imax I could find with the best sound. And it didn’t sound as good as this.

Just wondering if I am wrong. What is it that we would need to get atmos to work. Not talking about height I’m talking about object based.

From what I understand for Dolby atmos you need 7.1 speakers, Dolby gets the sound information and makes those 7.1 speakers sound like sound is coming from 3d space. So why can’t we just get Dolby hacked into the drivers and make Dolby think our hrir is actual Dolby 7.1 speakers.

Also when I installed the hacked driver for listened to device (lg cx), it now shows supported formats as Dolby atmos and Dolby mat 2.0/2.1 isn’t that what we need to decode atmos?
The height is object based with home media. So the two common formats of home Atmos is Dolby Digital with Atmos (streaming), and TrueHD with Atmos (UHD disc). Both are similar in that they have a base surround stream (5.1 with DD+ or 7.1 with TrueHD). Then there's a Joint Object Coding stream: up to 16 dynamic channels (channels that can change from a channel to positional object). There are differences between these home formats vs the Atmos you can hear in cinemas: cinemas have a lot more speakers: so their "track" of speaker channels and objects is quite larger. Cinemas used to have arrays of speakers that were all driven off one channel of sound. One of the advantages of Atmos was that an array of speakers would then get individualized sound with the object based rendering of the system.

I do a lot of 4K movie viewing off Plex myself. My preferred device is an Apple TV 4K: it ordinarily wouldn't support Atmos with Plex (it does with streaming services doing Dolby Digital with Atmos). But for 3rd party apps, it doesn't normally support Dolby Digital with Atmos: there's a special API that they have to use to "passthrough" DD+ or TrueHD. I have another app that syncs with Plex that supports MAT 1.0 (Atmos through lossy): I can watch Dolby Digital with Atmos through Plex on Apple TV 4K with Infuse. Plenty of users keep complaining about how they rip UHD discs and can't listen to TrueHD Atmos through that streamer. The developers can only recommend people petition Apple to go MAT 2.0 (Atmos through lossless). Since it seems Apple may never do this: I do still have a NVidia Shield. The NVidia is the main streamer that keeps it pretty easy about ripped movies playing TrueHD Atmos. For myself, it's connected to my receiver for a 7.1.4 speaker system. If you're connecting to TV, the TV has to be more recent generation to have a E-ARC port (IE supports lossless audio).

These are also developers who are paying Dolby to use the Atmos API, to be able to use the full Atmos renderer. You don't have to "hack" the track to get the 5.1/7.1 stream: that is the legacy surround stream that non-Atmos systems see as DD+ or TrueHD. So a virtualizer that isn't reading the full Atmos track may be reading that surround info, or it may be doing its own DSP with stereo.

Bringing this up since you ask what is Atmos, and the sub-forum seems to ask from time to time. When I try to look earlier in this thread, see empty posts. Would recommend messaging me if you'd like to know more about Atmos. When it comes to this thread on virtualization: from what I've read it's really limited with Dolby headphone through computer. The problem with Dolby is that they don't have a headphone calibration and just use a generic HRTF with the Atmos renderer.
 
Last edited:
Feb 3, 2024 at 1:47 AM Post #1,792 of 1,816
The height is object based with home media. So the two common formats of home Atmos is Dolby Digital with Atmos (streaming), and TrueHD with Atmos (UHD disc). Both are similar in that they have a base surround stream (5.1 with DD+ or 7.1 with TrueHD). Then there's a Joint Object Coding stream: up to 16 dynamic channels (channels that can change from a channel to positional object). There are differences between these home formats vs the Atmos you can hear in cinemas: cinemas have a lot more speakers: so their "track" of speaker channels and objects is quite larger. Cinemas used to have arrays of speakers that were all driven off one channel of sound. One of the advantages of Atmos was that an array of speakers would then get individualized sound with the object based rendering of the system.

I do a lot of 4K movie viewing off Plex myself. My preferred device is an Apple TV 4K: it ordinarily wouldn't support Atmos with Plex (it does with streaming services doing Dolby Digital with Atmos). But for 3rd party apps, it doesn't normally support Dolby Digital with Atmos: there's a special API that they have to use to "passthrough" DD+ or TrueHD. I have another app that syncs with Plex that supports MAT 1.0 (Atmos through lossy): I can watch Dolby Digital with Atmos through Plex on Apple TV 4K with Infuse. Plenty of users keep complaining about how they rip UHD discs and can't listen to TrueHD Atmos through that streamer. The developers can only recommend people petition Apple to go MAT 2.0 (Atmos through lossless). Since it seems Apple may never do this: I do still have a NVidia Shield. The NVidia is the main streamer that keeps it pretty easy about ripped movies playing TrueHD Atmos. For myself, it's connected to my receiver for a 7.1.4 speaker system. If you're connecting to TV, the TV has to be more recent generation to have a E-ARC port (IE supports lossless audio).
Thanks for the explanation that makes different atmos more clear. My pc sound options for the lg tv never showed atmos or Dolby digital with atmos mat 2.0/2.1. Since using the modded driver I now have these formats listed and when using with Dolby access on pc with height mix I’m getting subtle height sounds it actually makes my hair stand when I hear it.

