The term "chopping it off" actually should be used for clippers. Clipper is the one makes waveform that exceeds the threshold flat. Limiters is "compressing it down" while remaining the waveform as original as possible.
OK, no one explains the difference correctly, so I'll do it here. Imagine we have a 44.1KHz, 16-bit PCM digital audio format for export. Each of the 44,100 samples per second (one stream for left and one for right) has a value between -32768 to 32767 (signed, twos-complement 16-bit binary). But, our DAW uses 32-bit or 64-bit floating-point internally, so there are values allowed outside this range (that's why you can go into the red on your individual and even master tracks all over the place but not distort). In the final PCM file, however, such values aren't allowed. What do we do with values outside this range (from -96db to 0db, 6db per bit)? A clipper simply takes anything less than -32768 and makes it -32768; it takes anything greater than 32767 and makes it 32767. That is, a clipper throws aways information. This has the effect of saturation or distortion depending on how far out of range the sample values are. A limiter, in contrast, scales out-of-range values by multiplying by a value less than 1.0. Now, how this clipping and limiting is actually done is where the differences in individual clippers and limiters comes in (the more clever they are, usually the better, but cleverness induces latency). Some, actually most if not all, limiters look ahead, some clippers look ahead, and act on groups of sample values. What about that over-sampling value you can set on some clippers and limiters? So, if we have four sample values: 32766, 32767, 32767, 32766, you can guess what the value between the two 32767 values would be if we created, with oversampling, some intermediate value(s): it would be greater than 32767. That's a true peak value: it's what the waveform would look like when it's reconstructed by the D/A conversion on output. So, even though there's no sample in an audio file that's greater than 0db (32767 or less than -32768), there can still be a peak above 0db after the signal is reconstructed for presentation to a speaker (or over-sampled). So, clippers throw away information by just chopping sample values; limiters try to preserve dynamics and harmonics by scaling sample values (a compressor scales by the ratio, for example). Scaling can preserve more of the original information, though it clearly cannot preserve all of it (unless you scale all samples, which is the same as turning down the volume). Also, the issue here is crest factor, which is the difference between the peak value and RMS value of your sound. If you can get a low crest factor, you have the potential to make your final, exported mix loud without too much distortion from either clipping or limiting. And, clipping and limiting serve to reduce crest factor. That's the reason you can use saturation on a, say, cymbal sound and reduce it's peak value while actually making it sound louder: saturation takes the energy of the transient and spreads it around with harmonics that don't change the sound much but reduce the transient/peak energy: this serves to reduce the crest factor of the sound. There's plenty more to say, but, I've said my peace. (This is actually a comment I've found in another YT video and saved in my phone)
A fun thing a lot of anti clipping people don't realize is that the artifacts caused by most limiters (even the most popular one) when not placed into long(ish) lookahead, are similar to the artifacts caused by clipping. The difference is that any halfway decent clipper is going to be built with this in mind. For example, I know airwindows essentially attempts to shape the offset of the clipping, I think OnlyClip2 from him uses slew rate limiting to achieve this. Limiters main downside is that compression is rarely 'transparent', we are much more sensitive to amplitude changes over time (ie the pumping effect) than we are to the 'noise' clipping adds when dealing with transients.
I remember talking more than 20 years ago with one top notch jazz artist about how my synth guitars don't sound real enough, like the real guitar in the mix. And he said? Why do they have to sound real? They sound as they sound. It blew my mind. It was an "aha" moment right away ... The same with clippers, limiters etc., they do what they do, use their characteristics as they are, in your advantage.
Clippers act instantaneously (or at least very nearly if some Analogue modelling is done) limiters hold the signal down longer, reducing harmonics from rapidly modulating waveform, but introducing pumping and more loss of perceived peak level. In digital the limiter can start to reduce the signal before a peak happens (lookahead)
Hi David. The only piece of hardware that I use when mixing is the Zulu Tape emulator. I print groups through the Zulu and blend it with the original take 50%. If I do that with a lot of tracks do I get phase problems? I work with Samplitude Pro X8 and there is a plugin that calculates the latency from hardware units. Thanks....and your new videos look great. 🤘 Michael
Yes you will. First off I'm not a fan of printing in general, that's really not how analog or hybrid mixing works, but besides that, why blend? You're definitely asking for trouble doing that for multiple tracks
Distortion and saturation are not the same thing. The terms are used interchangeably because often non-linear signal processing results in both. Distortion usually leads to saturation, and saturation leads to distortion.
