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

[Ingonator]

Yesterday someone seemed to have published a video about a NKS template for NI Maschine 2.5 (download link for teh file at the YouTube link/page):

https://www.youtube.com/watch?v=ue4Ub5XuAzQ


I do not have the Maschine 2 hardware here but got the Maschine software and a NI Komplete Kontrol S61 keyboard controller where i also got my own NKS template for The Legend.

[AnX]

German products, German artists, German music.

[Ingonator]

Just had another look at the Gearslutz thread:
https://www.gearslutz.com/board/electro ... ng-18.html

5 days ago Richard posted this concerning the ADSR vs ADS envelopes discussion:

Actually this is simple, set both Release knobs to zero and you have the decay off switch emulated. The decay on position corresponds to identical (non-zero) Decay and Release settings. This is how we created our software-vs-hardware comparisons.

Technically, the extra Release knob does not mean the envelopes work different from the vintage hardware. If Legend were analog itself, then yes, going for ADSR would almost certainly change the behavior and spoil the emulation. But in software, anything is possible. We simply run different code depending on the Release knob setting, so you could see the Release knob as an extended switch giving you more freedom without losing anything.

We think ADSR is superior and more intuitive than the ADS concept, and since nothing stops us technically from doing it, we went ahead and used ADSR

Richard

I fully agree with Richard concerning the ADSR being superior as had been mentioend earlier here.

Another post at GS by Richard about the envelopes:

We thought about having ADS in some way during the early development stage, perhaps in a way where the GUI changes. Such an adaptive UI would have been a major issue with the Rack Extension version, however. Later on it turned out the gain is far too small to justify having two envelope systems and the ADSR felt very intuitive in practice.

-----------------------------------------------------------------------------------

Zeroscossing had posted this at the GS thread:

Another tangential point of Monark is that it later found itself dismantled and put into Reaktor 6 as separate Blocks. What would a Model D sound like if it had Waldorf wavetables? A few clicks and you can find out. I never found myself going to Monark itself very much, but the Blocks are fantastic.

This is exactly for what the FX version of The Legend could be ueful for, especially if you route a MIDI track to the FX plugin to make full use of the built-in envelopes and Oscs...

I'll try to prepare a few audio demos of using the FX plugin of The Legend with other synths soon.

[AnX]

Please do, im sure nobody is capable of doing it themselves....

[Ingonator]

nvm

[Yorrrrrr]

OK, found the reason why in my initial A/B comparisons I preferred the sound of Monark over the Legend. In my tests the sampling rate was set to 44KHz, and Monark was "cheating" by running at 88KHz internally. It's set up like this by default to minimize artifacts at high resonance values.

Now that the same A/B test, matching closely the patches, was repeated at 96KHz, the sound quality is equally good on both synths. And the Legend has the added bonus of 4x polyphony.

So, setting the sample rate to something higher than 441 KHz will improve sound quality?

[izonin]

For Monark, definitely. It even warns you if you try to switch to 44KHz from the default 88KHz.

Also some audio interfaces perform better at higher sample rates. Mine definitely does.

I was skeptical about the Legend in the beginning, because for some odd reason I only demoed it at the lower SR. But seems like the hype is well deserved.

[kmonkey]

OK, found the reason why in my initial A/B comparisons I preferred the sound of Monark over the Legend. In my tests the sampling rate was set to 44KHz, and Monark was "cheating" by running at 88KHz internally. It's set up like this by default to minimize artifacts at high resonance values.

Now that the same A/B test, matching closely the patches, was repeated at 96KHz, the sound quality is equally good on both synths. And the Legend has the added bonus of 4x polyphony.

So, setting the sample rate to something higher than 441 KHz will improve sound quality?

I believe he a bit does not understand what he is talking about and there is really no any cheating because Legend likely use oversampling where needed the same way Monark does (it's just you can't see it in Legend like in Reaktor/Monark panel).

It is a very good Chance Synapse already optimized code to rune just as good on any sample rate. I would be really surprised if that is not the case but only Richard of Synapse can tell us that.

