Recording Impulse Responses for Speaker Virtualization

[musicreo]

The new measurement is without that high pitched hiss and sounds a lot more like how i expected it to. That being said, i feel like there is a crazy amount of reverb and not much in the way of bass. It makes the music sound sort of like an unprocessed demo version of the mix for each song.

A strong reverb could be caused by a large distance to a speaker (I measured at 1.5m) or a low recording level.

 

[Kosorm]

A strong reverb could be caused by a large distance to a speaker (I measured at 1.5m) or a low recording level.

Was roughly 4-5 feet away from a kali lp6v2 with about 2db of headroom. Anything louder caused clipping. Should i move closer and turn the volume slightly down you think?

 

[Lion3]

but for now here is my my_hrir folder if anybody feels like poking around to see what might have went wrong this time around/what i should fix with this recording.

This your file graph.

There's not enough bass as you hear. (This is the influence of the room.)

It also feels like a reverb when you look at the impulse.

When looking at the opposite ear impulse, the initial reflection compared to the direct sound reacts up to almost -2db, so the change in tone is expected to be very large.

What I kind of wonder about is that your measurements have very low altitudes, and I'm worried that it might be an inaccurate high-pitched representation.
I can understand the measurements with open ears (such as dummy heads), but if you look at your previous post, it seems like you used a plug.
Like I drew a hypothetical line, it's different for each person's face, for each ear, but usually in blocked ear measurements, that's what it looks like.
But it's a good thing if it sounds the same as you hear the speaker.

RT60 is also very high overall. (Looks like 500 to 600ms and low range is experiencing room mode and errors)





https://drive.google.com/drive/folders/164WNZvPqXo00nrvNXKKPaHARNcXZZelC?usp=drive_link

Get hisuvi from the link and apply it to listen. I replaced the bass in your file.
I also limited the entire Decay to 200ms.
Since there seems to be no room treatment, I limited it because I don't think a reverb like me would give you a good feeling
Also testing and I'm listening to your original Hesuvi but it works fine for me.
Check your playback unit sample rate and the number of bits (24 bits 48000)

C:\Program Files\EqualizerAPO\config\HeSuVi\hrir

Please make sure that you put hisuvi.wav in this folder.

 

[Kosorm]

This your file graph.



There's not enough bass as you hear. (This is the influence of the room.)



It also feels like a reverb when you look at the impulse.




When looking at the opposite ear impulse, the initial reflection compared to the direct sound reacts up to almost -2db, so the change in tone is expected to be very large.






What I kind of wonder about is that your measurements have very low altitudes, and I'm worried that it might be an inaccurate high-pitched representation.
I can understand the measurements with open ears (such as dummy heads), but if you look at your previous post, it seems like you used a plug.
Like I drew a hypothetical line, it's different for each person's face, for each ear, but usually in blocked ear measurements, that's what it looks like.
But it's a good thing if it sounds the same as you hear the speaker.






RT60 is also very high overall. (Looks like 500 to 600ms and low range is experiencing room mode and errors)





https://drive.google.com/drive/folders/164WNZvPqXo00nrvNXKKPaHARNcXZZelC?usp=drive_link




Get hisuvi from the link and apply it to listen. I replaced the bass in your file.
I also limited the entire Decay to 200ms.
Since there seems to be no room treatment, I limited it because I don't think a reverb like me would give you a good feeling
Also testing and I'm listening to your original Hesuvi but it works fine for me.
Check your playback unit sample rate and the number of bits (24 bits 48000)

C:\Program Files\EqualizerAPO\config\HeSuVi\hrir

Please make sure that you put hisuvi.wav in this folder.

Thank you so much. Ill try this out soon and report back to you. Based on what you have seen, is there anything you would recommend me change about my recording process going forward to ease some of these issues?

 

[Lion3]

Based on what you have seen, is there anything you would recommend me change about my recording process going forward to ease some of these issues?

I'm not sure...
Not necessarily, but next time, if you record again, put on a microphone and take a picture of your ears

First of all, when you look at the location of the resonance point, you can infer that the insertion depth is not very deep.
Assuming that black is a plug and white is a microphone, I think this is what it looks like.
+++
But I'm Asian, and I'm not sure about this because there is a difference between Western and Asian ears.
But it's a story compared to a lot of curves with average resonance points.

But I do recommend a deeper depth of insertion in this way (or when I recommend it to others).
But if you have a certain level of depth, and if the resonance in your ears is stimulated well, it can be regenerated
As I told you before, I don't think you need to worry if it's similar to listening to the speaker.

ecause the rest of the early reflection is difficult unless you're room-treated, getting rid of any interiors that might be covered by the speaker's route to your ears? That might be best.