When I have scenes with rain instead of the rain being around me I can hear it in different places close and far to me like it’s falling near my shoulders and far away too. Also I can close my eyes and pin point where a sound is coming from in the scene. Like someone walking can be tracked exactly.

I know people have said I shouldn’t’ be able to get it. But sound wise it sounds so much better than what I had before not clarity but object based. I thought it might be placebo myself but when I updated the Nvidia drivers the mat2.0/2.1 disappeared and I lost the object based sounds until I fixed it again with the modded driver.

I have the modded driver if anyone wants to try it
 
Last edited:
Feb 3, 2024 at 2:12 AM Post #1,793 of 1,816
Thanks for the explanation that makes different atmos more clear. My pc sound options for the lg tv never showed atmos or Dolby digital with atmos mat 2.0/2.1. Since using the modded driver I now have these formats listed and when using with Dolby access on pc with height mix I’m getting subtle height sounds it actually makes my hair stand when I hear it.

When I have scenes with rain instead of the rain being around me I can’t hear it in different places close and far to me like it’s falling near my shoulders and far away too. Also I can close my eyes and pin point where a sound is coming from in the scene. Like someone walking can be tracked exactly.

I know people say I can’t get it. But sound wise it sounds so much better than what I had before not clarity but object based. I thought it might be placebo myself but when I updated the Nvidia drivers the mat2.0/2.1 disappeared and I lost the object based sounds until I fixed it with modded driver.
Well Atmos has also gotten pretty generic since Dolby also likes to market it as a hybrid system that gives you better sound with any number of speakers. Does your TV have E-ARC? If so, it can pass TrueHD (and TrueHD Atmos). So for true Atmos, all your devices have to be "Dolby Atmos" certified: in that they either pass the JOC metadata or they process it with the Atmos renderer. For home theater folks like myself, I just have devices going to my Atmos/DTS:X/Auro-3D receiver and then get Atmos through my speakers (and I would say the 3D effects I hear from my speakers sound better than any cinema I've been to here). I did see the latest Avatar with IMAX: and I really hated how loud their sound was. I've also heard some impressive 5.1 tracks up-mixed to 3D: like scenes of Master and Commander having clear above deck sounds on top with my default DTS:Neural.

Not sure what system you're referencing: if it is a particular headphone system that you think may be seeing some surround info. FWIW, one of the best headphone surround I've heard is a Dolby Pro Logic II processor that came with my HD 580s (many years ago). You just dial in a parabolic curve setting that sounds best to you....and it's still one of the clearest front to back sound fields I've heard (all with a stereo source). DSPs can be pretty awesome. Also growing up I had particular Sony Discmans that had a good "surround" processor. Not that they sounded 3D, but they had an interesting way of making studio albums sound live.
 
Last edited:
Feb 3, 2024 at 2:19 AM Post #1,794 of 1,816
Sounds good! I find interesting that you use free field HRTF since that's often perceived to be too bright having no tilt. What's your take on that?
"Free-field" here is technically closer to a windowed smoothing of my personal near-field response. https://www.audiosciencereview.com/...n-susvara-headphone-review.50705/post-1850477 (post #1,085) and https://www.audiosciencereview.com/...n-susvara-headphone-review.50705/post-1850739 (post #1,088) show my blocked ear canal measurements for my final "free-field" PEQ with SPARTA AmbiBIN including some comparison to my past "Harman-like" V3 PEQ, the HE1000se, and my in-room Genelec 8341A measurements from around 1 m away at my desktop. It initially seemed bright, but after acclimatizing, great recordings sound more "vivid" than ever while my old V3 PEQ can sound "relaxed" by comparison. Yes, some stuff will be more shouty or some strings may sound harsh (for some recordings, they stay harsh even after EQing down the ear gain region), but you may also come across recordings where the strings stay wonderfully smooth and textured despite having this much ear gain. Now, in post #1,088 of that thread, you can better see how SPARTA AmbiBIN's phase rendering yields the magenta trace for in-phase centered sounds, perhaps causing music to sound a bit smoother than true neutral near-field. I still need to see if Earfish can produce another version of my SOFA file for EQing on top of a flat blocked canal headphone response and whether that would improve some imaging or tonality or the phase response.

