Clip, then Limit? Or the opposite?

[bionic brown]

Quick question.

Yes, I did a search. However I did not find the answer I was looking for.

What's your common method of incorporating a clipper. Do you clip first then apply the limiter or the opposite?

[jayfrigo]

Originally Posted by bionic brown

Quick question.

Yes, I did a search. However I did not find the answer I was looking for.

What's your common method of incorporating a clipper. Do you clip first then apply the limiter or the opposite?

Different limiters sound OK with clipped material or awful, so that's one variable. Also, depending on where and how you want to clip, it may dictate the order. If you want to clip your A/D after the analog chain, that will be before the limiter. You never want to limit or clip before the processing chain. Also, since it's a good idea to keep a ceiling or -0.3 dB, a limiter at the end is a good way to do this, which means post clipping. I'd say the limiter at the end with just a touch would be the more common way, but by no means only way.

Of course, you know the caveat that clipping is using sledgehammer to drive a thumbtack, so be careful. You need a quality signal path and excellent monitoring environment to track the damage you are doing and try to keep it from being too obvious. Everybody seems to love talking about clipping these days, but it's not the thing you should default to for everything, and if you do use it, just do a little bit in conjunction with other techniques.

[loujudson]

Originally Posted by bionic brown

Quick question.

Yes, I did a search. However I did not find the answer I was looking for.

What's your common method of incorporating a clipper. Do you clip first then apply the limiter or the opposite?

WHAT are you talking about? Never heard of a clipper outside of a barber shop or trimming nails. Clipping in audio is bad distortion, and is to be avoided.

What do you mean by clipper?

[Kiwiburger]

Clipping is effectively extremely fast limiting, and often it's the best sounding limiting for certain things - e.g. drums.

Most forms of distortion induce clipping, and are worth trying instead of a limiter.

I generally use Voxengo Elephant as my limiter, and it has a clipping option, and depending on the material, clipping may or may not sound the best.

I would tend to use clipping last, because it's fairly extreme. Further compression after clipping would expose a lot of ugliness.

FWIW - the top mastering engineers in recent years have been pushing the loudness wars by clipping mixes with their high end A/D converters. Not digital algorithmns, but literally abusing the analog front end of their A/D for a clipping effect. That sort of confirms my belief that it's best left last in the chain.

But whatever sounds good, don't knock it. There is plenty of material where clipping must be avoided at all cost.

[loujudson]

"But whatever sounds good, don't knock it. There is plenty of material where clipping must be avoided at all cost.
__________________
I refuse to let common sense cloud my judgement."

These two statements go together well. I believe in correct use of the language, and to me, clipping is severe distortion, and should be avoided.

No wonder I can't listen to rock and radio any more! I do music with acoustic musicians and clean sound. Clipping is like wrecking the car, it is NOT music.

Have fun, and I'll never hear your "clipping." I just won't listen to it if that is the "new" standard... I DO knock it. And don't buy it. I'm just too old to be abused. Give me good old music, not new distortion!
;-)
L

[MAzevedo]

The thing is, limiting or compressing create distortions too, just different ones than clipping. In many cases, 1dB of clipping and 2dB of limiting create less audible distortion than 3dB of limiting. Especially on percussive signals, the harmonic distortion created by clipping is less audible than the smearing created by the attack and release of a limiter. There are a lot of people that have decided clipping is the ultimate answer to loudness and put out terrible sounding records. However, clipping can be an important tool in an overall compression scheme, and often sounds better than 'gentler' methods under the right circumstances.

Regarding the original question: Jay hit it on the head. Just given that your analog loop is almost certainly ahead of your brickwall limiter, clipping an AD almost always happens before digital limiting.

[bionic brown]

Originally Posted by loujudson

"But whatever sounds good, don't knock it. There is plenty of material where clipping must be avoided at all cost.
__________________
I refuse to let common sense cloud my judgement."

These two statements go together well. I believe in correct use of the language, and to me, clipping is severe distortion, and should be avoided.

No wonder I can't listen to rock and radio any more! I do music with acoustic musicians and clean sound. Clipping is like wrecking the car, it is NOT music.

