compression with Levelizer or channel strip

[Dave Labrecque]

So.... if I understand correctly... the "Cleanest" mixdown (in case there are overs) is with "SoftClip" off and the Levelizer patched into an output and appropriately adjusted. Right? (advice?)

I think the 'cleanest' mixdown is one with no clipping, regardless of SoftClip implementation. A recorded, clipped signal is not a happy thing. But the Levelizer could be used to preclude the possibility of clipping in certain mixdown situations. The trade-off is brick-wall limiting instead of clipping -- which still sounds amazingly transparent up to several dB of gain reduction for most program material in my experience. This is why my standard mix template has a Levelizer set to limit at 0 dB patched post-fader everywhere, and SoftClip turned off. Even so, those are measures intended to help me out when headroom runs out and my options are limited (no pun intended). I still try to keep my mixer gain structures under control as to limit (no pun intended) the use of this limiting.

[Grekim]

The way I use the Levelizer (and limiters in general) is to set the threshold to where it catches the few highest peaks in the song and then set the output (makeup gain) a little less than the threshold. So, if your threshold is -6.50 dB then your output gain might be set to +6.00 dB. This will give you a maximum peak of -0.50 dB. If you are trying for a certain RMS level then there's more involved, but that's a simple way to use it.
I don't usually leave soft clip on. My levels are always well below full scale until they've passed through the limiter on the master output.

[Carl G.]

Not really... since the soft-clip does not affect signal at all until it passes clipping... which it should never do... you can leave the soft-clip ON and it will not affect anything if your source signal levels do not clip... and if they do... you would probably want the soft-clip to flattop them rather than have them roll-over... the Levelizer will then take care of the final levels after processing and mixing.

Bob L

Very interesting protection you designed there - I remember now what brought about that feature. It reminds me that I should leave it on, while riding gain. Thanks, Bob.

[jmh]

I'm a bit rusty on the details as I've been focusing on a construction project for the last several months - going green! A while back, I had suggested a switch to place the limiter before the compressor within the levelizer. Of course some things that are easy to suggest might be difficult to do. But I think it would boost the levelizer's power. You can accomplish this already by using levelizer peak limit in front of a compressor. The compressor's response can be altered quite a bit (it is easy to do too much). You can use 2 levelizers but the interface becomes confusing or set the strip FX pre compressor which I find a bit cumbersome.

[Carl G.]

I'm a bit rusty on the details as I've been focusing on a construction project for the last several months - going green! A while back, I had suggested a switch to place the limiter before the compressor within the levelizer. Of course some things that are easy to suggest might be difficult to do. But I think it would boost the levelizer's power. You can accomplish this already by using levelizer peak limit in front of a compressor. The compressor's response can be altered quite a bit (it is easy to do too much). You can use 2 levelizers but the interface becomes confusing or set the strip FX pre compressor which I find a bit cumbersome.

Interesting - unusual, creative concept:
Lightly touch off the 'extreme peaks' before compression in order to limit (pardon the pun) compression artefacts reacting on large transients.

I always prefer leaving in "some degree" of transient peaks on most things in order to bring it to life.

[jmh]

Carl, I can't take as much credit as your comment. I got the idea for that setup after reading about the 50 year old Fairchild - which I've never had the pleasure of using. If I were to re-submit this, I would include your idea as a slider for raw signal bipass. And incorporate the Grekim's method of setting gain a small dB below the peak limit with a 'chain' button. I find the Levelizer is great as is, but those would be nice add ons that would expand the power and flexibility.

[jmh]

I'm going to experiment with routing the Levelizer's compressor key signal through a limiter, or a mix of raw and limited key. This might achieve mostly similar effect to pre-compressor peak limit? In this capacity, a simple limiter would probably work just as well as levelizer's. You can get a different kind of fattening of a track than simple compression. It is a little tedious tweaking all the parameters with multiple plugins that interact so much but at least you can evaluate if it is something worth asking for.

[Grekim]

Transients are most likely due to a specific instrument rather than the band as a whole, so consider treating that specific track with a limiter before it reaches the mix.

[Cary B. Cornett]

I would think that, if you want to have the compressor not react to very short transient peaks, the answer would be to just increase the attack time on the compressor. A common compromise that I use is a compressor attack time of anywhere from 20 to 100 msec or so, letting the limiter grab the really fast stuff that "gets away" from the compressor.

Am I missing something here?

[jmh]

A non-zero attack happens after the threshold level is exceeded and the compressor is triggered. Consider a picked acoustic guitar waveform - it generally has a spike (particularly with a pick) of percussive noise preceding the tone - this can be kind of irregular. If the peak level determines the compression depth, you get some irregularities in your compression. A RMS triggered compressor smooths out the response of the compressor as a momentary peak will not contribute as much to the time averaged energy that bosses the compression resulting in an overall smoother compression sound. Similarly, but not nearly the same, knocking down wild transients that trigger the compression change the response of the compression (and this might not be a bad thing with RMS). Anyway I found this signal path changes the nature of the compressor's sound and enabled a fattening of the sound in a way that you don't get with straight compressor. And as I said, it is easy to do to much. All of that aside, try it out and see if it works for you... Maybe I'm just being nostalgic of being raised on the sounds of analog circuits (kids ignore this) particularly the ones that sound real good be it saturation of transformer cores, saturation of magnetic tape or pushing a signal through a vacuum tube beyond it's linear range. And then there is all of those other beautiful circuits of old that evolved over many decades. Some of the stuff that is relatively straightforward in the real world, proves difficult to do in digital realm - although I don't miss the hiss... One of the funny thing the saturation issues I listed were probably regarded as problems.