saturator noise

[magmagwa]

Hi, I ran a low frequency sine wave, D1 (~40Hz) through Saturator (both Sat and SatMB) and got a bad static artifacts for Soft 2 and 4. It happens after holding it for at least a couple of seconds.

I ran higher sample rates and it actually got worse.

Input sound is peaking at -5. I'm using Vital, with a sine wave.

Without Sat it is completely clean.

Since Saturation is used everywhere, and I assume the same implementation, I think it would be key to make this clean in all situations.

Any feedback? File a bug report?

[dmbaer]

Hi, I ran a low frequency sine wave, D1 (~40Hz) through Saturator (both Sat and SatMB) and got a bad static artifacts for Soft 2 and 4. It happens after holding it for at least a couple of seconds.

I ran higher sample rates and it actually got worse.

Input sound is peaking at -5. I'm using Vital, with a sine wave.

Without Sat it is completely clean.

Since Saturation is used everywhere, and I assume the same implementation, I think it would be key to make this clean in all situations.

Any feedback? File a bug report?

So ... saturation causes harmonics to be added, the frequencies of which are integer multiples of the base frequency when you're using a sine wave test signal. These harmonics tend to get lower in amplitude as the frequency increases. If the harmonics are still of sufficient amplitude by the time you get to half the sample rate (plus a few more kHz, but let's not get into that here), you will get audible aliasing. That's what saturation does: adds harmonics. But it's not uncommon to have the additional harmonics drop to a non-audible level by the time the frequency reaches half the sampling rate. In any case, if this is causing problems, use the oversampling feature (the +1 button in the right-hand sub-panel).

I have no idea what Vital is or how it relates to this behavior.

[magmagwa]

I ran higher sample rates and it got worse.

[valankar]

I believe there are some bugs with this too. I made a video:



See the video description of the timepoints where you can see the problem.

I also mentioned similarly in:

viewtopic.php?t=578327

It seems related to the 'Antialiasing' setting. I am not yet able to reproduce it with that setting off.

[DarkStar]

it would be best to report your findings directly to MeldaProduction, via the Contact page on the web site.

[Starship Krupa]

I have no idea what Vital is or how it relates to this behavior.

Vital is an increasingly popular wavetable VSTi: https://vital.audio/

It's licensed similar to MFreeFX Bundle. Basic version is free, paying gets you more presets and wavetables. At the highest level, a couple of features are extended.

How it relates to this issue is that the OP was trying to use MSaturator to process the output of Vital when they noticed the ugly artifacts.

[dmbaer]

Vital is an increasingly popular wavetable VSTi: https://vital.audio/

Ah ... thanks. So I assume Vital was used as the source of the sine wave test signal. I've always just used MOscillator for my own tests in this area, but I think it's safe to assume the source of the sine was is not a suspect in this mystery.

[magmagwa]

Here it is in an MSF patch. Hold D1, about 10-15 seconds in it sounds like a digital spike/artefact.

I will send to Melda support, but I know they look in here a lot too.

Thanks for taking an interest.

dmbaer, got the </> code insert working, looks like preview does not work but it does look correct on submit, thanks

Code: Select all

$eNrFV11v4jgUfedXROnrbEPC50ghI0pDt5rSsg3T6T6a5AJWHTuyHbbtr1-ZJoGGbYaPkfYFJdfnHF8fX4dr-9trSqw1cIEZHdjuZdO2gMYswXQ5sHO5+KNvfwsa-iRiOU3GKJaMvwmQEtOlsC4GtmtbF8Pra8jkamA3Lzu2dXFD2BwRMUNzPRwmWDIe4XcY2K7X63-pu93daIjE2+7In3-9yARKM6JTaNrW3WOUESwlaLg7sN1es1MNe3quCXq9Gz9cERa-mAlbnm1Nbq9vp5wtOUpHK0SXEFI0J6C1p4gjQoDgdySNA7b1ATtCEpaM43fgmvBUWPW13+p0PDvwo3w+JfkS0407gW9eo48uHZ9aXRoREIglJGaqgd38olIxb6J4sK4QfRmxnEqdwTgn5B6lMLAfRIwJQZJxS2cmXm6plq0O29azjv+tfgNf6TWdTxZ48YNiwahZwYgRgjIBiRl6YiRXys3L3tff7UXgRyuUgRXNwun0MYzCmYo6ge98TLMMfGLPKOccqFTBil0Rkjkv3KqEdhzyfuXQb122s7Mep9z4-TXvlWdjp2Kf-6dibW6L1alP+jlo+EMpUfziWndAl+ozo4wO6RoIy8AaCoGX20TC+6eiCDzt0YZcynjnyHilTOscmVYp0z5Hpl3KdM6R6ZQy3XNkuqVM7xyZXinTP0emHzQUxQ38GYeUEfbZR0KDNNY7BOsZbOsQbMtg24dg2wbbOQTbMdjuIdiuwfYOwfYMtn8IVvlr-uhDuv4lYYts+HfjB7e6iduN3iNrgncsoXUsoX0soXMsoXssoXcsoX8U4RHRhKX6gCAqFoynugWqOV99vZcFr1DwTlTwSoXWiQqtUqF9ooI6dd-hTcSIwPFObJlbFe9kFeXHExAWY3lqQhV6Rc87T283vyq0sVWechAi51AHwTJeXQFNnmtAE5b8XAGQGsgMp3MOvXYN5IoDkqumVwMJXzOVMmbUdWtgUS4kwrQGcc8kPCwW+4gbghMwn8a9sRGjkjNCgAsrwnRJQPVQFEhxTIexxGtQroK0irZ6xsFs2TWscQwTkMDFx55bXbDUQ+Bv2rG3WwnpLV0wa4qE+IfxZGCrXS6GI4kkBP4NUA5O4BtzncAfconjnOiKUa9EtXZylTqBPwbVBoPu4Soym256P2PTP++kdca8epGmfF1dsBVNZ5NF4A+nxkHdIOw4qgzmmfWIJJg2Y4Ikx687YJ6p66822Av8nziRq0ngRyznMVhTDgtMJPDNXVUxCohTkBq+81-XZxE0-gU2t+9h

[Held]

I think this has already been fixed in 16.00f I held it for about 30 seconds and it seemed fine to me (Windows)

[dmbaer]

could not find the cool "inline code" element people used to use

It's the button with the "</>" symbol.

[valankar]

Without Sat it is completely clean.

Does the problem go away when you uncheck 'Antialiasing'?

[magmagwa]

No it does not.

[MeldaProduction]

I don't think this is noise. It's DC offset, quite natural thing, that doesn't really matter. On master you'll use a DC filter or HP anyways.

[valankar]

I don't think this is noise. It's DC offset, quite natural thing, that doesn't really matter. On master you'll use a DC filter or HP anyways.

Could you please look at this video and confirm the high frequencies that show up here are caused by DC offset and not some bug? I was under impression that was only low frequencies. See the video description for the interesting time points.

[magmagwa]

I don't think this is noise. It's DC offset, quite natural thing, that doesn't really matter. On master you'll use a DC filter or HP anyways.

It's in the audible range, it can't be filtered out without losing content.