----Messaggio originale----
Da: Anton Kholomiov <anton.kholomiov@GMAIL.COM>
Data: 25/03/2016 20.23
A: <CSOUND@LISTSERV.HEANET.IE>
Ogg: Re: [Csnd] Making hard limiter with compressor to prevent huge amplitudes

@vallste It's interesting way to do this. It can be optimized with ftable I think.

So that we don't need to calculate the taninv on each step. We can store the function and do lookup.

With this trick it's super lightweight algorithm to solve my task. 

As for compress vs soft clip. I've tried the compress and result is much better

than it was with clip. I don't hear the distortion when the sound is above the limit

I\ve set the arguments to:

~~~

aduck init 1

ithresh = 0

iloknee = 98

ihiknee = 98

iratio  = 100

iatt = 0

irel = 0

ilook = 0

aout1 compress aout, aduck, ithresh, iloknee, ihiknee, iratio, ilook

~~~

I still don't get the difference between iloknee and ihiknee.

Is the hiknee or loknee the same thing as threshold in this article:

http://www.homestudiocorner.com/what-is-knee-on-a-compressor/

But why do we need two of them?

Anton

2016-03-25 20:17 GMT+03:00 Justin Smith <noisesmith@gmail.com>:

as an aside about the need for this - I think this exposes a very bad design choice in operating systems.

If a user sets a system level volume control, the sensible thing is to set that level as the maximum level any application can produce. Instead, this is just used as a multiplication factor applied to the output of each application - so that there is no way for the user to control the actual max amplitude (unless their sound card has hardware controls).

This isn't just a poor choice in design, it's one that can cause actual physical injury.

On Fri, Mar 25, 2016, 09:31 Peter Burgess <pete.soundtechnician@gmail.com> wrote:

@vallste awesome! I might give that a try myself.

I'd not realised that clip did soft clipping. This all sounds like good advice!

If you are still interested in the compress opcode though, I'm not
sure about the amplitude converters, but it seemed that the dB range
needed for the knees was 0 -> 90 instead of -90 -> 0 (when working
with 0dbfs that is). For the hard limiter that you're after, you'd
want the knees the same, or close to the same, and both pretty much at
90.

As for the ratio, 100 sounds like it'll do the trick, that's really
harsh compression. As you suspected, I don't need such harsh
compression. Or do I? I've still got to do some experimentation
myself, lol ;)

