• 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!

Replicate imagined speaker sound to headphones. (With Impulcifer)

LIΟN

Active Member
Joined
Oct 17, 2023
Messages
159
Likes
172
Location
South Korea
Hi. I live in South Korea, and I enjoy exploring and studying various audio information from all over the world. I use a translator. I'm sorry the sentence wasn't smooth.
I used the speaker set up in the in-room and was quite satisfied. But I had no choice but to dispose of the speaker because of my family and various personal reasons.
And I started using various headphones. (Koss KSC75 / Sony MV1 / Sennheiser HD600 / Sennheiser HD800s)

While enjoying music like that, I learned about binaural and impulcifer things.
A few months ago, I carefully asked a senior user who has a very good room in Korea, so I enjoyed replicating the sound of the room. (Perfect REVEL speaker)
But I always had a desire for something clearer, something I could control... In order to solve it, i eventually measured it myself.

Of course, I was already experiencing a crazy experience by adding a measurement from the reflection angle separately to the response I was using and applying binaural reverb to it. But everyone is greedy.

9.png

6 (1).png

11.png



And the following is part of the process that I've been doing so far. There are so many things to do in the future. an uncharted world

002.jpg


I used a Generec speaker previously but disposed of it, and this time I bought a 305mkii speaker from JBL for binaural. (Of course this was not a very good choice.)
I measured Onaxis at a height of 2m from the rooftop and made my own filter in advance to correct the deviations and problems with my speakers.

And I waited for months to prepare and plan where and how I would measure and prepare again.

What I was able to try and be confident about was that I could mix the responses I wanted by mixing impulcifer with REW and audacity properly.
And you can apply as many corrections as you want to be more flexible than the situation in the in-room.
But the most important thing is that you rarely need that kind of correction, and it's best to get pure sound, right?
However, ordinary people cannot live in a fragrance-free room. Furthermore, we hear reflections as well as direct sounds.

So... I rented a soccer field and a big gym(or hall? whatever) in my town.

1700019973.jpg

1699949801.jpg

1699949801 (1).jpg


weather is really nice.
But there were a number of problems.
Software issue with topping E2X2 at the time. Noise floor outdoors (it was too noisy around). And there were many mistakes, so the measurement on this day was too difficult, but I couldn't use a single one. So I was in despair.

And I prepared and prepared for a few more days (body and mind). Information. Knowledge, everything I need)


15.jpg


It was quite quiet here. And this day was even quieter because I took the test at high schools all over Korea. (On this day, the plane doesn't even move at a certain time.)
I had written down the whole process on a piece of paper and kept checking.

1698749870.jpg


Btw, microphone I used is Samson's LM-10X.(With Rode VXLR) 2pair.
Also Umik-1 (for Check and Match Speaker Level, Impulse)

16.jpg


Measurements were made at 1m / 1.3m / 2.4m / 3m / 7m when held large , and specific angle arrangements and distances for experiments were also attempted in the middle.
Speaker height/measurement and ear height is 2m.

I think the picture above is about 1.3m.
I bought a small mirror the day before at Daiso and checked the reflection point with a mirror on the ladder to place the sound-absorbing material(basstrap), and about 1 to 1.5m, it was covered with one sound-absorbing material (1m). Glass wool

So, floor reflection absorption was performed with 500mm Glass wool, and after that, it was divided into 250mm at a distance above that.

11.jpg


I've already marked the protractor and the specific angles I want on the ladder.

14.jpg


1700208426.jpg


And i've also experimented with a lot of things that aren't normal stereo angles.
To effectively remove crosstalk (XTC), bring it closer to mono, but not at any angle
Referring to various data, we tried to arrange them in various ways, such as an 8 degree angle. I remember that the distance was about 2.7m? 3m.
As shown in the picture, all arrangements were laser-adjusted toe-in angles except for more than 3 meters, temporarily distance-adjusted with tape measure, and the speakers and speakers were adjusted accurately once again using a laser thermometer. (It's a part that helps ITD reunification)

10.jpg


12.jpg


13.jpg


It doesn't look that far in the picture, but it's about 7 meters. The XLR cable was 10 meters long, so I could afford it, but the speakers and stands were so unstable that I took it here. Even if you go more than that, the performance of 305 won't support you. (Compress Gurley used it all up to the maximum level and later found the Maginot line volume and increased the microphone amplification a little more. The headroom fell a bit, but it was indoors, so there was no problem with the file itself.)

acac.png


And it's one of the reflection angles collected at different distances. The reflection was measured by turning a 360 degree angle at a 5 degree angle, and I put an eyeglass case behind the speaker to tilt it
It was so unstable that I only did one height layer. (72 measurements of ear height, 72 measurements of height channel higher than ear height)
And as in the picture above, I shot the speaker directly on the wall or looked at the ceiling. I ran out of sound pressure and used up to the limit of 305 speakers today (I actually hear the distortion even in the sweep)

acac2.png

acac3.png


But the reflections were very satisfactory. The most important thing is the reflection measured to fit my ears and I really like the amount of the initial reflection itself.
I think it will be a reflection IR with a length of approximately 1.4s, and it would be good to process it well.

