[LAU] Limiters?

Hartmut Noack zettberlin at linuxuse.de
Sun Jul 10 22:20:16 UTC 2011


Am 09.07.2011 23:00, schrieb Fons Adriaensen:
> On Sat, Jul 09, 2011 at 06:17:55PM +0200, Gabbe Nord wrote:
>
>> I'm on a ardour 2.8.11-setup and I'm looking for the best possible limiter.
>> Rest of my plugin-use is linuxdsp and some calf, so I need a good limiter to
>> be able to crank it all up a notch compressionwise. I'm currently using TAP
>> Scalinglimiter, which is the best one I've found yet that dont give me
>> zippernoise etc, but I can't push that limiter as far as I want without
>> artifacts in the sound etc.
>>
>> Do you guys have some tips? I'd be most grateful!
>
> Every time I read a post like this (no offense intented to the
> poster), there is this desire creeping up my back to write a
> decent peak limiter, or just release the things I already have.
>
> What stops me is the simple fact that by doing that I'd be
> contributing to the IMHO completely misguided and even stupid
> fashion of increasing the apparent loudness of recordings by
> any means, at the expense of sound quality.

There is no such thing in the 21st century. The failure of all the 
enhanced super-CD efforts has proved that.

There is only a set of methods to make music distributable as a file or 
as a physical record. And to do this the HiFi/Audiophile way is only but 
one of these methods.

In the early 1960ies guitarists like Pete Townshend or John Lennon 
started to crank up the volume on their amps and thus introduced 
distortion. And the engineers startet to scratch their heads to find 
ways to make such bad distortion disappeare. When Jimi Hendrix sent his 
tapes of Electric Ladyland to the EMI-Masterlabs, the engineers took 
great efforts to eliminate the phasing-effects Hendrix had applied to 
some of his tracks and thus Hendrix was shocked, whe he heared the first 
copy of the record.

And this is all that matters: if the artist is unhappy with the result, 
then the recording and mastering has failed - period -

If a band *wants* to have their songs brick walled so that the 
wave-graphs look like toothpaste: let them have it! This is a perfectly 
sensible method of expression. The simple reason is: it is different. If 
an artist is not satisfied with a work, then it simply means, he/she 
hears a difference between the results and the vision in his/her head. 
To eliminate such differences is the job of the recording-engineer and 
he/she should be ready to eliminate by all means neccessary ;-)


Of course brick wall limiters are often used for the wrong reasons. To 
make it simpler for a bad mixing-technician to make it sound something 
like "loud and clear". Or to make sure, the track does not sound lower 
than the others on a stupid format-radio-programme.

But there is fantastic, great music out there that looks like 
tooth-paste and is spiced with tons of distortion.

So in the end: a hard limiter is but a tool -- it represents an 
opportunity for artistst to get certain effects to make their music 
sound different the way they like to. And such opportunities should be 
available.

I use the foo-limiter to handle peaks and some vocals, it works for me. 
If I really want to smash a mix to the wall, I try to do it as Fons 
recommends by balance the mix to make it sound louder. And than I use 
Jamin to go a step beyond...


btw: Yes I admit it! I LOVE the sound of the late-90ies/2000s Red Hot 
Chili Peppers. I do not care for technical details like distortion, I 
only hear great, powerful, lively pop-music.

That some bands not the same as great, powerful and lively believe, they 
could be the same as cool by applying the same loudness is another story 
though...

best regards

HZN


> Simple fact is this:
> if it isn't loud enough, turn up the volume. The result will be
> vastly superior to what you can achieve by squeezing dynamics
> to death.
>
> Regarding ScalingLimiter, I wonder how many peope are actually
> aware of what it is doing. Which is to measure the peak level
> of segments delimited by zero crossings and then apply a
> constant gain factor to each segment to adjust its peak level
> to close to the maximum. The idea seems to be that changing gain
> at a zero crossing doesn't introduce distortion. Which is wrong,
> it does generate gross amounts of intermodulation distortion,
> just having less HF energy than when switching gain at random
> points. This makes a complete joke of whatever follows in the
> reproduction chain - you could as well use the worst amplifier
> (in terms of IM distortion) you can find and things wouldn't
> sound any different.
>
> Ciao,
>



More information about the Linux-audio-user mailing list