Have fun, and I'll never hear your "clipping." I just won't listen to it if that is the "new" standard... I DO knock it. And don't buy it. I'm just too old to be abused. Give me good old music, not new distortion!
;-)
L

I don't see how clipping is severe distortion. You're removing information from the file so that it won't be distorted.

Clipping is the only thing I've found that will remotely give me the sound of the original mix at a much higher volume. Limiting seems to always tamper with the transients. If you know of a better way, I humbly ask that you show me. Yes, the loudness wars are part of the issue. If it makes you feel any better, i'm shooting for an average RMS level of about -14db, which is not as high as the average track these days. However, I want my tracks to not "sound" compromised by boosted volume as much as possible. And solely using a clipper, rather than a brickwall limiter has given me the best result so far.

I hate to make this thread yet another "the secret is out" discussion about clipping. And I'm not here to debate limiting vs. clipping.

I just want to know how certain gearslutz approach the usage of clipping and limiting, together.

[mworks]

IMHO clipping should never be a 'strategy'....maybe, only perhaps acceptable as an occasional event that takes place when you have the perfect gain staging set for a track. If you are clipping a track simply to get 1 db of added gain you're giving up way more than you're getting. If you're clipping cause you like the way it sounds then you're choosing your own poison.

Where you clip is most certainly determined by the application. If you clip for the sound of it you may want to NOT limit post-clip to avoid the smoothing effect a limiter can have so you might capture the clipping and then drop the level by some .3 or so.

Anyway, these are not the decisions I want to be making daily.

[jayfrigo]

Originally Posted by bionic brown

I don't see how clipping is severe distortion. You're removing information from the file so that it won't be distorted.

Sine wave is fundamental frequency and no harmonics. Square wave is a fundamental and all the odd harmonics - a bunch more stuff. Clipping takes your wave and starts turning it into a square wave, and where those sharp edges appear, you are adding a bunch of harmonics that weren't there in the original signal. This is adding info, not taking it away, and it is most certainly distortion.

Clipping is the only thing I've found that will remotely give me the sound of the original mix at a much higher volume. Limiting seems to always tamper with the transients. If you know of a better way, I humbly ask that you show me.

I (and most everybody else) don't like clipping and the loudness wars, but I understand that the war exists and some clients want it loud regardless of the negative consequences. So, in that case, clipping can indeed be part of the puzzle and perhaps a necessary evil.

As for ideas of how to do it better, that you can find in a search, but to give you a quick starting place, spread out your work among several stages so that each one's artifact will not dominate the landscape; therefore you will be able to get more gain before obtrusive side effects. Of course some times two or three techniques work better than all techniques together, so experiementation and good choices based on experience are key.

For an example, you may want to compres a bit, then a slight clip of A/D in analog, then a slight digital clip in an oversampled EQ (2X to avoid egregious amounts of in-band aliasing, and so that some of the junk gets filtered when it comes back to 1X rate Fs), and then two limiters of different manufacture in series, each doing a bit. If you get 1 dB from each of these places, it will likely sound a lot better than trying to get 5 dB all from one tool.

It's odd to suggest techniques on how to make music sound worse, but not quite as "worse" as it might have been otherwise... The world seems to have gone crazy, but there it is. A recipe for loud with a minimum of sour. Adjust to taste, rinse, repeat, and enjoy!

[lucey]

Originally Posted by bionic brown

I just want to know how certain gearslutz approach the usage of clipping and limiting, together.

No offense but whatever sounds best to you, in your chain, for any given track ... that's all that matters.

I clip then limit. And sometimes for extreme level: limit (analog), clip, limit.

[jamsmith]

Originally Posted by Kiwiburger

Clipping is effectively extremely fast limiting.

Damn, you walk away from audio for a decade or so the definitions change. Clipping always meant literally clipping the top of the wave form off. You would do this by using a diode to shunt the signal to ground above a certain threshold. Clipping was call hard limiting and used in radio you so you didn't inadverently overmodulate you signal - a violation of FCC regulations. The only strictly audio application I ever heard of was in distortion boxes. I have no earthly idea why you would want to in any matter clip your mix. But if that is what you, buy a couple of diodes and an op amp!

[Kiwiburger]

I suppose digital has made clipping easier to use/abuse.