To answer your question - yes increasing sample rate can sometimes produce better result especially with older plugins but new generation plugins already are optimized to run at 44khz. In short - avoid and stay away from plugins which behave differently on different sample rates - these are shit.

Legend tend to sound marvelous at any sample rate.

[Yorrrrrr]

The Legend is built with an '8x oversampled engine'..So, it should sound incredible even at 44.1 KHz, and..it really does...

Sorry, but I can't notice a difference when setting the sample rate to 96 Khz (FTR, I have a good audio interface, an RME Babyface Pro) It sounds great regardless.

[chk071]

OK, found the reason why in my initial A/B comparisons I preferred the sound of Monark over the Legend. In my tests the sampling rate was set to 44KHz, and Monark was "cheating" by running at 88KHz internally. It's set up like this by default to minimize artifacts at high resonance values.

Now that the same A/B test, matching closely the patches, was repeated at 96KHz, the sound quality is equally good on both synths. And the Legend has the added bonus of 4x polyphony.

So, setting the sample rate to something higher than 441 KHz will improve sound quality?

I believe he a bit does not understand what he is talking about and there is really no any cheating because Legend likely use oversampling where needed the same way Monark does (it's just you can't see it in Legend like in Reaktor/Monark panel).

Indeed. It says "8x oversampled" here: http://www.synapse-audio.com/thelegend.html

[izonin]

My point is that when you double the sample rate for the Legend, the CPU load doubles too. Which means you're processing at 8x44 or 8x88 respectively.

For Monark, and RePro-1, the CPU stays relatively the same, regardless of the sampling rate. They automatically change the internal oversampling multiplier.

[EnGee]

My understanding from Urs, aciddose and some other members in a thread about aliasing, is that what's important is the internal oversampling inside the coding of the synth, not really the oversampling of the host.

Personally, I run everything at 44.1 Khz and I found the Legend, RePro~1 and Monark sound great and are all at the same level of quality (for me). IMO, running at higher sample rate (96 Khz ..etc) is just a waste of CPU and storage!

[izonin]

My understanding from Urs, aciddose and some other members in a thread about aliasing, is that what's important is the internal oversampling inside the coding of the synth, not really the oversampling of the host.

Personally, I run everything at 44.1 Khz and I found the Legend, RePro~1 and Monark sound great and are all at the same level of quality (for me). IMO, running at higher sample rate (96 Khz ..etc) is just a waste of CPU and storage!

Yes, I remember that discussion. It was about a plugin wrapper that "oversamples" a vst. Not all synths benefit from that approach.


Say, we have a plugin that has a Full mode with x4 oversampling and an Eco mode with x1 oversampling. Full mode will use a relative 100% CPU, and Eco mode - 25% CPU. The sound in Eco will be slightly compromised, of course.

Now let's assume that Monark, RePro-1 and the Legend run natively at 88.1KHz and it's their Full mode with 100% relative CPU load.

Now what happens when we decide to lower the sampling rate?

At 44.1KHz the Legend will halve it's CPU load, entering a sort of an Eco mode. It will run at 50%, not changing it's oversampling.

Monark and RePro-1, on the other hand, will "cheat" by doubling their oversampling multiplier, and remain in Full mode even at 44.1KHz. They will use 100% of their relative CPU and retain the optimal sound quality.

[EnGee]

Yes, I remember that discussion. It was about a plugin wrapper that "oversamples" a vst. Not all synths benefit from that approach.


Say, we have a plugin that has a Full mode with x4 oversampling and an Eco mode with x1 oversampling. Full mode will use a relative 100% CPU, and Eco mode - 25% CPU. The sound in Eco will be slightly compromised, of course.

Now let's assume that Monark, RePro-1 and the Legend run natively at 88.1KHz and it's their Full mode with 100% relative CPU load.

Now what happens when we decide to lower the sampling rate?

At 44.1KHz the Legend will halve it's CPU load, entering a sort of an Eco mode. It will run at 50%, not changing it's oversampling.

Monark and RePro-1, on the other hand, will "cheat" by doubling their oversampling multiplier, and remain in Full mode even at 44.1KHz. They will use 100% of their relative CPU and retain the optimal sound quality.

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?

[izonin]

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.