BuzzCut (The Erosion Clipper)

[LusiD_Music_UK]

I need an energy drink version – shatters my quantization error energy and tilts me towards a higher frequency...

I'll record some freshly opened redbull and use the fizz as modulation just for you

[Dr.Gunjah]

After quite a bit of testing I still can't hear the magic tbh. Maybe I'm still using it wrong, but even at -3 LUFS (which sounds awful anyway) I can't really hear any difference between my usual clippers (Airwindows 'Adclip8' or the clipper in 'Mastering') and this. When comparing them I've used Buzzcut at 16x oversampling with various noise levels and buzz thresholds.

Is there any improvement to be expected anyway if you just aim for something like 7-8 LUFS?

[billinder33]

After quite a bit of testing I still can't hear the magic tbh. Maybe I'm still using it wrong, but even at -3 LUFS (which sounds awful anyway) I can't really hear any difference between my usual clippers (Airwindows 'Adclip8' or the clipper in 'Mastering') and this. When comparing them I've used Buzzcut at 16x oversampling with various noise levels and buzz thresholds.

Is there any improvement to be expected anyway if you just aim for something like 7-8 LUFS?

I would argue that trying to crush a mix into a clipper (BuzzCut included) for loudness rarely does anything other than create unwanted distortion.

In theory, you could use BuzzCut to increase loudness over a normal clipper by driving into the soft clipper before the hard clipper stage, and I think this is where the claims of "louder than a typical clipper" come from. That's basically like running Sonnox Inflator or UrsaDSP Boost into a hard clipper like KClip. And the Erosion function in BuzzCut might make the ensuing distortion a little more palatable, but that's a matter of taste.

IMO the better way to use BuzzCut is on the 2-buss just snipping the peaks where your Kick and Snare hit. When you compare that to a typical hard clipper, you'll notice that it softens the transients and won't sound as sudden or harsh as a typical hard clipper. IMO that's the magic here, but it's a very subtle effect and isn't really going to give you crazy loudness.

If you want crazy loudness, that has to happen in all stages of your mix with lots of different techniques before everything hits a 2-buss clipper. Hitting the 2-buss input at -8 LUFs and expecting a good sounding mix at -3 LUFs isn't very realistic. The most loudness you should expect to gain from 2-buss processing is maybe one extra LUF maximum. Anything beyond that is mostly going to flatten and damage your mix. If you want -3, figure out how to get to -4 before your 2-buss input. IMO

[LusiD_Music_UK]

After quite a bit of testing I still can't hear the magic tbh. Maybe I'm still using it wrong, but even at -3 LUFS (which sounds awful anyway) I can't really hear any difference between my usual clippers (Airwindows 'Adclip8' or the clipper in 'Mastering') and this. When comparing them I've used Buzzcut at 16x oversampling with various noise levels and buzz thresholds.

Is there any improvement to be expected anyway if you just aim for something like 7-8 LUFS?

Usually when things sound good or bad there is a technical reason why. There's two main DSP differences. BuzzCut's soft clip is using an exponential transfer curve meaning the details of your signal are always present in the file no matter how hard you push (within the limitations of available bit depth - is this noticeable? debatable, but I'd say it makes more logical sense). The erosion clipping serves to mask aliasing when used subtly (or provide vibe when pushed). Again, these are all fairly subtle differences backed by my intuitive understanding of what would be the best way to clip and sound good.

If you want to trim some peaks with a clipper.. BuzzCut is probably going to sound best. I'd say these differences logically compound to become more obvious the more you destroy your signal like we do in DNB and Dubstep. Is it magic? No. But, is it better than all the other clippers out there? For the reasons above and many other "non DSP reasons" I'd say yes. I built it for myself.

[LusiD_Music_UK]

I would argue that trying to crush a mix into a clipper (BuzzCut included) for loudness rarely does anything other than create unwanted distortion.

As a DNB producer, I actually mix loud (-3LUFS) into the clipper (now BuzzCut) from the get go. I don't add a clipper to the mix afterwards whenit's finished. That's just asking for nasty surprises. I wouldn't recommend the "mixing into it" approach for anything outside of EDM/Bass Music though.

The Erosion function in BuzzCut might make the ensuing distortion a little more palatable, but that's a matter of taste.

Exactly. Logically, in the real world sounds don't just square off at some stable mathematical value. Adding some randomness should in theory sound a bit more natural. The theory is true (in my subjective, biased, opinion).

[Dr.Gunjah]

Usually when things sound good or bad there is a technical reason why. There's two main DSP differences. BuzzCut's soft clip is using an exponential transfer curve meaning the details of your signal are always present in the file no matter how hard you push (within the limitations of available bit depth - is this noticeable? debatable, but I'd say it makes more logical sense). The erosion clipping serves to mask aliasing when used subtly (or provide vibe when pushed). Again, these are all fairly subtle differences backed by my intuitive understanding of what would be the best way to clip and sound good.

If you want to trim some peaks with a clipper.. BuzzCut is probably going to sound best. I'd say these differences logically compound to become more obvious the more you destroy your signal like we do in DNB and Dubstep. Is it magic? No. But, is it better than all the other clippers out there? For the reasons above and many other "non DSP reasons" I'd say yes. I built it for myself.

