[morgin]

You still need a way of getting the wire to come out of the ear because it will always be on the back of the mics. I don’t think that housing will help much unless you’re using that housing to glue onto the foam plugs so you don’t have to glue straight onto the mics.

 

[reter]

You still need a way of getting the wire to come out of the ear because it will always be on the back of the mics. I don’t think that housing will help much unless you’re using that housing to glue onto the foam plugs so you don’t have to glue straight onto the mics.

Ok this is my second attempt and i'm happy with it, they fits in the ears deep enough and are somewhat stable (hot glue does miracles); tomorrow i will begin with the measurements

this foam in particular is awesome, it stretches very slowly so you can fit them deeper, the orange ones were very hard, so even the choice of the foam matters

i cut 2-3 fibers of copper but the mic is still working, hopefully this will not impact the overall efficiency, i'm still thinking about insulate that critical part with some insulating tape (just a little), but i'm not sure

 

[morgin]

Ok this is my second attempt and i'm happy with it, they fits in the ears deep enough and are somewhat stable (hot glue does miracles); tomorrow i will begin with the measurements

this foam in particular is awesome, it stretches very slowly so you can fit them deeper, the orange ones were very hard, so even the choice of the foam matters

i cut 2-3 fibers of copper but the mic is still working, hopefully this will not impact the overall efficiency, i'm still thinking about insulate that critical part with some insulating tape (just a little), but i'm not sure

That looks very close to mine. I cut more of the housing all the way to the mic otherwise it still irritates the ear. Plus with the bare thin wires you can loop them behind the foam and back out the ear.

 

[reter]

OK done my first measurements, just wow, the first impact is absolutely awesome, it's really like i'm hearing inside my room with speakers, but i don't know if i done it right

The first graph shows my headphones measurement and i see the error line is well upper than the target line, is this normal? i tried several times and i couldn't get under the 10.5db headroom (mixing both mics and headphones volumes), and 7-10 headroom for the speaker measurements

What's that "difference" line is showing? is this a bad thing?

This is my folder showing all the graphs: https://drive.google.com/file/d/1ZGqYEXMDoHcHFRSNDMG8VChHIl-MW9wL/view?usp=sharing

This is how equalizer apo shows my hrir,

another thing is that the volume is too low, and i have to set my amp liteally way way up, as shows in the measurements i have a peak gain of 1.3db that's too low i think...?

also, a lot of reverb, i was expecting some reverb from my room but holy! this is way too much!

 

[morgin]

OK done my first measurements, just wow, the first impact is absolutely awesome, it's really like i'm hearing inside my room with speakers, but i don't know if i done it right

The first graph shows my headphones measurement and i see the error line is well upper than the target line, is this normal? i tried several times and i couldn't get under the 10.5db headroom (mixing both mics and headphones volumes), and 7-10 headroom for the speaker measurements

What's that "difference" line is showing? is this a bad thing?

This is my folder showing all the graphs: https://drive.google.com/file/d/1ZGqYEXMDoHcHFRSNDMG8VChHIl-MW9wL/view?usp=sharing

This is how equalizer apo shows my hrir,

another thing is that the volume is too low, and i have to set my amp liteally way way up, as shows in the measurements i have a peak gain of 1.3db that's too low i think...?

also, a lot of reverb, i was expecting some reverb from my room but holy! this is way too much!

That actually looks really good for your first attempt. I got to where you are after a fair few trials and changes.

 

[morgin]

What's that "difference" line is showing? is this a bad thing?

No that looks normal to me. Mine was very similar

another thing is that the volume is too low, and i have to set my amp liteally way way up, as shows in the measurements i have a peak gain of 1.3db that's too low i think...?

you will need to change the target level. This was mine.

python impulcifer.py --test_signal="data/sweep-6.15s-48000Hz-32bit-2.93Hz-24000Hz.pkl" --target_level=10 --dir_path="data/my_hrir" --plot

also, a lot of reverb, i was expecting some reverb from my room but holy! this is way too much!

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.