@@mixbustv Not really, you said those are the same thing, only used to describe different levels of subtlety. But really saturation is spectral density, distortion is a process of deforming signal.
You make limiting, clipping etc. the part of the song, as an effect and there's no problem. Every tool you use or you know you'll use (or you experiment and it's sounds good to you) is a part of artistic expression. Choping, limiting, doing crazy stuff, distortion etc. Of course you have to know first what the f... you're doing (or what some tool does) in the first place, but it really doesn't matter; you hear your messy mix or master at the end anyway, haha.
I think the main difference between clippers and limiters is that clippers only look at the current signal level, while limiters operate over time as they may have lookahead and release settings. In other words, clippers work instantly and limiters are compressors with infinite ratios.
No they are not. Limiters have zero attack. That's literally the defining features of brickwall limiters. They have one job and one job only, not let signal pass above a certatin threshold. Without zero/instant attack that they would NOT be limiters.
@@mixbustv They have zero attack, yes. I did not say they don't. I said they may have a lookahead, which means they may even have "negative" attack, meaning they will start processing before the treshold limit. Also they may have a release, which means they will still process the signal after the signal drops below the threshold. The point is: they work over time, while clippers work at the moment.
I discovered a few months back that my old dbx 166xl I bought 20 years ago has a soft clipper circuit named as a “ peak stop limiter “. So after reading this in the manual. I dusted it off and experimented with it to replace my pluggin clippers and ….. I found a use for it. lol
As far as I understood, in the digital realm clippers work on a sample-by-sample basis. This means there is (like a waveshaper) a clipping function (from basic -1 to 1 hard clipping to arctan or tanh or more complex stuff) where each incoming sample is mapped to (incoming sample is located on the X-axis whereas the output sample is taken from the calculated function value at that functional value from X). Each deviation of the transfer function (clipping function) from linear is going to introduce new harmonics in the output signal according to Fourier/Laplace theory. So the transient material doesnt get enhanced by simply pushing gain (or like the enhancement through the attack phase of a compressor/limiter) but by chopping of the wave there are more harmonics generated for the transient what makes it kinda "enhanced/louder". More harmonics mean more RMS, thats why saturators/clippers are essential to achieve modern loudness standards. A limiter on the other hand works basically like an compressor with infinite ratio. It relies on the envelope of a signal which is why there are some "release/attack" controls or lookahead features on some limiters. Based on envelope detection these things try to reduce gain in a more linear fashion in contrast to clippers.
I always thought it was opposite, that clippers cut off the shape (and therefore distorts) and limiters retain the shape but lower the volume. Learned something again. 👍
Yoooo Can you do a video on K system and best levels and balance of a mix. I know it’s basic but I find that do the basics over and over you use less overall.
@@mixbustv You can argue that release isn't happening, but any limiting with immediate attack and release is practically clipping. That is the crucial difference. Of course clippers can have more complexity than that (and limiters can incorporate clipping) but release is what practically makes the difference.
Clippers are a way of manipulating and responding to resolve the issue of the confusion and google made it possible for you to see if someone is available for your support for your message and I will send you the address of that.
The term "chopping it off" actually should be used for clippers. Clipper is the one makes waveform that exceeds the threshold flat. Limiters is "compressing it down" while remaining the waveform as original as possible.
@@YungBear so if we wanr to mimic old school sound, it best to use clippers last instead of limiter?
