Get an open source clipper plugin here...

[Download SOphist]

Oba! thanks again Christian!

now, let me try to figure your explanation out...

cheers!

[shogger]

Thanx Christian,

sounds cool. Great clipper plugin.

Shogger

[everlast]

Wow, it's nice to find that someone thinks in the same way as you. For some time I use Cyanide 2 as my master clipper, and I use it in two passes - I reverse file for the second time - to compensate for group delay. I find that this way it is much better than simply rasing gain and much closer to clipping of ADC.

Is there a chance that you make FIR linear phase oversampling stage?

Can you explain a little bit more why do you reverse file and do you clip reversed file?

[JCM]

The two stages are necessary, because depending on the filter order, still some clipping can occur in the first stage.

The diagram is:

Input * InputGain > Upsampling Stage 1 (with the given filter order and the hidden transition bandwidth) > hard clip between +1/-1 > Downsampling Stage 1 (again with the given filter order and the hidden transition bandwidth) > Upsampling Stage 2 > hard clip between +1/-1 > Downsampling Stage 2 > Output Gain

If the oversampling factor is only 1 and the filter order 0, only a classic hardclip takes place (with all the aliasing).

I was also thinking about adding a second input gain, so that you can drive the second stage even more into clipping, but I discarded that idea. Anyway, the plugin is open source, so whoever likes to add more features is welcomed to do so.

Christian

Thanks for the explanation , now I undestand it better.

Having used it for a couple of days I can see how having more gain available could be useful , but would rather not lose the gain control fine tuning precision it has now.

The second input gain would have been after upsampling?

[Rottweiler]

is it more gentle than pure digital clipping?

or that is hard clipping too?

[Christian Budde]

is it more gentle than pure digital clipping?
or that is hard clipping too?

pure digital clipping = hard clipping
This is actually used, but on the oversampled material (if oversampling is activated). The downsampling however might lead to overs as well, that's why you have the second stage available to clip those as well.
In comparison to pure clipping (without oversampling) this will lead to a slightly more gentle limiting, but significantly reduce the alias if oversampling is carried out. Anyway you should always know a bit about this topic in order to use this tool properly.

Christian

[Rottweiler]

is it more gentle than pure digital clipping?
or that is hard clipping too?

pure digital clipping = hard clipping
This is actually used, but on the oversampled material (if oversampling is activated). The downsampling however might lead to overs as well, that's why you have the second stage available to clip those as well.
In comparison to pure clipping (without oversampling) this will lead to a slightly more gentle limiting, but significantly reduce the alias if oversampling is carried out. Anyway you should always know a bit about this topic in order to use this tool properly.

Christian

yea, i think i will use this then, i always liked digital clippin cause i can get some severe volume increase with a high transparency.

off-topic: or i should use some soft clipping stuffo to get a more musical sounding tone?

[michat]

Great!!!
Thanks Christian.
Boy, 2008 is a great year for new and really good freeware!!!

[meloco_go]

Can you explain a little bit more why do you reverse file and do you clip reversed file?

By applying non-linear phase filter to the reversed file again you compensate group dealy from the first time you've applied the filter. That effectively makes filter linear-phase and twice the order. I usually do same setting on revrsed processing, so if it clips first time it clips second time.
That trick - processing reversed file with the same filter again, is quit neat, you can make any EQ behave like linear phase that way.

[Christian Budde]

Can you explain a little bit more why do you reverse file and do you clip reversed file?

By applying non-linear phase filter to the reversed file again you compensate group dealy from the first time you've applied the filter. That effectively makes filter linear-phase and twice the order. I usually do same setting on revrsed processing, so if it clips first time it clips second time.
That trick - processing reversed file with the same filter again, is quit neat, you can make any EQ behave like linear phase that way.

We did the same trick in one mode of our Electri-Q. You can then benefit from all the filters and simply make all IIR-Filters you can find there linear phase (at the expense of a higher order and latency).

Christian

[everlast]

Can you explain a little bit more why do you reverse file and do you clip reversed file?

By applying non-linear phase filter to the reversed file again you compensate group dealy from the first time you've applied the filter. That effectively makes filter linear-phase and twice the order. I usually do same setting on revrsed processing, so if it clips first time it clips second time.
That trick - processing reversed file with the same filter again, is quit neat, you can make any EQ behave like linear phase that way.

Thanx for explaining that - very useful info there

[daxliniere]

Hi Christian,
Nice work, mate, just tried it out during on some test mastering.

The output, even when set to <0db seems to clip, isn't the whole point of this plugin to hardclip to the specified output level? My settings were: input +2dB, O/S(1) 16x, Order(1) 6, O/S(2) 16x, Order(1) 6, output -0.2dB

Seems to sound very nice, other than the digital overs.

Also, I would really love to see a linear-phase version of this plugin. Would you consider making it? (I know it's open source, but unfortunately I know NOTHING about programming in any language).

Great stuff, mate!
Keep up the good work,
Dax.

[Christian Budde]

Hi Christian,
Nice work, mate, just tried it out during on some test mastering.

The output, even when set to <0db seems to clip, isn't the whole point of this plugin to hardclip to the specified output level? My settings were: input +2dB, O/S(1) 16x, Order(1) 6, O/S(2) 16x, Order(1) 6, output -0.2dB

If you choose a higher filter order (6 in your case), the oversampler may clip itself. It also doesn't make sense to have identical settings on both stages. The initial idea was to have the first stage running fairly oversampled and probably with a high order (e.g. 8x oversampling, filter order 4) and apply a real hard clip (no oversampling, filter order 0) in the second stage. With this you will not get any more clipings.

I just added an additional parameter to the plugin that reads 'hard clip output' and performs an additional and not oversampled hard clip right before the output volume.

It can be downloaded from the same place.

Christian

PS: If you have access to the hidden parameters, note that you should not set the transition bandwidth too high, because this will probably result in more overshoots and intersample clippings.

[Christian Budde]

Also, I would really love to see a linear-phase version of this plugin. Would you consider making it? (I know it's open source, but unfortunately I know NOTHING about programming in any language).

Forgot to answer this question: Right now the open source project does not feature any FFT. Although it is possible to create and calculate linear phase filters in the time domain, it is better done in the frequency domain (for what an FFT is necessary).
Alternatively I could offer the use of bessel filters. They have a nearly linear phase transition band. So you would have something like a linear phase behaviour up to 20kHz or so. The only problem is, that they are a bit unstable, but as long as you would keep the filter order low...

Christian

[lazerkind]

Big thanks for this one Christian!
Your Noname Eq is used quite often here so it's about time I show some gratitude.

//L