[musicreo]

Never heard about such a strange hearing problem but you could measure the speakers normaly and add the delay later in eq-APO to the virtual speaker.

 

[yosoro]

Never heard about such a strange hearing problem but you could measure the speakers normaly and add the delay later in eq-APO to the virtual speaker.

I tried normal measurement and added delay
The result is a low frequency shift to the left

I also tried to put the speaker left channel -10DB
But the result is also the low frequency shift to the left

 

[castleofargh]

I encountered a problem
The position of the sound is not accurate

Because my ears are inconsistent left and right, I have to adjust the delay to listen to the speakers or headphones to enjoy the stereo.


But after using impulcifer, it will cause the low frequency to shift to the left, mid to high frequency shift to the right, and even if the delay on the left and right is adjusted, the stereo sound cannot be normal.

You need to explain things a little more(what you're trying to achieve?). I can't imagine that applying the same delay on speakers and headphones would give a similar experience. On speakers, the one with an added delay will still reach both ears. So the brain will still get the same interaural delay it's getting with any other sound source. It's just that one speaker comes in "late".
While on headphones, it would clearly delay all sounds of one ear. those are 2 pretty different situations with different consequences for something like Impulcifier.
We must know precisely what you need, to even be able to consider if that can be properly done here.

 

[yosoro]

You need to explain things a little more(what you're trying to achieve?). I can't imagine that applying the same delay on speakers and headphones would give a similar experience. On speakers, the one with an added delay will still reach both ears. So the brain will still get the same interaural delay it's getting with any other sound source. It's just that one speaker comes in "late".
While on headphones, it would clearly delay all sounds of one ear. those are 2 pretty different situations with different consequences for something like Impulcifier.
We must know precisely what you need, to even be able to consider if that can be properly done here.

The stereo sound that my ears hear is on the left (you can imagine that there is only left mono)
So add a delay to the left channel so that the stereo MID is in the middle

This method is very effective when impulcifer is not used. Adding a delay after impulcifer will result in only the middle and high frequencies in the middle and the bass on the left.

 

[sander99]

Adding a delay after impulcifer

Do you mean a delay of one channel of the already binauralised signal? That you should never do as it would totally destroy the binauralisation.
If you want to emulate the situation in which one real speaker is delayed then you should delay that channel in the input signal to the binauralisation process [Edit: with the binauralisation process I mean HeSuVi or whatever you use for that]: then one of the virtual speakers will play a delayed signal.

 

[yosoro]

Do you mean a delay of one channel of the already binauralised signal? That you should never do as it would totally destroy the binauralisation.
If you want to emulate the situation in which one real speaker is delayed then you should delay that channel in the input signal to the binauralisation process [Edit: with the binauralisation process I mean HeSuVi or whatever you use for that]: then one of the virtual speakers will play a delayed signal.

I hope you can use pictures to illustrate your correct usage
Google Translate is not very accurate.

 

[sander99]

I hope you can use pictures to illustrate your correct usage
Google Translate is not very accurate.

1. This is what I mean:

2. Unfortunately I don't know how to do it. Sorry.

 

[musicreo]

2. Unfortunately I don't know how to do it. Sorry.

That is easy in EQ-APO
Channel: L
Delay 50.5 ms

If you use HESUVI this must be loaded in Beforehand.

 

[Benik3]

So I played again with Impulcifier. I made even room calibration, but still the sound is bad.
I totally lost bass, the sound is like from the cheapest headphones :/ Also it's VERY quiet. Even after adding 20dB preamp it's still much quieter the without hesuvi.
Here is frequenccy response in REW before and after apply of Hesuvi

 

[johnn29]

If you want to lift the bass you can add a manal EQ to boost bass frequencies.

The quiet HRIR's are an issue I experience too. Equalizer APO seems to run out of headroom for all the EQing Impulcifer needs. I can't just simply preamp boost in EAPO because you get clipping. My only solution is to use voicemeeter. So then I can get Voicemeeter to apply a 12db boost. Gain chaining I guess.

Would be real nice to get a loud HRIR though like the out of the box Dolby Headphone etc. that HeSuVi ships with.

 

[musicreo]

@Benik3 Can you show a screenshot of the frequency response shown in the analyse panel of EQ-APO? Did you use the Harman target for room correction? I use room correction with the Harman target and I have more bass then my headphones would provide without any pre processing.

Is it possible to upload your measurement somewhere?
I'm wondering if this would be generally a good idea to upload some measurements to get a feeling how different measurements are (more in terms of noise balance problems and different measurement equickment) and maybe to find problems that occur during measurements?