And........

006.png


I got really clean sound fit my ears. -47db lol (This messurement IR file from 2 or 2.4 meter)

007.png


And it's a response when my speakers and direct bass are combined. very clean. (Never mind the low-frequency boost or not. I can raise it as much as I want.)

001.png


And this is my ears messurement. It's a different response from the umik mic, right?But this is how I hear (face, shoulder, etc.).
I previously thought of a 50 Hz crossover, but I changed it to 90 Hz. The target is LR4.

002.png

And it's my response that I got really cleanly. Too clean and sharp.


003.png


004.png




Decay is Perfect!
And most importantly, this is not the end. I have 72 messages for each distance at a 360 degree angle (there's another height channel!) This is that I can apply as many reflections as I want.
I can create "the space I want" to be precise.
And this clean direct sound is just a white drawing paper for it.
unspoiled paper.


008.png


Phase is also good!

This is what some users think. (In Korea, too)
Why don't you listen to it on the speaker?
Why would you want to replicate speakers with headphones?

Not everyone can have an ideal room. But everyone has a desire for sound.
And this is one of the ways to relieve that desire a lot.
This is the same as a speaker system under the assumption that it is well-measured.

I'm under the illusion that I'm listening to a large JBL horn speaker, not a small book shelf speaker 305 on JBL anymore. File measured at a distance of 3 meters.
In fact, it's not at all 3m away from that space. Nearfield.
But nevertheless, the sound is so huge and clear that it wraps around me.
Now that all the ingredients I can play with are ready, I can mix as many reflections as I want.
 
Last edited:
OP
LIΟN

LIΟN

Active Member
Joined
Oct 17, 2023
Messages
159
Likes
172
Location
South Korea
Always nice to see that some people spare no effort and have my utmost respect!
Thank you for your insight and kindness.
This has been quite a difficult, complicated and long process, and I'm just getting the ingredients ready now. I'm excited to see how much more I can do in the future.
 

LTig

Master Contributor
Forum Donor
Joined
Feb 27, 2019
Messages
5,833
Likes
9,571
Location
Europe
I see one problem in almost all measurement setups. The speaker position on the stand is recessed so there will be reflections which influences the frequency response. To prevent this move the speaker forward so its front is flush to the plate it is sitting on.
 
OP
LIΟN

LIΟN

Active Member
Joined
Oct 17, 2023
Messages
159
Likes
172
Location
South Korea
I see one problem in almost all measurement setups. The speaker position on the stand is recessed so there will be reflections which influences the frequency response. To prevent this move the speaker forward so its front is flush to the plate it is sitting on.
Thank you for the comments.
Yes. I already knew that. However, I bought a cheap stand and had to put it in the center because the center of gravity was very unstable.
But still, thank you for your advice.
 
OP
LIΟN

LIΟN

Active Member
Joined
Oct 17, 2023
Messages
159
Likes
172
Location
South Korea
KakaoTalk_20231116_161156728_13.jpg


Here's what it looked like when I wore the earmic. (In case you were wondering. Sorry if not)
Push in as far as you can while pulling on the top of the ear.
And when you get to the point where there's a secondary bend in the ear canal, you can pull down on the ear and push it in deeper with a slight upward flick, and you'll hear a rustling sound and you'll be able to get it in almost all the way.
The microphone I used is a 3mm microphone and I used Final Audio's E500 E-type SS size ear tips. If you push the microphone all the way in, it will be almost completely invisible in the mirror. I ran the line over the ears and attached it with medical tape behind the ears and around the neck to keep it from flopping around.
 

Tim Link

Addicted to Fun and Learning
Forum Donor
Joined
Apr 10, 2020
Messages
766
Likes
656
Location
Eugene, OR
Wow, that's quite an undertaking! You'll have a super customized system for your ears!
 
OP
LIΟN

LIΟN

Active Member
Joined
Oct 17, 2023
Messages
159
Likes
172
Location
South Korea
Wow, that's quite an undertaking! You'll have a super customized system for your ears!
Thanks. I think in your room, if you measure using a microphone like this yourself, you can see that your XTC is working effectively. I'm curious about your sound
 
OP
LIΟN

LIΟN

Active Member
Joined
Oct 17, 2023
Messages
159
Likes
172
Location
South Korea
Amazing effort and perseverance.
Thank you. I've been through this whole process and realized once again that measurement professionals around the world, with you at the forefront, have it so hard.
Putting a speaker on a stand to listen to a song is very different from setting it up for measurement, and it's a lot of work.
It's really great that you do that every time.
Thank you for your dedication to your speaker and headphone reviews.
 
