[Zenvota]

I'm thrilled to see people being interested in my humble project. Perhaps a separate thread would be in order, I'm very interested in hearing how it works for other people. There is already a thread in sound science but that discussion is probably a bit too technical for some people.

Ah yup
https://www.head-fi.org/threads/recording-impulse-responses-for-speaker-virtualization.890719/

 

[jaakkopasanen]

Sounds very interesting. Now if you could figure out a way to include head tracking in the package.

I'm going to look into it. I just hope I don't have to fork and extend EqualizerAPO myself. They head tracking part of shouldn't be a problem but updating EqualizerAPO live fast enough might. In any case this is not a immediate future thing.

 

[Zenvota]

I use a sound card with optical out into a 5 way optical splitter.

This works great btw.

I set up a very simple affordable system using a $20 usb to optical converter, a $10 optical splitter, a $10 optical dac, and a $25 Lepai LP-168HA which has a separate subwoofer output with a crossover for the tactile transducers. Zip tied 2 $10 shakers to the frame of a desk chair and good to go.

Thanks for mentioning that, don't know why I didn't know about them Xd Nice not having to use voicemeter(much increased latency) or worry about any ground loops with rca y cables.

 

[johnn29]

Around 2015 I had collected a ton of optical equipment - cables, splitters caps etc. Since my home theater / loud speaker setup had migrated to HDMI I thought I was never going to use it and threw them all out. This year I re-ordered everything because it's the best way of getting signal to everything with no latency or ground issues! Glad you've tweaked your setup. I finally got my mini quake put on my office chair - really like it for music.

2019 seems to be the age where speaker virtuslisation really took off. It laid sanguine for years - since the 90s really with synthesized solutions. OOYH came out but it was very difficult to find a preset that worked. Smyth research's products just aren't available but Super X-FI and Impulcifier have kicked things into gear!

 

[Zenvota]

. I finally got my mini quake put on my office chair

I've wondered about better and more powerful transducers on lightweight chairs, but I guess levels being the same you're getting a smoother response and deeper extension so there shouldnt be any hesitation to upgrade ;]

OOYH came out but it was very difficult to find a preset that worked

Truly I feel every preset works, but for really impressive fidelity some efforts required with the headphone eq, some hrtf/channel balance, and using equipment capable of rendering all that information with a completely noise free system.

Glad you've tweaked your setup

I had asked around a while back about splitting/duplicating a digital signal, but no one was aware of the optical route apparently. Its definitely a great option for getting the signal to shakers, and even to your headphone systems dac, but I notice a big big boost in detail and naturalness of the virtualization using an isolated well powered usb to i2s stage, so my main system stays with voicemeter, but my little backup mini theater that I use occasionally is perfectly enjoyable with the optical connection.

 

[johnn29]

The reason I was saying it doesn't work is because I've heard what the different ear mappings on SXFI get you in terms of sound signature. Localisation for me is there on most of the OOYH presets - some better than others - but the sound signature is off. It's the same with SXFI - on some of the mappings its created it just sounds like a really bad room with poor speakers. On a couple it sounds extremely good. The problem with all this is that it's so dependent on your ears. Impulcifier has the clear advantage here - but the weakness is your room. But I understand virtual "room correction" is forthcoming. I suspect the measurement process for any of these systems doesn't capture everything so your brains ability to process out reflections is compromised (Toole's work), so the headphone experience will always sound worse than the real loud speakers unless some sort of room correction is done.

With transducers it really depends on your seats too. I have dual Quake 10B's on my main sofa which reclines. When you're reclined you get a nice even distribution of shaking and the power that the XJ-700R feeding them has is incredible for movies. I used to strap two of the big quakes on a single Ikea pong chair - it had the power to move you forward during intense bass scenes like the IMAX Hubble space shuttle take off. It's changed my movie watching forever. For music a mini is sufficient though.

 

[Zenvota]

I wonder how much is just the sound signature of the speakers in the room and how much is the differences in hrtfs. When I look at graphs like these, I feel that its the former, but you're experience with Impulcifier may contradict that.

