• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Headphone Correction & Spatialization

dasdoing

Major Contributor
Joined
May 20, 2020
Messages
4,209
Likes
2,674
Location
Salvador-Bahia-Brasil
I like your filter better than Waves, but none of these make me think I'm listening to speakers. I think I prefer CanOpener.

I will let you chose a song for my 2.0 version. actualy I kind of treated the first version like a beta, since I wasn't realy fully happy with it
 

whazzup

Addicted to Fun and Learning
Joined
Feb 19, 2020
Messages
575
Likes
486
If you're looking at speaker virtualization, I've personally had the best results from Out Of Your Head (OOYH) software, though it's more of an end-user software, not DAW plugin.
https://fongaudio.com/out-of-your-head-software/

There's a discussion on OOYH (where the developer drops by as well) at Head-fi:
https://www.head-fi.org/threads/out-of-your-head-new-virtual-surround-simulator.689299/

The software is finicky, but when it works, the presets (recorded with the Realiser A8 hardware) do allow that feeling of listening to speakers from headphones. It's not going to cut it when you start to listen critically, but for me, of the similar technologies out there, it's the most successful so far in removing the sound pressure you 'feel' on your ears when listening to audio through headphones, and create the impression of sound coming from the front. Within OOYH, I feel the Sonus Faber Elipsa and Genelec presets get the closest to mimicking speakers.

Haven't had the fortune to hear the Realiser A16 though. And also haven't tried NX other than listening to their web demos.
 
Last edited:

whazzup

Addicted to Fun and Learning
Joined
Feb 19, 2020
Messages
575
Likes
486
I tryed all kinds of crossfeed/vitual room aplications and they all sound shit. Best I found is Abbey Road Studio 3, but it still colorizes af.
So I took some binaural IRs from WDR studios and tweeked it until I got most colorization out.
at the time I had a wide angle setup in my main speakers so I used a wide angle for this. I will create a regular stereo triangle version in the future since I went back to the standard:

The voice still sounds like it's inside my head, just more diffused / crossfed (?) when compared to the original track. And the telltale signs of feeling the 'sound pressure' on my ears from the headphones. It's definitely not easy, good luck!
 

dasdoing

Major Contributor
Joined
May 20, 2020
Messages
4,209
Likes
2,674
Location
Salvador-Bahia-Brasil
The voice still sounds like it's inside my head, just more diffused / crossfed (?) when compared to the original track. And the telltale signs of feeling the 'sound pressure' on my ears from the headphones. It's definitely not easy, good luck!

personaly I only hear the hard panned sounds "in my head"
the other song I posted with my filter is not hard panned and I don't hear it "in my head".
creating a filter like this is not easy, because if you leave a lot of "virtual room" it sucks, if you remove it stays "in the head".
my personal goal was to create not a filter that represents real speakers in a real room (it sucks on headphones), but to create a filter that represents utopic perfect speakers in a utopic perfect room. I apreciate the critics. I will try to leave more room in the next version.
this is if I don't like OOYH, which I will test, since if I like it I have no reason to make my filter better
 

whazzup

Addicted to Fun and Learning
Joined
Feb 19, 2020
Messages
575
Likes
486
Sure. The installation process can be a headache, as the developer creates a driver that sorts of rides on top of your current audio device. If you switch between audio devices like multiple usb dacs / onboard sound, you can encounter issues like no sound coming out. Reinstalling and restarting worked for me. But my computer is close to 10 years, the usb ports may not be very robust / well supported in Windows 10, so you may get a smoother experience.
 

Guermantes

Senior Member
Joined
Feb 19, 2018
Messages
484
Likes
561
Location
Brisbane, Australia
I was listening to a classical album this morning during my daily commute:
https://www.deccaclassics.com/au/cat/4758237

I was listening on IEMs and my DAP has no cross-feed or similar function but I was struck by how the soundstage of the Henry Cowell piece was less "in-my-head" and more panned across the front though rather close -- much like what I would expect of a chamber orchestra -- but not necessarily equivalent to speakers-in-a-room. This is a recording that seems to translate well to headphone listening. I am listening to it now on a pair of Genelec 8260As and it still has a nice natural image. The engineer was Stan Goodall and he probably used a Decca Tree -- three omni microphones in a T configuration.

So this made me think: As a content creator, should I be considering headphone listening as a target? In other words, should mixing for speaker listening still be the primary reference when consumers are possibly spending more and more time with headphones? And how do I get my recordings to sound satisfying on both speakers and headphones? I consider both to be necessary compromises with attendant idiosyncrasies, some of which we have become quite psychologically attached to.

