Repro-1 (out now)

[Urs]

Well, for a start my personal choice (for practical use) is not most expensive/accurate model, but rather one of "mystran" models, probably one without corrective step.

While Mystran's method works exceptionally well on this filter, it still requires more CPU than the iterative method if we wanted to reach equal stability. It needs 16x or 32x oversampling at 44.1/48kHz to behave as stabil as Newton's method at 8x oversampling. That said, from all methods I've tried, Mystran's (aka "pivotal" in my blog) works best, much better than the "tangential" method I proposed, or anything Heun/Euler.

[urosh]

(aka "pivotal" in my blog)

Like it, this should be official. Biggest issue I see with pivotal is that I don't know of any criterion of convergence for this approach, unlike Newton.

[EvilDragon]

Yep, Urs - post a short version here then do a longer one on your blog or in a PDF or something. Would be great - I'd like to know both short and long explanations

[PAK]

In RePro topic on u-he subforum you have some. This is not all, but I think that long descriptions in vague terms (which is any describing timbre of sound) are boring for other folks.

edit - Ok. Noticed on page 2 of that thread you chose 4.

The only reason I asked specifically is I don't consider any of them as being able to really copy the sound of what Urs posted, for varying reasons, at least with what was presented in the current alpha.

[urosh]

Noticed on page 2 of that thread you chose 4.

I'm somewhat leaning now to 1 for particular range of patches relevant for me. Right out of the box I didn't like 5. I have theory on why I like 1 and 4, but I'll wait for post-reveal discussion.

The only reason I asked specifically is I don't consider any of them as being able to really copy the sound of what Urs posted

Pro One is IMHO a bit lousy choice to match one physical unit to another or to model, compared to lets say SH101 (although I did once heard 101 that was a bit off compared to norm).

[db3]

I'm somewhat leaning now to 1 for particular range of patches relevant for me. Right out of the box I didn't like 5. I have theory on why I like 1 and 4, but I'll wait for post-reveal discussion..

Rather than any scientific method, or through pushing to extremes, I selected #4 back on page 4, purely because a couple of patches I made sounded more alive to me. I felt #1 was in the same ball park, but just too exaggerated.

[zerocrossing]

I can force Diva into high CPU usage (50% in Studio One's CPU meter) if I use Divine mode, set the voices to 16 and stack to 4, jack up the resonance and filter FM, and play 4 to 6 note chords with a long release. In that regards, yeah, I guess Diva is intensive, but I'm usually playing it in Great quality, with 8 voices and little to no stacking, with less brutal filter settings. In those conditions, Diva's CPU usage will be around 10% while playing pads, which is excellent. I've created projects with a Diva bass, a lead (and a double), and 2 keys/pads tracks and I didn't have to bounce down until it came time do a mix with all the other tracks in the large project needing effects.

Right, but 16 voices... Well the only analog synth I can think of that had 16 voices is the Andromeda A6. All the "classics" were 5-8 voices max. So, right there you're no longer in vintage synth land anyway. If you actually use Diva as a monophonic synth or an 8 voice poly, the CPU usage isn't that bad even in devine mode.

[zerocrossing]

Since Diva's release we suffered stress from two directions: There was some from the side of the people who insisted it didn't sound analogue. Those were easily ignored. Then there were the people who thought it took too much CPU, and that wasn't warranted. This thread is about latter. It's about showing that implementing the same physical/analogue model sounds different if done with CPU saving shortcuts.

Very strange complaints. Diva does indeed sound 100% analog and somebody recently did a matched Oberheim OB-X patch bank proving it. As far as CPU complaints, multicore is working fine and composing in Standard or Great then exporting in Divine is just common sense.

Diva does not sound 100% analog. If it did, the test would have been about 50/50 as people would just be guessing but there was a preference. Maybe the preference wasn't analog, but there was a difference. Also, it was a commercial for a sound bank. Do you think all the sounds that didn't quite work ended up in the sound bank/test?

I'm not saying Diva is bad. It's great. 100% perfect emulation of analog? No. Maybe you don't hear it or don't care, but I was doing all my monophonic synthesis in hardware and using Diva and some other plug ins for my polyphonic analog tasks. Then I bought a Prophet 6 because I kept hearing demos that haunted my dreams (in a good way) I sold Diva a few weeks later. The difference was night and day.