On Fri, Mar 25, 2016 at 4:07 PM, vallste@libero.it
<00000008a49663bc-dmarc-request@listserv.heanet.ie> wrote:
> Hi,
> I normally use arctan limiter (I don't rememeber who suggested it on this
> list, I owe her/him a beer), I've encapsulated in an UDO:
> opcode AtanLimit, a, a
> ain xin
> aout = 2 * taninv(ain) / 3.1415927
> xout aout
> endop
> Simple and quite effective
>
>
>>----Messaggio originale----
>>Da: Oeyvind Brandtsegg <oyvind.brandtsegg@NTNU.NO>
>>Data: 25/03/2016 16.45
>>A: <CSOUND@LISTSERV.HEANET.IE>
>>Ogg: Re: [Csnd] Making hard limiter with compressor to prevent huge
> amplitudes
>>
>>Just my 2c, if you want a simple limiter, the clip opcode may work
>>reasonaby well. It does soft clipping, and the default (Bram de Jong)
>>method works nicely IIRC. I've used it a lot earlier, but not
>>recently.
>>I seem to recall I put the clipping start point at 0.75, instead of
>>the default 0.5.
>>
>>2016-03-25 12:40 GMT+01:00 Anton Kholomiov <anton.kholomiov@gmail.com>:
>>> Also in the manual the suggested value for ratio is 100 (if you want to
>>> implement the limiter). Do you have any thoughts on that? reason behind
> your
>>> being 10
>>> you don't want to have a lot of compression?
>>>
>>> 2016-03-25 14:04 GMT+03:00 Anton Kholomiov <anton.kholomiov@gmail.com>:
>>>>
>>>> @Pete Thanks for the solution! I think I'll integrate it into my lib.
>>>> I need to use `compress`. I'd like to not to use very recent opcodes for
>>>> default things.
>>>> In that case with `compress` do I need to use ampdbfs to convert your
>>>> values for knees?
>>>> or it have to be `ampdb`?
>>>>
>>>> 2016-03-24 23:32 GMT+03:00 Peter Burgess <pete.soundtechnician@gmail.com>:
>>>>>
>>>>> Hi Anton, I was also in the same boat recently, so I will share what I
>>>>> have learnt this far.
>>>>>
>>>>> compress or compress2 are probably your best bet. Apparently Dam was
>>>>> an experiement and never meant to be used. John recently added
>>>>> compress2 to Csound, which uses dB values that work with any 0dbfs.
>>>>> What you do is set the threshold of the output using the loknee hiknee
>>>>> values, and set ratio to determine how much the sound is squashes
>>>>> above those thresholds. 0 on either knee will represt 0dbfs, and -90
>>>>> represents near to silence.
>>>>>
>>>>> This is the compressor setting I've got going on the main output of
>>>>> algorythm radio at the moment:
>>>>>
>>>>> ithresh = -90
>>>>> iloknee = -4
>>>>> ihiknee = -2
>>>>> iratio = 10
>>>>> iatt = 0
>>>>> irel = 0
>>>>> ilook = 0
>>>>> ablank init 1
>>>>> acompL compress2 amixL, ablank, ithresh, iloknee, ihiknee, iratio,
>>>>> iatt, irel, ilook
>>>>> acompR compress2 amixR, ablank, ithresh, iloknee, ihiknee, iratio,
>>>>> iatt, irel, ilook
>>>>>
>>>>> I'm still not sure it's ideal, but it's working at the moment while I
>>>>> develop. If I have this right (and I still might not):
>>>>>
>>>>> my loknee + hiknee value mean I'm only starting to compress a little
>>>>> from -4dB up, and compressing fulling from -2dB up.
>>>>> ratio of 10 means all sounds above -2dB are being reduced by 10dB,
>>>>> which means by a magnitude of 10.
>>>>>
>>>>> I'm still not sure what threshold actually does, it might have
>>>>> something to do with the sidechain function, it doesn't seem to make
>>>>> much difference when I change it, so I'm just leaving it at -90.
>>>>> ablank is where a sidechain modulator would be if I was using one, so
>>>>> I just set this to 1 to represent a constant signal to keep the
>>>>> compressor active.
>>>>>
>>>>> Feel free to ask more questions if you have any, and hope this helps.
>>>>>
>>>>> Pete
>>>>>
>>>>> On Thu, Mar 24, 2016 at 7:30 PM, Dr. Richard Boulanger
>>>>> <rboulanger@berklee.edu> wrote:
>>>>> > Good luck working on this Anton and please share your results.
>>>>> >
>>>>> > This "feature" should be an optional setting in Csound (from the
>>>>> > command-line or a check-box in CsoundQt
>>>>> > And
>>>>> > It should, by default, always be set to "on".
>>>>> >
>>>>> > Dr.B.
>>>>> >
>>>>> > On Thu, Mar 24, 2016 at 3:26 PM, Anton Kholomiov
>>>>> > <anton.kholomiov@gmail.com>
>>>>> > wrote:
>>>>> >>
>>>>> >> I'm developing a haskell lib that uses Csound.
>>>>> >> It generates the Csd's. Maybe you have heard about it.
>>>>> >> I'd like to save the ears and speakers of my users from
>>>>> >> huge amplitude values. I'd like to build the limiter that's always on.
>>>>> >> So that the user doesn't need to know anything about it.
>>>>> >>
>>>>> >> I'm using the clip by 0dbfs value right now.
>>>>> >> But there is a desire to make the algorithm more musical.
>>>>> >> Clipping often produces unwanted distortion so I'd like to
