Synapse Audio Minimoog emulation "The Legend" for VST/AU and RE released!

[EnGee]

Are you sure about the Legend when operating at 44.1KHz? I really don't know. Is there any source to read about it or maybe Richard had said that somewhere?

It's a simple test. Run the synth at 88.1KHz, play a note, mark the CPU meter. Now switch your DAW to 44.1KHz, play a note on the same patch. The CPU load has halved, which means that the plugin hasn't compensated internally. It still uses the same oversampling.

But isn't the CPU load related to the host's sample rate? What I understood that the sample rate of the host and the plugin are independent. Or do you mean that the Legend won't oversample internally at 44.1 Khz while Monark does that at 88.2 Khz?

What I know (and I could be wrong of course) that the sound when oversampled (or not) and leaves the plugin, then you can do nothing about its quality in the host whatever upsampling you do. So that some plugins (like LuSH 101) give you the choice of 'high quality' in play time or/and rendering time.

[izonin]

I'm just saying that the Legend won't switch to a higher oversampling multiplier when the SR is reduced. It's internal sampling rate, of course, is independent of the host. At 88.1KHz it's 8x88 or 704KHz. But at at 44.1KHz it runs at 8x44 or 352KHz internally. And thus uses less CPU resources.

Monark and RePro-1 will run at a constant internal sampling rate. When you force them to work at 44.1 they will simply increase their oversampling multiplier by a factor of 2.

[EvilDragon]

You mean 88.2k, not 88.1.

[izonin]

You mean 88.2k, not 88.1.

Hehe, those funny numbers. 96KHz is what I usually use when recording and mixing.

[Ingonator]

I have done tests with Monark and The Legnd using a signal analyzer and playing at high octaves. The tests were done with the host set to 44.1 kHz.

Monark when forced to 44.1 kHz indeed shows a lot of "noise" which is much better with 88.2 kHz and again slightly better at the the highest setting which is 176.4 kHz. With Monark you could not just see the alissing when forced to 44.1 kHz but you could also hear it and with 88.2 kHz it is much better.
In Moanr k whe doing this test you show also set the "Lekage" parameter at the B page to 0 as this produces additional noise (i turned this off in Monark anyway when i used it).

With the Legend the result at 44.1 kHz alraedy looks very close to that of Monark at the highest setting and also better than that of Monark at 88.2 kHz. If you switch the host to 96 kHz a differnce with The Legend seems to be hard to notice with the signal analyzer and yu could not hear a difference anyway.

The internal oversampling of The Legend seems to do it's job nicely so even at 44.1 kHz set in the host it sounds great and/or "clean" so no need to use the host at higher sample rate for it. I would inded not recommend to force Monark and/or Reaktor to use only 44.1 kHz but the default 88.2 kHz instead.

At 96 kHz the CPU use of The Legend is indeed double or more as i already mentioned with the CPU use for my Improvisation track (that used 9 instances including poly and Unison patches) yesterday.

[izonin]

Hmm, I can definitely hear a difference when switching between sample rates on my system. The same patch sounds brighter and more present @96KHz. Also there is more audible distortion at 44.1

And to make sure it's not my audio interface playing tricks on me, I took a couple of screen shots from Span @ 44.1 and 96k.





The upper harmonics are definitely louder, compared to the fundamental in the second graph.

[Ingonator]

My tests were for aliasing and i used high octaves and/or frequencies for that (up to around 4000 Hz).

Monark when forced to 44.1 kHz indeed has serious and audible aliasing opposing to the 88.2 kHz setting while The Legend sounds "clean" at 44.1 kHz (like Monark at 88.2 kHz or even better).

As The Legend at 96 kHz has more than the double CPU use the audible sound difference does not really justify the increased CPU load IMO but you are free to use it at as high sample rates as you like. Concerning aliasing as alraedy mentioned The Legend sounds clean enough at 44.1 kHz.