 

[Kosorm]

I'm not sure...
Not necessarily, but next time, if you record again, put on a microphone and take a picture of your ears




First of all, when you look at the location of the resonance point, you can infer that the insertion depth is not very deep.
Assuming that black is a plug and white is a microphone, I think this is what it looks like.
+++
But I'm Asian, and I'm not sure about this because there is a difference between Western and Asian ears.
But it's a story compared to a lot of curves with average resonance points.




But I do recommend a deeper depth of insertion in this way (or when I recommend it to others).
But if you have a certain level of depth, and if the resonance in your ears is stimulated well, it can be regenerated
As I told you before, I don't think you need to worry if it's similar to listening to the speaker.

ecause the rest of the early reflection is difficult unless you're room-treated, getting rid of any interiors that might be covered by the speaker's route to your ears? That might be best.

Wow! I got hesuvi working properly and that edited version you sent me finally made me fully understand just how incredible impulcifer can be. There are a few spots were i can see room for improvement but i can already appreciate this more than the headphone on its own. What did you do exactly to elevate the bass this way? I would like to do that myself for future recordings as well. The change to decay fixed the reverb issues basically entirely as well. As for the fitting into my ears, im unsure if i can insert them deeper but i will try to figure something out and report back. I will also take a picture when i go for the next set of recordings. I will also see about room treatment in the future as per your suggestion. Thank you a ton! Music brings tears to my eyes when played with this virtualization

 

[Lion3]

I got hesuvi working properly

good !
I'm glad it works.
Of course, you can also convolve directly in the apo with the command that reo told you, but it will be more complicated if it becomes 7 channels instead of 2 channels, and it is easier to write with one hesuvi.txt file.

What did you do exactly to elevate the bass this way?

There are two (or more options).
1. One is doing EQ. - This can fix the tone balance to some extent, but you can't delete the room mode in the room or change the reverberation at all. (Until group delay)
2. The other one is a synthesis. It's like aligning a subwoofer with a speaker in real life, aligning a different bass response to your response.

The bass is hardly affected by hrtf, and there is no individual variation.
And it's a bit more convenient to synthesize because it's quite expensive to effectively control the bath in a normal space. (Of course, you can use the decay function of the impulcifer to cut some reverberations, but it's a bit of a different story because reverberations are cut off and the problem with the room doesn't disappear.)

The change to decay fixed the reverb issues basically entirely as well

good.
Is your space big?
Of course, some reflections can be understood if the recording distance exceeds approximately 3-5 m.
But when I converted the distance feet you mentioned into m units, it was close.
So we decided that the original recordings had quite a lot of reverb.
Generally, 100ms to 200ms is enough to give you some feeling.
It would be more convenient to manipulate and synthesize by applying an additional vst to a response of 100 ms.

I will also see about room treatment in the future as per your suggestion.

It's just my recommendation. Don't worry too much. I and other users are just saying things that will most benefit you in their own experience based on the limited information you've given us.
If you think you can hear it well now, you can keep it as it is.

Music brings tears to my eyes when played with this virtualization

lol
I often listen to brir in the forest.
It may be a sudden story, but it also comes out of the blue in the paper.
Apparent source width (ASW) is heavily influenced by visuals as well.
The size of the display I'm seeing (such as a smartphone, a 27 inch monitor, a screen size like a movie theater) or the open plains, forests, or closed eyes at all to provide visual information. ---> Our brains are smart, but they are more stupid than you think. So we have to cheat well.
The experience will be different because the brain correction will change depending on this.
I have a personal opinion that the final destination of BRIR is how much better to trick the brain after all.
I can understand the appreciation of your tears, and I'm hoping even more that you'll hear unreal sounds in the future that you can't get in real life.
Congratulations


First of all, congratulations once again on how well it plays.
Not only stereo, but since I'm measuring it by turning my body around
It would be fun to record the center (0 degree mono), the side (90 to 110 degrees), and the rear (140 to 150) together.
On top of that, you can play virtual 7.1 using VB cable and have a fun experience through upmixing stationery.
Lastly, I'm answering using a translator, so my words might be a little weird. Please understand!

 

[Kosorm]

good !
I'm glad it works.
Of course, you can also convolve directly in the apo with the command that reo told you, but it will be more complicated if it becomes 7 channels instead of 2 channels, and it is easier to write with one hesuvi.txt file.


There are two (or more options).
1. One is doing EQ. - This can fix the tone balance to some extent, but you can't delete the room mode in the room or change the reverberation at all. (Until group delay)
2. The other one is a synthesis. It's like aligning a subwoofer with a speaker in real life, aligning a different bass response to your response.