>>>>> >> substitute it with a limiter. I know that it have to be implemented
>>>>> >> with compressor. I know there are dam and compress opcodes.
>>>>> >> But  unfortunately I don't understand the arguments of those opcodes.
>>>>> >>
>>>>> >> Can you help me to solve this task?
>>>>> >>
>>>>> >> How can I build the limiter that doesn't change the signal if it's
>>>>> >> below
>>>>> >> the 0dbfs
>>>>> >> and prevents the sound from becoming way too huge?
>>>>> >> Maybe it have to be the combo with limiter and clip after that.
>>>>> >>
>>>>> >> Thanks,
>>>>> >> Anton
>>>>> >>
>>>>> >>
>>>>> >>
>>>>> >>
>>>>> >> Csound mailing list Csound@listserv.heanet.ie
>>>>> >> https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND Send bugs reports to
>>>>> >> https://github.com/csound/csound/issues Discussions of bugs and
>>>>> >> features can
>>>>> >> be posted here
>>>>> >
>>>>> >
>>>>> >
>>>>> >
>>>>> > --
>>>>> > _____________________________________________
>>>>> > Dr. Richard Boulanger
>>>>> > Professor of Electronic Production and Design
>>>>> > Professional Writing and Music Technology Division
>>>>> > Berklee College of Music
>>>>> > ______________________________________________
>>>>> > President of Boulanger Labs - http://boulangerlabs.com
>>>>> > Author & Editor of The Csound Book -
>>>>> > http://mitpress.mit.edu/books/csound-book
>>>>> > Author & Editor of The Audio Programming Book -
>>>>> > http://mitpress.mit.edu/books/audio-programming-book
>>>>> > ______________________________________________
>>>>> > about: http://www.boulangerlabs.com/about/richardboulanger/
>>>>> > about:
>>>>> > http://www.csounds.com/community/developers/dr-richard-boulanger/
>>>>> > music:
>>>>> >
>>>>> > http://www.csounds.com/community/developers/dr-richard-boulanger/dr-
> richard-boulanger-music/
>>>>> > ______________________________________________
>>>>> > email: rboulanger@berklee.edu
>>>>> > facebook: https://www.facebook.com/richard.boulanger.58
>>>>> > Csound mailing list Csound@listserv.heanet.ie
>>>>> > https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND Send bugs reports to
>>>>> > https://github.com/csound/csound/issues Discussions of bugs and
>>>>> > features can
>>>>> > be posted here
>>>>>
>>>>>
>>>>>
>>>>> --
>>>>> http://algorythmradio.com
>>>>> https://soundcloud.com/algorythmradio
>>>>>
>>>>> Csound mailing list
>>>>> Csound@listserv.heanet.ie
>>>>> https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND
>>>>> Send bugs reports to
>>>>>         https://github.com/csound/csound/issues
>>>>> Discussions of bugs and features can be posted here
>>>>
>>>>
>>>
>>> Csound mailing list Csound@listserv.heanet.ie
>>> https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND Send bugs reports to
>>> https://github.com/csound/csound/issues Discussions of bugs and features
> can
>>> be posted here
>>
>>
>>
>>--
>>
>>Oeyvind Brandtsegg
>>Professor of Music Technology
>>NTNU
>>7491 Trondheim
>>Norway
>>Cell: +47 92 203 205
>>
>>http://www.partikkelaudio.com/
>>http://soundcloud.com/brandtsegg
>>http://flyndresang.no/
>>http://soundcloud.com/t-emp
>>
>>Csound mailing list
>>Csound@listserv.heanet.ie
>>https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND
>>Send bugs reports to
>>        https://github.com/csound/csound/issues
>>Discussions of bugs and features can be posted here
>>
>
> Csound mailing list
> Csound@listserv.heanet.ie
> https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND
> Send bugs reports to
>         https://github.com/csound/csound/issues
> Discussions of bugs and features can be posted here



--
http://algorythmradio.com
https://soundcloud.com/algorythmradio

Csound mailing list
Csound@listserv.heanet.ie
https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND
Send bugs reports to
        https://github.com/csound/csound/issues
Discussions of bugs and features can be posted here

Csound mailing list Csound@listserv.heanet.ie https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND Send bugs reports to https://github.com/csound/csound/issues Discussions of bugs and features can be posted here

Csound mailing list Csound@listserv.heanet.ie https://listserv.heanet.ie/cgi-bin/wa?A0=CSOUND Send bugs reports to https://github.com/csound/csound/issues Discussions of bugs and features can be posted here

That's sounds like a good idea to me, though it would probably be better if it was off by default cos otherwise it's toying with your sound without you knowing. Maybe it could be turned on and off with an option like sr and ksmps and that, or have a limit level set with an option like that, that way if you did have it on, it would be right there in your code