shanonpearce / ash-listening-set

A dataset of filters for headphone correction and binaural synthesis of spatial audio systems on headphones

Audio Spatialisation for Headphones

The Audio Spatialisation for Headphones (ASH) Listening Set is a continuation of the ASH-IR-Dataset project. It includes a set of binaural room impulse responses (BRIRs) suitable for everyday listening, a comprehensive set of headphone correction filters (HpCFs), and a set of configuration files for Equalizer APO. The dataset can be used to create spatial surround sound on headphones by convolving an audio stream with a set of BRIRs and a HpCF.

Binaural Room Impulse Responses

The dataset includes a set of BRIRs derived from a range of publicly available BRIR datasets. The BRIRs were measured using head and torso simulators (HATS) in a variety of reverberant rooms, each containing unique acoustical properties. For each room, a set of BRIRs are provided for a range of source directions around the head on the horizontal plane. The BRIRs have been equalised to remove undesired spectral colouration and to make the BRIRs compatible with diffuse-field equalised over-ear and on-ear headphones. An additional compatibility filter is supplied to provide compatibility with in-ear headphones. The BRIRs are provided as 2 channel WAV files with a sampling rate of 44100Hz. A set of true stereo BRIRs and HeSuVi compatible BRIRs are also provided.

Headphone Correction Filters

The dataset also includes correction filters for a wide range of headphones. The filters can be used to equalise individual headphones to the diffuse-field target frequency response. Individual headphone equalisation is recommended to remove undesired spectral colouration introduced by the listener's headphones and to improve the plausibility of the binaural simulations.

The filters are provided as finite impulse response (FIR) filters and Graphic EQ filters. The FIR filters are single channel WAV files with a sampling rate of 44100Hz and are compatible with IR convolution software such as Equalizer APO. The Graphic EQ HpCFs are graphic equalizer filters with a total of 127 frequency bands and are compatible with Equalizer APO, HeSuVi, and Wavelet. Magnitude response plots of the filters are also provided.

Equalizer APO Configuration Files

The dataset can be used with Equalizer APO, an Audio Processing Object (APO) for windows featuring convolution capabilities. Configuration files for BRIR convolution are provided for each room in the dataset and for a range of common speaker configurations. Sample configuration files for HpCF convolution are also provided. Configuration steps for Equalizer APO can be found in the wiki.

License

Unless otherwise stated, all files in this repository are licensed under Creative Commons Attribution-NonCommercial-ShareAlike 4.0 https://creativecommons.org/licenses/by-nc-sa/4.0/

Further Information

Refer to the Wiki for more information about the dataset including configuration instructions.

ash-listening-set's Issues

Pipewire configuration problem

Here I don't have filter-chain.conf.d/virtual-surround.conf

The content of that folder is:

demonic.conf sink-make-LFE.conf sink-virtual-surround-5.1-kemar.conf source-rnnoise.conf
sink-dolby-surround.conf sink-matrix-spatialiser.conf sink-virtual-surround-7.1-hesuvi.conf
sink-eq6.conf sink-mix-FL-FR.conf source-duplicate-FL.conf

Question about stereo upmixing

All of the audio content I listen to only uses 2 channels, but I would still like to use all of the BRIR .wav's instead of just the A-30 and A30 .wav's for stereo content. However, this project doesn't have a suggestion for how to upmix stereo for this purpose. I have been able to find an idea from HeSuVi (Copy: L=0.5*L R=0.5*R C=0.2*L+0.2*R SUB=0.0 RL=0.3*L+-0.2*R RR=-0.2*L+0.3*R SL=0.45*L+-0.25*R SR=-0.25*L+0.45*R, though I just multiplied all of the values by 2 to result in Copy: C=0.4*L+0.4*R SUB=0 RL=0.6*L+-0.4*R RR=-0.4*L+0.6*R SL=0.9*L+-0.5*R SR=-0.5*L+0.9*R), but I don't know if this makes sense/is valid for this purpose, nor do I know if this specific upmix is for narrow 7.1 or wide 7.1.

Here is what my current configuration looks like:

Linux support

EasyEffects with PipeWire made Linux more audio capable & easy to use than ever before. But not for virtual headphone surround yet. Although, virtual surround support is possible with Pipewire, it is finnicky to set & implement.

In the future, wiki implementation would be nice to address easy & user-friendly Linux support.

Why removed reverberation?

I'm curious if there's a possibility that larger rooms could also come alongside a version with their original reverberation, like the Small broadcast studio

Request for more info on the virtual rooms

I can't exactly discern the source of R18 (Small Room I) and R20 (Small Room A) to R33 (Studio B). What database(s) do these rooms come from? R32 (ASH Listening Room) is definitely directly from this project, so I also wonder how it is made and what its goals are supposed to be. I seem to be preferring it over R03 (Control Room 7).

Great Project, any recommendations for beginners?

Hi, any recommendations on where to start? Lot of measurements here!
Are there any crowd favorites?

Also, can I use these with regular old convolution effect plugin?

Generate HeSuVi-compatible WAVs

Is there a possibility the BRIRs can be processed as HeSuVi wav files?
HeSuVi is the generally the most used method for surround sound with Equalizer APO. There may however be issues as HeSuVi takes 48KHz files, with 44.1KHz being secondary.

Differences in peak gain

Hi again, I'm really loving this project and how far it has come. However, there's on thing I would like to bring up. Each HRIR has it's own peak volume and so in order to eliminate clipping, you have to change the preamp according to the room every time you switch them. I know, there's tools like peace UI that solve that problem but I was wondering if you could make the room files have a more equal peak gain. Thanks in advance

Question on some HPCF file

Hello. Could you point me please what source of measurements you have used to generate EQ for Sennheiser 650_Sample_B?