Anyway of course this also depends on the overall CPU load of your project. If enough CPU load is left you could of course use the highest available quality.
With my Improvisation for The Legend where i used 9 instances and a maximum of 20 voices at once as mentioned yesterday the maximum CPU load was "only" around 33% at 44.1 kHz while it was at 70+% at 96 kHz (dispaly in Live 9).

[izonin]

Yep, it's a personal thing. The effect to me is like putting a fresh set of strings on my guitar. The sound is instantly clearer and crisper at 96k.


Some beautiful synths in the recent years! Now waiting for the RePro-5 from U-He.

[woodsdenis]

Quick shout out to Richard @ Synapse for incredible support as usual. Great company.

[Dasheesh]

Some beautiful synths in the recent years! Now waiting for the RePro-5 from U-He.

Oh the irony of an oxymoron.

[izonin]

Some beautiful synths in the recent years! Now waiting for the RePro-5 from U-He.

Oh the irony of an oxymoron.

Ah, RePro-5 will be beautiful, too. You'll see.

And hopefully Andy from Cytomic will release some analog emu's based on that gorgeous filter he coded, the Drop.

[waltercruz]

I noticed that The Legend uses another base note for filter key tracking compared to Monark.

On Legend it's G1.
On Monark it's F2.

If I'm understanding it right, from the review of Gordon Reid on the Minimoog reissue, even the new Minimoog has some quirkiness on keytrack

By this time, I was itching to stop analysing and begin playing the new Minimoog properly, so I created some of my favourite Minimoog patches on both synths and, with appropriate care, I was able to duplicate them on each to within a gnat’s wotsits of one another. This was when I uncovered another small difference between the two synths. On a vintage Minimoog, the bottom note outputs 0V, which means that switching on the keyboard control (tracking) switches has no effect on the filter cutoff frequency when playing this note. But the bottom note on a Minimoog is an F, which doesn’t conform to the modern V/Oct standard that decrees that 0V should be generated by a C, so Moog decided to make the bottom key of the new Minimoog output 0.5V. This means that, if you hook it up to another V/Oct synth, they’ll play in tune with one another, but is also means that switching on keyboard tracking when playing the bottom F causes the filter cutoff frequency to shift upwards by up to five semitones. To overcome this, Moog has added a Power On Command (see box) that allows you to determine which note will generate 0V, but this shifts the pitch CV throughout the whole of the instrument so, while it can recreate the historical filter response, it simultaneously causes the oscillators to detune by five semitones! That isn’t much use, and it’s much easier to leave the CV alone and set the cutoff frequency knob to a slightly different position if you’re going to recreate a patch with filter tracking on.

http://www.soundonsound.com/reviews/moo ... og-model-d

[Chris-S]

Thanks, so the base note for keytrack on a vintage MM is the leftmost key. (F2)

Seems that Monark is right.

[Ingonator]

http://www.soundonsound.com/reviews/moo ... og-model-d

Finally found the procedure there that is used for the test:

So how did the two Minimoogs compare? I switched off the oscillators on both synths and caused the filter to oscillate by maximising the emphasis, switched on both Keyboard Control switches so that the filter tracked the keyboard 1:1, and then tuned the top A on the keyboard to A=440. Then I played. Initially, the results were disappointing; I could play downward on my Minimoog for fewer than 20 semitones (to around 150Hz) before the emphasis diminished to the point that the self-oscillation was gone but, on the new model, I could play down to the lowest note on the keyboard and the self-oscillation was still going strong at 50Hz and below.

So i guess with "tuning" he means adjusting the Cutoff to receive 440 Hz in a tuning plugin? At pure self-oscillation tuning is only possible with the Cutoff.

Will try to check this with The Legend soon.

[Chris-S]

Hi Ingo, C4 is middle C (midi #60, 260 hz). 8' foot pitch is fundamental (no octave switch).
If you have a standard 5 oct midi keyboard, the MOOG keyboard would start at the leftmost F.
Playing that note, pressing the Keytrack switches should have no effect, but on LEGEND they have.