The bass is hardly affected by hrtf, and there is no individual variation.
And it's a bit more convenient to synthesize because it's quite expensive to effectively control the bath in a normal space. (Of course, you can use the decay function of the impulcifer to cut some reverberations, but it's a bit of a different story because reverberations are cut off and the problem with the room doesn't disappear.)


good.
Is your space big?
Of course, some reflections can be understood if the recording distance exceeds approximately 3-5 m.
But when I converted the distance feet you mentioned into m units, it was close.
So we decided that the original recordings had quite a lot of reverb.
Generally, 100ms to 200ms is enough to give you some feeling.
It would be more convenient to manipulate and synthesize by applying an additional vst to a response of 100 ms.



It's just my recommendation. Don't worry too much. I and other users are just saying things that will most benefit you in their own experience based on the limited information you've given us.
If you think you can hear it well now, you can keep it as it is.


lol
I often listen to brir in the forest.
It may be a sudden story, but it also comes out of the blue in the paper.
Apparent source width (ASW) is heavily influenced by visuals as well.
The size of the display I'm seeing (such as a smartphone, a 27 inch monitor, a screen size like a movie theater) or the open plains, forests, or closed eyes at all to provide visual information. ---> Our brains are smart, but they are more stupid than you think. So we have to cheat well.
The experience will be different because the brain correction will change depending on this.
I have a personal opinion that the final destination of BRIR is how much better to trick the brain after all.
I can understand the appreciation of your tears, and I'm hoping even more that you'll hear unreal sounds in the future that you can't get in real life.
Congratulations


First of all, congratulations once again on how well it plays.
Not only stereo, but since I'm measuring it by turning my body around
It would be fun to record the center (0 degree mono), the side (90 to 110 degrees), and the rear (140 to 150) together.
On top of that, you can play virtual 7.1 using VB cable and have a fun experience through upmixing stationery.
Lastly, I'm answering using a translator, so my words might be a little weird. Please understand!

Thank you for this, its incredibly useful for the future. I think i have discovered why my data looks so...strange. i believe that my Wiim Ultra is the culprit.i was running an optical cable out of my pc to the Wiim Ultra which then ran out to the speakers. The speakers were then connected through rca to an rca switcher that ran to the wiim ultra and i would switch between my headphone amp and the speakers. When listening out through my headphones with the hesuvi on, everything sounded very thin and compressed unlike the dac/amp hooked directly to my pc.

This leads me to believe something in the chain is heavily compressing sound and that what my speakers played was warped by this so the microphones picked up that strangeness.i will have to experiment more with this in the morning but i hope i can come to some sort of solution with this.

Either way though, my listening experience is already greatly elevated and this just leaves me excited to see what else I can extract out of this setup!

 

[castleofargh]

lol
I often listen to brir in the forest.
It may be a sudden story, but it also comes out of the blue in the paper.
Apparent source width (ASW) is heavily influenced by visuals as well.
The size of the display I'm seeing (such as a smartphone, a 27 inch monitor, a screen size like a movie theater) or the open plains, forests, or closed eyes at all to provide visual information. ---> Our brains are smart, but they are more stupid than you think. So we have to cheat well.
The experience will be different because the brain correction will change depending on this.
I have a personal opinion that the final destination of BRIR is how much better to trick the brain after all.

The eyes are an integral part of audio spatial location. The brain trusts sight more than hearing, and learns to interpret some audio cues based on how they correlate with vision. And because we just suck at estimating distances with our ears, the brain is naturally going to rely on the easiest way to get an answer(easiest in terms of processing power I think?) which of course is to just look at how far something is.
I also get some nice surprises listening to processed tracks on the go. I have a thing for large rivers, or the beach, but right now I'm living in the mountains, so the forest is my best option.
At home, I still use "dummy" speakers to anchor the sound, because without them, my brain decides that sounds come from my computer screen. It's still a small room, but I manage to just about double the perceived distance with that. Also, those 2 speakers somehow define where other virtual speakers are in a 5.1 or 7.1 playback(the closer the speakers I see, the closer the ones I don't). The brain is a funny thing.

 

[Lion3]

hen listening out through my headphones with the hesuvi on, everything sounded very thin and compressed unlike the dac/amp hooked directly to my pc.

This leads me to believe something in the chain is heavily compressing sound and that what my speakers played was warped by this so the microphones picked up that strangeness.i will have to experiment more with this in the morning but i hope i can come to some sort of solution with this.

interesting. If that was the cause, I hope it will be resolved

Either way though, my listening experience is already greatly elevated and this just leaves me excited to see what else I can extract out of this setup!

