Unfiltered Audio Zip: Creative Compression with Modulation

[masterhiggins]

Thanks for the bug report! We'll look into that.

For the analysis modes, we're using various Music Information Retrieval algorithms (https://en.wikipedia.org/wiki/Music_inf ... _retrieval) to match the audio up against various descriptors.

Hi there. Sorry to derail the thread, but do you plan to incorporate the modulation system into g8? That'd be wicked awesome.

Yep! We have to finish the G8 1.5 patch first. Adding the modulation system will need a ground-up rewrite, but it's on our agenda for next year.

THANK YOU!

In addition to creating out-of-this-world plugins, Unfilted Audio also fill the gaps that other companies totally ignore or pretend aren't there. For example, a certain company releases plugins with the names "Pro-C/G/L" and completely miss the opportunity to incorporate their modulation system with them (for creative purposes). Big miss on their part.

[sqigls]

it's an amazingly creative and versatile plugin (as always).
but i feel the CPU tax will prevent me from buying this compressor.

Yeah CPU is pretty light here, you're on a mac right? Have you tried it without
opengl? Typically opengl is faster and more efficient, "unless" your system
has trouble with it.

-Cheers

is that an option in the menu? I'll have a look.
but compared to other compressors, if i just need some bread'n'butter compression, Zip uses 2-3 times more juice.

[edit] no, doesn't make any difference. @48kHz, Zip uses 12% of my CPU on the 'standard parallel' preset.

[audiosabre]

I moved on to other analysis modes & noticed an issue with consistency. Using the 'Noisiness' mode on a oneshot snare looped ad infinitum. Maybe I'm understanding it wrong, but I'd expect the compression behaviour to be fairly similar each time, similar to the amplitude mode. However, I can even see by looking at the blue readout of the analysis signal that each 'hit' can produce a audibly different analysis signal ; in peak mode you can get little spikes occasionally, and RMS causes it to suddenly drop the output signal level slightly, then creep back up. Not close to constant. Which is sad, because the noisiness mode is actually really decent for tightening up noisy or tape crunched snare sounds.

Is this by design, and if so, can it be improved upon? Thanks.

[tapiodmitriyevich]

Unfiltered Audio, dev of the year ?

I wish I could afford all their stuff EUR 100 is always a hard punch. Well... one by one... Hey UA, just leave PA, do pricing more like in the < EUR 50 area, sell plugins to anybody and his cat, get wealthy, rich, prostitutes, time magazine man of the year,,, yadda dadda

[jasco]

Unfiltered Audio, dev of the year ?

I wish I could afford all their stuff EUR 100 is always a hard punch. Well... one by one... Hey UA, just leave PA, do pricing more like in the < EUR 50 area, sell plugins to anybody and his cat, get wealthy, rich, prostitutes, time magazine man of the year,,, yadda dadda

The only 'complete' PA bundle I bought was the UA one and then they discontinued it and added all these new cool plugins

[pekbro]

no, doesn't make any difference. @48kHz, Zip uses 12% of my CPU on the 'standard parallel' preset.

Ok, yeah... after making some comparisons of my own, its pretty similar to Pro C2 or Ozone Dynamics, in regard to cpu use. E.g. not the most economical, still perfectly usable for me. No doubt this is due to the sophisticated analyses and probably some over sampling as well. *both of those are very high quality compressors I might add, as is Zip...

-Cheers

*Just a further note: The UA guys are PHD level researchers / developers. No doubt,
every optimization has been made here.

[SinclairX]

Very versatile tool, really becoming the fixer of lots of mixing needs.

Great at removing the harshness of high hats.

Also using the modulation sequencer to attenuate the high freq on vocals in the start of a verse and open up towards the end of verse for a rise FX.

Anyone else got some good ways I could use it?

Thanks

[audiosabre]

I moved on to other analysis modes & noticed an issue with consistency.

Any word on this? They say an image is worth a thousand words:

Zip-Inconsistency.png

There's nothing in the settings that I'd expect to cause any dynamic behaviour like this (auto threshold, AGC, etc). I've tried literally every parameter and I can't get Zip not to do this in noisiness mode. It's literally the same snare sample over and over again

Man I was loving the demo until this happened. As you can see from the image, your mix will be all over the place if levels change on their own like that

[audiosabre]

I literally said literally twice

[Amusesmile]

I moved on to other analysis modes & noticed an issue with consistency.

Any word on this? They say an image is worth a thousand words:

Zip-Inconsistency.png

There's nothing in the settings that I'd expect to cause any dynamic behaviour like this (auto threshold, AGC, etc). I've tried literally every parameter and I can't get Zip not to do this in noisiness mode. It's literally the same snare sample over and over again

Man I was loving the demo until this happened. As you can see from the image, your mix will be all over the place if levels change on their own like that

Hey this is Josh from Unfiltered. If you send me the project and stems I can take a detailed look tonight or over the next few days. I have a pretty good guess about what's happening though. In amplitude and quietness mode we're determining the level sample by sample, whereas in the four FFT modes we're determining brightness (etc.) level every FFT bin size, then interpolating the analysis level to that point. So essentially you're seeing the FFT sampling rate affect the transient response. Depending on where in the bin the hit happens each pass, the result will be slightly different. We can include different FFT sizes and/or develop sample-by-sample MIR detection in an update if it's possible and enable that as an option. I can think of a few ways that would work at least for brightness/darkness. Just to be clear, this is mainly an issue only because having silence in between discrete drum hits is kind of a very particular weakness to these algorithms (since silence always registers fully high or fully low) If you run a loop through the thing you should see consistent results each pass or at least nearly so.

TLDR: sampling rate issue with FFT based algorithms. Only an issue in situations involving a lot of silence on four FFT modes.

[audiosabre]

Hi Josh. I don't think stems or a project will be necessary, as the information you've provided correlates to what I'm experiencing. It's not like there's anything special about the snare sample or settings I used to produce the results shown in the image. It pretty much does that with any kind of one-shot, or tracks with silence as you now informed me about. Should be fairly trivial to replicate for yourself; a hi-hat one-shot will produce even more dramatic results

I've tried running loops as suggested and it does seem to be consistent, so I'd say you've nailed it. Many thanks for the reply & insights as to what's happening

[Jedinhopy]

And does it work on windows xp ?

You gotta suck it

SUCK WHAT????

[risome]

Unfiltered Audio, dev of the year ?

I wish I could afford all their stuff EUR 100 is always a hard punch. Well... one by one... Hey UA, just leave PA, do pricing more like in the < EUR 50 area, sell plugins to anybody and his cat, get wealthy, rich, prostitutes, time magazine man of the year,,, yadda dadda

I picked this up for $49 US with my monthly discount voucher plus the current discount on it.Always unique good stuff from Unfiltered Audio

[thecontrolcentre]

Unfiltered Audio, dev of the year ?

I wish I could afford all their stuff EUR 100 is always a hard punch. Well... one by one... Hey UA, just leave PA, do pricing more like in the < EUR 50 area, sell plugins to anybody and his cat, get wealthy, rich, prostitutes, time magazine man of the year,,, yadda dadda

I picked this up for $49 US with my monthly discount voucher plus the current discount on it.Always unique good stuff from Unfiltered Audio

Waiting patiently for my November PA coupon

[pekbro]

There’s a $50 off coupon floating around with which you can get it for 49, so long as you grab something else that’s on sale. The total has to exceed 100. If you happen to want something else they have on sale atm.

Just sayin, In case anyone didn’t see it.

*in the bargain, gossip thread, sry on my cell phone so difficult to link.