[loujudson]

Originally Posted by Kiwiburger

I suppose digital has made clipping easier to use/abuse.

Yes, but all you need to do to get analog clipping is turn it up too far. I mean TOO far. I just don';t gte how this kind of defect in engineering technique can be useful - it is to be avoided entirely, for good sound. But maybe that is not everyone's goal...

I do everything I can to avoid distortion in waveforms, but then guitarists actually buy things to distort their sound.

Whatever blows your skirt up, I guess. Or floats your boat. Whatever.

L

[jamsmith]

Originally Posted by loujudson

I do everything I can to avoid distortion in waveforms, but then guitarists actually buy things to distort their sound.

I love extreme clipping on my guitar signal. I have even been known to distort my vocals (put my MXR distortion plus in the fx loop on the board). But I alway prided myself on my -3db peaks on meters of my digital recording systems. But I have to admit, I hit my DAT real hard one day just to see what it sounded like - wasn't near as bad as everyone made it out to be! (But is was a Sony consumer deck and may well have had an analog clipping circuit just to prevent a digital overdrive.)

[Nordenstam]

Originally Posted by bionic brown

If it makes you feel any better, i'm shooting for an average RMS level of about -14db, which is not as high as the average track these days.

-14 is so natural (non squished) that you can probably mix to that level!

Try some limiting/clipping/whatever on individual (offensive) tracks instead. You can adress the most dynamic sounds alone, without affecting all the mix at once. It will probably sound tons better than clipping and limiting everything. Both of these distortion effects usually sounds way much better on individual drum sounds(and such) than when applied on the whole mix.


Andreas N

[radiated]

If music is clipped or squshed I always turn if off as my ears get tired after 30 seconds. Even -14db is madly loud when you have a massive 96 db availible. Even LPs with much less available headroom sound perfectly good even on untouched symphonic orchestra recordings.
Why not keep it at -20 or more? A high end stereo sound no better that a boomblaster with that crap you are putting out.

[colonel_claypoo]

necrothread but didn't want to start a new one.

i'm having problems thinking through an answer for the following question:

how is it possible to use both a limiter and clipping in the same chain?

my answer currently is that it isn't.
how can you clip a signal that will be limited by the limiter in the first place? the signal will never exceed the limiter to reach the clipper.

is that correct?

please help me understand.

thanks

[Waltz Mastering]

Originally Posted by colonel_claypoo

i'm having problems thinking through an answer for the following question:

how is it possible to use both a limiter and clipping in the same chain?

my answer currently is that it isn't.
how can you clip a signal that will be limited by the limiter in the first place? the signal will never exceed the limiter to reach the clipper.

is that correct?

please help me understand.

thanks

I think most ppl who employ both clipping and limiting place the limiter last in the chain, so you would just adjust the threshold on the limiter to engage gain reduction after detecting the clipped signal. A limiter's function would still be the same even if the peaks from the incoming signal were diminished by clipping.

[(un)reason]

Originally Posted by colonel_claypoo

necrothread but didn't want to start a new one.

i'm having problems thinking through an answer for the following question:

how is it possible to use both a limiter and clipping in the same chain?

my answer currently is that it isn't.
how can you clip a signal that will be limited by the limiter in the first place? the signal will never exceed the limiter to reach the clipper.

is that correct?

please help me understand.

thanks

More makeup gain between each stage. To get to that brutal -7 RMS that appears in far too many pop productions, it is not unknown for people to use multiple layers of compressing, limiting, clipping, eqing, and repeat, each device shaving a db or two off the dynamic range until the track is just a solid block of black.

[engmix]

Originally Posted by loujudson

Yes, but all you need to do to get analog clipping is turn it up too far. I mean TOO far. I just don';t gte how this kind of defect in engineering technique can be useful - it is to be avoided entirely, for good sound. But maybe that is not everyone's goal...

I do everything I can to avoid distortion in waveforms, but then guitarists actually buy things to distort their sound.

Whatever blows your skirt up, I guess. Or floats your boat. Whatever.

