Pushing the limit..er

[scott2000]

I've heard this covered in several more recent threads in some form or another... just can't find them.....

@ Ruari......It would be interesting to hear what you or others listen out for when pushing into your limiter. Is it more the sub frequencies acting up? Certain instruments that stand out by acting funny? I can hear the obvious stuff and stop before , but notice when I turn things down, I've missed the mark by quite a bit and need to pull back the limiter more. I know it'll take a lot of listening .

I hope this makes sense. Just looking for ideas to feel more confident when using them more. More ...Like just under the most..lol

 

[Gold]

I listen for any distortion or pumping. Low end can do both distortion and pumping. Midrange and high end is usually more like distortion. If you are getting pumping in the midrange it's probably best to remix/remaster because something ain't right. Many of the artifacts end up most obvious in the side channel. Soloing the side channel can be enlightening.

Look and listen for things that are 'pokey'. Large peak to average ratios will make the limiter work too hard. All that needs to be taken care of before hitting the limiter. Hope that helps.

 

[ruairioflaherty]

Yep, what Paul said.

When a limiter isn't working it can present in a variety of ways and for a variety of reasons.

Pumping/punching holes and distortion are the big tells. Both will be more apparent if the music has sustained vocal/piano/strings. If the EQ balance is wrong on the track these uglies are more likely to present.

But there are lots of other ways things can get ugly - aliasing, release chatter on 808s, oversampling blurriness (can be not fun on acoustic music).

I avoid multi band limiters and have a strong like for FFT based limiting like Elevate etc.

 

[scott2000]

Many of the artifacts end up most obvious in the side channel. Soloing the side channel can be enlightening.

Gold...Thank you

Look and listen for things that are 'pokey'. Large peak to average ratios will make the limiter work too hard. All that needs to be taken care of before hitting the limiter. Hope that helps.

Yes I think this was brought up before. Makes sense. I guess, in some ways, running another
limiter for this before could be an option if not having other choices apart from maybe manually turning down some peaks?

Pumping/punching holes and distortion are the big tells. Both will be more apparent if the music has sustained vocal/piano/strings.

This isn't to say adding a track with any of the above along side would help reveal anything as it relates to where you should be? Obviously that would change things. Would be cool though...

But there are lots of other ways things can get ugly - aliasing, release chatter on 808s, oversampling blurriness

I've heard things like this and know it's not good. It's just weird that when I think I'm safe, after turning things down, I can hear I wasn't as safe as I thought. Which leaves me short of where I know I could be. Maybe it is the source material that needs to be mixed....
It's interesting too that. and I could be wrong, bringing it out of the box after limiting, at a reasonable level into a hardware piece, sounds not good/funky sometimes and it makes me think some of these are the artifacts of pushing the limiter too hard because if, without limiting, the levels being the same wouldn't sound off.. Not talking level being too hot coming out, it can be quite low, but sounds funky. I just don't understand why, when in the box, I can't hear it do this..
That might not make sense....and I haven't gone further into this apart from thinking this is what I was hearing. Would need to experiment.

Thank you very much!

I avoid multi band limiters and have a strong like for FFT based limiting like Elevate etc.

Thanks. I'm just starting to play with limiting so have a lot of listening to do.

As far as multiband, have you had a chance to check out the Leapwing one? Dyn something?
I embarrassingly don't have the processing power I would think unless I did some changing around but, when I was at Bob Katz playing around with Heikki's compressor and his other compressors, at the very end, he threw that plugin on and, if I'm honest, it sounded better to me than anything. .
But the project he auditioned Heikki's and the others on was one he had already been working on with that plug in afaik... could have had some other stuff going on.

 

[Gold]

It's interesting too that. and I could be wrong, bringing it out of the box after limiting, at a reasonable level into a hardware piece, sounds not good/funky sometimes and it makes me think some of these are the artifacts of pushing the limiter too hard because if, without limiting, the levels being the same wouldn't sound off..

You are right. It sounds terrible. I never do that.

Not talking level being too hot coming out, it can be quite low, but sounds funky. I just don't understand why, when in the box, I can't hear it do this..
That might not make sense....and I haven't gone further into this apart from thinking this is what I was hearing. Would need to experiment.

All correct. The pre limited waveform tends to get reconstructed. If you really want to see level go crazy try putting in an hpf or lpf.

he threw that plugin on and, if I'm honest, it sounded better to me than anything. .

If you were comparing apples to apples then fine. But it’s rarely apples to apples. In mastering you should really have a reason to do what you’re doing. First what, then why, then how. Like ‘the vocal is muddy’. Why change it? Because it’s clouding the instrumental arrangement. How to fix it? EQ, compression? Then you start turning knobs.

 

[scott2000]

You are right. It sounds terrible. I never do that.

All correct. The pre limited waveform tends to get reconstructed. If you really want to see level go crazy try putting in an hpf or lpf.

Thank you!

Would compressing things to mp3 be some kind of test to see if I've done something wrong that, for whatever reason, I can't hear until after I start doing unconventional things to the audio? I'm starting to confuse myself really.
Something tells me I need to practice listening and it will click eventually....