Here, a speaker that essentially measures flat in an anechoic chamber before and after room correction, still has a far from flat response


and here various hrtfs for the most part are fairly close below 5khz


I don't want it to seem like Im diminishing the impact of custom measurements though for sure its the one major thing thats missing in my system. But after upgrading equipment and doing some manual eq, the different tonal signature of the various presets are enjoyable. What would be great is if we can apply the measurements from impulcifier to ooyh and have access to all these rooms. Like the realisers exchange program.

Heres the eq for my audeze el8s with the dekoni velour pads, and aside from the headphone eq I think I'm probably adjusting a little bit for hrtf with the genelc preset, the left channel has a bit of bass roll off, the right treble, and im reducing the right channel by 2db in ooyh, but with those changes the bass is even and the center channel is center, this eq carries over to other presets as well with only doing channel balance in ooyh.

Spoiler: Audeze EL8

 

[Zenvota]

I suspect the measurement process for any of these systems doesn't capture everything so your brains ability to process out reflections is compromised (Toole's work), so the headphone experience will always sound worse than the real loud speakers unless some sort of room correction is done.

Yes our brains and ears can process an insane amount of information, but heres where I come back to equipment. This is the very thing that has improved drastically with specific circuits, transducers, and digital/power conditioning, the naturalness of the reflections, as the system is able to render more information. Think in terms of a feedback tube amp with heavy dynamic drivers not rendering all the detail due to slower circuits and slower transducers.

and maybe much in the way that i keep my budget shakers at lower levels and you can strap 2 quake 10bs to a paper chair lol

 

[johnn29]

I tried that out with my Audeze Mobius - so Planar Magnetic headphones with an integrated amp - I couldn't hear any difference compared with OOYH over BT or DT990's driven by a Xonarr USB card. But I know Darrin mentioned on the Home Theater Geeks podcast he was on that Electrostats were the best at rendering all the information.

The Realiser Exchange program type thing would be amazing, I'm sure something like that will be done.

What's good though is the externalization is definitely there in OOYH, Impulcifier, Super X-Fi. The rest just seems to be what sound signature I prefer. Even with SXFI I find that the preset that has crazy realistic channel placements is great for movies, but the bang on imagining almost sounds like running my AVR in Dolby Surround:Music without the rear effects - it puts the center stereo into the center speaker so some mixes just collapse to that. But I prefer another mapping for music that kind of muddies the stereo image but is worse for good stereo mixes or movies. So the additional processing would be very welcome when it's here with Impulcifier. I don't think SXFI will ever offer that.

 

[Zenvota]

I tried that out with my Audeze Mobius - so Planar Magnetic headphones with an integrated amp - I couldn't hear any difference compared with OOYH over BT or DT990's driven by a Xonarr USB card. But I know Darrin mentioned on the Home Theater Geeks podcast he was on that Electrostats were the best at rendering all the information.

I have the Mobius as well, its the only bluetooth headphone I use, with a pad change a little extra eq and a little ambience boost they sound very good, but cant do what either of the desktop setups do, even with OOYH instead of waves. And I haven't heard a modeled renderer that sounds as natural and spacious as the measured stuff, so looking forward to trying impulcifier.

Im more refering to using crazy technical setups xD balanced non feedback current signal class a dac/amps, with a usb to i2s stage isolated from the pc regenerated regulated and reclocked, isolation transformers and quadrapole cabling. I'll detail it maybe you'll find it interesting =/

My home office rig that gets all the music time sounds more or less perfect, incredibly natural sounding virtualization, perfect tonality, massive imaging, all the descriptors ;]
This system consists of:
Headphone: Audeze EL-8 balanced with Dekoni LCD Velour pads
DAC: Audio-GD NFB-7(single es9018 chip)
Amp: Audio-GD NFB-1AMP, current signal interconnects between the 2.
the usb to i2s stage consists of an uptone iso regen, powered by an Uptone Ultracap and the Uptone SMPS, with a quadrapole(star quad) dc cable, and the ultracap also powers the dacs usb to i2s board which has an isolator on the i2s output.
power conditioning with a Topaz Ultra Isolator Line Noise Suppressor and quadrapole ac power cabling(VH Audio Flavor 4

