Alternatives to Sonnox Inflator?

[Ploki]

It's been proven over and over that it's just a clipper.

Maybe the main function could be described as "just a clipper", but that doesn't mean any old clipper will do exactly what it does. The DIY versions like the JS Inflator have carefully tweaked parameters to make it actually null on all settings, instead of just being in the same ballpark with fixed examples.

No, many clippers will sound different that's why people use different clippers.
OVC-128, SIR StandardClip, Venn VClip,New Fangled saturate, Inflator, Kazrog KClip, Signum SKYE, IK Classic - have different (some very similar) clip shapes (some adjustable, with adjusting having different intermittent rounding shape even if end is sine) and consequently slightly different sound. They also have different parameters available so you interact with them differently. They have different oversampling options so they have more or less aliasing.
they also have different oversampling filters so they have more or less preringing or phase distortion.

But they are all clippers. None is "any old clipper", but none is an upward-downward physics bending whatevermachine

[ramseysounds]

Confused here. If all a clipper does is literally clip the tops off a waveform usually to prevent it going over 0db, how can they ‘sound’ different? How can a clipper alter the sound of a track?

[gaggle of hermits]

Confused here. If all a clipper does is literally clip the tops off a waveform usually to prevent it going over 0db, how can they ‘sound’ different? How can a clipper alter the sound of a track?

if you're only clipping the tops of hi-hat transients, it won't make much of a difference. if you're slamming a kick or bass tone to the point where you're changing the core wave shape, then it is going to make a difference. some of these things, such as clipshifter, move the clip threshold while it's clipping, either relaxing it or tightening it. so that's going to make quite a difference on a sustained bass versus something that just lops the top xdB off.

[ramseysounds]

I need to look more closely into this. Usually with correct gain staging you shouldn’t need either. I only ever use a limiter as a last resort on a 2 buss for any pesky overs I can’t contain. I can’t think of an instance where I’ve ever used a clipper. Just turn it down

[gaggle of hermits]

hi-hats can create big transient spikes that will trigger a limiter when you don't really want that to happen. just turning them down doesn't work that well. clipper's are great for handling this, and probably best not on the master or even the drum bus.

outside this, in a lot of cases, the clipper's being used as a sound-design tool not a mixing tool.

[El°HYM]

nvm

[dionenoid]

Confused here. If all a clipper does is literally clip the tops off a waveform usually to prevent it going over 0db, how can they ‘sound’ different? How can a clipper alter the sound of a track?

When you use hard clipping they all sound the same. When using soft clipping is where they start to sound different. Hard clipping only affects the peaks above a set threshold, soft clipping usually also affects signal lower than the threshold.

But basically all hard clippers null with each other, there are only slight differences because of the used oversampling methods and/or values. But even these differences are very small when you only use them to shave off peaks.

in the case of Inflator, it's not crazy that others null with Inflator when it's set to a hard curve (hard clipping). As they all do.

And clipping can alter the sound of a track, simply because shaving off peaks affects all further processing ; a snare with a 10db initial peak will sound very differently going through a compressor than that same snare with a 5db peak. Or a mix with a clipper catching some 1db peaks will sound different through a limiter, because the limiter has 1db less work to do. A clipper can be very useful in these cases.

[Ploki]

I need to look more closely into this. Usually with correct gain staging you shouldn’t need either. I only ever use a limiter as a last resort on a 2 buss for any pesky overs I can’t contain. I can’t think of an instance where I’ve ever used a clipper. Just turn it down

Most clippers also have some form of soft clipping. The "shape" of the curve affects what kind of harmonics in what kind of proportions are produced.

sc-f1.gif

If it's hard clipping, most sound similar. However even with hard clipping:

Oversampling:
- depending on oversampling factor (can go from 2x to 128x (Voxengo) or 256x (VClip) or more 1024x (Melda i.e.), it will have more or less aliasing which also affect sound.
Hardclipping without oversampling will mirror A LOT.

And finally AA filters:
Minimal phase vs Linear phase - preringing or phase distortion,
AA filter frequency, order, Q and topology.

DC block filters
Clipping can cause DC.
Low frequency content filters, same parameters as AA filters.

This is what affects sound of the clipping without considering other factors a clipper might employ.

----

Re: turn it down:
you don't use a clipper incidentally. You do it deliberately because it has a certain (popular) sonic signature, esp with Pop and Electronic music. Because it's not a dynamic processor like Limiter, it won't "pump" or squash, but overshoots cause distortion - which can augment the tracks character, or, kill it. That's for mastering.
Clipping drums is common even in mixing, and not only electronic. Clipping (soft clipping) is what "driving the console" was on analog recordings.