 

[kags]

Gold...Thank you


Yes I think this was brought up before. Makes sense. I guess, in some ways, running another
limiter for this before could be an option if not having other choices apart from maybe manually turning down some peaks?


This isn't to say adding a track with any of the above along side would help reveal anything as it relates to where you should be? Obviously that would change things. Would be cool though...


I've heard things like this and know it's not good. It's just weird that when I think I'm safe, after turning things down, I can hear I wasn't as safe as I thought. Which leaves me short of where I know I could be. Maybe it is the source material that needs to be mixed....
It's interesting too that. and I could be wrong, bringing it out of the box after limiting, at a reasonable level into a hardware piece, sounds not good/funky sometimes and it makes me think some of these are the artifacts of pushing the limiter too hard because if, without limiting, the levels being the same wouldn't sound off.. Not talking level being too hot coming out, it can be quite low, but sounds funky. I just don't understand why, when in the box, I can't hear it do this..
That might not make sense....and I haven't gone further into this apart from thinking this is what I was hearing. Would need to experiment.

Thank you very much!


Thanks. I'm just starting to play with limiting so have a lot of listening to do.

As far as multiband, have you had a chance to check out the Leapwing one? Dyn something?
I embarrassingly don't have the processing power I would think unless I did some changing around but, when I was at Bob Katz playing around with Heikki's compressor and his other compressors, at the very end, he threw that plugin on and, if I'm honest, it sounded better to me than anything. .
But the project he auditioned Heikki's and the others on was one he had already been working on with that plug in afaik... could have had some other stuff going on.

The Leapwing stuff is good, but not magic (like everything). It is a plug-in that really needs to be studied to use to good advantage. Like any multiband compbused for mastering, you can quickly get yourself into trouble. I try to only use 2 bands if I feel the need for multiband- usually to control either lows or mids without affecting the other range. I use multiband probably less than 5% of the time and mainly on bass-heavy electronic genres.

I also liken limiting to making toast. The right amount is great but even a little too much and you’ve got useless burnt toast. Learn to find that sweet spot, which is unique to every song. I also find that sometimes serial limiting (or clipping + limiting ) provides less artifacts than pushing one limiter past the sweet spot. That is useful for loudness-hungry clients.

 

[ruairioflaherty]

Contrary opinion incoming. The owner of Leapwing is a friend but I really disliked how it sounded.

Scott - zooming out. It sounds like some of your issues are just down to monitoring quality, if you're getting surprised after the fact by artifacts then it's likely that something is up with your reproduction or your playback levels (too quiet or too loud). Of course it could be that you are getting lost in overcomplicated processing too. In that spirit see below...

Assuming we're talking about fairly typical rock, pop and Hip Hop then this 3 part strategy covers a lot of ground

1) Make sure the EQ is as balanced as you can get it - most of your perceived loudness comes from the arrangement and mid range focus (2k to 8k). Too much low end will sound bad when you limit. Make sure the relationship between 80, 160 and 320 makes sense (think of them a little like fundamental, 1st and 2nd harmonic).

2) Consider shaving snare transients (and maybe kick transients) with a clipper. This spreads that peak energy harmonically across the spectrum, keeping energy, hiding in the noisey nature of snares and lets the limiter work less hard (I like SIR Standardclip - $25, great metering and responds intuitively). a dB or two of clipping will help the limiter work less.

3) Pro-L2 on the settings below is a great start point for clean and predictable limiting, raising input gain until you get the desired level

 

[scott2000]

It sounds like some of your issues are just down to monitoring quality, if you're getting surprised after the fact by artifacts then it's likely that something is up with your reproduction or your playback levels (too quiet or too loud). Of Cours eit could be that you are getting lost in overcomplicated processing too. In that spirit see below...

Thank you..
Yes.. All of the above for sure. I trust my room as a second opinion at most. Maybe good for checking low end.
Really can't see ever trusting my room the way I need to but am hopeful.
There was a time I didn't think I could trust my plug ins.....
Thank you for the settings. When I get back into town the first thing I'll do is upgrade my Pro-L1..lol
Told you I was just getting into understanding limiting

 

[gyraf]

..you ever tried the G21, Scott?

/Jakob E.
(who has commercial interests in said unit)

 

[JohnRoberts]

Back in the 70s/80s I wasted too much time developing dynamic processors, mostly companding NR, to improve the signal to noise ratio of analog delay lines and/or tape recordings. To help parse out audible artifacts I developed my own signal burst generator, not unlike a sine wave tone burst generator but with more controls and capable of gating full range audio input signals.

By coordinating the on/off gate with zero crossings I eliminated HF clicks and LF thumps (gating half a sine wave adds a DC component). By feeding some dry signal around the gate I was able to create a step function with larger crest factor than the original signal. With this tool I could make almost any dynamic path misbehave by stressing the weakest link (usually cheap gain elements).

JR

 

[scott2000]

.you ever tried the G21, Scott?

no I haven't..
Much for the same reason I don't listen on ATC SCM150s...lol

but I have heard it spoke highly of.... iirc Ruari has it ?