@andreasmaaan Perhaps you can try your mixes with and without speaker-simulation and see if they are commensurable between the two listening paradigms. I'm thinking about replicating a Decca Tree effect on the master bus and see how that affects imaging, though the success or failure of this might depend on how much it violates generic expectations.

Just some thoughts.
 

Bob-23

Senior Member
Joined
Jul 3, 2019
Messages
425
Likes
379
Location
Berlin, Germany
should mixing for speaker listening still be the primary reference
That's the point, we need recordings that are, at least, mixed for headphones, and/or, even better, particularly recorded for headphone listening. All the measures we take in order to correct for recordings that are mixed on boxes for listening on boxes will rest somehow imperfect, however good the are. That said, headphone listening, corrected by crossfeed, nevertheless, has some advantages over listening on speakers: you are always in the sweet spot, you get all the details, and no room reflections fake the sound. So, listening on boxes is, mostly, imperfect, too.
 

Bob-23

Senior Member
Joined
Jul 3, 2019
Messages
425
Likes
379
Location
Berlin, Germany
I was listening to a classical album this morning during my daily commute:
https://www.deccaclassics.com/au/cat/4758237

I was listening on IEMs and my DAP has no cross-feed or similar function but I was struck by how the soundstage of the Henry Cowell piece was less "in-my-head" and more panned across the front though rather close -- much like what I would expect of a chamber orchestra -- but not necessarily equivalent to speakers-in-a-room. This is a recording that seems to translate well to headphone listening. I am listening to it now on a pair of Genelec 8260As and it still has a nice natural image. The engineer was Stan Goodall and he probably used a Decca Tree -- three omni microphones in a T configuration.

So this made me think: As a content creator, should I be considering headphone listening as a target? In other words, should mixing for speaker listening still be the primary reference when consumers are possibly spending more and more time with headphones? And how do I get my recordings to sound satisfying on both speakers and headphones? I consider both to be necessary compromises with attendant idiosyncrasies, some of which we have become quite psychologically attached to.

@andreasmaaan Perhaps you can try your mixes with and without speaker-simulation and see if they are commensurable between the two listening paradigms. I'm thinking about replicating a Decca Tree effect on the master bus and see how that affects imaging, though the success or failure of this might depend on how much it violates generic expectations.

Just some thoughts.
I forgot: "Longtemps, je me suis casqué de bonne heure."
 
Last edited:

Guermantes

Senior Member
Joined
Feb 19, 2018
Messages
484
Likes
561
Location
Brisbane, Australia
I forgot: "Longtemps, je me suis casqué de bonne heure."

marcel-proust-tin-foil-hat.jpg


:D
 

dasdoing

Major Contributor
Joined
May 20, 2020
Messages
4,209
Likes
2,674
Location
Salvador-Bahia-Brasil
my sub broke, so I figured I do this project today (testing OOYH and creating new version of my filter)

first song is Spanish Harlem by Rebecca Pidgeon.
I choose this song since OOYH has a processed version (for demoing) on their website

processed with my new filter https://drive.google.com/file/d/1VeTZ0Y0sU4o-OU9YA8l8DmhiuRV5hgYO/view?usp=sharing

processed with ooyh (volume brought down to match +or-) https://drive.google.com/file/d/1vbyTvOlPVfnHJHpalv472ts5BivEDm7J/view?usp=sharing

since many hear don't like treated rooms (and prefer bass boosted house curves) I processed another one with a filter (Room1) of this dataset (Ash-Ir) of a less treated room https://github.com/ShanonPearce/ASH-IR-Dataset
https://drive.google.com/file/d/14LCNbPJcf3GmFNJSOOqThsNVbhlE71aM/view?usp=sharing

second example is Dvorak 9 (Viena Phil)

processed with my new filter https://drive.google.com/file/d/10QY8lxadRyIybXqEZFG6ikphBzlSxmPC/view?usp=sharing

Ash-Ir_Room1 https://drive.google.com/file/d/1pbCsxwB1Zo1T2IUwYJCV8gYkUy1ONdHQ/view?usp=sharing

EDIT: so what do I think of OOYH? it is very colourized (mid heavie), like all the other solutions I tested. visualy comparing Spanish Harlem files in Audacity, my filter produces waveforms very similar to original, OOYH on the other hand is very diferent
 

phoenixdogfan

Major Contributor
Forum Donor
Joined
Nov 6, 2018
Messages
3,297
Likes
5,075
Location
Nashville
If you're looking at speaker virtualization, I've personally had the best results from Out Of Your Head (OOYH) software, though it's more of an end-user software, not DAW plugin.
https://fongaudio.com/out-of-your-head-software/

There's a discussion on OOYH (where the developer drops by as well) at Head-fi:
https://www.head-fi.org/threads/out-of-your-head-new-virtual-surround-simulator.689299/

The software is finicky, but when it works, the presets (recorded with the Realiser A8 hardware) do allow that feeling of listening to speakers from headphones. It's not going to cut it when you start to listen critically, but for me, of the similar technologies out there, it's the most successful so far in removing the sound pressure you 'feel' on your ears when listening to audio through headphones, and create the impression of sound coming from the front. Within OOYH, I feel the Sonus Faber Elipsa and Genelec presets get the closest to mimicking speakers.