My work office system is a bit less and while it sounds very good( much better than the mobius for instance) it isnt able to render the perfect naturalness of the main system:
Headphone: Hifiman HE-400i, single ended, dekoni lcd velour pads, sorbothane lined cups, custom rear grills with large open cell foam, melamine fuzzors on the rear magnets. This eliminated alot of resonance and really improved the tonality/timber of the virtualization.
Dac/Amp: Audio-GD NFB-11
usb stage is an Uptone USB Regen powered by a Jameco regulated linear psu into a lt3042, with a quadrapole dc cable to the regen, to a Breeze DU-U8 usb to optical converter, and optical to the dac
power conditioning is a topaz ultra isolator and vh audio cable.

I also have an Audio-GD nfb12 with the optical splitter setup to the shaker system as a backup for occasional movie watching when I can't use the surround system and that unit has no conditioning and suffers from sibilance, boomy voices, etc. all the negative effects of ac noise optical jitter phy noise phase noise etc. But it still sounds much better than using the pc motherboard, a denon x4200w, a jds odac/o2, any topping/aune/cambridge feedback op amp ic w.e, due to the audio gd amplifiers. I prefer this unit with just an optical connection, with OOYH and the alpha primes to the Mobius as well.

But hey, Im not encouraging anyone to go out and spend thousands of dollars on this stuff, only sharing for those that find it interesting, and encouraging people to look at circuit design, signal integrity and bandwidth, and all the various forms of noise. Some simple tweaks and choosing the right components can make a world of difference. Just as you hear an improvement with a better pair of headphones, adding an eq, or correcting hrtf, the circuit from the pc to the transducers is equally important.

Its interesting that you couldnt hear any difference between the mobius and dt990s + xonar card, was this with no eq? With the Mobius over bluetooth, what codec was used to connect to the pc? regardless this adds compression which will reduce the effectiveness of the virtualization. I tried a plethora of dynamics and not one came close to the hifimans, for me they actually sounded bad in comparison, and I generally prefer dynamics for normal headphone listening(I like the HD650s).

 

[johnn29]

Wow that's an intricate setup! I suspect my ears/brain don't hear the power conditioning issues in my setup. Unless it's ground hum. But I also can't hear the difference between 320kbps mp3 and FLAC on any of my setups - so it's my brain.

The DT990 is always EQ'd. I use oratory1990's result to EQ to Harmen - otherwise they have horrific sounding treble. Whenever I test between different headsets I make sure I EQ to the Harmen Target. That's likely why I have to use the simple treble bell cuts to make them sound decent - because the Harmen EQ isn't flat for HRTF. Mobius was over USB. I love the warm preset on that with Waves just for music - doesn't have a great center channel but works well for music. I should add - I'm using impulse recording of OOYH (all digital) because I simply couldn't get the software working on my laptop without blue screening. It also makes A vs B much easier to do in HeSuVi.

On my other Bluetooth setups I just use an optical USB transmitter that supports Apt X HD or LL plugged into a cheap Turtle Beach USB sound card I had laying around that has optical out.

What I've really understood since playing with SXFI and Impulcifier is how important the Headphone Compensation really is. SXFI really doesn't sound "right" without a supported pair of headphones. I tried to use my WI-1000x's with it (unsupported) with various EQ settings and it sounded off. Soon as I used my MDR-1000x, which is supported, boom - it was bang on. In using Impulcifier I've got mutliple headphone compensations now - and when cycling quickly in HeSuVi you can really tell what a dramatic different it makes. Even the L and R being level matched has a big impact on the location of the stereo image and even sound signature. From oratory1990's measurement of my WI-1000x I know there's a channel imbalance, but only at 2khz-4khz. I've tried to do it before with tones but it's so much easier when it's automated or you have measurements. It's a shame OOYH doesn't have that as a feature.

