1) I would check to highest amplitude and see what the peaks look like, and listen to it (to see/hear if I find excessive distortion).

2) I usually hard-clip and then limit.Then limit and hard-clip at the end. But that's just me I guess.

 

that was very enlightening, tnx a lot

 

clipped audio is a cut off waveform showing as flat,
it's impossible to fix something which is lost, you can only rebuild some sort of estimation
https://www.pro-tools-expert.com/production-expert-1/whats-the-best-audio-restoration-software

 

I've already mentioned in a couple of posts above why the "find clipping" stuff on audacity won't work.

And no, it's not the case that a declipper won't do nothing if it doesn't identify clipped material, a good example: if u just lower the threshold in RX's de-clip, it will start its process irrespective of whether or not the audio is clipped, the same case applies with the declipper module in Thimeo Stereo Tool: it will start its process automatically when enabling it, at least if used in advanced mode.

 

Last edited: Jul 26, 2022

ya, but that's orthogonal to the topic at hand.

 

does that recognizes clipped samples of an audio whose max true peak value is below 0bdfs?

 

1- yeah, but again, doing that for every peak is tedious and it's worst if we talk about lengthy audio.

2- lol

 

i think you are over thinkin! it's a case of tops,mids,bass. or upward compression
https://plugins4free.com/plugin/857/

 

Last edited by a moderator: Jul 27, 2022

Try not to fry ya brain on peaks,which i have for years! as long as it sounds warm,you're good to go! ANALOG WINS!

 

I don't like to sweep stuff under the carpet, especially with stuff like this, hence why I asked my question. I'm pretty sure there's a solution that I'm not aware of.

 

also, I'd appreciate if u answer my question and don't ignore it. tnx!

 

Lots of good answers here, not sure what else you'd want to know.

[Overly simplified:] Digital audio samples are just snapshots of a wave's amplitude taken several thousand times per second, that amplitude is stored as a single value between -1 and 1, nothing else of the wave remains. Your ear can't hear bleep bleep blop 00110110 digital samples, so you have to go back to analog waves. To do that, you draw a sine between samples. A sine is not a line, a sine has a round bottom and a round top. Here's the problem - try it on paper. What happens if you draw a sine between two points that are both at 1 (the max value in the digital domain)? You end up overshooting 1, which leads to distortion - even though the sample's amplitude is at a perfectly fine 1. That's where clippers come into play. But most clippers do not start working at 1, they start earlier. So here's a graph of how clippers/waveshapers affect amplitude using different clipping/waveshaping functions.



The black line is your baseline amplitude from -1 to 1. Purple and orange are the tanh() and sin() functions applied to amplitude. Essentially what many guitar amplifiers and distortion/clipping/saturation/waveshaping plugins use as their clipping function. The other lines are random cubic functions.

The sooner clippers/waveshapers kick in, the softer they are and the harder they'll gobble up transients. Not all clippers shape the bottom and top ends of your waves equally, some are asymmetrical.

Estimating how much something clipped/was reshaped and if it was deliberate or accidental is not a hard science (unless you have a baseline on hand). Half the plugins you have in your chain are on some level reshaping the wave and thus 'clipping' because they start kicking in well before you hit 1.

 

I tried, seriously tried after your original post. I listened to about eight videos but to no avail. If you could suggest a few, recommend those which are not wubby for my poor Boomer ears.

I listen to everything, have a 17 year old daughter who is constantly turning me on to things that I would never hear and am open to quality in any and every genre. Hit me with your rhythm stick.

 

so, it's possible to find hard clipping digital distortion (clipped samples) with some sort of analysis tool even if the audio's true peaks are normalized to values below 0dbfs, correct?

do u mean non-linear plugins? or are u also including linear plugins?

 

Same way thousand albums were mastered by clipping a TC Finalizer
That's why good mastering engs say "TRUST YOUR EARS" as first rule.
And second one "Know your gear".

 

No, what I meant by 'it's not a hard science' is that it's just a guess, even if it's a good guess. 'probably' or 'probably not' is as good as it gets. If it says 'analog', it distorts. If it says 'colourful' or 'character', it distorts. If it says 'sound gooderizer', it distorts. Everything distorts. Your compressor, your saturator, your clipper, your limiter, your guitar pedal, your amp, your waveshaper, your transient shaper, your sub enhancer, your exciter, they all probably distort.

All big synths offer filters that distort. Most EQs don't deliberately distort, unless they provide 'ladder', 'diode', 'transistor' or similar filters.

IR loaders shouldn't distort. I suppose most filter-derived effects don't saturate? I don't know.

Most reverbs and delays probably only distort the wet signal.

 

no, only non-linear systems can distort the signal, linear systems aren't capable of doing that due to the principles of additivity and homogeneity. reverbs and delays are linear systems.

they won't saturate unless they have a non-linear system added, u gave a good example when u mentioned "synths that offer filters that distort", since they have distortion algorithms implemented that will change the waveshape of the signal (create partials), ergo, the following is a bit unaccurate:

"they all probably distort" is unaccurate. they will distort is correct, since all of the processors u mentioned are non-linear, but also saying that "everything distorts" is a bit of a strech bc that sort of implies that linear systems are included in that statement.

that's a bit dissapointing, but tnx for the reply!

 

Last edited: Jul 29, 2022

Only the most basic verbs and delays. Here, I use this delay and reverb quite often. They literally come with 'drive' and 'distortion' knobs. Most verbs and delays distort their wet signal, even if they don't say so in their UI.

 

nah, that's totally not the case if we are talking about plugins. the majority of them don't provide any distortion at all, and the ones that do, usually offer separate controls within the plugin, just as the one in the picture u attached.

either way, in most cases it's better to just use a separate distortion plugin with its own oversampling options and distortion modes, since it provides way more alternatives to choose from.

do u think that if delays or verbs don't have a distortion knob/s, that makes them basic? I guess altiverb, pro-r, seventh heaven, and a large etc are basic in ur view. ¯\_(ツ)_/¯

I think the proper way to frame this is just to say that reverbs and delays are linear systems in principle.

 

Last edited: Jul 29, 2022

we were talking about hard-clipping all the time right? in my head it makes more sense in mastering than soft-clipping, I'm experimenting with it and hard-clipping is better at shaving little peaks without affecting rest of the signal like soft-clipping (or saturation) would do