L

I don't think that the op is talking about clipping and limiting the new Joan Baez re-mastering of her greatest hits....which in all seriousness would be terrible sounding. To say something should never be done is really not what the op is asking as well. I'm sure your approach works for you, but having to work within the madness of the LW, one has to figure out as many ways to get the job done in the most effective way possible. I certainly know if my clients want a certain sounding mix, and i don't give them that, they will go else where to get it, and the potential return business will be history.

If my clients want their steak well done i'm not going to convince them that they will like it better rare.

Personally i employ a host of different boosts of level throughout my signal chain so not one thing is picking up most of the slack. But sometimes that doesn't work either. In that case i start trying different approaches.

[Cheebs Goat]

Originally Posted by bionic brown

If it makes you feel any better, i'm shooting for an average RMS level of about -14db, which is not as high as the average track these days. However, I want my tracks to not "sound" compromised by boosted volume as much as possible.

Hold on here. You could probably hit an RMS of -14 without a limiter or a clipper. What exactly are you doing where you need to use both to hit that level?

[Lagerfeldt]

There are cases where clipping or limiting earlier in the processing chain makes sense. The effect may be partly negated at the reconstruction stage or during any subsequent processing, but whether the artifacts are worse than the benefits depend on the situation.

If you have a mix with one extremely loud peak, e.g. +6 to +12 dB above average peak level, you're usually better off taking care of that in advance. If not, you risk that the rest of the processing chain will react violently to that transient. Another option is to use a restoration algorithm or destructively attenuate or redraw the waveform at the offending peak.

Regarding clipping into limiting vs. limiting into clipping, that too depends entirely on the situation, material and limiter. In many cases I find that the latter sounds better.

Entirely substituting limiting for clipping can sound better in some cases, but there's always a price to pay. Especially since you're likely to cause uncontrollable overloads at the D/A stage or confuse the phase rotator in the broadcasting processing chain. I very rarely use clipping this particular way, though any amount of clipping can cause trouble, I'd like to add.

So I generally avoid clipping if I can, but if the client demands extreme loudness, and distortion masking is possible, there's no way around it. Just another tool in the box.

When masking is not possible at certain points in the mix you can sometimes splice the clipped version with an additionally limited version (another pass with volume matched limiting) in places where the clipping distortion is most obvious. If the extra limiting causes audible pumping artifacts you'll either have to live with it or bring the overall level down again.

[Eli Resnick]

To clarify a bit here, let's define "distortion." It comes form latin roots meaning "not" and "shape," with an ending that makes it a noun. It means changing the shape of the existing sound waves. If it were adding new content it would be something else--most likely "overdubbing," "harmonizing" or both.

Of course, harsh, clipping distortion makes such a difference to the way that our ears hear the sound that it sounds like something else is going on. And, of course, without pefect speaker damping, a couple other sounds and rattles are likely to originate at the point where the speaker is moving at top velocity and then the signal just suddenly stops changing. But that doesn't mean anything amazing and unusual is happening in one particular clipping method that doesn't happen in another.

Any $20 stompbox guitar distortion works by using diodes to clip the tops off of overamplified sine waves (that have been amplified through, as mentioned above, an op-amp) and make them into square waves.

If someone charges a lot because they pull off the same effect using a limiter or the input buffer of any type of equipment, whether it's a preamp, an ADC, an EQ or a toaster that's been wired for balanced audio, they're still doing the same thing to the sound wave form. They're turning it up and clipping off the top to make it resemble a square wave.

That is, technically, a very fast form of limiting. But it's still something you can do for $20 by turning up a distortion pedal, and it's not traditionally what you pay a top mastering engineer to do to your music. It's actually quite the opposite. Usually you hire a mastering engineer on the basis of who has the knowledge, skills, equipment and loving patience to get your track loud enough that it's all audible without clipping anything. That's where the war is.

But it's certainly a lot easier to just crank everything up until it's totally clipped and tell you that you win. If that's what you're going for, I've got a ton of great clippers on my pedalboard. Heck, I'll even run each side of your track independently through three consecutive "clipping boxes" into my Ampeg half-stack with the high and low EQs dimed to finish the mastering chain with some frequency dependent stereo clipping. For a little extra, I'll mic the cab low over a tile floor to give you some killer natural reverb. Just PM me with a link to this thread for a substantial discount off my already competitive mastering rates.

