Oberheim Clone - Eight Voice from Cherry Audio

The sound blows you out of the water, but some patches need CPU!

Yes, but it is Mac only…

1 Like

There’s a quite interesting comparison on this German page:

1 Like

OK, So OB-E is Mac only … Well that settles that LOL!

Damn, I couldn’t resist to get the Cherry Audio Eight Voice for ~ 25€ yesterday :grimacing:

This ist the first plugin where I experience higher CPU load and audio dropouts in the AU version on my 2015 MBP 13". VST is just fine.

Did somebody experience something similar with this or other plugins? Do we already have a list here at the GP community forum regarding differences in AU/VST/VST3 of plugins?


There is no reason why one particular format should be more intensive than others although there may be differences depending on the tools use to build them but as a matter of course, I tend to avoid AU and just use VST as I’ve found the latter to be much more reliable in general, regardless of the host I’m using.

I can’t tell… i only use VST or VST3 and as far as i’ve seen, there never was a remarkable diffrence between them.

I also would expect nearly the same performance/behaviour independent from the format.
However, yesterday this was my first experience with this plugin - audio dropouts in some of the presets in AU - same presets in VST everythuing fine. I will recheck this evening.

Why? Suppose the two versions were implemented by two different developers? They would have had the same front end but the back end (audio handling) could be completely different, depending on how it was developed. Or there could be an implementation issue in a library that impacts one format more than another. Mistrust the obvious.

Well, I hope that even if more than one developer are implementing the different versions of a plugin rely on the same underlying principles and algorithms behind. It would be strange if we have to discuss “the AU version of ‘FancySFX’ has a little bit more bottom as the VST3 and sounds harsh”, as the implementations would found on different DSP models (e.g. IIR vs. FIR filters in worst case).

My first (and last) VST plugin coding has now 20h anniversary - I went back to hardware development in my professional life - but what I see in SW development in general is a lot of reuse and working on a common code base. Of course there are slightly different requirements for the AU/VST/VST3 framework. In the end they should provide the more or less same exact result!

But yes, you are right, you could see differences. This was the first time experiencing this, therefore my question :wink:

Sorry, I wasn’t talking about issues around different algorithms (although that’s certainly possible). I was thinking of more “stupid” things where one developer added some GUI stuff with a lock inside the process block and another one didn’t. That kind of thing could lead to noise issues if a real-time callback was taking too long.

I won’t mention any names but we once saw a problem where some plugins were occasionally license checks inside the real-time audio processing thread, thereby creating occasional glitches.

You can’t just assume that everybody gets it right!