 

[musicreo]

No that looks normal to me. Mine was very similar

you will need to change the target level. This was mine.

python impulcifer.py --test_signal="data/sweep-6.15s-48000Hz-32bit-2.93Hz-24000Hz.pkl" --target_level=10 --dir_path="data/my_hrir" --plot

​

--target_level=10 means 10db above 0db! This produces a clipped impulse response.

 

[morgin]

--target_level=10 means 10db above 0db! This produces a clipped impulse response.

Yikes I best not give advice I’m still a learner. But it’s the only way I get decent volume at 80% in windows and 70% on the media players.

 

[musicreo]

Yikes I best not give advice I’m still a learner. But it’s the only way I get decent volume at 80% in windows and 70% on the media players.

For the "--target_level" it is better to use negative values to have some headroom. If it is to silent it is better to use the amplification in EQ-APO to amplify the input.

 

[reter]

ok test 2 done

even better headrooms for speaker and 9db for headphones, is this enough or should i go lower?

I went to target level -10, it's pretty good for me at the moment

MY HRIR TEST 2: https://drive.google.com/file/d/1Bqxy_yXTXnaBE385DJem1gQcQjJRJuX7/view?usp=sharing

now i'll try the decay, hopefully i will achieve good results

 

[morgin]

For the "--target_level" it is better to use negative values to have some headroom. If it is to silent it is better to use the amplification in EQ-APO to amplify the input.

Thanks man. I keep saying this but I’m glad you’re part of this with so much solid advice

 

[castleofargh]

For the "--target_level" it is better to use negative values to have some headroom. If it is to silent it is better to use the amplification in EQ-APO to amplify the input.

Digital boost wherever it is, will give the same clipping for over 0dB signal before going to the DAC. Even floating points can't save us.

If it looks like the mic or headphone might be failing at either end of the frequency range, change them(move the mic if it might be obstructed somehow, or get a headphone with better FR extension in the area where the impulse tries to give so much boost). Or add another EQ over the impulse to attenuate the area that's abnormally high. Anything above ... let's say 13 or 15kHz is probably not accurate anyway and not very important for music(less so if you're an adult and you've lost so much sensi in those ranges), so if there is some big boost only there in the final impulse, maybe EQ down that area. What amplitude you save with that, you can apply as global digital gain and not clip.
If the peak is an abnormal boost only in the low end, a high pass(or a more subtle EQ)might help do the same and hopefully it will still sound nice. In some cases it could even improve the sound if the headphone distorts too much when trying to make loud subs it was never able to make in the first place.

If you clip in the low freqs with some digital settings errors, it might not always sound wrong(up to a certain amount of clipping!!!!) because specifically for subs, there's a lot we don't actually hear on headphone, and there's also a psychoacoustic phenomenon where our brain makes up the missing top of the slow sine by itself. We feel like we're hearing the entire sine without clip. I don't remember the percentage until it just sounds horrible to us, but it's something abused in the loudness war and absurdly overused in too many rap albums.

For those struggling too much with loudness, you should consider a more sensitive headphone(if it has a smoother FR and better extension that probably wouldn't hurt). Or get an amp with higher gain(double the voltage gain gives +6db).

 

[reter]

Actually the only thing i'm struggling at the moment is the reverb. i tried the decay value that i think won't h+go under 200, i was 1m80cm from the speaker when i recorded which is me almost near the wall, and the speaker almost to the other side of the wall, i think i should try to record with the speaker closer, i don't know

 

[morgin]

I had a lot of reverb with my speaker in one corner of the room and me in the other corner to try and make the distance as large as possible. I found a significant decrease when I used the speaker and sat in the middle of the room with 5-6ft distance.

 

[musicreo]

Digital boost wherever it is, will give the same clipping for over 0dB signal before going to the DAC. Even floating points can't save us.

I think there is a difference. When I boost the impulse response over 0db I cut off some information of the impulse response. This can't be undone. But if I boost in EQ-APO I can always change the amplification depending on the input to prevent clipping.