Recording Impulse Responses for Speaker Virtualization

Apr 1, 2022 at 2:08 AM Post #1,066 of 2,034
I actually made the same folder as yours in the location of the text file and copied the hesuvi.wav changing it to the same name as what you have. Same error.

I also tried changing it to another location but didn’t work. Maybe I’m missing something in python
 
Apr 1, 2022 at 2:15 AM Post #1,067 of 2,034
I do some equalizing for channel balance correction. The reason for that you find in post #960. I could merge the equalizing into the wav file now but for testing it was really helpful to have the option to change filters on the fly.
very interesting! I didn't know that after the impulcifer measurements you can still have channel unbalance, i will try too after done my personal hrir

ordered all the stuff, i should start in a week or so hopefully...
 
Apr 1, 2022 at 3:08 AM Post #1,068 of 2,034
I actually made the same folder as yours in the location of the text file and copied the hesuvi.wav changing it to the same name as what you have. Same error.

I also tried changing it to another location but didn’t work. Maybe I’m missing something in python
The only package used is SoundFile. You can run "pip install SoundFile" to check if you have installed it.
 
Last edited:
Apr 1, 2022 at 3:29 AM Post #1,069 of 2,034
very interesting! I didn't know that after the impulcifer measurements you can still have channel unbalance, i will try too after done my personal hrir

ordered all the stuff, i should start in a week or so hopefully...
Good luck, hope you get good results. Keep a track of what you tried and what worked for and let us know
 
Apr 1, 2022 at 4:12 AM Post #1,070 of 2,034
Good luck, hope you get good results. Keep a track of what you tried and what worked for and let us know
will i have to repeat the recording of the hrir most of the time or some of the impulcifer functions are post recording? i'm trying to think about how i should prepare the working area, hopefully i will not have to do the recording and turn around many times

do you use the stereo and 5.1 upmix in hesuvi? i mean that function which plays all the 7 channels together even if the format you're playing is in stereo
 
Apr 1, 2022 at 4:41 AM Post #1,071 of 2,034
will i have to repeat the recording of the hrir most of the time or some of the impulcifer functions are post recording? i'm trying to think about how i should prepare the working area, hopefully i will not have to do the recording and turn around many times

do you use the stereo and 5.1 upmix in hesuvi? i mean that function which plays all the 7 channels together even if the format you're playing is in stereo
Whilst recording you will need to turn your head. I’d recommend a few in one go. Each time you complete the recording save the folder somewhere else, delete the recording, adjust a little (change volume, push the mics a little further, change speaker position a little closer/further away) and do another recording.

The balance and other tweaks can be done another time. I was like you and wanted it done in the first sitting. I was wrong and it took me a few days

I upmix 5.1 to 7.1 in hesuvi

Also whilst waiting, you might as well set up and try the demo so your computer is all ready for the full measurements
 
Last edited:
Apr 1, 2022 at 6:12 AM Post #1,072 of 2,034
Whilst recording you will need to turn your head. I’d recommend a few in one go. Each time you complete the recording save the folder somewhere else, delete the recording, adjust a little (change volume, push the mics a little further, change speaker position a little closer/further away) and do another recording.

The balance and other tweaks can be done another time. I was like you and wanted it done in the first sitting. I was wrong and it took me a few days

I upmix 5.1 to 7.1 in hesuvi

Also whilst waiting, you might as well set up and try the demo so your computer is all ready for the full measurements
i already installed impulcifer a while ago to try to cut some reverber in some hrir but i didn't like how it sounded, in fact my only concern about my future hrir recording is "how much reverb i will have?", i was even thinking to buy some acoustic foam to help reduce the eventual reverb lol but too much money so whatever
 
Last edited:
Apr 1, 2022 at 6:48 AM Post #1,073 of 2,034
i already installed impulcifer a while ago to try to cut some reverber in some hrir but i didn't like how it sounded, in fact my only concern about my future hrir recording is "how much reverb i will have?", i was even thinking to buy some acoustic foam to help reduce the eventual reverb lol but too much money so whatever
You can reduce the reverb with one of the balance options it worked really well for one of my measurements in the past.

Decay Time Management​

The room decay time (reverb time) captured in the binaural room impulse responses can be shortened with --decay parameter. The value is a time it should take for the sound to decay by 60 dB in milliseconds. When the natural decay time is longer than the given target, the impulse response tails will be shortened with a slope to achieve the desired decay velocity. Decay times are not increased if the target is longer than the natural one. The decay time management can be a powerful tool for controlling ringing in the room without having to do any physical room treatments.
 