OK, no one explains the difference correctly, so I'll do it here. Imagine we have a 44.1KHz, 16-bit PCM digital audio format for export. Each of the 44,100 samples per second (one stream for left and one for right) has a value between -32768 to 32767 (signed, twos-complement 16-bit binary). But, our DAW uses 32-bit or 64-bit floating-point internally, so there are values allowed outside this range (that's why you can go into the red on your individual and even master tracks all over the place but not distort). In the final PCM file, however, such values aren't allowed. What do we do with values outside this range (from -96db to 0db, 6db per bit)? A clipper simply takes anything less than -32768 and makes it -32768; it takes anything greater than 32767 and makes it 32767. That is, a clipper throws aways information. This has the effect of saturation or distortion depending on how far out of range the sample values are. A limiter, in contrast, scales out-of-range values by multiplying by a value less than 1.0. Now, how this clipping and limiting is actually done is where the differences in individual clippers and limiters comes in (the more clever they are, usually the better, but cleverness induces latency). Some, actually most if not all, limiters look ahead, some clippers look ahead, and act on groups of sample values. What about that over-sampling value you can set on some clippers and limiters? So, if we have four sample values: 32766, 32767, 32767, 32766, you can guess what the value between the two 32767 values would be if we created, with oversampling, some intermediate value(s): it would be greater than 32767. That's a true peak value: it's what the waveform would look like when it's reconstructed by the D/A conversion on output. So, even though there's no sample in an audio file that's greater than 0db (32767 or less than -32768), there can still be a peak above 0db after the signal is reconstructed for presentation to a speaker (or over-sampled). So, clippers throw away information by just chopping sample values; limiters try to preserve dynamics and harmonics by scaling sample values (a compressor scales by the ratio, for example). Scaling can preserve more of the original information, though it clearly cannot preserve all of it (unless you scale all samples, which is the same as turning down the volume). Also, the issue here is crest factor, which is the difference between the peak value and RMS value of your sound. If you can get a low crest factor, you have the potential to make your final, exported mix loud without too much distortion from either clipping or limiting. And, clipping and limiting serve to reduce crest factor. That's the reason you can use saturation on a, say, cymbal sound and reduce it's peak value while actually making it sound louder: saturation takes the energy of the transient and spreads it around with harmonics that don't change the sound much but reduce the transient/peak energy: this serves to reduce the crest factor of the sound. There's plenty more to say, but, I've said my peace.
(This is actually a comment I've found in another YT video and saved in my phone)
Based response.
A fun thing a lot of anti clipping people don't realize is that the artifacts caused by most limiters (even the most popular one) when not placed into long(ish) lookahead, are similar to the artifacts caused by clipping. The difference is that any halfway decent clipper is going to be built with this in mind. For example, I know airwindows essentially attempts to shape the offset of the clipping, I think OnlyClip2 from him uses slew rate limiting to achieve this. Limiters main downside is that compression is rarely 'transparent', we are much more sensitive to amplitude changes over time (ie the pumping effect) than we are to the 'noise' clipping adds when dealing with transients.
Great explanation, thanks !
I remember talking more than 20 years ago with one top notch jazz artist about how my synth guitars don't sound real enough, like the real guitar in the mix. And he said? Why do they have to sound real? They sound as they sound. It blew my mind. It was an "aha" moment right away ... The same with clippers, limiters etc., they do what they do, use their characteristics as they are, in your advantage.
I like your frankness and simplicity of explanation. Despite your accent (as the American I am) you are clearer understood than most.
Clippers act instantaneously (or at least very nearly if some Analogue modelling is done)
limiters hold the signal down longer, reducing harmonics from rapidly modulating waveform, but introducing pumping and more loss of perceived peak level.
In digital the limiter can start to reduce the signal before a peak happens (lookahead)
I'm really only just starting to use clipping, so it's good to hear these things clarified. 🙂👍
Hard clip everything.
Hi David. The only piece of hardware that I use when mixing is the Zulu Tape emulator. I print groups through the Zulu and blend it with the original take 50%. If I do that with a lot of tracks do I get phase problems? I work with Samplitude Pro X8 and there is a plugin that calculates the latency from hardware units. Thanks....and your new videos look great. 🤘 Michael
Yes you will. First off I'm not a fan of printing in general, that's really not how analog or hybrid mixing works, but besides that, why blend? You're definitely asking for trouble doing that for multiple tracks
@@mixbustv thanks David. That helped a lot
Distortion and saturation are not the same thing.
The terms are used interchangeably because often non-linear signal processing results in both.
Distortion usually leads to saturation, and saturation leads to distortion.
Which is exactly what I said.
@@mixbustv Not really, you said those are the same thing, only used to describe different levels of subtlety.
But really saturation is spectral density, distortion is a process of deforming signal.