You also generally can't finish with a clipper, at least if you don't want to have peaks in the end, because a lot can happen with aggressive clips (3+ dB of clipping)

I need to look more closely into this. Usually with correct gain staging you shouldn’t need either. I only ever use a limiter as a last resort on a 2 buss for any pesky overs I can’t contain. I can’t think of an instance where I’ve ever used a clipper. Just turn it down

You can't get a modern loud master without limiting and/or clipping. No matter how well "gain staged" the mix is. It doesn't have much to do with gain staging anyway, it's a tool like every other.
You don't always use a compressor just to "level' the amplitude inconsistencies (which is what it's original intent was), you sometimes use it to impose a specific sonic signature.

[diroxe7660]

hard clipping sets the amplitude of samples that exceed a threshold to exactly that same threshold, but will leave samples below the threshold completely untouched. usually that means whenever the sample exceeds 1, it will be set to 1. hard clipping adds a ton of harmonics but best preserves dynamics.

soft clipping will alter samples according to whatever function the developer chose and this function is usually applied regardless of amplitude of the sample (there is no threshold). tanh is very popular, but so are many proprietary functions.

most soft clipping functions will lower the amplitude ever increasingly the more amplitude comes in, like a compressor.
many soft clipping functions will not allow an incoming signal to leave with an amplitude of more than 1.

the inflator soft clip needs to additionally hard clip because its soft clip function would otherwise result in the wave starting to fold in on itself.

inflator's innovation is its curve value that will slant added harmonics more towards the low end or top end depending on setting.

as far as I know, the various reimplementations out there (rc inflator jsfx, melda, hotair) null perfectly with the original.

[subterfuge]

inflator's innovation is its curve value that will slant added harmonics more towards the low end or top end depending on setting.

as far as I know, the various reimplementations out there (rc inflator jsfx, melda, hotair) null perfectly with the original.

Yes, at least with RC Inflator they spent some time trying out and testing to find the correct curves. Keeping in mind, that was supposed to be as close as possible to Inflator to begin with. Maybe an alternative to Inflator doesn't necessarily have to be identical though.

[ramseysounds]

Ironically this was posted today

[plexuss]

Ironically this was posted today

He addresses why I don't use clippers - for my music I do not like the sound at all unless I am going for that brittle high end and reduced dynamic range soaking distortion that seems popular in popular music productions. I don't make popular music and prefer a clean smooth transparent high quality sound.

[sqigls]

clippers are often good for snare drums

[Uncle E]

huh? It's been proven over and over that it's just a clipper.

The Effect and Curve controls have a lot of effect on the frequency response. It's easily measurable, not that you'd need to measure because it's obvious when you hear what they do. Fairer to say that Inflator is a clipper and enhancer.

FWIW, in my 20+ years of using Inflator, I've never used it the way Paul Frindle recommends (turn Effect to 100% and play with input and output). 100% is always too much for me.

[MrBeagleton]

The Effect and Curve controls have a lot of effect on the frequency response. It's easily measurable, not that you'd need to measure because it's obvious when you hear what they do. Fairer to say that Inflator is a clipper and enhancer.

Sorry, but no they don't. You may perceive the distortion that way, but it's just a clipper. Also, Inflator boosts the input by as much as 6 dB as you turn up Effect which throws off one's ability to evaluate it fairly.

FWIW, in my 20+ years of using Inflator, I've never used it the way Paul Frindle recommends (turn Effect to 100% and play with input and output). 100% is always too much for me.

I have to agree with keeping Effect at 100%.

The Effect control doesn't create a proper soft-knee how you might expect it to. It just does a linear crossfade between a hard-clipper (E=0) and a soft-clipper (E=100). So for any Effect value other than 100, you're getting some amount of hard clipping in your sound.

Inflator's curve shown here vs. GClip (which does a proper soft-knee):
https://www.desmos.com/calculator/oagvhrtjw3



Notice how GClip has no slope discontinues. Low inputs are linear and gradually transition into the clip ceiling as input level increases. Compare with Inflator's hard discontinuity at x=1.


Same graph, but removing the input boost from Inflator.
https://www.desmos.com/calculator/n9fmtv0dvy



Notice: when Effect=100 and Curve=50, Inflator and GClip are identical.

Conclusion, GClip is actually a better soft-clipper than Oxford Inflator.