[Th3_uN1Qu3]

Originally Posted by Eli Resnick

But it's certainly a lot easier to just crank everything up until it's totally clipped and tell you that you win. If that's what you're going for, I've got a ton of great clippers on my pedalboard. Heck, I'll even run each side of your track independently through three consecutive "clipping boxes" into my Ampeg half-stack with the high and low EQs dimed to finish the mastering chain with some frequency dependent stereo clipping. For a little extra, I'll mic the cab low over a tile floor to give you some killer natural reverb. Just PM me with a link to this thread for a substantial discount off my already competitive mastering rates.

Pure win. :D

Originally Posted by Lagerfeldt

If you have a mix with one extremely loud peak, e.g. +6 to +12 dB above average peak level, you're usually better off taking care of that in advance. If not, you risk that the rest of the processing chain will react violently to that transient.

You are right. However, most people assume that what applies for ONE loud peak in the whole song is also valid for 100 peaks. tutt Hence the destruction.

Or you could buy a FM radio processor and call it a day. AGC, multiband compression, multiband limiting, multiband clipping, final clipper, final limiter, and even composite clipping of the broadcast signal on top of that for that lovely stereo intermodulation.

These threads will never end but here's my take on it again - yes, to some extent, clipping sounds better than limiting. However, peoples' stereos don't generally have limiters! So you should leave to the listener's extent if he wants to turn the volume up on your song till it clips. If you clip it before, i bet you the customer won't turn it as loud. Clipping is acceptable when you have peaks to clip in the first place.

[Lagerfeldt]

I'm sure we can agree on the fact that people have a volume control, but the fact is that people listen to shuffled music on iPods or from a library on a computer.

[wado1942]

I only glanced through the thread, but here's what I do.

My clients know up front that I'm fairly conservative with loudness maximization. Most of the time, I'll just use some limiting (compression, EQ to taste of course). If that doesn't quite do it, I'll add a SOFT clipper in front of the limiter to take down the absolute highest peaks. It doesn't shear off the tops of the waves, just gradually smooths them over 2-3dB. This is done very carefully, and always oversampling by either 2X or 4X. I do not ever use hard clipping, unless it's purely for a special effect.

BTW, Elephant, which is a GREAT limiter, doesn't seem to gain any benefit from having a clipper in front of it. You can fine tune how much distortion vs level to tolerate in such great detail, it somewhat reduces the need for other destructive tools.

[Cheebs Goat]

Originally Posted by Lagerfeldt

I'm sure we can agree on the fact that people have a volume control, but the fact is that people listen to shuffled music on iPods or from a library on a computer.

But the thing people hate is "inconsistent". Nobody hates "proper head room".

Look at television. In that case, the majority of the sound is reasonable programming and the minority is loud commercials. Nobody says "finally, the commercials are on and I can have good sound". They get pissed about the volume difference and mute it.

If we made "sane" the majority, nobody would have a psycho acoustic edge from using today's volume. They would just be annoying.

[jeff_free69]

Originally Posted by wado1942

BTW, Elephant, which is a GREAT limiter, doesn't seem to gain any benefit from having a clipper in front of it. You can fine tune how much distortion vs level to tolerate in such great detail, it somewhat reduces the need for other destructive tools.

Based on high praise in this forum I just started playing with the Elephant. But I have a question about the Clipper algo - WHen set as the last plug with just 1-2 db GR, 4X oversampling and output GR = -.2, Cubase 5 still shows clipping on its bus output. Adding something else after like UAD Prec Lim or T-Racks Brick Lim takes care of it, as usual,

They call it a "clean clipper" , which seems a bit of a contradiction. Any thoughts on what this actually does compared about the other limiter algorthyms ?

[Herbeck]

Voxengo plugins has a bypass knob, however the oversampling is still engaged.
So it's quite simple to hear or to do a null test to see how much mess the oversampling does.


Cheers,

Herbeck

[Lagerfeldt]

Originally Posted by Cheebs Goat

But the thing people hate is "inconsistent". Nobody hates "proper head room".

Indeed.

If we made "sane" the majority, nobody would have a psycho acoustic edge from using today's volume. They would just be annoying.

Too late, it seems.