By coordinating the on/off gate with zero crossings I eliminated HF clicks and LF thumps (gating half a sine wave adds a DC component). By feeding some dry signal around the gate I was able to create a step function with larger crest factor than the original signal. With this tool I could make almost any dynamic path misbehave by stressing the weakest link (usually cheap gain elements).

I think I'm following this...
At least as far as limiting, I'm really trying hard to find it inferior to anything I've tried. Even some gain staging to get level isn't something I'm confident is doing anything a digital limiter can't.
Look ahead controls,attack and release...etc.... it's pretty impressive if not daunting.... and these are just what's on the surface to know about...

Probably going to have more space in my room in the future after coming to terms with some wrong moves over the last few years if being honest.

 

[Gold]

2) Consider shaving snare transients (and maybe kick transients) with a clipper.

I do this on many masters. I just hard clip it in the DAW before hitting the limiter. I've been using the TDL Limiter no6 for some things too.

 

[ruairioflaherty]

I do this on many masters. I just hard clip it in the DAW before hitting the limiter. I've been using the TDL Limiter no6 for some things too.

Indeed, it's just part off the game for modern levels!

 

[JohnRoberts]

I think I'm following this...
At least as far as limiting, I'm really trying hard to find it inferior to anything I've tried. Even some gain staging to get level isn't something I'm confident is doing anything a digital limiter can't.
Look ahead controls,attack and release...etc.... it's pretty impressive if not daunting.... and these are just what's on the surface to know about...

Probably going to have more space in my room in the future after coming to terms with some wrong moves over the last few years if being honest.

The unfortunate reality about dynamics processing is that all gain manipulations effectively generate cross products. The goal is to make these cross products sound inoffensive (musical?). Slow moving gain changes generate low frequency products that are generally hard to hear. Very fast moving gain changes can sneak past our ability to perceive them briefly, but large recurring gain changes require management.

As I already shared, back in the 70s I spent a lot of time parsing out audible artifacts from inferior gain elements of the time. Modern digital multipliers do not introduce any distortion of their own, but we still cannot escape the cross modulation products from gain changes multiplying the dry signal.

Different kinds of signals have different problems to manage. Brief very narrow transient overloads can sometimes just be clipped because the distortion produced by clipping a narrow transient is likewise very HF. Simple pure tones can be more difficult because gain manipulations can generate harmonic distortion. Gain manipulation of LF tones creates distortion products in the mid band where our hearing is most sensitive. In complex waveforms the signal envelope is dominated by LF content so not unlike modulating tones envelope modulation can be audible.

This is a gross oversimplification and there is much more to it. This is just touching on a few broad strokes.

JR

 

[scott2000]

but we still cannot escape the cross modulation products from gain changes multiplying the dry signal.

I understand this somewhat.
It would be nice to fully understand what is happening but, I'm ok not getting into the weeds if it can't be talked about more....

Stupid question, is this to mean older digital technology would struggle more? What about slower computers?
Is that a relevant question to what you're describing?
Fascinating....

 

[JohnRoberts]

I understand this somewhat.
It would be nice to fully understand what is happening but, I'm ok not getting into the weeds if it can't be talked about more....

Stupid question, is this to mean older digital technology would struggle more? What about slower computers?
Is that a relevant question to what you're describing?
Fascinating....

I recall the early days of digital compressors and they were pretty much science fair projects limited by early A/D-D/A. Now with sophisticated fully digital environments that is no longer a concern.

I guess the short answer is that there is no free lunch, even with ideal digital multipliers we can not escape the the gain change artifacts. That said modern technology can make just about anything we can imagine.

JR

 

[gyraf]

there is no free lunch, even with ideal digital multipliers we can not escape the the gain change artifacts

..actually, we can wait a bit, then apply changes at zero-cross only - thereby get ourselves a bite of free lunch. Not easy by any means, but possible, and NTP did this (in the 179-160 IIRC).

For the first half-wave over, we off course have to hard-clip - unless we run look-ahead with the associated delay problems. But if clipping is less than some 1-2mS, we tend to not hear it as clipped..

/Jakob E.

 

[JohnRoberts]

..actually, we can wait a bit, then apply changes at zero-cross only - thereby get ourselves a bite of free lunch. Not easy by any means, but possible, and NTP did this (in the 179-160 IIRC).

Non-realtime processing opens up multiple opportunities. This can be as simple as a look ahead delay now that cheap high quality delay exists for signals already inside the digital domain.

A classic old school example of non-realtime processing exists for recordings. By playing a given recording backwards the loud transient impulses do not emerge from dead silence but are preceded by a slowly growing tail (their former sound decay). Of course real time processing is limited by how much look ahead delay can be afforded. IIRC building in short look ahead delay was first popular in premium noise gates to allow closed gates to fully open up in time for transient signals. These early built in delays were not free or even cheap.

For the first half-wave over, we off course have to hard-clip - unless we run look-ahead with the associated delay problems. But if clipping is less than some 1-2mS, we tend to not hear it as clipped..

/Jakob E.

yup...

JR

 

[Script]

By playing a given recording backwards the loud transient impulses do not emerge from dead silence

Which is easy to do ITB, albeit not in (musically meaningful) real time.