[jaakkopasanen]

What if my headphones arent on the list and I want to tune the EQ to my liking?

You can follow this guide if you happen to find the frequency response graph somewhere: Equalizing Headphones the Easy Way. If there is no frequency response for the headphones then AutoEQ can't help you.

 

[Dogway]

A few questions, is it possible to add negative bass boost?

I had this thing on my mind for some time, if AutoEQ is compensating FR for a Harman diffuse field, and HRIRs are doing the same, am I not doubling the compensation?
Reading here, it implies it isn't, and one only needs to set bass-boost to 0. For Harman 2018 that's fine, but for Harman 2013 I don't know how much less (negative bass-boost) I have to set. Common sense tells me the same delta than for Harman 2018 (-4 dB) but any info on this regard is welcome.

For games I use cmss_game which is quite neutral so I don't need to disable bass-boost BUT, here comes another question: Am I supposed to be listening the game audio as if I were in a room with loudspeakers? or within the game environment (no Harman compensation)? In this second case what kind of compensation should I use? I observed in the Road to PS5 video that they capture the HRIRs in an anechoic room, so a transparent FR as it can be should be preferable, despite not knowing what that means.

 

[jaakkopasanen]

A few questions, is it possible to add negative bass boost?

I had this thing on my mind for some time, if AutoEQ is compensating FR for a Harman diffuse field, and HRIRs are doing the same, am I not doubling the compensation?
Reading here, it implies it isn't, and one only needs to set bass-boost to 0. For Harman 2018 that's fine, but for Harman 2013 I don't know how much less (negative bass-boost) I have to set. Common sense tells me the same delta than for Harman 2018 (-4 dB) but any info on this regard is welcome.

For games I use cmss_game which is quite neutral so I don't need to disable bass-boost BUT, here comes another question: Am I supposed to be listening the game audio as if I were in a room with loudspeakers? or within the game environment (no Harman compensation)? In this second case what kind of compensation should I use? I observed in the Road to PS5 video that they capture the HRIRs in an anechoic room, so a transparent FR as it can be should be preferable, despite not knowing what that means.

A little context would welcome. Are you creating eq profiles to be used with HeSuVi? If that's the case then there's no simple answer to the bass boost question. Different HRIRs in HeSuVi have different levels of built in bass boost. Some have about 6dB some have none. Best would be for you to simply set the bass level by ear to where it sounds the best with the HRIR you're using.

 

[Dogway]

Yes, my question is about AutoEQ and HRIRs with HeSuVI. You say that HRIRs only have boosted bass but what I mean is that typically HRIRs do also implement some type of diffuse field on the high range (among other things), that is a roll off in the treble. Compensating with harman 2018 and using an HRIR would then not only double the bass but double the roll off. So that was my question, how do you optimally EQ your headphone for HeSuVi.

And related to that, do games still need a compensation curve? as if it was listened through loudspeakers in a room? Is there such thing as a compensation curve for loudspeaker in the open -anechoic room- (no roll off, no reflections, no bass boost) which would better suit a neutral HRIR for games?

PD: I was wondering if crinacle had .csv results, couldn't find in the repo

Thanks for the help.

 

[jaakkopasanen]

Yes, my question is about AutoEQ and HRIRs with HeSuVI. You say that HRIRs only have boosted bass but what I mean is that typically HRIRs do also implement some type of diffuse field on the high range (among other things), that is a roll off in the treble. Compensating with harman 2018 and using an HRIR would then not only double the bass but double the roll off. So that was my question, how do you optimally EQ your headphone for HeSuVi.

And related to that, do games still need a compensation curve? as if it was listened through loudspeakers in a room? Is there such thing as a compensation curve for loudspeaker in the open -anechoic room- (no roll off, no reflections, no bass boost) which would better suit a neutral HRIR for games?

PD: I was wondering if crinacle had .csv results, couldn't find in the repo

Thanks for the help.

The HRIRs in HeSuVi are captured from commercial products and those products are meant to be used with headphones. So it would make sense that the HRIRs are designed to work well with some type of headphone, probably a neutral one. That means Harman target (or similar neutral target) would be the best option for eq with HeSuVi HRIRs.

I would think no game audio designer expects the players to be playing the game in an anechoic chamber. I assume games sounds are designed to be listened on speakers in a room, just like everything else. Some games have separate headphone mode which usually means surround virtualization but you shouldn't use that with HeSuVi or you'll have double virtualization.

It's pretty much impossible to reason beyond these arguments. We don't know the exact design principles that the HRIR and game creators took. If something doesn't sound good to you, really the only option is to tweak it by ear. Of course you could record your own HRIR with Impulcifer.

Numerical data for Crinacle's measurements are not available in AutoEq repo because he asked me not to publish it. Crinacle has a Patreon and a certain subscription tier there gives you the access to the raw data.

 

[Dogway]

Yes, sorry for my thinking-out-loud annoying questions. I just felt it counterproductive or self-limiting (for games) to simulate how audio sounds in a room, as opposed to a be-there (simulate binaural recordings).