OP
LIΟN

LIΟN

Active Member
Joined
Oct 17, 2023
Messages
159
Likes
172
Location
South Korea
머리터진다에.png


머리터진다에2.png


aa.jpg


Near mono placement (8-degree placement to be exact, not the usual 30-degree stereo) is Main Response. (FL-L, FR-R)

14.jpg

something like this.

And
I combined the measurements from the angles that maximize ASW and LEV.
And when I applied the XTC, the 8 is also much wider than the XTC.
The ASW/LEV angles measured at 3 meters, so there is a lack of early reflections, but the sound at that angle is still enough to trick my brain. It stimulates very well.

머리터진다에3.png


I removed the crosstalk information from about 250 to 8000hz.

Now what I need to do is add to that another measurement from an additional distance (rich in early reflections) and merge, merge, duplicate, merge again.
How much?

As much as I want.
 
OP
LIΟN

LIΟN

Active Member
Joined
Oct 17, 2023
Messages
159
Likes
172
Location
South Korea
1.png


I was previously using a direct base with align. and asked several senior users in Korea for their in-room measurement responses(messured Umik or Earthwork things) and received a few.
I've combined several in-room responses, including quad subwoofer and octa-subwoofer setups, and they work perfectly. My ear delay tricks my brain into thinking I'm hearing bass in the room. Below about 150-200Hz, you can synthesize and use this way. Of course, you can measure with your own ears, but it's not easy for anyone to get a setting close to perfect bass. (Their dedication and hard work in removing most room modes)
The above capture combines the octa-subwoofer responses. (90hz LR4 target)

2.png


And this bass is driving me crazy. It hits me like it's really hard to breathe. The bass is clean and punchy, something you don't get with a typical setup (single or dual subwoofers).

3.png


Of course, as I further calibrated the subwoofer integration and the front response at my favorite distance (3 meters), I accumulated too many graphs. It's a bit overwhelming.
The problem is that I have 72 measurements at each distance, 360 degrees of ear height, and 72 measurements of height channels higher than the ear. When do I finish this?
omg... :facepalm:
 

Hayabusa

Addicted to Fun and Learning
Joined
Oct 12, 2019
Messages
833
Likes
575
Location
Abu Dhabi
Wow! impressive!

Now: the atmos version with 13 channels :cool:
 
OP
LIΟN

LIΟN

Active Member
Joined
Oct 17, 2023
Messages
159
Likes
172
Location
South Korea
A rigorous measurement and science-based approach to achieving a specific personal subjective goal. I love it.

And thanks for taking all the extra time and effort to write it up and share you photos and graphs.

Thanks! I know I have a very long way to go, but it's still exciting and fun to think that much of what I imagine can be achieved by digitally synthesizing analog recordings!
Incredibly impressive work!

Thanks for sharing.

Thank you. I'll post some progress from time to time!I'll be sure to post the course in the future.
Wow! impressive!

Now: the atmos version with 13 channels :cool:
I have the materials for an Atmos, but I need a device that can shoot a signal to it. Omg... =( I remember listening to Apple Music spatial sounds with Atmos, it was very interesting and I felt like I was inside the music. It was really refreshing. I envy your experience.
 
Last edited:
OP
LIΟN

LIΟN

Active Member
Joined
Oct 17, 2023
Messages
159
Likes
172
Location
South Korea
Tonight, I had a pretty interesting test.

1.png


Green is Direct Sound (messured UMIK 1)
Pupple is My HRIR. (Include HRTF : Ear,Head,Shoulder...etc)

And I tried to extract this.

2.png


It was accurately extracted using Inversion.

3.png


And it looks the same when I multiply it by my original Umik measured response.

4.png


This is combined with Dirac pulse so that it doesn't affect below 900 Hz for testing earlier.

5.png


And I applied shelf filter, etc. according to the shape of my crosstalk channel since it is a test first. (I didn't even work on the left and right separately. I copied it and used it)
Just test!!

6.png


And I looped it out of REW and set up a random delay in audacity. (Delays that vary depending on the distance to the speaker, the distance to the speaker, and the angle I know myself)
270us in this case (tried with 13 samples)
4ch True Stereo

7.png


And i did loopback this virtual response back to the four-channel True Stereo.

8.png


And of course, when i run the impulcifer, the filter is made well.

And I heard this to solve my curiosity.
It's better than I thought. What I was concerned about was that if we applied the HRTF response curve to any response as it is, wouldn't we hear a number of side effects when we misapplied EQ in the in-room or when we applied it at a sharp angle.
But so far, I don't hear it at all. It's similar to what I actually recorded.

