BuzzCut (The Erosion Clipper)

[LusiD_Music_UK]

Firstly, masks, covers up, replaces. I use these terms very interchangeably. I think you’re being pedantic here. Secondly, the image compares normal clipping to BuzzCut “at its best”. While the oversampling is the reason for the reduction in aliasing (erosion for smoothing) I don’t specifically claim otherwise. Although I do admit in the context of talking about erosion. The omission of that fact is a bit misleading. At the end of the day I chose to show my plugin in the best light possible with a simple A/B on a page designed for selling a product (as would any marketing team). I think I will own this mistake though. It is probably over the line. I’ll switch it to the 16x vs 16x + erosion. I doubt it will have a noticeable negative impact on sales.

[Zeisner]

Firstly, masks, covers up, replaces. I use these terms very interchangeably.

Then you're using both wrong because they're very different things.

I think you’re being pedantic here.

See, you don't realize the scope here. You claimed that "stochastic erosion" is a solution against aliasing which is not the case, only oversampling is. But with oversampling comes latency. With latency comes issues with live performance - the higher the latency, the more difficult performing gets. A possible customer who performs live and who is not as familiar with audio engineering as me might look at your plugin and wrongly think "Hey, stochastic erosion shatting aliasing? Cool! No more screeching and droning sounds when I play my synths live through that other clipper!" and then try it out and realizes it's useless for live performance. No big deal if it was just a demo. But if they bought it right away they will be pissed and will want their money back. And they will write not positively about your plugin on the web. You can't afford that, especially not if you want to get a foot in the door.

You only need to change the sales pitch. Seperate the "stochastic erosion" (which is not a good term anyway, that's just technobabble) from the oversampling, done. Something like "Tired of clippers crushing your music? Let your music crush the clipper!". If you have a function that combines input level with the random generator to control the threshold it would be much closer to the real thing. And at the end you can still mention oversampling. "Don't let aliasing crush your music! Features x16 oversampling." I'm sure there is a better word than shattering or crushing but you should get the direction I'm pointing at. I would also empathize the asymmetrical function a lot more because this is something many developers still ignore, I often have to use additional plugins to add bias to get the even:odd ratio right.

My criticism is valid. I explained it over and over again and even added a better sales pitch. For free. I'm not losing money but you might lose money. I'm literally trying to help you to make more money. It's up to you to listen or not.

[Dr.Gunjah]

@lusid do not feed the troll

[hey212]

Just to be clear, we're saying the audio in this image was NOT oversampled:



...but the audio in this image WAS oversampled?



If that were the case, it would certainly seem a bit dishonest to me.

That's the whole point about "stochastic erosion shattering aliasing". It's the oversampling, not the clipping.

Well clearly, if you wanted to demonstrate the difference between two different clipping algorithms, you would need to hold all other variables equal, to have an apples to apples comparison. In this case that would mean oversampling BOTH clipping algorithms, or oversampling NEITHER.

Those two images and the marketing copy heavily suggest an apples to apples comparison, which if it actually were, would be extremely impressive, because it would mean that Buzzcut's clipping algorithm had eliminated the aliasing. Sadly, it seems to be as you suggest, that the oversampling alone is what eliminated the aliasing. Pretty shady.

[Zeisner]

because it would mean that Buzzcut's clipping algorithm had eliminated the aliasing. Sadly, it seems to be as you suggest, that the oversampling alone is what eliminated the aliasing. Pretty shady.

Correct. And it's not even necessary to claim that "stochastic erosion" reduces aliasing! There is oversampling anyway and it does its job! So why even connect both? Let the "stochastic erosion" just be a particular special effect that does its own thing with oversampling on top. Have two pictures with and without oversampling, have two animations with fixed and randomized thresholds. Use a different sales pitch without technobabble and claims about a new antialiasing method, done. It's that easy, way less work than coding the plugin.

[jinxtigr]

Ah, KVR. I missed ya, never change

I know dithering but I don't know if applying this technique with a big trowel makes it BE dithering. Dithering does indeed replace truncation distortion: if you've decorrelated both of the moments of truncation, you've produced a thing where statistically you're revealing information (like low frequencies) that would be GONE when truncated, but are easily retrievable (for instance through filtering) if you've dithered.

It is THIS that LusiD's thing brings to over-loud dithering, and that makes perfect sense given that it's for over-loud bass music at impossibly high levels. I have no idea if it'll actually break up high frequency aliasing: I'll experiment with that as I've got something that could try it, but it won't do what 'erosion clipping' does.

I think that's a fine term, especially since there's already a BuzzCut plugin, so I say stick with the 'erosion clipping' term, that can be 'how to deliver a bit of LF information while clipping beyond 0 dBFS'. Welcome to the plugin biz, don't let it stress ya too much. KVR will not sugar coat things

[Zoopy]

i just wanted to let everyone know i understand everything in this thread but i just dont really feel like chiming in right now

[onerob]

I'm an 'Average Joe' user, not a DSP guy, and I get the deal with masking – not technically removing – the aliasing. I've not been misled.

I didn't look at the comparison images for more than a millisecond but I think the decision to update them with a fairer apples-to-apples version is right.

So, all in all, I don't think there's a need to persecute a new developer for not using exactly the concretely defined terminology you might prefer.

[Zeisner]

Ah, KVR. I missed ya, never change

I see you still haven't corrected the description of Flipity. Because it's actually polarity inversion...

[Vortifex]

Weird thread...

[LusiD_Music_UK]

Alright guys. I’ve changed the image on my website for a more apples to apples comparison. Instead of the more marketable “normal clipper” vs “my clipper with maxed out oversampling”. I can see how that crossed a line. it’s now strictly A/B ing the effect of erosion clipping itself. I still want to stress that beyond this visual interpretation, I did not ever actually claim (to my knowledge) that erosion clipping “solves” aliasing and I’ve always claimed that it “masks/replaces” it with arguably more pleasing noise. (Using dither for truncation as an analogy) The comparison wasn’t a fair one and while I do believe showing your plugin in “the best light possible” is important, the omission of oversampling in the comparison in the context of a novel idea implied it is way more impressive than it actually is. Obviously the concept is still sound and the plugin is still an amazing tool for disrespectful clipping.
I have now fixed it. Hopefully I’ve not f**ked up my reputation already I’m on my own here and I was just trying to do “effective marketing”. I hope you’ll appreciate the “realising I was in the wrong” on this one and fixing it quickly for what it’s worth.
Ie. You can stop bullying me now Zeisner
Much appreciated!

[Zeisner]

Forget it.

[LusiD_Music_UK]

KVR will not sugar coat things

Yea. I’ve just learned that today

[pekbro]

Not for the squeamish certainly.

[BCollins]

Being the most average among all the average Joes - while making a living with these kind of tools - the only thing I can say is that the results I get from using this clipper sound more open and "clean" than any of my - much more expensive - clippers.

I´m not using it for extreme gain reduction and mostly with 16x os, though...