With Impulcifier you still record the in-room acoustics (unless you compensate for that with REW, not sure if that's a thing). For example in the Road to PS5 video you see they capture the HRTFs in an anechoic chamber so there's no room coloring the sound. That's also what I thought, recreating a head in 3D (there are addons to streamline the process) and shooting audio sweeps to in-ear virtual microphones. E.A.R. is long outdated and I don't think it works with intricate meshes as CG ears (need to test).

On compensations for HRIRs I did yesterday another run for my Sundara with updated AutoEq and indeed Harman 2018 sounds great without the bass boost. Only bested by a B&K EQ using H2018 compensation. Since I didn't have an error column for B&J I used as input_dir oratory1990 Sundara measurement summed to the "Sundara vs Sundara Stealth" difference equalization curve. Now I want to try with an X-Curve EQ if I'm not mistaken with my methods and by extension the Crinacle signatures since I don't have those errors either.

 

[jaakkopasanen]

Yes, sorry for my thinking-out-loud annoying questions. I just felt it counterproductive or self-limiting (for games) to simulate how audio sounds in a room, as opposed to a be-there (simulate binaural recordings).

With Impulcifier you still record the in-room acoustics (unless you compensate for that with REW, not sure if that's a thing). For example in the Road to PS5 video you see they capture the HRTFs in an anechoic chamber so there's no room coloring the sound. That's also what I thought, recreating a head in 3D (there are addons to streamline the process) and shooting audio sweeps to in-ear virtual microphones. E.A.R. is long outdated and I don't think it works with intricate meshes as CG ears (need to test).

On compensations for HRIRs I did yesterday another run for my Sundara with updated AutoEq and indeed Harman 2018 sounds great without the bass boost. Only bested by a B&K EQ using H2018 compensation. Since I didn't have an error column for B&J I used as input_dir oratory1990 Sundara measurement summed to the "Sundara vs Sundara Stealth" difference equalization curve. Now I want to try with an X-Curve EQ if I'm not mistaken with my methods and by extension the Crinacle signatures since I don't have those errors either.

It could be that game audio engines do some kind of adjustments when user has selected the headphone mode. If that's HRTF simulation then it would indeed make sense that the audio engine would have more of the games ambient reverb in the signal and with speakers those would be turned down a bit to give space for the room's own reverb. However I'm not entirely convinced that games would take audio design this far. Hardly anyone would pay any attention to minute details like these when the visuals and interaction in games take so dominant precedence in the player's attention.

 

[Dogway]

Well yeah, audio mixing in games is still done and assessed in an studio environment with speakers and all its implications. I will try to read a bit more if mixing in anechoic chamber is a thing or not, and try to think about it. Thank you.

 

[Gil80]

I use viper4android but I have a DDC correction profile for my earbuds (Philips TX1). Applying a correction profile negates the need for EQ correction, isn't it?

 

[billqs]

I applied AutoEQ first to Peace then to the Equalizer APO Editor per your directions (Each done independently with the other EQs removed from the chain). I really appreciate all this work you've done! I previously had downloaded the free version of Audeze's Reveal plug in. Is it attempting to do the same thing that Auto EQ is?

The results I got on my LCD MX4 through an Oppo HA-1 (which uses the ESS 9018 Sabre as a dac) were good but quite surprisingly bright. I have tinnitus and also a bit of hypercuses and the treble was too strong for me. Are there particular bands in EQ to lower to help bring some of the treble "bite" back down?

 

[Dogway]

@billqs
I would suggest maybe trying the inverse of "gender-related presbycusis deafness" EQ curve? You can EQ your headphone to Harman with this EQ as sound-signature so you distort less.

frequency,raw
40,0.0
100,0.0
500,0.0
1000,0.7
10000,5.0
20000,5.0

 

[jaakkopasanen]

I applied AutoEQ first to Peace then to the Equalizer APO Editor per your directions (Each done independently with the other EQs removed from the chain). I really appreciate all this work you've done! I previously had downloaded the free version of Audeze's Reveal plug in. Is it attempting to do the same thing that Auto EQ is?

The results I got on my LCD MX4 through an Oppo HA-1 (which uses the ESS 9018 Sabre as a dac) were good but quite surprisingly bright. I have tinnitus and also a bit of hypercuses and the treble was too strong for me. Are there particular bands in EQ to lower to help bring some of the treble "bite" back down?

You could add a parametric filter with Q of 0.7 at 4000 Hz. Turn down the gain into negative numbers as much as you want.

 

[billqs]

Thanks, I'm going to try these out after my workday ends and I can have my "headphone time". I'm still really new to Equalizer APO and utilizing AutoEQ so this advice is very helpful!

 

[billqs]

Hi! I use this all the time. Thanks! I was wondering if you were going to add a profile for the Hifiman HE1000SE. There are ones for the preproduction v.1 and v.2 but not for the SE. Keep up the great work!

 

[jaakkopasanen]

Hi! I use this all the time. Thanks! I was wondering if you were going to add a profile for the Hifiman HE1000SE. There are ones for the preproduction v.1 and v.2 but not for the SE. Keep up the great work!

Unfortunately there's not a lot I can do. Only the models which have been measure by one of the supported sources can be added to AutoEq.