You have all the sound filters in the wrong order. Always goes first noise reduction, then compressor and finally EQ. You may add a gain after this but in normal situations isn't needed like the limiter (you already use a compressor).
I'm guessing the issue is in this filter order you have.
I thought it was okay:
13:24:00.362: - source: 'Shure Mic' (wasapi_input_capture)
13:24:00.362: - filter: 'Gain' (gain_filter)
13:24:00.362: - filter: 'Noise Suppression' (noise_suppress_filter_v2)
13:24:00.362: - filter: '3-Band Equalizer' (basic_eq_filter)
13:24:00.362: - filter: 'Compressor' (compressor_filter)
13:24:00.362: - filter: 'Limiter' (limiter_filter)
Or does it report in reverse order? I've seen some things do that.
Anyway, Noise Suppression should be before any dynamic processing (Compressors, Limiters, Gates, anything that changes volume in response to the signal that it's processing), because the noise also follows their change in volume. Noise Suppression needs constant noise, so it goes before anything that changes volume over time.
The Limiter should be last, as a "safety net" to make sure you don't clip.
--
Otherwise, there's no technical reason for the order, but there might be an artistic one.
Some things make no difference at all, like the order of EQ and Gain, because those are both "linear" processes in the sense of being described entirely by linear calculus. Or in more layman's terms, it makes no difference in what order you apply multiple Gains, and EQ is only a frequency-dependent Gain. The frequency-dependence doesn't break that.
Noise Suppression doesn't really care if it's preceded by some linear functions, but it's easy to remember to have it first.
Non-linear things do make a difference in order. If you have an EQ boost, for example, before a Compressor, then the Compressor will react more strongly to the boosted frequency range than if you swapped their order and kept the same settings. Or if you have an EQ cut to get rid of something, the Compressor will react more strongly to that thing if it comes before the EQ than if it comes after. For that reason, most digital consoles allow you to swap the order of the channel EQ and channel Compressor.
Since you're building your own processing chain here, nothing says you can't have multiple of the same thing. EQ -> Comp -> EQ, for example, where the first EQ cuts out annoying things and the second EQ boosts nice things. Normally, you could combine the two EQ's and have fewer processors (always a good thing!), but the non-linear thing in between prevents that.
Also, the Compressor has a Gain control built into it. So you might absorb the Gain up front into that. It won't be *exactly* the same setting because it's an *output* gain as it leaves the Compressor, but it's not all that hard to figure out:
- Start with the Compressor's Threshold lower by the amount that the removed Gain was.
- Then boost the Compressor's Output Gain by that much.
- Then divide that amount by the Compressor's Ratio, and bring the Output Gain back down by that.
Not having a "pre-boost" like that, also makes it harder to clip early in the chain, which *might* actually be the problem here. I don't see a recording link, only logs, but I *have* had people describe clipping as "muffled".
