[yosoro]

I found the built-in speaker frequency response of Realiser A8 in the discussion about Realiser A16.
Is it possible to run them in Impulcifer?

 

[yosoro]

This is the file sent by A16 user
https://mega.nz/file/GZsxxa6a#Su6w6ipONLn-jOsy67knY6H4nePNLja8V-ijPPerDCE

 

[johnn29]

So I tried the --decay parameter on the speaker recordings I did without a sub, which then would rely on virtual room correction to boost the bass. Initially these were the ones that had that virtual channel bleed on channel ID call outs. Happy to say it's completely fixed them and none of the issues with the reverb crop. With a simple crop it just sounded way more in your head.

Also I'm constantly amazed at our perceptions. I took a BRIR in front of my speakers the same place with headphones. Sounded amazing and subjectively much better than any of the previous BRIR's I did. I go to sleep, go to my desk where there's no speakers - and another BRIR sounds better. We absolutely do listen with our eyes. It sounds insane but I think this has really helped me realize why there's so much quackery with audiohphilla.

Also anyone getting errors plotting charts make sure you update your dependencies

pip install -U -r requirements.txt

Happened to me on one of my recordings (not the other) but soon as I updated it went through fine

 

[musicreo]

I tested the decay option. It improves the sounding of my measurement a little. I like it.

Also anyone getting errors plotting charts make sure you update your dependencies

pip install -U -r requirements.txt

Happened to me on one of my recordings (not the other) but soon as I updated it went through fine

No, this didn't fix the error.

 

[johnn29]

@jaakkopasanen. Is eq.csv still supported? Trying to transform a recording for my Air Pods Pro - but the output doesn't sound right and I can't see it doing anything on the terminal. It's been like this way for a while - I was assuming I just couldn't get a decent headphone transform done anymore. But I remember early on, when it was eq.wav it used to work fine.

(venv) C:\Users\dushy\Impulcifer>python impulcifer.py --test_signal="data/sweep-6.15s-48000Hz-32bit-2.93Hz-24000Hz.pkl" --dir_path="data/my_hrir " --decay=300  --fs=44100
Creating impulse response estimator...
Running room correction...
Running headphone compensation...
Creating headphone equalization...
Creating frequency response target...
Opening binaural measurements...
Cropping impulse responses...
Equalizing...
Adjusting decay time...
Normalizing gain...
Plotting results...
Resampling BRIR to 44100 Hz
Writing BRIRs...

 

[kalstein]

Thanks. I debugged this a bit but actually got some other errros than what you reported, at first. My problems was caused by a bug which I found and fixed. I'm using branch reverb-management which is going to be merged to the master branch soonish so I'm not eager to spend a lot of time debugging and fixing the master branch at this point.

Your measurements are looking quite odd. There is a massive amount of harmonic distortion in all of your measurements. Impulcifer's algorithms aren't necessarily robust with this level of THD. What kind of speakers are you using for the purpose? How about microphones and other components of the recording chain?

Here are the plots: https://imgur.com/a/smhBuTv

For an example FL-left measurement graphs look like this:

While my own measurements look like this:

For some reason your decay graph shows 10 dB bump after the initial (fundamental) impulse has already decayed. Also the third harmonic impulse is just under -10 dB while my own has the highest harmonic at under -50 dB. Also the spectrogram graph for you measurements looks like nothing I've seen before. Your room measurement graphs actually looked a lot better although they do have the crazy spectrogram behaviour.

Another problem is that in this case the decay time detection algorithm fails with one of the impulse responses resulting in a very long BRIR file. My demo hesuvi.wav is less than 2 MB while the hesuvi.wav generated from your measurements is over 13 MB. This doesn't mean it wouldn't work but will introduce unnecessary CPU load, which might or might not be noticeable on your machine. The noise tail can always be cropped out in Audacity if that seems necessary.

All this doesn't necessarily mean that you wouldn't get results out of it that don't sound good or at least create the illusion of speakers. Especially the harmonic distortion is (in theory) negated entirely by Impulcifer as long as it doesn't mess up too badly with the different heuristic algorithms I'm using there.

