Dolphin

Official support for: signaldust.com
Locked New Topic
RELATED
PRODUCTS

Post

Just to say "Thanx*10000 for Dolphin".

Tried 2 days ago, not expecting a lot, ended completely floored. Currently, my favorite VSTi, tonwise, and that's compared to Diva and SEM V (another level compared to other TB emus). I don't care too much for accuracy of TB emulation, I simply like synths that sound and behave good, and I'm very picky, and so far, this little synth is completely effortless.

Kudos Mystran

Post

Glad you like it.

There are a few things in Dolphin that deviate from how the real deal: the accent is currently just a crude behavioral model, and the square is a band-limited approximation. But I guess I have to admit that I've certainly spent quite a bit of time comparing it both against some hardware and various other emulations.

Post

mystran wrote: There are a few things in Dolphin that deviate from how the real deal: the accent is currently just a crude behavioral model
Well, I had hotrodded xoxbox for a while, and I almost never used it with controls set to behave exactly like TB. For example, I would usually drive VCF 12dB hotter than in stock TB. I don't have xoxbox here ATM, but from memory, I like behavior of accent in Dolphin more than real deal.
mystran wrote:and the square is a band-limited approximation.
You mean, you use wavetable instead of blit/blep or you figured out how to make bandlimited dynamic waveshaper (if not exactly like "sqaure" in TB)?

Post

urosh wrote:
mystran wrote:and the square is a band-limited approximation.
You mean, you use wavetable instead of blit/blep or you figured out how to make bandlimited dynamic waveshaper (if not exactly like "sqaure" in TB)?
In a TB the square is made from the native saw waveform by a waveshaper that doesn't really result in anything approximating a proper square-wave in practice, there's a bunch of things wrong with it, most obviously frequency dependent pulse-width and some asymmetric "bleed-through" or .. whatever.

Anyway, the point is it doesn't look like a square at all, and while I could run a similar waveshaper, that would run into aliasing issues, so instead I currently synthesize a crude approximation of the resulting wave (or at least it's most obvious features) using BLEP techniques directly. It's one of that things that could be improved though, I think.

Locked

Return to “Signaldust”