Unfiltered Audio Zip: Creative Compression with Modulation

[Ah_Dziz]

Sooooo is the sidechaining available to the modulation system (input follower)? If so will this be added to your other products? It would be useful for very many things.

Yes, although we can't make any promises on timelines quite yet. Since adding in a sidechain input will change each plugin's channel configuration, it could lead to us breaking backwards compatibility.

That is awesome. I know you know this, but making a new version of the plugin (like waves and Zynaptic and some others) so that the sidechain enabled version shows up as a different plugin is a good workaround for backward compatibility. I'm sure I'll be checking this compressor out in the near future even though I can't even guess how many compressor plugins I already own.

[audiosabre]

I caved and purchased Zip. Surprised there isn't much more buzz around it. Guess we're spoiled for choice. Now I've used it for a while, I find it really fantastic; desert island good. In amplitude mode, it does it's thing with style & attitude. Zip's transient response & how it grabs the audio sounds really authentic and makes it easy to dial in. The envelope styles aren't subtle tweaks, they're nicely distinct. Love how they implemented the lookahead too. Looking forward to how they can build on this, and make some of the other modes more predictable. Feel like I haven't scratched the surface yet

[Amusesmile]

I caved and purchased Zip. Surprised there isn't much more buzz around it. Guess we're spoiled for choice. Now I've used it for a while, I find it really fantastic; desert island good. In amplitude mode, it does it's thing with style & attitude. Zip's transient response & how it grabs the audio sounds really authentic and makes it easy to dial in. The envelope styles aren't subtle tweaks, they're nicely distinct. Love how they implemented the lookahead too. Looking forward to how they can build on this, and make some of the other modes more predictable. Feel like I haven't scratched the surface yet

Hey, I can't tell you how happy I am to hear that you're enjoying it. I still view these algorithms as somewhat experimental so I was mildly nervous to see how they'd perform in the wild. Luckily the response has been hugely positive. Rest assured we will be expanding on these ideas and I will try to develop a solution to that "inconsistent single shot" issue you identified earlier in the thread. It's funny- after writing the response I realize the simplest way to solve it in situations where there are large gaps of silence would just be to detect the silence->onset and reset the FFT from the first non-zero sample. So there are a bunch of ways to improve that, we just need to decide on the correct approach. Thanks again for the note.

[Scarlet Pumpernickel]

Is it just me, if I engage loockahead, the signal looses high-end even without any gain reduction and no colour? I'm using Reaper64.

[Amusesmile]

Is it just me, if I engage loockahead, the signal looses high-end even without any gain reduction and no colour? I'm using Reaper64.

You'll hear a pitch drop as the delay line introduces the lookahead but then it goes back to normal. As a triple sanity check I just confirmed that with a sawtooth wave going through Zip and into a spectrogram. It's identical (minus the slight slight fluctuations of the actual spectrogram reading).
See these:

[pekbro]

This compressor has been making it difficult to test a few other notable compressors
that have shown up lately due to a particular fascination it seems to impart on the user.

Creative compression aside, it seems as though it "is" fast becoming my favorite compressor.
So currently, we have the compressor and a great clipper, any chance we will see a dedicated filter
to round out the Unfiltered Audio "mix chain" nicely?

Or are you guys truly "anti-filter"

Thanks for your amazing work guys...

-Cheers

[Twrogstudio]

Got my November coupon, no guesses for what it’s going towards!

[Scarlet Pumpernickel]

Is it just me, if I engage loockahead, the signal looses high-end even without any gain reduction and no colour? I'm using Reaper64.

You'll hear a pitch drop as the delay line introduces the lookahead but then it goes back to normal. As a triple sanity check I just confirmed that with a sawtooth wave going through Zip and into a spectrogram. It's identical (minus the slight slight fluctuations of the actual spectrogram reading).

I tried it again with 100ms like you did and and high frequencies stay unchanged, but they do drop with loockahead times around 1-2ms (even if I engage "bypass" button on the plugin). Must be something with Reaper or just the "overweight" project itself.

[Ah_Dziz]

Is it just me, if I engage loockahead, the signal looses high-end even without any gain reduction and no colour? I'm using Reaper64.

You'll hear a pitch drop as the delay line introduces the lookahead but then it goes back to normal. As a triple sanity check I just confirmed that with a sawtooth wave going through Zip and into a spectrogram. It's identical (minus the slight slight fluctuations of the actual spectrogram reading).
See these:

So what's under that black censor label? You can tell me. I'm a doctor.