 

[castleofargh]

Harmen

Harman(or maybe we can call it Samsung curve now ^_^), from https://en.wikipedia.org/wiki/Sidney_Harman

 

[Zenvota]

I'm using impulse recording of OOYH (all digital) because I simply couldn't get the software working on my laptop without blue screening. It also makes A vs B much easier to do in HeSuVi.

Ah ok thats it, OOYH sounds much much better than hesuvis ooyh preset. I believe it's due to the preset was captured from the output of ooyh as opposed to using the actual impulse response.

Had you tried asking Darin about the blue screens?

I've tried to do it before with tones but it's so much easier when it's automated or you have measurements. It's a shame OOYH doesn't have that as a feature.

Ya its tricky, manually eqing the el8s I used innerfidelitys graph, clarityfidelity/speakerphones graph, a graph of the difference of the lcd2s with and without the dekoni lcd velour pads, frequency sweeps, and listening to music. In some cases even 2db off was noticeablely distracting for some frequencies, mostly 2khz to 10khz.

 

[darinf]

I should add - I'm using impulse recording of OOYH (all digital) because I simply couldn't get the software working on my laptop without blue screening. It also makes A vs B much easier to do in HeSuVi.

Hi @john29 Not sure if you reached out to me already or not regarding your problems with Out Of Your Head.
But usually BSOD's are a result of a corrupted installation or having multiple instances of the Out Of Your Head driver when attempting to install and uninstall multiple times. It's not supposed to happen, but sometimes you can get into a weird state where the Out Of Your Head driver does not uninstall completely.

In any case, we are usually able to solve any BSOD problems. I am also happy to schedule a remote support session where I can control your computer remotely to troubleshoot or fix the issue.
Thanks,
-Darin

 

[phoenixdogfan]

The reason I was saying it doesn't work is because I've heard what the different ear mappings on SXFI get you in terms of sound signature. Localisation for me is there on most of the OOYH presets - some better than others - but the sound signature is off. It's the same with SXFI - on some of the mappings its created it just sounds like a really bad room with poor speakers. On a couple it sounds extremely good. The problem with all this is that it's so dependent on your ears. Impulcifier has the clear advantage here - but the weakness is your room. But I understand virtual "room correction" is forthcoming. I suspect the measurement process for any of these systems doesn't capture everything so your brains ability to process out reflections is compromised (Toole's work), so the headphone experience will always sound worse than the real loud speakers unless some sort of room correction is done.

With transducers it really depends on your seats too. I have dual Quake 10B's on my main sofa which reclines. When you're reclined you get a nice even distribution of shaking and the power that the XJ-700R feeding them has is incredible for movies. I used to strap two of the big quakes on a single Ikea pong chair - it had the power to move you forward during intense bass scenes like the IMAX Hubble space shuttle take off. It's changed my movie watching forever. For music a mini is sufficient though.

That attempt to capture real systems playing in real space which both OOYH and Smyth and now (I think) impulsifier are doing could be the real weakness of these systems. Think about it. Your headphones play full range without crossovers, and in most cases people are extremely happy with the sonic signature/voicing of those phones, and while they don't create the OOYH effect of something playing in a real room, they don't have the drawbacks regarding directivity/room interaction that all those real speakers have. So, why would you want to change the voicing of your phones by having them mimic a transducer which probably has a poor crossover, bad room acoustics, poor directivity, etc.

Isn't the next step forward to create a digital room and virtualize your own headphones with their exact sonic signature into that space with as many virtual speakers as you want? Moreover, why use in-ear microphones with their terrible S/N and distortion performance to capture your HRTF. HRTF is all about the topology of your head, ears, and body, and there is no theoretical reason why it can't be captured with the utmost precision with some kind of visual measurement system (and it would only have to be done once if we separate the HRTF from the speaker/room emulation).

A system modeled on these features could be a great deal simpler as well. Obviously headtracking would still be needed, and room virtualization and HRTF correction, and multichannel binaural mixing would be needed--but all that stuff is now available from one vendor or another. It just needs to be packaged as a single piece of software that runs on a Windows box and a MAC and sold along with a headtracking unit that connects to you pc via usb.