For example here is one my latest measurements with one of my primo pairs :

Impulcifer one drive link

It has some balance problems when you analyse the wav files (always higher amplitude on the right channel) but impulcifers correction compensates thats very good for me.

 

[castleofargh]

So I played again with Impulcifier. I made even room calibration, but still the sound is bad.
I totally lost bass, the sound is like from the cheapest headphones :/ Also it's VERY quiet. Even after adding 20dB preamp it's still much quieter the without hesuvi.
Here is frequenccy response in REW before and after apply of Hesuvi

if this 50Hz rockets are a bleeding of sort from the electrical grid, I would guess that the SPL values showed in REW are not the right ones and that at least one of your measurements needs to be done at higher SPL.
No idea if you have other issues.

 

[Benik3]

If you want to lift the bass you can add a manal EQ to boost bass frequencies.

The sound is so bad that I wasn't even able to fix it by EQ. Whole soundstage is weird.

@Benik3 Can you show a screenshot of the frequency response shown in the analyse panel of EQ-APO? Did you use the Harman target for room correction? I use room correction with the Harman target and I have more bass then my headphones would provide without any pre processing.

Is it possible to upload your measurement somewhere?
I'm wondering if this would be generally a good idea to upload some measurements to get a feeling how different measurements are (more in terms of noise balance problems and different measurement equickment) and maybe to find problems that occur during measurements?

Where do you mean to apply Harman correction? I did room calibration as mentioned in wiki of Impulcifer with stereo setup and run the processing command as mentioned on first page on github.
Here is my measuring: https://drive.google.com/file/d/1gg2zaKjE62Q8QMnc23-cnE1xblznHYJt/view?usp=sharing

if this 50Hz rockets are a bleeding of sort from the electrical grid, I would guess that the SPL values showed in REW are not the right ones and that at least one of your measurements needs to be done at higher SPL.
No idea if you have other issues.

The 50Hz peak is really weird, but I don't hear anything like this and when I measure my stereo, it's not there:

BTW when I do processing with --plot option I got this error:

(venv) C:\Users\benik\Desktop\Impulcifer>python impulcifer.py --test_signal="data/sweep-6.15s-48000Hz-32bit-2.93Hz-24000Hz.pkl" --dir_path="data/my_hrir" --plot
Creating impulse response estimator...
Running room correction...
 ** On entry to DLASCLS parameter number  4 had an illegal value
 ** On entry to DLASCLS parameter number  4 had an illegal value
Traceback (most recent call last):
  File "impulcifer.py", line 556, in <module>
    main(**create_cli())
  File "impulcifer.py", line 60, in main
    plot=plot
  File "C:\Users\benik\Desktop\Impulcifer\room_correction.py", line 118, in room_correction
    fr.smoothen_fractional_octave(window_size=1/3, treble_window_size=1/3)
  File "C:\Users\benik\Desktop\Impulcifer\venv\lib\site-packages\autoeq\frequency_response.py", line 1197, in smoothen_fractional_octave
    treble_f_upper=treble_f_upper
  File "C:\Users\benik\Desktop\Impulcifer\venv\lib\site-packages\autoeq\frequency_response.py", line 1155, in _smoothen_fractional_octave
    y_normal = savgol_filter(y_normal, self._window_size(window_size), 2)
  File "C:\Users\benik\Desktop\Impulcifer\venv\lib\site-packages\scipy\signal\_savitzky_golay.py", line 337, in savgol_filter
    coeffs = savgol_coeffs(window_length, polyorder, deriv=deriv, delta=delta)
  File "C:\Users\benik\Desktop\Impulcifer\venv\lib\site-packages\scipy\signal\_savitzky_golay.py", line 139, in savgol_coeffs
    coeffs, _, _, _ = lstsq(A, y)
  File "C:\Users\benik\Desktop\Impulcifer\venv\lib\site-packages\scipy\linalg\basic.py", line 1218, in lstsq
    raise LinAlgError("SVD did not converge in Linear Least Squares")
numpy.linalg.LinAlgError: SVD did not converge in Linear Least Squares

 

[musicreo]

The error during plotting is a bug. I have it too and I can only plot the pre processed files.

A first look into your upload shows me two things:
1) The measured response of the headphones shows some extreme bass at 50-150Hz. What headphones are you using?
2) The wav files show a relative low amplitude and in some of the speaker measurements you have a noise problem on the left channel. Is it possible that there was some low pressure on the left mic? Did you perform the sweeps as load as it is still comfortable for your ear?

 

[Benik3]

I use Corsair HS70 Pro. Maybe is problem that Impulcifier always measure on max volume? (if I got it right from description)
The mic was pretty well fitted in my ear and tried to make sure that they didn't changed over measuring.