Haven't had the fortune to hear the Realiser A16 though. And also haven't tried NX other than listening to their web demos.
I actually own the A16, and had it shipped to me. It locked up and had to be sent back for service, and I'm awaiting it's return. I've not had a chance to test it out. I also own and have used OOYH for five years, and I find it to be the most impressive of the software based solutions. I also owned the ToneBoosters plug in and never thought it was very good. OOYH, however, did get the image to a point where it seemed to surround my head, but at a very close distance--kind of like near field surround sound. Darren Fong is also great to work with--very nice guy and very service oriented.

OOYH only works for some people, and only with certain presets. The software has 23-25 presets and only two worked for me. Some were absolutely horrendous--like the one taken in a theater--while others seem like the speakers are playing 50 feet away in an underground cave. So never was the phrase YMMV more true--but he does offer free try out, so you can find out if it will work for your individual head, as well as which presets will work for you.
 

3125b

Major Contributor
Joined
May 18, 2020
Messages
1,352
Likes
2,204
Location
Germany
so what do I think of OOYH?
Sounds pretty horrible.
Honestly, they are all worse than just the original file.

If I want to listen to music with speakers, then I use my speakers, if I want to listen with headphones, I use my headphones. But to each his own I guess.
 

whazzup

Addicted to Fun and Learning
Joined
Feb 19, 2020
Messages
575
Likes
486
first song is Spanish Harlem by Rebecca Pidgeon.
I choose this song since OOYH has a processed version (for demoing) on their website

processed with my new filter https://drive.google.com/file/d/1VeTZ0Y0sU4o-OU9YA8l8DmhiuRV5hgYO/view?usp=sharing

processed with ooyh (volume brought down to match +or-) https://drive.google.com/file/d/1vbyTvOlPVfnHJHpalv472ts5BivEDm7J/view?usp=sharing

I'm not a hardware/software engineer, but I'll try my best to give my subjective comments. I alternated between the 2 versions you uploaded, as well as the various presets on the OOYH app itself (while playing Spanish Harlem from Spotify), and of course playing it from my speakers.

If it matters, I'm using the Sennheiser HD598 headphone.

Your filtered version: The reverb is too strong? Sounds too much like a cave, and it makes her voice sound more hollow.

The OOYH version: I tried testing the various presets, but I don't know which preset was used exactly, sounds like the Revel Ultima or Marten Momento V2 preset was used. However, I can say that while there's still reverb (both in the pre-recordable file and (some of) the better presets), her voice is still clear, and 'closer' to what I can hear from the speakers in front of me.


EDIT: so what do I think of OOYH? it is very colourized (mid heavie), like all the other solutions I tested. visualy comparing Spanish Harlem files in Audacity, my filter produces waveforms very similar to original, OOYH on the other hand is very diferent

I do not think producing waveforms closer to the original means anything....you can think of it this way:
Source file (Waveform A) > Speakers Output 'recorded' by our ears (Waveform B) <---- desired

With headphones, now the 'speakers' are strapped to our ears, it becomes...
Source file (Waveform A) > Headphone Output 'recorded' by our ears (Waveform C) <---- undesired (*from virtual surround perspective)

Then the virtual surround filters come in and they seek to make...
Source file (Waveform A) > Find how Waveforms B and C are related, plus consider the equipment (speakers / room / headphones) then somehow magically make a filter that produces > Headphone Output (Waveform D), that sounds like Waveform B

I am obviously grossly understating the process here, but my feeling is while Waveform D will still look like a variation of Waveform A, it probably needs to be transformed in weird ways just so that it gets 'interpreted' by our ears as Waveform B when it comes out from headphones.

You may want to talk to the developer for the Impulcifer software as well
https://www.head-fi.org/threads/recording-impulse-responses-for-speaker-virtualization.890719/
 

whazzup

Addicted to Fun and Learning
Joined
Feb 19, 2020
Messages
575
Likes
486
I actually own the A16, and had it shipped to me. It locked up and had to be sent back for service, and I'm awaiting it's return. I've not had a chance to test it out. I also own and have used OOYH for five years, and I find it to be the most impressive of the software based solutions. I also owned the ToneBoosters plug in and never thought it was very good. OOYH, however, did get the image to a point where it seemed to surround my head, but at a very close distance--kind of like near field surround sound. Darren Fong is also great to work with--very nice guy and very service oriented.