[bibz1st]

For anyone interested (and it seems like there may be a few) this place has a few Pro One (and other synths) sound examples.
http://greatsynthesizers.com
Just head to "The Listening Room"

[louderr]

All right, I attempted to rethink things and re-tried everything on my Mac Book Pro i7, in Ableton and watching CPU usage (which i know is large because of all filters running, etc). I used some of the patches found here and a good one I made, called "Slime Wow". I got the same artifacts as before on #3 and #4, so was unable to get a very good idea with anything swept or heavily modulated or in high freq/res ranges. But I got the same impressions as before. #5 sounded bland, although it had the least artifacts and stayed clean throughout the ranges. #1 was fat and badass and had good "tingle' and "excitement", with nice grit. #2 failed to reproduce the "fun" lo-res 8-bit effect that had attracted me to it as an optional 2nd choice. So now I'm still convinced that #1 sounds best overall. I have demos of every U-He synth, Ace and Diva run very well on my computer and sound wonderful. Repro is promising indeed but I have yet to hear what everyone likes about #3 and #4. Just for the record, something about filters #3 and 4 also "polluted" other plug ins in Ableton when I invoked them with Repro and passed the identical aliasing and other crud into them as well. I'm sure that will corrected, but in case no one has caught that yet it was interesting to note.

[fmr]

All right, I attempted to rethink things and re-tried everything on my Mac Book Pro i7, in Ableton and watching CPU usage (which i know is large because of all filters running, etc). I used some of the patches found here and a good one I made, called "Slime Wow". I got the same artifacts as before on #3 and #4, so was unable to get a very good idea with anything swept or heavily modulated or in high freq/res ranges. But I got the same impressions as before. #5 sounded bland, although it had the least artifacts and stayed clean throughout the ranges. #1 was fat and badass and had good "tingle' and "excitement", with nice grit. #2 failed to reproduce the "fun" lo-res 8-bit effect that had attracted me to it as an optional 2nd choice. So now I'm still convinced that #1 sounds best overall. I have demos of every U-He synth, Ace and Diva run very well on my computer and sound wonderful. Repro is promising indeed but I have yet to hear what everyone likes about #3 and #4. Just for the record, something about filters #3 and 4 also "polluted" other plug ins in Ableton when I invoked them with Repro and passed the identical aliasing and other crud into them as well. I'm sure that will corrected, but in case no one has caught that yet it was interesting to note.

If you were unable to get a good idea with anything swept or heavily modulated or in high freq/res ranges with #3 and #4, I wonder how did you get through with #1 (which, in my experience, was even worse).

[wagtunes]

FWIW, with the exception of the volume glitch between 6 and 7, I like filter 1. I think it has the most "character" of all the filters. It's completely unusable because of the glitch, but if I could put some of that character in the other filters, I'd be really happy. That dirt sounds really good for what I would use it for.

[hakey]

The difference was night and day.

When I know that I'm listening to an analog synth, I can hear the obvious difference. When I don't know, then I'm not so sure.

Confirmation something or other.

[beely]

Then I bought a Prophet 6 because I kept hearing demos that haunted my dreams (in a good way) I sold Diva a few weeks later. The difference was night and day.

You just bought a (beautiful) new P6, sat in front of the lovely physicality of the thing, tweaked it's custom physical control surface, and had fun making noise - and you came to the conclusion it's wonderful and much better than a softsynth?

No surprise there...

[zerocrossing]

The difference was night and day.

When I know that I'm listening to an analog synth, I can hear the obvious difference. When I don't know, then I'm not so sure.

Confirmation something or other.

Oh, don't get me wrong, Diva can fool me a lot of the time. I guessed wrong on that OB8 vs Diva test, but I heard a clear difference. I'm not analog purest, that's for sure. I use what works best for the job. I love love love Zebra 2 and HZebra. In a perfect world I'd have kept Diva for all the things it does that aren't like an analog synth, but I had to pay for that Prophet 6 somehow.

RePro-1 seems like a pretty big step forward from Diva in terms of modeling and I was really impressed.