A hardware 500 series or rack clipper? Do they exist? Where do I buy one?
Better lighting and video quality! Finally! 🎉
You make limiting, clipping etc. the part of the song, as an effect and there's no problem. Every tool you use or you know you'll use (or you experiment and it's sounds good to you) is a part of artistic expression. Choping, limiting, doing crazy stuff, distortion etc. Of course you have to know first what the f... you're doing (or what some tool does) in the first place, but it really doesn't matter; you hear your messy mix or master at the end anyway, haha.
I think the main difference between clippers and limiters is that clippers only look at the current signal level, while limiters operate over time as they may have lookahead and release settings. In other words, clippers work instantly and limiters are compressors with infinite ratios.
No they are not. Limiters have zero attack. That's literally the defining features of brickwall limiters. They have one job and one job only, not let signal pass above a certatin threshold. Without zero/instant attack that they would NOT be limiters.
@@mixbustv They have zero attack, yes. I did not say they don't. I said they may have a lookahead, which means they may even have "negative" attack, meaning they will start processing before the treshold limit. Also they may have a release, which means they will still process the signal after the signal drops below the threshold. The point is: they work over time, while clippers work at the moment.
I discovered a few months back that my old dbx 166xl I bought 20 years ago has a soft clipper circuit named as a “ peak stop limiter “. So after reading this in the manual. I dusted it off and experimented with it to replace my pluggin clippers and ….. I found a use for it. lol
DBX's always had that. I used to use an old 165's limiter for mono crush drums in parallel and it was amazing
As far as I understood, in the digital realm clippers work on a sample-by-sample basis. This means there is (like a waveshaper) a clipping function (from basic -1 to 1 hard clipping to arctan or tanh or more complex stuff) where each incoming sample is mapped to (incoming sample is located on the X-axis whereas the output sample is taken from the calculated function value at that functional value from X). Each deviation of the transfer function (clipping function) from linear is going to introduce new harmonics in the output signal according to Fourier/Laplace theory. So the transient material doesnt get enhanced by simply pushing gain (or like the enhancement through the attack phase of a compressor/limiter) but by chopping of the wave there are more harmonics generated for the transient what makes it kinda "enhanced/louder". More harmonics mean more RMS, thats why saturators/clippers are essential to achieve modern loudness standards. A limiter on the other hand works basically like an compressor with infinite ratio. It relies on the envelope of a signal which is why there are some "release/attack" controls or lookahead features on some limiters. Based on envelope detection these things try to reduce gain in a more linear fashion in contrast to clippers.
Thank you
Any thoughts on the undertone audio pyra sum? multiple bus summing mixer, like a console.
It's a cool idea
UA-cam ads are responsible for me no longer watching contents.. do you have alternative channels?
Yes you can buy courses and become a member 👍
I always thought it was opposite, that clippers cut off the shape (and therefore distorts) and limiters retain the shape but lower the volume. Learned something again. 👍
You are correct.
You were correct. What he says in this vid is wrong, he mixed it up
Clipper - punchy crunchy
Limiter - Smoothy softy
Haha the shots fired clip 😂
Haha. The way you switched to the song at the end with those lyrics. Fire bro! Hope them copycats get the point. 😊
The sexy choking song? That's at the end of most of his vids...
Yoooo Can you do a video on K system and best levels and balance of a mix. I know it’s basic but I find that do the basics over and over you use less overall.
Both do the limitting, the diference is the fast release of clippers producing harmonics.
It's not the release
@@mixbustv You can argue that release isn't happening, but any limiting with immediate attack and release is practically clipping. That is the crucial difference.
Of course clippers can have more complexity than that (and limiters can incorporate clipping) but release is what practically makes the difference.
You are just really good in explaining concepts.
Thanks for the explanation. That was very helpful.
🔥
Clippers are a way of manipulating and responding to resolve the issue of the confusion and google made it possible for you to see if someone is available for your support for your message and I will send you the address of that.
Why do people think of the spectra sonics hardware from 1969 ... 90 nanoseconds
Sorry dude,
Saturation and Distortion are not the same thing
Sure "dude"
Day 42 of me asking to abandon my life in new orleans and come be your studio intern rat boy