Is it possible to make an impulse response file of a specific graphic equalizer curve? I'm looking to reproduce an equalizer curve without having to use the equalizer itself in the chain.
yes it's possible and easy. take an identity impulse response (1 sample = 1.0, all other surrounding samples = 0.0), process it with your EQ of choice (if it's a foobar2000 DSP, then use converter to do this), then use an audio editor to trim excess silence, and the end result is what you want. it's more important to trim silence at the beginning, because that's the amount of delay you will get. excess silence at the end isn't too bad, it'll just waste computing power a bit more.
I'm not sure, but it's probably preferable to use the same sampling rate as the material that you're going to listen. or just always use it with resampler, to one fixed sample rate, 48000 makes sense since it will be a no-op when playing Opus files.
edit: attached an example, good starting point. if you need a stereo impulse (different EQ to left and right channels), first clone the channel (convert to stereo) because my impulse is mono.
also, it's preferable to do processing (preparation of the impulse) in floating point because EQ may produce samples outside of normal range -1..1, and if you clip your impulse, it'll no longer be an accurate representation of the EQ.
alternatively, just make sure that your EQ doesn't produce clipped samples when used on the identity impulse.
An even easier way to do it would be by using a free program called rePhase (https://sourceforge.net/projects/rephase/).
Make an EQ you that you need in Parametric Gain EQ window then generate an impulse response wav file in the impulse settings part.
It also allows you to use linear phase filters which is pretty cool if avoiding phase distortion is important to you.
> It also allows you to use linear phase filters which is pretty cool if avoiding phase distortion is important to you.
Same is true for many different EQ effects (VST or others).
Benefits of linear phase just for listening (where you don't usually care about correlated signals canceling out when mixing) are not so clear btw, humans don't hear small differences in phase, it's mostly about altered peak levels vs. delay
linear phase tends to generate higher peaks less (so less chance of clipping if you're pushing the volume), but it has delay (also pre-echo) but in practical listening use cases, none of these 2 things really matter
automatic exporting of impulse response is a cool feature regardless. I wish more audio software had that.