Help understanding host sample rates on synths sound quality

VST, AU, AAX, etc. plug-in Virtual Instruments discussion
yul
KVRian
940 posts since 26 Sep, 2002 from Montreal, CANADA

Post Sun Sep 29, 2019 4:37 pm

Hello, I am looking to understand the impact of various sample rates of my host on synths.

I know the general idea that oscillators would alias more or less in high ranges.

Filters otoh I am not sure as my experiments with older synths lead to changes in cutoff ranges mostly and perhaps better quality in filter modulation (high rates). Is there another impact on overall quality i am missing?

I also know some synths have internal oversampling which may make it somewhat neutral to the hosts sample rate. Is that accurate? How can I tell if that's the case? DIVA and a few others seem to have independent sample rate.

What about the older synths? Is there a impact here? I am looking at Korg Legacy (Which does not seem to oversample). Will there be a difference in quality or cutoff range depending on my sample rate?

Tal U-NO-LX here does mention of filter quality related to sample rate. What would the impact be?

Thanks

User avatar
Tj Shredder
KVRAF
3263 posts since 6 Jan, 2017 from Outer Space

Re: Help understanding host sample rates on synths sound quality

Post Mon Sep 30, 2019 2:30 am

Its mainly the aliasing which will be different. Any non-linear process, as distortion for example, will create frequencies which can exceed the nyquist frequency, which is half of the sample rate. In synths its the oscillators and sample playback which have to take this into account. If you have an arbitrary sample it will contain frequencies up to the nyquist. If you play it back at a higher rate, those will/can exceed the nyquist. The easiest way to avoid that it doubling the sample rate, which need roughly twice the CPU (a bit less if done properly...) You gain more than an octave headroom until the nyquist, as part of the mirrored frequencies are still out of the hearing range... There are a lot of techniques in DSP to lower the impact on the CPU...
The impact on recording audio is higher, because the analog filter needed on the input will create more artefacts on lower sample rates, as it can roll off more soft for the higher sampling rates...

User avatar
DJ Warmonger
KVRAF
3365 posts since 7 Jun, 2012 from Warsaw

Re: Help understanding host sample rates on synths sound quality

Post Mon Sep 30, 2019 2:56 am

Sampling rate has impact on many effects. Phase alignment, chord and pitch resolution, dynamic response time (especially limiters).

Aliasing is a bigger concern that it may seem, as pretty much any modulation creates some extra partials - even simple ADSR is not neutral operation from a mathematical point of view. Especially with extreme FM you may fall well into MHz range.
Now, if you stack a number of nonlinear processes (modulation, distortion, compression, gate) - their effect will stack, messing up the sound completely with aliased components.

Another benefit from high sample rates are quicker plugin responses - especially filters, which add ringing over a number of samples. Higher sample rate -> shorter ringing.
http://djwarmonger.wordpress.com/
Tricky-Loops wrote: (...)someone like Armin van Buuren who claims to make a track in half an hour and all his songs sound somewhat boring(...)

yul
KVRian
940 posts since 26 Sep, 2002 from Montreal, CANADA

Re: Help understanding host sample rates on synths sound quality

Post Mon Sep 30, 2019 4:08 am

Thanks a lot guys!! So aliasing and overall quality are top of list.

So what is the deal with filter cutoff ranges being different on some synth when I switch rates? Is this a common consequence or are some synths not affected by it (haven't checked all of mine yet).

Thanks!!

User avatar
Tj Shredder
KVRAF
3263 posts since 6 Jan, 2017 from Outer Space

Re: Help understanding host sample rates on synths sound quality

Post Mon Sep 30, 2019 8:48 pm

If this is the case, an amateur has programmed it and did not adapt the algorithm to the sample rate. Filter code needs to take the actual sample rate into consideration. It would be a bug, or at least has to be documented (works only for 44.1 correctly for example)...
Some might only read out the sample rate at instantiation and work correctly after a reload...

yul
KVRian
940 posts since 26 Sep, 2002 from Montreal, CANADA

Re: Help understanding host sample rates on synths sound quality

Post Mon Oct 21, 2019 5:59 pm

Hello! A bit late to reply back but a clarification is needed (after double checking)!

The synth in question's cutoff does not change but rather the aliasing in the higher ranges which is quite normal for most synths!

Cheers!

izonin
KVRian
1199 posts since 27 Oct, 2009

Re: Help understanding host sample rates on synths sound quality

Post Tue Oct 22, 2019 9:35 am

I think the synth that is the best at illustrating the effect of aliasing is Reaktor. It allows you to change the internal sampling rate up to 768MHz. But to be honest, anything over 384KHz is just a waste of CPU cycles.

There's also the free little vst oversampler plugin. You can stack instances to explore the advantages of higher sample rates with any softsynth.

mystran
KVRAF
5441 posts since 12 Feb, 2006 from Helsinki, Finland

Re: Help understanding host sample rates on synths sound quality

Post Wed Oct 23, 2019 7:02 pm

DJ Warmonger wrote:
Mon Sep 30, 2019 2:56 am
Another benefit from high sample rates are quicker plugin responses - especially filters, which add ringing over a number of samples. Higher sample rate -> shorter ringing.
Filtering ringing time depends on the response shape and the wavelength at the frequencies of interest. Tuning a given response one octave down will double the ringing time (because the wavelengths involved are twice as long). Doubling the sampling rate shifts the relative frequency (with respect to sampling rate) down by one octave, so the ringing time will double in samples, but since in this case we're also playing back those samples twice as fast, the effective change in terms of wall-clock time is net-zero.
If you'd like Signaldust to return, please ask Katinka Tuisku to resign.

Return to “Instruments”