Some people may wonder why I'm bothering this process now.
It's because if I'm right in my mind, I'll get even more perfect sound if I meet a few conditions.

1. Binaural ear mic measured responses should also be close to almost direct sound.
= I think this will increase the probability of generating a clean response.
2. There should be no difference between the microphone wearing stage and the speaker setting.
= Of course, there is a difference between the ear and the body. It's obvious because people are not symmetrical.
But there should never be a deviation in the category of "setting."

If I'm right... If this works well (it's still working well now)
1. It can surpass the performance of recorded speakers. (distortion and various problems)
= We will be able to create a speaker sound that is almost perfect. Depending on your preference, dip or peak... I think I can design the sound.
2. In the in-room, you can implement the hard-to-make sound as you want.
= If you have a precise extraction of your own personalized response curves that respond to distance and angle, you can mix as many direct notes as you want or listen as you want. The same goes for reverberations. If you already have a response to the limit of the individual HRTF, regardless of space, you can apply the desired pattern of reversals to combine them perfectly as much as you want.

001.png


In addition, the cumulative channel can be completely controlled. It can also perfectly correct imbalances that inevitably appear on the speaker or where you record it. (I think the Crosstalk channel is the sound of a shadowy place.

002.png


Isn't the response so lovely and beautiful?
I didn't prepare and record for this purpose for the past few months.
But I was curious today, so I tried... It's much better than I thought.
All variables, deficiencies in the speaker itself, and deficiencies in the recording can be removed.
 
Last edited:

jimbomb

Member
Joined
May 13, 2020
Messages
21
Likes
15
For anyone interested I've been using Impulcifer for about 3 years. I did all the configuration in my living room using relatively inexpensive equipment. The results for me are incredible, really beyond magical. I've owned very good speakers over the decades and listened to many others, what I'm hearing with my Senn HD600 headphones and Impulcifer is easily the best sound I've ever heard. The illusion is literally as if I'm listening live or at least to very high end speakers 8 ft in front of me in a virtually perfect room. The soundstage is as wide and deep as the recording, the imaging/detail is amazing, and the tonal balance is just exceptional. I still sometimes take off my headphones to make sure I don't have my speakers on disturbing the neighbors. I only use it for listening to stereo music recordings so can't comment about multi-channel surround.
The setup, measurement and tuning does take some DIY effort, there are equipment costs, and command line details to learn. But if you want to use headphones with spatial realism better than you imagine, and truly exquisite sound quality, it's more than worth the effort.
 
OP
LIΟN

LIΟN

Active Member
Joined
Oct 17, 2023
Messages
159
Likes
172
Location
South Korea
For anyone interested I've been using Impulcifer for about 3 years. I did all the configuration in my living room using relatively inexpensive equipment. The results for me are incredible, really beyond magical. I've owned very good speakers over the decades and listened to many others, what I'm hearing with my Senn HD600 headphones and Impulcifer is easily the best sound I've ever heard. The illusion is literally as if I'm listening live or at least to very high end speakers 8 ft in front of me in a virtually perfect room. The soundstage is as wide and deep as the recording, the imaging/detail is amazing, and the tonal balance is just exceptional. I still sometimes take off my headphones to make sure I don't have my speakers on disturbing the neighbors. I only use it for listening to stereo music recordings so can't comment about multi-channel surround.
The setup, measurement and tuning does take some DIY effort, there are equipment costs, and command line details to learn. But if you want to use headphones with spatial realism better than you imagine, and truly exquisite sound quality, it's more than worth the effort.
Nice to meet another impulcifer user!
There are too few users even if you look at articles from all over the world.

Have you tried it with IEM?
I've tried this with ZeroRed and AppleEarpods and Weston's W10 and I'm very happy with it.
You can use the Headphones response in combination with the GRAS response.

1702829976.png


something like this (This headphones.wav files for Weston W10)

1702831270.jpg
1702993778.jpg




I also have HD800s, but I mostly use IEM these days. The fine details of ultra-high frequency are also better heard and superior. (Of course that's how much the reflection is affected too lol)
 

jimbomb

Member
Joined
May 13, 2020
Messages
21
Likes
15
Nice to meet another impulcifer user!
There are too few users even if you look at articles from all over the world.

Have you tried it with IEM?

I agree, I've often posted in different forums and encounter very few other users. I do realize there is a fairly significant hurdle in making the measurements, and of course we can't demo the results to anyone else. One of my motivations was being able to listen at any time of day or night at realistic volumes without being concerned about disturbing anyone else, as well as being a lifelong seeker of the best sound. When I happened across Impulcifer and ended up with the stunning results it was a dream come true. I often think if everyone could hear what I'm hearing there would be millions of users.
I haven't tried it with IEM, but will seriously look into it.
Thanks.
 
Top Bottom