Regarding the below, Harman Linear In-Room is pretty close to diffuse-field (I don't know how the upper treble and top octave stuff was decided upon, whereby I just keep the full treble extension without issue). Free-field there exhibits the treble null that I do measure in my own HRTF, and has a bit more 3 kHz canal resonance than diffuse-field. My blocked canal measurements per the lack of canal resonance already EQ in a dip similar in character to my V3 PEQ where I with sine sweeps and pink noise heard too much 3 kHz, so I don't find that part too bright. As for why Harman Over-Ear EQs things down more, I suspect it may have to do with the lack of crossfeed and differences in 3 kHz canal resonance between a 0-degree or 30-degree speaker placement (I don't know the actual Harman Linear In-Room configuration off the top of my head) versus the 90-degree HRTF. My personal measurements also confirmed that for my ears, the 90-degree HRTF contributing to my in-ear headphone responses causes them to have excessive 6 kHz (post #1,088 gold trace), though I did originally EQ that down too much in my V3 PEQ, losing out on some string "zizz" or vibrance. I afterward apply a 5 to 7.4 dB bass shelf to taste.

1706944146237.png
 
Feb 3, 2024 at 4:53 AM Post #1,795 of 1,816
To fully enjoy Dolby Atmos from Windows OS to height channels, you need a multi-channel audio interface and a receiver. Each convolution must be applied at the DAW stage.
Tracking or extracting an individual's HRTF is a good thing.
If you have a chance, measure it from various distances.
The HRTF that comes in (up to 7m) is also different for each distance.
I like 3m.
In addition, I recommend using IEM to utilize HRTF, HRIR, and BRIR properly.
Headphones cannot ignore wearing deviation, and there are already many variables from the response of the headphones themselves.
I also sold a used HD800s, and now I only enjoy IEM.
 
Feb 3, 2024 at 8:51 PM Post #1,796 of 1,816
To fully enjoy Dolby Atmos from Windows OS to height channels, you need a multi-channel audio interface and a receiver. Each convolution must be applied at the DAW stage.
Tracking or extracting an individual's HRTF is a good thing.
If you have a chance, measure it from various distances.
The HRTF that comes in (up to 7m) is also different for each distance.
I like 3m.
In addition, I recommend using IEM to utilize HRTF, HRIR, and BRIR properly.
Headphones cannot ignore wearing deviation, and there are already many variables from the response of the headphones themselves.
I also sold a used HD800s, and now I only enjoy IEM.
Isn't using IEMs itself a bad practice for binaural audio because of the lack of pinna interaction?
 
Feb 3, 2024 at 9:01 PM Post #1,797 of 1,816
Isn't using IEMs itself a bad practice for binaural audio because of the lack of pinna interaction?
I think it depends on how the HRTF is arrived at. The limitation of using Dolby Atmos's headphone alone is that it just has one generic HRTF. There are systems that take the Atmos data and then map to a custom HRTF. There's a few systems out there for that: popular ones including taking photos of your ears for AI mapping...or using microphones in the ear cups to sense ear type.
 
Feb 4, 2024 at 4:39 AM Post #1,798 of 1,816
Isn't using IEMs itself a bad practice for binaural audio because of the lack of pinna interaction?
It's similar in the end. While measuring, the mics will affect the ear's acoustic anyway, so we will ideally require one extra EQ no matter what. That often ends up being massively subjective because we don't necessarily have a solid method to deal with it(switching back and forth between speakers and headphones/IEMs while using test tones or sweeps is not easy IMO. The usual solutions relate to some reference of equal loudness where we have to hope for no serious hearing loss (excellent super young ear might be another problem as it doesn't fit in the average curves).

I personally don't agree with the IEM being better because in general, a nice headphone will simply measure better. For the HD800 I think we have no doubt about that. And because, at least in my case, cutting out the outside noise seems to interfere with the simulation. I can only guess why, as I really don't know, but maybe my brain goes something like "how come those outside sounds are clear while those other outside sounds are muffled?".
 
Feb 4, 2024 at 4:40 AM Post #1,799 of 1,816
Isn't using IEMs itself a bad practice for binaural audio because of the lack of pinna interaction?
I don't think so.
In Korea, there is already a trend of using brir as an IEM.

They have a much better response, lower THD, and if worn correctly, a more consistent sound, and if soundproofed well, you can listen to songs in the same silence as you would in an anechoic chamber (more dynamic range).
Headphones, on the other hand, were an easier approach, but imperfect: when it came to sound reproduction (like the real thing), headphones were always 2% off.
 
Feb 4, 2024 at 5:09 AM Post #1,800 of 1,816
I don't believe that THD is very meaningful for subjective experience (until the values explode), but I suspect some simple testing errors if someone came to the conclusion that IEMs at large have lower distortions than full size headphones. It's probable that the measurements got drowned in ambient noises and only showed that the IEMs have better isolation.
I've had some single dynamic driver that measured quite well, but I would still argue that most IEMs have rather poor distortions figures. More so if we consider the balanced armatures.
 

Users who are viewing this thread

Back
Top