Yes. Welcome to the world of BRIR
As I said before, consider a 7ch recording. It takes longer than a 2ch recording, but it would be quite useful to record.
And by the time your recording is done with stability, and everything is somewhat stabilized (when you're mostly satisfied with it without any big issues), try IEM instead of headphones at that time.
Headphones are great, of course, but BRIR's experience listening on an IEM is incredible. Undistinguishable from reality. (Extremely low noise floor, exterior noise blocking, linear distortion cleaner than headphones)
I use several IEMs now
Use Apple's Earpods 3.5mm for comfortable listening
I use my ciem when I want to concentrate and listen.

 

[Kosorm]

interesting. If that was the cause, I hope it will be resolved


Yes. Welcome to the world of BRIR
As I said before, consider a 7ch recording. It takes longer than a 2ch recording, but it would be quite useful to record.
And by the time your recording is done with stability, and everything is somewhat stabilized (when you're mostly satisfied with it without any big issues), try IEM instead of headphones at that time.
Headphones are great, of course, but BRIR's experience listening on an IEM is incredible. Undistinguishable from reality. (Extremely low noise floor, exterior noise blocking, linear distortion cleaner than headphones)
I use several IEMs now
Use Apple's Earpods 3.5mm for comfortable listening
I use my ciem when I want to concentrate and listen.

Ooo that sounds great. I plan to buy a pair once everything is said and done. Never went down the iem rabbit hole but i might get a pair of etymotics for this

 

[Lion3]

The eyes are an integral part of audio spatial location. The brain trusts sight more than hearing, and learns to interpret some audio cues based on how they correlate with vision. And because we just suck at estimating distances with our ears, the brain is naturally going to rely on the easiest way to get an answer(easiest in terms of processing power I think?) which of course is to just look at how far something is.
I also get some nice surprises listening to processed tracks on the go. I have a thing for large rivers, or the beach, but right now I'm living in the mountains, so the forest is my best option.
At home, I still use "dummy" speakers to anchor the sound, because without them, my brain decides that sounds come from my computer screen. It's still a small room, but I manage to just about double the perceived distance with that. Also, those 2 speakers somehow define where other virtual speakers are in a 5.1 or 7.1 playback(the closer the speakers I see, the closer the ones I don't). The brain is a funny thing.

Thank you for the nice words, and I agree with you.
To make the brain more confusing, some users have even considered Tactile Bath and VR devices.
I've tried Oculus' VR device, but I can't use it because my eyes are very bad and the fit is a little uncomfortable.
If you have a subwoofer, it wouldn't be a bad way to leave it behind your back as a nearfield subwoofer and feel it.
As a head tracking, it can help improve the reproducibility to some extent, but personally, I didn't feel the need for head tracking.

 

[Kosorm]

I'm not sure...
Not necessarily, but next time, if you record again, put on a microphone and take a picture of your ears




First of all, when you look at the location of the resonance point, you can infer that the insertion depth is not very deep.
Assuming that black is a plug and white is a microphone, I think this is what it looks like.
+++
But I'm Asian, and I'm not sure about this because there is a difference between Western and Asian ears.
But it's a story compared to a lot of curves with average resonance points.



But I do recommend a deeper depth of insertion in this way (or when I recommend it to others).
But if you have a certain level of depth, and if the resonance in your ears is stimulated well, it can be regenerated

Hi XingYu, this is roughly as deep as i can physically push them into my ears. It doesnt hurt but i cant physically push them further in. Is this an okay depth?

Also ive confirmed when i went into the Wiim Home app that somehow i must have accidentally turned up around 4khz by multiple db haha. Everything else was completely flat except for that, so knowing this I can do a more accurate measurement next time

 

[Lion3]

Hi XingYu, this is roughly as deep as i can physically push them into my ears. It doesnt hurt but i cant physically push them further in. Is this an okay depth?

Also ive confirmed when i went into the Wiim Home app that somehow i must have accidentally turned up around 4khz by multiple db haha. Everything else was completely flat except for that, so knowing this I can do a more accurate measurement next time

good . That seems to be enough. Also, it's good news for you that you've found a 4 kHz emphasis EQ.
I think we can measure it as it is.

And to tell you one more thing
Record several times at the same angle.
For example
Let's say we record the L channel on the front.
Then as you record, you can move and breathe.
Keep that in mind and record two or three times in one position. Also, keep your back straight and stay comfortable without lowering your chin. (It's easy to say, but it's actually hard.)

 

[Kosorm]

good . That seems to be enough. Also, it's good news for you that you've found a 4 kHz emphasis EQ.
I think we can measure it as it is.

And to tell you one more thing
Record several times at the same angle.
For example
Let's say we record the L channel on the front.
Then as you record, you can move and breathe.
Keep that in mind and record two or three times in one position. Also, keep your back straight and stay comfortable without lowering your chin. (It's easy to say, but it's actually hard.)

Good to know, thank you. trying to do a measurement now and i noticed every one of my headphone recordings look like this
Is there anything wrong with my right microphone that its picking up so much static?