OOYH only works for some people, and only with certain presets. The software has 23-25 presets and only two worked for me. Some were absolutely horrendous--like the one taken in a theater--while others seem like the speakers are playing 50 feet away in an underground cave. So never was the phrase YMMV more true--but he does offer free try out, so you can find out if it will work for your individual head, as well as which presets will work for you.

Same, some of the presets don't really work for me too, I think the room where recording was done makes a large difference. Hope when you get the working A16, give us a review.
 

whazzup

Addicted to Fun and Learning
Joined
Feb 19, 2020
Messages
575
Likes
486
Sounds pretty horrible.
Honestly, they are all worse than just the original file.

If I want to listen to music with speakers, then I use my speakers, if I want to listen with headphones, I use my headphones. But to each his own I guess.

True, a lot of people still feel DSP and equalizers are evil....:D
 

dasdoing

Major Contributor
Joined
May 20, 2020
Messages
4,209
Likes
2,674
Location
Salvador-Bahia-Brasil
Sounds pretty horrible.
Honestly, they are all worse than just the original file.

If I want to listen to music with speakers, then I use my speakers, if I want to listen with headphones, I use my headphones. But to each his own I guess.

If you like the headphone sound you have nothing to fix, and obviously any alteration/compensation, even the perfect one if it exists, will sound worse to you.


Your filtered version: The reverb is too strong? Sounds too much like a cave, and it makes her voice sound more hollow.

You are right. it sounds crap. made the mistake to produce the result on the same day, without hearing it on fresh ears. When I first created a filter last year I could only solve this be producing a minimum phase version of it in REW. But this destroys the spacial ilusion. Also cutting out reverb with a window on Rew destroys the ilusion. atm I am stuck.


The OOYH version: I tried testing the various presets, but I don't know which preset was used exactly, sounds like the Revel Ultima or Marten Momento V2 preset was used. However, I can say that while there's still reverb (both in the pre-recordable file and (some of) the better presets), her voice is still clear, and 'closer' to what I can hear from the speakers in front of me.




I do not think producing waveforms closer to the original means anything....you can think of it this way:
Source file (Waveform A) > Speakers Output 'recorded' by our ears (Waveform B) <---- desired

With headphones, now the 'speakers' are strapped to our ears, it becomes...
Source file (Waveform A) > Headphone Output 'recorded' by our ears (Waveform C) <---- undesired (*from virtual surround perspective)

Then the virtual surround filters come in and they seek to make...
Source file (Waveform A) > Find how Waveforms B and C are related, plus consider the equipment (speakers / room / headphones) then somehow magically make a filter that produces > Headphone Output (Waveform D), that sounds like Waveform B

I am obviously grossly understating the process here, but my feeling is while Waveform D will still look like a variation of Waveform A, it probably needs to be transformed in weird ways just so that it gets 'interpreted' by our ears as Waveform B when it comes out from headphones.

You may want to talk to the developer for the Impulcifer software as well
https://www.head-fi.org/threads/recording-impulse-responses-for-speaker-virtualization.890719/

It's not that I tried to maintain waveform, I just looked for a reason why the OOYH proccesed file sounds so thin.
I wasn't aware of the various presets of OOYH. I sould try the demo version. The spatial ilusion is great, I just didn't like the strong colorarion of the demo song
 

whazzup

Addicted to Fun and Learning
Joined
Feb 19, 2020
Messages
575
Likes
486
You are right. it sounds crap. made the mistake to produce the result on the same day, without hearing it on fresh ears. When I first created a filter last year I could only solve this be producing a minimum phase version of it in REW. But this destroys the spacial ilusion. Also cutting out reverb with a window on Rew destroys the ilusion. atm I am stuck.

Haha, with regards to creating filters, I'm useless to you.

It's not that I tried to maintain waveform, I just looked for a reason why the OOYH proccesed file sounds so thin.
I wasn't aware of the various presets of OOYH. I sould try the demo version. The spatial ilusion is great, I just didn't like the strong colorarion of the demo song

Yes, personally I don't think the pre-recorded demo song is a good representation, BUT it could also be a matter of headphones. There're so many variables unfortunately.
 

odyo

Addicted to Fun and Learning
Joined
Oct 30, 2019
Messages
636
Likes
320
I also find that auto eq and oratory eq really bad for my taste. Sonicworks as well.
 
Top Bottom