[Scarlet Pumpernickel]

So, I tried it in the fresh project with just white noise. As you can see in the attached picture the output signal (red line) is lower than the input - it's the missing HF by the sound of it.

[Amusesmile]

So, I tried it in the fresh project with just white noise. As you can see in the attached picture the output signal (red line) is lower than the input - it's the missing HF by the sound of it.

Good test. Let me duplicate it to try to see what's going on. Maybe there's a ghost in the code.

[Amusesmile]

Ok, I see what’s going on and it’s so obvious in retrospect. I thought that I had included code for rounding our delay line targets to whole-sample numbers instead of allowing for fractional delays. That’s why I was puzzled because with a whole-sample delay (i.e. delaying the signal by 44 samples) you just have a time shifted version of the original signal.

I didn’t actually add that section of code in one necessary part to the release version of Zip however, so we in fact are allowing fractional delay time targets. Linear interpolation on fractional delay lines causes slight low pass filtering depending on the fraction between samples, which is what you’re hearing. (For a very quick explanation of why that would occur, imaging the highest possible frequency signal you could have, so 1.0, -1.0, 1.0, -1.0, if we try to delay that by 0.5 samples using linear interpolation you’ll get every sample averaged to 0.0.)

This also explains why I wasn’t able to reproduce it when testing at exactly 100ms, since that’s a whole number of samples (4410 samples at 44.1k) but of course when it’s non-whole I see it instantly. Adding the code to round to the nearest whole sample target (aka instead of allowing 1.0ms delay to equal 44.1 samples we'd round it to 44 samples or 0.9977324263 ms delay) fixes the issue as expected.

This should be the default behavior so I think what we’ll do is release a quick update with a note that in order to keep EXACT session compatibility you’d need to manually turn on an option that disables the rounding.

Thanks for finding this! Send me an email or DM with your PA email address and I’d be happy to give you a full NFR of one of our products either now or in the future if you’re interested.

Cheers.

[Scarlet Pumpernickel]

Ok, I see what’s going on and it’s so obvious in retrospect. I thought that I had included code for rounding our delay line targets to whole-sample numbers instead of allowing for fractional delays. That’s why I was puzzled because with a whole-sample delay (i.e. delaying the signal by 44 samples) you just have a time shifted version of the original signal.

I didn’t actually add that section of code in one necessary part to the release version of Zip however, so we in fact are allowing fractional delay time targets. Linear interpolation on fractional delay lines causes slight low pass filtering depending on the fraction between samples, which is what you’re hearing. (For a very quick explanation of why that would occur, imaging the highest possible frequency signal you could have, so 1.0, -1.0, 1.0, -1.0, if we try to delay that by 0.5 samples using linear interpolation you’ll get every sample averaged to 0.0.)

This also explains why I wasn’t able to reproduce it when testing at exactly 100ms, since that’s a whole number of samples (4410 samples at 44.1k) but of course when it’s non-whole I see it instantly. Adding the code to round to the nearest whole sample target (aka instead of allowing 1.0ms delay to equal 44.1 samples we'd round it to 44 samples or 0.9977324263 ms delay) fixes the issue as expected.

This should be the default behavior so I think what we’ll do is release a quick update with a note that in order to keep EXACT session compatibility you’d need to manually turn on an option that disables the rounding.

Thanks for finding this! Send me an email or DM with your PA email address and I’d be happy to give you a full NFR of one of our products either now or in the future if you’re interested.

Cheers.

This is quite comprehensive explanation and the very generous offer that I am not able to refuse. If I understand correctly, another way to implement the fix would be adding a sample mode to loockahead (although it would change depending on SR).
I will PM you my email shortly.
Thank you.

[hivkorn]

I'd really like to have an eq with there coloration algo and a reverb with lfo and coloration.... I ask to myself : " what would be a Unfiltered Audio synth ?....a huge screaming badass motha fu**a ...! "

[simon.a.billington]

I caved and purchased Zip. Surprised there isn't much more buzz around it. Guess we're spoiled for choice. Now I've used it for a while, I find it really fantastic; desert island good. In amplitude mode, it does it's thing with style & attitude. Zip's transient response & how it grabs the audio sounds really authentic and makes it easy to dial in. The envelope styles aren't subtle tweaks, they're nicely distinct. Love how they implemented the lookahead too. Looking forward to how they can build on this, and make some of the other modes more predictable. Feel like I haven't scratched the surface yet

It might be simply a case of how people like their emulations. Since it doesn’t emulate anything in particular and the UI is a modern one to some resepect, people haven’t probably giving it a second thought.

...it would be their loss though.