Try out the reverb-management branch:

git fetch
git checkout reverb-management
git pull

I think Creative Sxfi dosen`t fit to measure audio signal.
Below plot is measured by Topping D50s today.

Thanks a lot.

 

[johnn29]

I use my SXFI amp and it works perfect.

 

[jaakkopasanen]

I think Creative Sxfi dosen`t fit to measure audio signal.
Below plot is measured by Topping D50s today.



Thanks a lot.

Looking very good! How does it sound?

 

[jaakkopasanen]

@jaakkopasanen. Is eq.csv still supported? Trying to transform a recording for my Air Pods Pro - but the output doesn't sound right and I can't see it doing anything on the terminal. It's been like this way for a while - I was assuming I just couldn't get a decent headphone transform done anymore. But I remember early on, when it was eq.wav it used to work fine.

(venv) C:\Users\dushy\Impulcifer>python impulcifer.py --test_signal="data/sweep-6.15s-48000Hz-32bit-2.93Hz-24000Hz.pkl" --dir_path="data/my_hrir " --decay=300  --fs=44100
Creating impulse response estimator...
Running room correction...
Running headphone compensation...
Creating headphone equalization...
Creating frequency response target...
Opening binaural measurements...
Cropping impulse responses...
Equalizing...
Adjusting decay time...
Normalizing gain...
Plotting results...
Resampling BRIR to 44100 Hz
Writing BRIRs...

It's supposed to work. Is the result any different if you use eq.csv vs not using it?

 

[kalstein]

I use my SXFI amp and it works perfect.

I don't know why... that's just result.

 

[kalstein]

Looking very good! How does it sound?

Normal case has same sound i think.
But,'Harman Room Target' is completely different.
Previously, only unpleasant reverb sounds in the bass range were heard.
This time, the correct bass is played.

 

[johnn29]

It's supposed to work. Is the result any different if you use eq.csv vs not using it?

Yep they are different, so I guess it is processing it.

I don't get it though - early on when you first came out with the transform method I used a Bose 700 measurements from rtings with their Air Pods Pro measurements and it sounded bang on right. Now i just can't get them to sound right. Whenever I use my Air Pods Pro I still use one of the first BRIR's i made. I wish I kept the damn AutoEQ output!

This time I'm trying to use a recording I made with the Creative Aurvana Live and running this in AutoEQ

python autoeq.py --input_dir="C:\Users\dushy\AutoEq\oratory1990\data\earbud\Apple Airpods Pro" --output_dir="my_results/APP (Aurvana)" --compensation="compensation/harman_over-ear_2018_wo_bass.csv" --sound_signature="C:\Users\dushy\AutoEq\results\oratory1990\harman_over-ear_2018\Creative Aurvana Live!\Creative Aurvana Live!.csv" --equalize --parametric_eq --max_filters=5+5 --ten_band_eq --bass_boost=4

But it sounds nothing like it. Same when I try and use rtings as a source.

Odd. will experiment more and with more IEM's.

Edit: Can confirm it's all working. Ran a bunch of them just now. Still getting best results from the rtings source. I don't think oratory's Air Pods Pro measurements were solid - I know that he didn't publish an EQ because the FR varied wildly on the dummy head depending on insertion. Perhaps rtings on this occasion is just better.

 

[kalstein]

I think everyone`s room left and right fr is different.
so, how about putting the same speaker data on the left and the right?
then can the result of room acoustic be applied relatively well?

 

[johnn29]

You should try the --balance modes to see if any work well.

I found that while they worked well for restoring channel balance - they did have an impact on HRTF. I guess because your ears aren't actually identical so it messed with the localization and you got an unusual feeling from it.

 

[kalstein]

You should try the --balance modes to see if any work well.

I found that while they worked well for restoring channel balance - they did have an impact on HRTF. I guess because your ears aren't actually identical so it messed with the localization and you got an unusual feeling from it.

I think balance is affect to headphone record. Dosen't it? I said about speaker record.