Compressor/Limiter

[Albert H]

Having done this for 40+ years, I have yet to find a truly good-sounding software solution. They're all uniformly horrid!

My audio processing is all done in the analogue domain (though I do use A to D and D to A for the delay lines). I apply slow acting 2 : 1 (ish) compression to bring up the low level content somewhat. I apply a little tonal shaping at this point.

I then apply the audio to a delay-line based limiter, with (effectively) zero attack time: it "sees the peaks coming" and pulls down the gain appropriately. The gain cells I use are based on the THAT2180, with two cascaded cells in each channel to eliminate the possibility of clipping. I do not include clippers in the path (though there's a "guard" composite clipper after the stereo coder - just as a brickwall limit at 75kHz deviation).

The beauty of a delay-line limiter is that it doesn't overshoot, doesn't "pump" or "breathe", and is pretty transparent in operation. It's possible to overdo the limiting - drive the limiter too hard and it sounds horribly "squashed" - but if the levels are sane, you can't hear it in operation. The "density" can be increased, and it's possible to reduce the dynamic range for people in their cars, but I don't ever pander to the silly programme directors who insist that their station must be the "loudest" on the band!

All clippers - when active - introduce distortion. I don't care what Robert Orban says - even his expensive multi-band efforts just add huge amounts of distortion!

If anyone wants to try building their own "feed-forward" delay line limiter, I can provide the details. I designed stuff for Inovonics (the "David" range of gear), and set up correctly, they're some of the best-sounding and most cost-effective processors / stereo coders on the market. They were designed to be competitive, and capable of sounding "loud", but that wasn't the principal aim - we just wanted a high quality, easy-to-use processor that would blow all the others out of the water. I think that (within the price constraints) we won!

[radium98]

History

[Bton-FM]

Hasn't jump up been like that for about 2 decades?

I haven't been around long enough to know

...But, Warhead by Krust, which I think is an early jump track, Is smashed with compression after the initial drop. The waveform looks like a solid block.

That's from 1997... so 23 years of no dynamics and the same foghorn on every track!

[Maximus]

Here's a better version of my preset without the clipping.