Page 1 of 5

Echo bug (incorrect term, it is not a bug)

Posted: Sat May 07, 2016 10:51 am
by giancarlo
I'm starting a topic, in a very visible forum section so we can discuss freely about it

Several days ago we refunded an user just because it, than I received an email few minutes ago.
Really lately there aren't many complains, but I want to understand if there are products where the situation is not acceptable, so we can fix it. I think it is not something we can discuss using tickets, because it is more about a compromise which is the result of plural opinions

Quickly: normally plugins have noise. Nebula not. At the opposite the noise creates a sort of echo, which sums as soon as you stage more bands.
It is not a serious trouble, since we can "fade out" noise. Exactly like you do with audio. We cannot remove noise or denoise, since we would damage the impulse response or the kernel. Deal with it, hardware is noisy.
Now: fading out is a good technique, and we can mask our very well, but at the expense of ripple. Many of you will never experience it, but lately edm bottom drums are technically uncorrect, they have a lot of energy till almost zero Hz, till having dc offset. In this case this situation is more visible than, let's say, mastered and band limited rock.

Now: we can reduce this effect and increase ripple. Or do the opposite. My question is about the current situation of our products so we can set an acceptable compromise which makes everybody happy.

Re: Echo bug (incorrect term, it is not a bug)

Posted: Sat May 07, 2016 3:47 pm
by Avgatzeblouz
If the issue is only with dc offset, then don't waiste your time on this. IMO.

Re: Echo bug (incorrect term, it is not a bug)

Posted: Sat May 07, 2016 4:10 pm
by giancarlo
Well, it is common in edm music. We had less than 10 refunds cause this issue in all our history, but we are ready to discuss or tune our presets more aggressively. I see there are developers with a very aggressive fade out, for example 70 ms for a pultec and it is still considered a good library.
We are lucky, we are almost the only developer using fir filters, they are finite and limited by definition. So it is just a matter of deciding the length.
Shorter kernels mean less cpu intensive emulations, smaller plugin data size and less echo, at the expense of the sound

Re: Echo bug (incorrect term, it is not a bug)

Posted: Sat May 07, 2016 5:29 pm
by CamoKrooked
This issue does not just happen with sounds going down to a very low frequency range, it is rather transient dependant. So any kind of sound with hard transients will end up being echoed, which can be quite audible if you use a compressor after an Acqua EQ, or a few Acqua plugins in a row.
Pink has a very loud echo for example, especially with both preamp and a few EQ bands turned on. We haven't noticed it on Scarlet 3 for example.

I think it's fair to say it can't just be ignored. Giancarlo, what about being able to choose between an economy mode with a fadeout of the impulses, which doesn't create echo, and a HQ mode?

Re: Echo bug (incorrect term, it is not a bug)

Posted: Sat May 07, 2016 5:44 pm
by Brian
CamoKrooked wrote: I think it's fair to say it can't just be ignored. Giancarlo, what about being able to choose between an economy mode with a fadeout of the impulses, which doesn't create echo, and a HQ mode?
Exactly my question - how feasible is this to create? I suppose it could be done with a separate vector set, but would that require another plugin binary (ie normal, econo, zero latency) or could the full quality set (longer kernel vectors with a gentle fade) be loaded with a toggle button from the same plugin instance?

The downsides may include longer loading time, nearly double the download/storage size requirements, but this would let users decide when to play the longer kernels on a case-by-case basis when they notice the 'echo'.

Re: Echo bug (incorrect term, it is not a bug)

Posted: Sat May 07, 2016 6:06 pm
by mightymosaic
I've been using pink with allot of edm sounnds and heavy kicks and haven't had this problem. I haven't heard the echo since magenta (first release) once it was fixed it never came back in any aqua.

Re: Echo bug (incorrect term, it is not a bug)

Posted: Sat May 07, 2016 7:35 pm
by giancarlo
Yes it could be flexible, but it adds complexity... And buttons... And things to explain :)

Re: Echo bug (incorrect term, it is not a bug)

Posted: Sat May 07, 2016 7:55 pm
by CamoKrooked
It would be great to hear a comparison between a non-echo and the echo version of pink for example, with the same band settings, most preferable all bands turned on and boosted quite a bit. It's the only way knowing how much quality loss this would actually cause.
Any chance giancarlo?

Re: Echo bug (incorrect term, it is not a bug)

Posted: Sat May 07, 2016 8:12 pm
by giancarlo
Mmmm too much work and too many things to do yet
If you experience something not acceptable on a particular product this is the right topic where to discuss it. So we know we need to improve in a further version. Otherwise actual presets are already tuned at the best comprise possible for us and betas

Re: Echo bug (incorrect term, it is not a bug)

Posted: Sat May 07, 2016 11:13 pm
by Avgatzeblouz
I must say I never heard it.

Re: Echo bug (incorrect term, it is not a bug)

Posted: Sat May 07, 2016 11:57 pm
by nverxion
I've never heard it either, and I do a lot of post compression in mastering. So it seems to be case-specific depending on computer hardware and software configurations. Is this issue Mac-related? AAX-related? Windows-related? I'm on Windows and like I said I have not heard this echo.

My advice will be not to make any adjustments to any of the current or future releases, but rather to issue alternate configs of the acquas to fix the issue for anyone having this specific problem. Offer it as an alternative option, not a replacement. I'd hate to have my versions of the plugins compromised in any way related to sound quality considering that there is nothing wrong with them in the first place. In my case it ain't broke. So don't fix it.

Re: Echo bug (incorrect term, it is not a bug)

Posted: Sun May 08, 2016 12:47 am
by Avgatzeblouz
nverxion wrote:I'd hate to have my versions of the plugins compromised in any way related to sound quality considering that there is nothing wrong with them in the first place. In my case it ain't broke. So don't fix it.
I'm with you on this.

Re: Echo bug (incorrect term, it is not a bug)

Posted: Sun May 08, 2016 1:33 am
by nverxion
I should specify:

I'm on Windows 10.
I use Studio One, Sequoia, and Reaper.
I use VST.

Re: Echo bug (incorrect term, it is not a bug)

Posted: Sun May 08, 2016 1:48 am
by RJHollins
to help explain ... this is a topic in the beta-test section.

We try to provide observations back to the Dev team, assisting in fine-tuning this balance.

The desired goal is that this 'balance' not be an issue for the commercial Users, and may explain why it goes unnoticed.

Notwithstanding ... <G> has raised this point of concern. He wants to be certain of this fine-tuning balance.

... just a bit of 'behind the scenes' 8-)

Re: Echo bug (incorrect term, it is not a bug)

Posted: Sun May 08, 2016 8:38 am
by markgalup
I recently submitted a ticket to support, which I've detailed below. Is this related to the echo scenario you are describing?

When I was compressing a high hat cymbal with Amethyst1, using a medium fast attack and a fairly fast release, the noise on the track began sounding like it was starting and stopping - like looping - on/off or mute/unmute behavior on JUST the noise (hiss). Is this an issue with the fading of the samples? Or with the compressor in Amethyst vs Amethyst2? I believe I can send you an audio sample of this, as I was only able to solve it via LPF on amthyst, so I will take that out and export it for you if that helps.

Curious what you all think of that.

MG