Apr 1, 2022 at 2:40 PM Post #1,074 of 2,034
Hi, firsts tests, first measurement - stereo -, i've got a warning FT measurement has lower delay to left ear than to right, 1,8125... But everything seems correctly set up : i'm using a zomm H5 for input, the test signal is heared on left then on right, on headphones or speakers. Any idea ?
I've tested with Dire Straits Sultans of Swing, when i activate in Hesuvi, i loose all reverbs on guitars.
results.png
headphones.png
 
Last edited:
Apr 1, 2022 at 3:55 PM Post #1,075 of 2,034
Hi, firsts tests, first measurement - stereo -, i've got a warning FT measurement has lower delay to left ear than to right, 1,8125... But everything seems correctly set up : i'm using a zomm H5 for input, the test signal is heared on left then on right, on headphones or speakers. Any idea ?
I've tested with Dire Straits Sultans of Swing, when i activate in Hesuvi, i loose all reverbs on guitars.
I was getting similar graphs and my volume was way too loud. I set mine to just above normal loud listening and set the gain on both mics to just about below where they were clipping. Both mics had a slightly different gain setting.

Try lowering your volume or the gain on your in ear mics
 
Last edited:
Apr 1, 2022 at 8:21 PM Post #1,076 of 2,034
I was getting similar graphs and my volume was way too loud. I set mine to just above normal loud listening and set the gain on both mics to just about below where they were clipping. Both mics had a slightly different gain setting.

Try lowering your volume or the gain on your in ear mics
if both mics have different gain does this mean that you have to lower the gain on one of the two? then how can i understand the exact gain to balance both the mics?
 
Last edited:
Apr 2, 2022 at 12:21 AM Post #1,077 of 2,034
Hi, firsts tests, first measurement - stereo -, i've got a warning FT measurement has lower delay to left ear than to right, 1,8125... But everything seems correctly set up : i'm using a zomm H5 for input, the test signal is heared on left then on right, on headphones or speakers. Any idea ?
It's likely that you reversed the left and right mics. When the right speaker is playing, sound should arrive earlier in the right ear than the left. That's what the warning means.
 
Last edited:
Apr 2, 2022 at 3:15 AM Post #1,078 of 2,034
It's likely that you reversed the left and right mics. When the right speaker is playing, sound should arrive earlier in the right ear than the left. That's what the warning means.
Humm yes, i didn't checked that perhaps on zoom H5, the input 1 not corresponding at L channel, i will test that.
For the levels, has said Morgin, i've increased them so that i was very close to 0db headroom, has suggested in the wiki, so that the S/B his high.
 
Apr 2, 2022 at 1:01 PM Post #1,079 of 2,034
Humm yes, i didn't checked that perhaps on zoom H5, the input 1 not corresponding at L channel, i will test that.
For the levels, has said Morgin, i've increased them so that i was very close to 0db headroom, has suggested in the wiki, so that the S/B his high.
Ok i switched the mics, but i still have the warning, but the delay is lower, 0,29 ms. The plots are very similar to the first one, i decreased the levels. First listening, the lower delay has a impact: there is something here, this far from perfect, but that the very first time i have no more the feeling that the sound with the headphones is in my head, neither around the head but almost coming from my speakers that are mute. No virtualisation (dolby headphone, ooyh and others) never gave me this feeling, I understand now the benefit of personal HRIR. Still have a lot of work to make it better but it's a good beginning.
Any other advice to correct this delay et my bad graphs ?

Edit : grrr i've got a "hummmm" with one of the vxlr+ adaptator, not very huge, but i can hear it when i increase the input level of the mic. So more tests are useless, i've got to resend the bad vxlr+
 
Last edited:
Apr 2, 2022 at 7:44 PM Post #1,080 of 2,034
Ok i switched the mics, but i still have the warning, but the delay is lower, 0,29 ms. The plots are very similar to the first one, i decreased the levels. First listening, the lower delay has a impact: there is something here, this far from perfect, but that the very first time i have no more the feeling that the sound with the headphones is in my head, neither around the head but almost coming from my speakers that are mute. No virtualisation (dolby headphone, ooyh and others) never gave me this feeling, I understand now the benefit of personal HRIR. Still have a lot of work to make it better but it's a good beginning.
Any other advice to correct this delay et my bad graphs ?

Edit : grrr i've got a "hummmm" with one of the vxlr+ adaptator, not very huge, but i can hear it when i increase the input level of the mic. So more tests are useless, i've got to resend the bad vxlr+

"hummmm"? explain me, now i'm doubting too with my vxlr+ because one of the two had the package opened, still my mics are in shipment so i can't test
 

Users who are viewing this thread

Back
Top