I honestly can't confirm it's the best clipper. It sounds neither better nor worse than Adclip8 or Kclip zero to me. I couldn't even hear a difference between erosion on and off which makes me believe I've probably still not set it up correctly. The UI is not quite intuitive to me. Is there a manual now?
My demo ran out already but maybe when reading about the correct setup I can figure out what went wrong

[LusiD_Music_UK]

I honestly can't confirm it's the best clipper. It sounds neither better nor worse than Adclip8 or Kclip zero to me. I couldn't even hear a difference between erosion on and off which makes me believe I've probably still not set it up correctly. The UI is not quite intuitive to me. Is there a manual now?
My demo ran out already but maybe when reading about the correct setup I can figure out what went wrong

Did you get the 72hour extension too? You must have been very early. Manual is my next priority! Won't be long.

So erosion isn't on or off? It's a gradient. That makes me wonder if you missed the thresholds on the oscilloscope? (they do disappear to keep the UI clean when your cursor isn't close to them...) If you drag "BUZZ" down a lot to -10dB for example you will 100% hear what it's doing. Obviously that's very extreme and in that case is just going to sound like noisy wave shaping (which can be useful itself in certain scenarios, like making sub bass audible at higher frequencies).

If you click the delta button and compare that to the delta of K-clip

dry signal running in parallel, invert phase of one signal. Just mentioning how in case you (or anyone reading) doesn't know. . Most other clippers don't have a built in delta, because they suck

Comparing delta's you'd 100% hear the difference even using normal settings.

You could even delta the deltas and hear nothing but the difference. Deltaception

Better or worse is of course subjective. It just provides another option that no other clipper (that I'm aware of) provides.

[Dr.Gunjah]

I don't know about an extension. I had the standard 14 days from what I can tell.
If it's possible to get an extension, I'd gladly use that once the manual is out.

In regard to erosion, I actually missed the controls on the oscilloscope for the first few days since the labels only show up on mouse-over. My last tests were using all controls though. I just didn't find a setting for buzz and noise level that I preferred over simply clipping with 16x OS.
With erosion on/off I don't mean an actual button btw. I assumed a noise level fully left means no noise is injected at all, but that may be a misconception on my side. Generally while using the plugin I've never had the feeling that I know what I'm doing (still no idea what the buzz threshold is supposed to do, for example) so a manual which describes the workflow would help me a lot.

[Funky40]

ok, gave it a go. I make no Drum&Bass, i make live play experimental electronic.
Here would be my special feature request. So, that´s re. creative uses.....for the fun of it :

- wish it had a output 3+4 => and the option to run the delta to that out 3+4 (probably with a own button then to select where the delta is routed to).


- even more fun: add a crossfader to crossfade the regular signal with the Delta. It would probably require to have a own volume control for the delta re. how load it would go into said X-fade

Would be huge fun to get such

[LusiD_Music_UK]

I'd gladly use that once the manual is out.

I've just "finished" the user manual today!

https://lusidmusicuk.com/pages/buzzcut- ... per-manual

Please let me know if there's anything that's still not quite clear enough, could be explained better etc!

[LusiD_Music_UK]

- wish it had a output 3+4 => and the option to run the delta to that out 3+4 (probably with a own button then to select where the delta is routed to).

That's a fun idea for sure! I'd be quite scared trying to add new outputs though in case certain DAWs don't like it.

- even more fun: add a crossfader to crossfade the regular signal with the Delta.

Good news! This feature already exists! (The mix knob)

[Dr.Gunjah]

I've just "finished" the user manual today!

https://lusidmusicuk.com/pages/buzzcut- ... per-manual

Please let me know if there's anything that's still not quite clear enough, could be explained better etc!

Good job from what I can see

The interactive manual definitely looks helpful and is easier to navigate than a traditional manual.
Is it possible to get a demo extension now that it's out?

[billinder33]

I'd gladly use that once the manual is out.

I've just "finished" the user manual today!

https://lusidmusicuk.com/pages/buzzcut- ... per-manual

Please let me know if there's anything that's still not quite clear enough, could be explained better etc!

The signal flow is very useful. One question, if I used the soft clip with downward erosion, would the hard clip set at 0db ever be reached?

[LusiD_Music_UK]

I've just "finished" the user manual today!

https://lusidmusicuk.com/pages/buzzcut- ... per-manual

Please let me know if there's anything that's still not quite clear enough, could be explained better etc!

Good job from what I can see

The interactive manual definitely looks helpful and is easier to navigate than a traditional manual.
Is it possible to get a demo extension now that it's out?

Thank you! It would have been so much easier to do a PDF but this is so much cooler. Hope it was worth it

Unfortunately I don't actually have this power
(At least not without A LOT of headache)

The trial is tied to your machine ID so you can get a new trial on a 2nd machine.
BPB30 at checkout will bring the price down to like $34
I've also set up a path to get a free license here:
https://lusidmusicuk.com/pages/buzzcut-bounty-board

[LusiD_Music_UK]

The signal flow is very useful. One question, if I used the soft clip with downward erosion, would the hard clip set at 0db ever be reached?

Thanks I appreciate that!

Love this question <3

If digital audio had perfect resolution. The answer would be "No purely because the soft clip is exponential".

The erosion does serve to reduce the likelihood of a sample reaching 0 but it's still very common. The noise modulator wasn't rectified to maintain the filter shape, so my guess would be that (depending on noise frequency tilt) ballpark 10-50% of total samples that would have clipped still make it to 0dB with erosion alone. (I'm pretty sure you can actually stack erosion in series to make it less common if you wanted. Been considering this as an update, so you could do that internally. Need to test that.)

In practical reality, with the exponential soft clip which is the real thing preventing true 0dB, we're working within the limitations of the available bit depth which isn't infinite.

So yes. 0dB (or at least within like 0.01% of 0dB) will get reached quite easily.
I'd be interested to actually check this and give a proper answer at some point.