Satin, different sound at different sample rates, be aware of it

[Traverser]

Thank you, Urs, for the detailed reply. A little disappointing but, I totally get the current focus on Zebralette 3….that, and the fact that there aren’t enough hours in the day.

What is really disappointing is that this issue still not mentioned on the official product info and sale page, and the manual still informs us about "Internal sample rate: 352-384kHz (depends on project sample rate, 8 x oversampling 44.1kHz)".

Most misleading, most dishonest.

[Urs]

What's misleading about it and why would it be dishonest? The internal samplerate *is* 352-384kHz. The problem is that the oversampling factor depends on the host samplerate. Dunno how you figure that it isn't so...?

Also, I'd be interested in examples where a company similar to us keeps an up-to-date list of issues that were found by their users on a product page. Can't quite recall ever having seen that, but I'd be happy to be set straight if that was a common thing that I simply missed.

[Traverser]

What's misleading about it and why would it be dishonest? The internal samplerate *is* 352-384kHz. The problem is that the oversampling factor depends on the host samplerate. Dunno how you figure that it isn't so...?

Also, I'd be interested in examples where a company similar to us keeps an up-to-date list of issues that were found by their users on a product page. Can't quite recall ever having seen that, but I'd be happy to be set straight if that was a common thing that I simply missed.

1. You wrote beforehand:
"The pre-emphasis highpass filter is not oversampled. ... It's easily fixed by oversampling it, but then we'd get the response of 88.2/96 kHz. However, I'm sure Satin was designed to sound correct at 44.1/48kHz. So the physical constraint of the filter may be a feature." ... OTOH not oversampling it was done to save CPU" and regarding the analyzer: "Unrelated to this is an issue where the visual feedback in Satin differs vastly between samplerates"

So Satin is not completely oversampled as the manual suggests.

2. Many software companies provide release notes to their software with a section titled "Know Issue". Even you can do that if and when you care about doing that:

in: https://u-he.com/products/satin/releasenotes.html

Known Issues:
Context menus need plugin focus to work (Apple issue)

.

These problems that I reported created really annoying obscure problems for me. A "Known Issues" section mentioning this problem, and a detailed manual on where and what kind of oversampling happens and not happens in Satin could helped me sparing all my frustration with this plugin. Even a note that "Satin was designed to sound correct at 44.1/48kHz" ...

[Urs]

Yeah, Release Notes are updated as the plug-in is updated. Like pretty much all other companies we don't keep a blog advertising shortcomings that we are eager to resolve. Because we are. I did sit down and spend several weeks on this right as you started this thread. It's unfortunate that we weren't able to publish this.

I find it hardly "dishonest and misleading" if a *linear* component that does not benefit from oversampling isn't oversampled in a plug-in. Dozens if not hundreds of other components are, because they are non-linear and benefit from a high sample rate.

[Traverser]

Then find it hardly dishonest and misleading on your side of the story.
On my side it is dishonest and misleading.
Even if I would like to buy it now for multi-sample rate uses or just for 96k projects all the presets are misleading and dishonest.

[Urs]

I think your exaggerating.

By far most presets vary within the typical boundaries at different sample rates. One would not use different sample rates if it wasn't for some audible improvement, hence minor variations are expected. For most settings, unless the input material has obvious spectral differences at different sample rates, there's hardly anything audible added or changed at all. The input material needs quite some energy beyond 22.5/24kHz to make much of an audible difference here (this is true for all dynamic effects of course, not just for Satin, regardless if they're oversampled or not).

The built-in analyzer is a different story. It uses equal energy across the spectrum, such that at 96kHz the spectrum beyond 24kHz is as loud as the whole spectrum between 0Hz and 24kHz - at 192kHz even three times as much. This is the main culprit why the visuals are off. Another issue is that Dirac impulses are not suited for visualising dynamic effects such as companders, and that's more so true at different samplerates. I guess the perception that lead to this thread mostly stems from this.

The fix we'll publish is therefore mainly a fix for the built in analyser, as good as it gets, and it also fixes the issue that energy beyond 22.4/24kHz may be pronounced with certain settings of pre-Emphasis. It furthermore improves the quality of the EQs in Satin, which is something that we did in Filterscape, MFM and Uhbik as well.

We'll be beta testing the AAX version soon, and that'll also contain the fixes described within this thread.

That'll hopefully be the end of that "story".

[JohnZ]

Thank you Urs. I find your transparency around this area refreshing and I'll be buying the Satin as soon as the next budget is approved.