search

ShanonPearce / ASH-Listening-Set

A dataset of filters for headphone correction and binaural synthesis of spatial audio systems on headphones
Other
58 stars 2 forks source link

Great Project, any recommendations for beginners? #3

Open mmontag opened 1 year ago

mmontag commented 1 year ago

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?

ShanonPearce commented 1 year ago

Hi,

I would recommend starting out by trying each of the headphone correction filters available for your headphones. It can often be a matter of personal preference but if there are multiple filters available for your headphones I would recommend using the averaged filter (e.g. HpCF_Audio-Technica_ATH_M50x__Avg). In some cases there are even multiple averaged filters due to large amounts of variance between measurements. Similarly the BRIRs (used for room simulation) can be a matter of personal preference. I would recommend trying each one.

All of the filters provided in this dataset are WAV files so you will need to use convolution software to apply them. I highly recommend using Equalizer APO if you are using Windows. The BRIRs can also be loaded in HeSuVi (interface for E-APO) which some users have found that to be more convenient. Convolution plugins should also work. I have not done any testing on other software/systems but it should work on all audio processing software supporting IR convolution. One limitation is the software/plugin will need to support stereo or true stereo processing in order to apply the BRIRs.

More information can be found in the wiki

If you have any other questions feel free to drop them here.

mmontag commented 1 year ago

Thanks. I'm on macOS - I tried some of the BRIRs and quickly realized that there was no channel crosstalk (cannot hear the left channel in the right ear, and vice versa) because my convolution plugin (LAConvolver) does not support stereo summing.

Seems simple but I will keep looking.

mmontag commented 1 year ago

Also, another quick question: In a file like BRIR_Studio.wav which has 14 channels, what is the typical channel order? i.e.,

  1. L source to L ear
  2. R source to L ear
  3. L source to R ear
  4. R source to R ear
  5. L source to Arbitrary mic position 1
  6. R source to Arbitrary mic position 1
  7. L source to Arbitrary mic position 2
  8. R source to Arbitrary mic position 2 ...

or something else?

ShanonPearce commented 1 year ago

Hi,

In regards to the channel mapping/ordering, I have added the numbering details to the wiki here.

As for the issue of multichannel convolution on macOS, I did some searching around and found X-MCFX, a multichannel convolver VST plugin that can be used on a suitable plugin host. I did some brief testing using the windows version of the X-MCFX VST-3 plugins on the Kushview Element host, which also has a macOS version available.

I managed to successfully apply BRIR convolution on 2 input channels (L and R) by convolving each input channel with an azimuth-specific BRIR, then mixing the result and applying headphone correction by convolving with a HpCF filter. Below is my workflow. This could potentially be optimised further. It could also be scaled to process additional channels (such as 7.1 surround).

Element + X-MCFX, BRIRs and HPCFs sample

mmontag commented 1 year ago

Thank you, this helps a lot!